diff options
| author | Taru Karttunen <taruti@taruti.net> | 2011-03-30 15:46:40 +0300 |
|---|---|---|
| committer | Taru Karttunen <taruti@taruti.net> | 2011-03-30 15:46:40 +0300 |
| commit | e5888a1ffdae813d7575f5fb02275c6bb07e5199 (patch) | |
| tree | d8d51eac403f07814b9e936eed0c9a79195e2450 /sys/src/ape/cmd/diff | |
Import sources from 2011-03-30 iso image
Diffstat (limited to 'sys/src/ape/cmd/diff')
37 files changed, 23922 insertions, 0 deletions
diff --git a/sys/src/ape/cmd/diff/COPYING b/sys/src/ape/cmd/diff/COPYING new file mode 100755 index 000000000..a43ea2126 --- /dev/null +++ b/sys/src/ape/cmd/diff/COPYING @@ -0,0 +1,339 @@ + GNU GENERAL PUBLIC LICENSE + Version 2, June 1991 + + Copyright (C) 1989, 1991 Free Software Foundation, Inc. + 675 Mass Ave, Cambridge, MA 02139, USA + Everyone is permitted to copy and distribute verbatim copies + of this license document, but changing it is not allowed. + + Preamble + + The licenses for most software are designed to take away your +freedom to share and change it. By contrast, the GNU General Public +License is intended to guarantee your freedom to share and change free +software--to make sure the software is free for all its users. This +General Public License applies to most of the Free Software +Foundation's software and to any other program whose authors commit to +using it. (Some other Free Software Foundation software is covered by +the GNU Library General Public License instead.) You can apply it to +your programs, too. + + When we speak of free software, we are referring to freedom, not +price. Our General Public Licenses are designed to make sure that you +have the freedom to distribute copies of free software (and charge for +this service if you wish), that you receive source code or can get it +if you want it, that you can change the software or use pieces of it +in new free programs; and that you know you can do these things. + + To protect your rights, we need to make restrictions that forbid +anyone to deny you these rights or to ask you to surrender the rights. +These restrictions translate to certain responsibilities for you if you +distribute copies of the software, or if you modify it. + + For example, if you distribute copies of such a program, whether +gratis or for a fee, you must give the recipients all the rights that +you have. You must make sure that they, too, receive or can get the +source code. And you must show them these terms so they know their +rights. + + We protect your rights with two steps: (1) copyright the software, and +(2) offer you this license which gives you legal permission to copy, +distribute and/or modify the software. + + Also, for each author's protection and ours, we want to make certain +that everyone understands that there is no warranty for this free +software. If the software is modified by someone else and passed on, we +want its recipients to know that what they have is not the original, so +that any problems introduced by others will not reflect on the original +authors' reputations. + + Finally, any free program is threatened constantly by software +patents. We wish to avoid the danger that redistributors of a free +program will individually obtain patent licenses, in effect making the +program proprietary. To prevent this, we have made it clear that any +patent must be licensed for everyone's free use or not licensed at all. + + The precise terms and conditions for copying, distribution and +modification follow. + + GNU GENERAL PUBLIC LICENSE + TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION + + 0. This License applies to any program or other work which contains +a notice placed by the copyright holder saying it may be distributed +under the terms of this General Public License. The "Program", below, +refers to any such program or work, and a "work based on the Program" +means either the Program or any derivative work under copyright law: +that is to say, a work containing the Program or a portion of it, +either verbatim or with modifications and/or translated into another +language. (Hereinafter, translation is included without limitation in +the term "modification".) Each licensee is addressed as "you". + +Activities other than copying, distribution and modification are not +covered by this License; they are outside its scope. The act of +running the Program is not restricted, and the output from the Program +is covered only if its contents constitute a work based on the +Program (independent of having been made by running the Program). +Whether that is true depends on what the Program does. + + 1. You may copy and distribute verbatim copies of the Program's +source code as you receive it, in any medium, provided that you +conspicuously and appropriately publish on each copy an appropriate +copyright notice and disclaimer of warranty; keep intact all the +notices that refer to this License and to the absence of any warranty; +and give any other recipients of the Program a copy of this License +along with the Program. + +You may charge a fee for the physical act of transferring a copy, and +you may at your option offer warranty protection in exchange for a fee. + + 2. You may modify your copy or copies of the Program or any portion +of it, thus forming a work based on the Program, and copy and +distribute such modifications or work under the terms of Section 1 +above, provided that you also meet all of these conditions: + + a) You must cause the modified files to carry prominent notices + stating that you changed the files and the date of any change. + + b) You must cause any work that you distribute or publish, that in + whole or in part contains or is derived from the Program or any + part thereof, to be licensed as a whole at no charge to all third + parties under the terms of this License. + + c) If the modified program normally reads commands interactively + when run, you must cause it, when started running for such + interactive use in the most ordinary way, to print or display an + announcement including an appropriate copyright notice and a + notice that there is no warranty (or else, saying that you provide + a warranty) and that users may redistribute the program under + these conditions, and telling the user how to view a copy of this + License. (Exception: if the Program itself is interactive but + does not normally print such an announcement, your work based on + the Program is not required to print an announcement.) + +These requirements apply to the modified work as a whole. If +identifiable sections of that work are not derived from the Program, +and can be reasonably considered independent and separate works in +themselves, then this License, and its terms, do not apply to those +sections when you distribute them as separate works. But when you +distribute the same sections as part of a whole which is a work based +on the Program, the distribution of the whole must be on the terms of +this License, whose permissions for other licensees extend to the +entire whole, and thus to each and every part regardless of who wrote it. + +Thus, it is not the intent of this section to claim rights or contest +your rights to work written entirely by you; rather, the intent is to +exercise the right to control the distribution of derivative or +collective works based on the Program. + +In addition, mere aggregation of another work not based on the Program +with the Program (or with a work based on the Program) on a volume of +a storage or distribution medium does not bring the other work under +the scope of this License. + + 3. You may copy and distribute the Program (or a work based on it, +under Section 2) in object code or executable form under the terms of +Sections 1 and 2 above provided that you also do one of the following: + + a) Accompany it with the complete corresponding machine-readable + source code, which must be distributed under the terms of Sections + 1 and 2 above on a medium customarily used for software interchange; or, + + b) Accompany it with a written offer, valid for at least three + years, to give any third party, for a charge no more than your + cost of physically performing source distribution, a complete + machine-readable copy of the corresponding source code, to be + distributed under the terms of Sections 1 and 2 above on a medium + customarily used for software interchange; or, + + c) Accompany it with the information you received as to the offer + to distribute corresponding source code. (This alternative is + allowed only for noncommercial distribution and only if you + received the program in object code or executable form with such + an offer, in accord with Subsection b above.) + +The source code for a work means the preferred form of the work for +making modifications to it. For an executable work, complete source +code means all the source code for all modules it contains, plus any +associated interface definition files, plus the scripts used to +control compilation and installation of the executable. However, as a +special exception, the source code distributed need not include +anything that is normally distributed (in either source or binary +form) with the major components (compiler, kernel, and so on) of the +operating system on which the executable runs, unless that component +itself accompanies the executable. + +If distribution of executable or object code is made by offering +access to copy from a designated place, then offering equivalent +access to copy the source code from the same place counts as +distribution of the source code, even though third parties are not +compelled to copy the source along with the object code. + + 4. You may not copy, modify, sublicense, or distribute the Program +except as expressly provided under this License. Any attempt +otherwise to copy, modify, sublicense or distribute the Program is +void, and will automatically terminate your rights under this License. +However, parties who have received copies, or rights, from you under +this License will not have their licenses terminated so long as such +parties remain in full compliance. + + 5. You are not required to accept this License, since you have not +signed it. However, nothing else grants you permission to modify or +distribute the Program or its derivative works. These actions are +prohibited by law if you do not accept this License. Therefore, by +modifying or distributing the Program (or any work based on the +Program), you indicate your acceptance of this License to do so, and +all its terms and conditions for copying, distributing or modifying +the Program or works based on it. + + 6. Each time you redistribute the Program (or any work based on the +Program), the recipient automatically receives a license from the +original licensor to copy, distribute or modify the Program subject to +these terms and conditions. You may not impose any further +restrictions on the recipients' exercise of the rights granted herein. +You are not responsible for enforcing compliance by third parties to +this License. + + 7. If, as a consequence of a court judgment or allegation of patent +infringement or for any other reason (not limited to patent issues), +conditions are imposed on you (whether by court order, agreement or +otherwise) that contradict the conditions of this License, they do not +excuse you from the conditions of this License. If you cannot +distribute so as to satisfy simultaneously your obligations under this +License and any other pertinent obligations, then as a consequence you +may not distribute the Program at all. For example, if a patent +license would not permit royalty-free redistribution of the Program by +all those who receive copies directly or indirectly through you, then +the only way you could satisfy both it and this License would be to +refrain entirely from distribution of the Program. + +If any portion of this section is held invalid or unenforceable under +any particular circumstance, the balance of the section is intended to +apply and the section as a whole is intended to apply in other +circumstances. + +It is not the purpose of this section to induce you to infringe any +patents or other property right claims or to contest validity of any +such claims; this section has the sole purpose of protecting the +integrity of the free software distribution system, which is +implemented by public license practices. Many people have made +generous contributions to the wide range of software distributed +through that system in reliance on consistent application of that +system; it is up to the author/donor to decide if he or she is willing +to distribute software through any other system and a licensee cannot +impose that choice. + +This section is intended to make thoroughly clear what is believed to +be a consequence of the rest of this License. + + 8. If the distribution and/or use of the Program is restricted in +certain countries either by patents or by copyrighted interfaces, the +original copyright holder who places the Program under this License +may add an explicit geographical distribution limitation excluding +those countries, so that distribution is permitted only in or among +countries not thus excluded. In such case, this License incorporates +the limitation as if written in the body of this License. + + 9. The Free Software Foundation may publish revised and/or new versions +of the General Public License from time to time. Such new versions will +be similar in spirit to the present version, but may differ in detail to +address new problems or concerns. + +Each version is given a distinguishing version number. If the Program +specifies a version number of this License which applies to it and "any +later version", you have the option of following the terms and conditions +either of that version or of any later version published by the Free +Software Foundation. If the Program does not specify a version number of +this License, you may choose any version ever published by the Free Software +Foundation. + + 10. If you wish to incorporate parts of the Program into other free +programs whose distribution conditions are different, write to the author +to ask for permission. For software which is copyrighted by the Free +Software Foundation, write to the Free Software Foundation; we sometimes +make exceptions for this. Our decision will be guided by the two goals +of preserving the free status of all derivatives of our free software and +of promoting the sharing and reuse of software generally. + + NO WARRANTY + + 11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY +FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN +OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES +PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED +OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF +MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS +TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE +PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, +REPAIR OR CORRECTION. + + 12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING +WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR +REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, +INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING +OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED +TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY +YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER +PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE +POSSIBILITY OF SUCH DAMAGES. + + END OF TERMS AND CONDITIONS + + Appendix: How to Apply These Terms to Your New Programs + + If you develop a new program, and you want it to be of the greatest +possible use to the public, the best way to achieve this is to make it +free software which everyone can redistribute and change under these terms. + + To do so, attach the following notices to the program. It is safest +to attach them to the start of each source file to most effectively +convey the exclusion of warranty; and each file should have at least +the "copyright" line and a pointer to where the full notice is found. + + <one line to give the program's name and a brief idea of what it does.> + Copyright (C) 19yy <name of author> + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + +Also add information on how to contact you by electronic and paper mail. + +If the program is interactive, make it output a short notice like this +when it starts in an interactive mode: + + Gnomovision version 69, Copyright (C) 19yy name of author + Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'. + This is free software, and you are welcome to redistribute it + under certain conditions; type `show c' for details. + +The hypothetical commands `show w' and `show c' should show the appropriate +parts of the General Public License. Of course, the commands you use may +be called something other than `show w' and `show c'; they could even be +mouse-clicks or menu items--whatever suits your program. + +You should also get your employer (if you work as a programmer) or your +school, if any, to sign a "copyright disclaimer" for the program, if +necessary. Here is a sample; alter the names: + + Yoyodyne, Inc., hereby disclaims all copyright interest in the program + `Gnomovision' (which makes passes at compilers) written by James Hacker. + + <signature of Ty Coon>, 1 April 1989 + Ty Coon, President of Vice + +This General Public License does not permit incorporating your program into +proprietary programs. If your program is a subroutine library, you may +consider it more useful to permit linking proprietary applications with the +library. If this is what you want to do, use the GNU Library General +Public License instead of this License. diff --git a/sys/src/ape/cmd/diff/ChangeLog b/sys/src/ape/cmd/diff/ChangeLog new file mode 100755 index 000000000..6cfc7791a --- /dev/null +++ b/sys/src/ape/cmd/diff/ChangeLog @@ -0,0 +1,1766 @@ +Sat Oct 1 05:24:19 1994 Paul Eggert <eggert@twinsun.com> + + * Version 2.7 released. + + * configure.in (AC_HEADER_SYS_WAIT): Add. + (AC_CHECK_HEADERS): Remove sys/wait.h. + (AC_CHECK_FUNCS): Add tmpnam. + * system.h (<sys/wait.h>, WEXITSTATUS): Use simpler scheme + now that HAVE_SYS_WAIT_H is not set on hosts + that are incompatible with Posix applications. + + * util.c (dir_file_pathname): Use filename_lastdirchar not strrchr. + * sdiff.c (expand_name): Likewise. + (private_tempnam): Use tmpnam if HAVE_TMPNAM; this simplifies porting. + (exists, letters): Omit if HAVE_TMPNAM. + + * diff3.c (read_diff): If STAT_BLOCKSIZE yields zero, + adjust it to a more reasonable value. + +Sat Sep 24 20:36:40 1994 Paul Eggert <eggert@twinsun.com> + + * sdiff.c (exists, private_tempname): Adopt latest GNU libc algorithm. + (private_tempnam): Specialize for sdiff to avoid portability problems. + +Thu Sep 22 16:47:00 1994 Paul Eggert <eggert@twinsun.com> + + * configure.in (AC_ARG_PROGRAM): Added. + (AC_OUTPUT): Add [date > stamp-h]. + + * Makefile.in (DEFAULT_EDITOR_PROGRAM, DIFF_PROGRAM, LIBOBJS, + NULL_DEVICE, PR_PROGRAM, PROGRAMS): New variables. + (check, stamp-h.in, cmp.o, util.o): New targets. + (edit_program_name): New variable; replaces old binprefix method. + (install, uninstall): Use it. + (binprefix): Removed. + (distfiles): Add stamp-h.in. + (clean): Clean stamp-h. + (config.hin, config.h): Use time stamp files. + (cmp_o): Add $(LIBOBJS). + (install): Install info files from srcdir if they're not in `.'. + + * cmp.c, io.c (word): Don't define if already defined. + + * comp.c (main): Use setmode, not open(..., O_BINARY); this gets stdin. + Use NULL_DEVICE instead of "/dev/null". + (cmp): Use %lu instead of %ld when it is more likely to be right. + + * diff.h (PR_FILE_NAME): Rename to PR_PROGRAM and move to Makefile.in, + util.c. + + * diff3.c (main): Give proper diagnostic if too many labels were given. + (read_diff): Use SYSTEM_QUOTE_ARG. + + * system.h: <string.h>: Include if HAVE_STRING_H, too. + <ctype.h>: Include here. All includers changed. + (CTYPE_DOMAIN, ISDIGIT, ISPRINT, ISSPACE, ISUPPER): New macros that + work around common <ctype.h> problems. + (O_BINARY): Remove. + (SYSTEM_QUOTE_ARG): New macros. + + * diff.c: Add comment. + + * util.c (PR_PROGRAM): Moved here from diff.h. + (begin_output): Use SYSTEM_QUOTE_ARG. + + * io.c (read_files): Set mode to binary before returning 1. + + * sdiff.c (TMPDIR_ENV): New macro. + (DEFAULT_EDITOR_PROGRAM): Renamed from DEFAULT_EDITOR for consistency. + (expand_name): Change `isdir' to `is_dir' to avoid theoretical ctype + namespace contamination. + (main): Use SYSTEM_QUOTE_ARG. + (private_tempnam): Don't access "/tmp" directly; use PVT_tmpdir. + +Tue Sep 13 18:46:43 1994 Paul Eggert <eggert@twinsun.com> + + * configure.in (AC_FUNC_MEMCHR): Remove. Autoconf didn't adopt this, + since we need not worry about an old experimental library + where memchr didn't work. + (AC_FUNC_MEMCMP): Not needed, since we only test for equality. + (AC_REPLACE_FUNCS): Add test for memchr. + (AC_CHECK_FUNCS): Check for memchr, not memcpy, since it'll be cached. + (AC_CHECK_HEADERS): Add string.h; regex.c uses on some old hosts. + + * system.h (memcmp): Define in terms of bcmp. + Use HAVE_MEMCHR to test for all mem* routines. + + * Makefile.in (srcs): Remove memcmp.c. + We use bcmp if memcmp doesn't work, since we only test for equality. + +Mon Sep 12 15:52:22 1994 Paul Eggert <eggert@twinsun.com> + + * configure.in (AC_CONFIG_HEADER): Rename config.h.in to config.hin. + (AC_ISC_POSIX, AC_MINIX): Go back to these old names for Autoconf 2. + (AC_CHECK_HEADERS): Remove now-redundant check for <string.h>. + (AC_CHECK_FUNCS): Check for strchr. + (AC_FUNC_MEMCHR, AC_FUNC_MEMCMP, AC_CHECK_FUNCS): Use special-purpose + macros when suitable. + * memcmp.c: New file. + * Makefile.in (CPPFLAGS, DEFS, CFLAGS, LDFLAGS, prefix, exec_prefix): + Default to autoconf-specified strings. + (COMPILE): Use the defaults. + (srcs): Add memcmp.c. + (distfiles): Rename config.h.in->config.hin, install.sh->install-sh. + (Makefile, config.h, config.hin, config.status): Rework for + compatibility with Autoconf 2. + * io.c (binary_file_p): Assume non-broken memchr. + * memchr.c: Assume compiler understands void *; otherwise + we don't match GCC's internal declaration of memchr. + * system.h: Use more modern autoconf approach to standard C headers. + * version.c: Include <config.h>, not "config.h". + + * diff.c, diff.h (ignore_some_line_changes): + New variable; replaces `length_varies'. + (line_end_char): Replace with '\n'; it wasn't being used consistently. + + * io.c (find_and_hash_each_line): Fix inconsistencies with -b -w -i and + incomplete lines. Put incomplete lines into their own bucket. + This means line_cmp no longer needs line length arguments, + and equivalence classes' line lengths no longer need to include \n. + Invoke line_cmp only if ignore_some_line_changes. + (prepare_text_end): -B no longer ignores missing newlines. + (read_files): Allocate another bucket for incomplete lines. + + * util.c (line_cmp): Now takes just two arguments. No longer + optimizes for common case of exact equality; the caller does that + optimization now. The caller is changed accordingly. + Optimize for the common case of mostly equality. + Use isupper+tolower instead of islower+toupper, for consistency. + + * waitpid.c (waitpid): Fix typo with internal scoping. + +Thu Sep 8 08:23:15 1994 Paul Eggert <eggert@twinsun.com> + + * configure.in: Revamp for Autoconf 2. + * memchr.c, waitpid.c: New source files for substitute functions. + * Makefile.in (diff_o, diff3_o, sdiff_o): Add $(LIBOBJS). + (srcs): Add memchr.c, waitpid.c. + (distfiles): Add install.sh, memchr.c, waitpid.c, install.sh. + * system.h: Use Autoconf 2 style HAVE_DIRENT_H etc. macros for dirs. + * dir.c (dir_sort): Prefer NAMLEN (p) to strlen (p->d_name). + Change VOID_CLOSEDIR to CLOSEDIR_VOID for Autoconf 2. + * sdiff.c, util.c (memchr, waitpid): Remove; use new substitutes. + * diff3.c (read_diff): Use new waitpid substitute. + + * cmp.c, diff.c, diff3.c, sdiff.c (check_stdout, try_help): New fns. + (usage): Just print more detailed usage message; let caller exit. + * diff.c (option_help): New variable. + (filetype): Add Posix.1b file types. + +Fri Sep 2 16:01:49 1994 Paul Eggert <eggert@twinsun.com> + + * configure.in: Switch to new autoconf names. Add sys/file.h test. + * Makefile.in (distclean): Clean config.cache, config.log + (used by new autoconf). + + * diff.c, diff3.c, (main), sdiff.c (trapsigs): If we'll have children, + make sure SIGCHLD isn't ignored. + + * diff3.c (DIFF_CHUNK_SIZE): Removed. Get size from STAT_BLOCKSIZE. + (INT_STRLEN_BOUND): New macro. + + * ifdef.c (format_group, groups_letter_value): + Use * instead of [] in prototypes. + + * system.h: Include <sys/file.h> only if HAVE_SYS_FILE_H. + (S_IXGRP, S_IXOTH, S_IXUSR): Remove unused macros. + + * util.c (begin_output): Check fdopen result. + + The following changes simplify porting to non-Posix environments. + * cmp.c, diff.c, diff3.c, sdiff.c, (main): Call initialize_main first. + * diff.c (binary_I_O): New variable for --binary option. + (main, usage, compare_files): Support --binary option. + (compare_files): Use filename_lastdirchar to find last + directory char in a file name. + * cmp.c (main), diff.c (compare_files), dir.c (compare_names, + diff_dirs): Use filename_cmp to compare file names. + Use same_file to determine whether two files are the same. + * context.c (print_context_label): Check whether ctime yields 0. + * diff3.c (read_diff), sdiff.c (cleanup, main, waitpid), + util.c (begin_output): Use popen+pclose if !HAVE_FORK. + * io.c (sip): If HAVE_SETMODE, test for binary files in O_BINARY mode. + * sdiff.c (ck_fdopen): Function removed. + (edit): Use system if !HAVE_FORK. + (execdiff): Now assumes caller has pushed all args, plus trailing 0. + All callers changed. + (private_tempnam): Try TMP if TMPDIR isn't defined. + Fit temporary filenames into 8.3 limit. + * system.h (STAT_BLOCKSIZE): Don't define if already defined. + (min, max): Undef if already defined. + (filename_cmp, filename_lastdirchar, HAVE_FORK, HAVE_SETMODE, + initialize_main O_BINARY, same_file): New macros. + +Fri Jun 17 11:23:53 1994 David J. MacKenzie (djm@geech.gnu.ai.mit.edu) + + * Makefile.in (info, dvi, diff.dvi): New targets. + (clean): Remove TeX output files. + +Fri Jun 17 05:37:52 1994 Paul Eggert (eggert@twinsun.com) + + * cmp.c, io.c (word): Change from typedef to #define, to avoid + collision with Unicos 8.0 <sys/types.h>, which also typedefs `word'. + +Thu Apr 15 00:53:01 1994 Paul Eggert (eggert@twinsun.com) + + * diff3.c (scan_diff_line), util.c (print_number_range): Don't + rely on promotion to make the old-style parameter type agree + with the prototype parameter type; this doesn't work on + Apollos running bsd4.3. + +Mon Jan 3 02:05:51 1994 Paul Eggert (eggert@twinsun.com) + + * Makefile.in (LDFLAGS): Remove -g. Change all link commands + to use both $(CFLAGS) and $(LDFLAGS). + +Mon Dec 13 12:23:27 1993 Paul Eggert (eggert@twinsun.com) + + * system.h: Don't assume dirent.h exists just because + _POSIX_VERSION is defined. + +Fri Dec 3 18:39:39 1993 Paul Eggert (eggert@twinsun.com) + + * diff.c (main): allow -pu. + +Tue Nov 23 03:51:08 1993 Paul Eggert (eggert@twinsun.com) + + * Makefile.in (distclean): Remove config.h. + +Wed Nov 10 00:28:27 1993 Paul Eggert (eggert@twinsun.com) + + * Version 2.6 released. + + * analyze.c (too_expensive): New variable, for heuristic to + limit the worst-case cost to O(N**1.5 log N) at the price of + producing suboptimal output for large inputs with many differences. + (diff_2_files): Initialize it. + (struct partition): New type. + (SNAKE_LIMIT): New macro; merely documents already-used number 20. + (diag): New `minimal' arg; all callers changed. Put results into + struct partition. Apply `too_expensive' heuristic. Tune. + (compareseq): New `minimal' arg; all callers changed. Tune. + (shift_boundaries): Improve heuristic to also coalesce adjacent runs + of changes more often. + + * diff.c (long_options, main, usage): Add `--help'. + (main): Send version number to stdout, not stderr. + (usage): Send usage to stdout, not stderr. + (compare_files): Initialize `inf' properly. + + * io.c (word): Change to `int'; it makes a big difference on x86. + (sip, slurp): Put off allocating room to hold the whole file until we + have to read the whole file. This wins if the file turns out + to be binary. + + * util.c (xmalloc, xrealloc): "virtual memory" -> "memory" + (primes): Omit large primes if INT_MAX is small. + + * sdiff.c (usage): Send usage to stdout, not stderr. + (long_options, main, usage): Add `--help'. + (main): Send version number to stdout, not stderr. Exit afterwards. + + * diff3.c (usage): Send usage to stdout, not stderr. + (long_options, main, usage): Add `--help'. + (read_diff): Detect integer overflow in buffer size calculations. + + * cmp.c (word): New type. All uses of `long' for + word-at-a-time comparisons changed to `word'. + (long_options, main, usage): Add `--help'. + (usage): Send usage to stdout, not stderr. + (main): Add `-v'. Send version number to stdout, not stderr. + + * configure.in (AC_HAVE_HEADERS): Add unistd.h; remove AC_UNISTD_H. + +Mon Sep 27 07:20:24 1993 Paul Eggert (eggert@twinsun.com) + + * diff.c (add_exclude_file): Cast memchr to (char *) + to suppress bogus warnings on some nonstandard hosts. + + * Makefile.in (cmp): Add version.o. + + * analyze.c (diff_2_files): Work around memcmp bug with size=0. + + * cmp.c (main, usage, version_string): Add --version option. + + * system.h (malloc, realloc): Declare only if !HAVE_STDLIB_H. + (memchr): Declare only if !HAVE_MEMCHR. These changes are + needed to keep some nonstandard hosts happy. + + * util.c (memchr): Make first arg char const * + to match standard. + (xmalloc, xrealloc): Cast malloc, realloc + to (VOID *) to suppress bogus warnings on some nonstandard hosts. + + * diff3.c (xmalloc, xrealloc): Cast malloc, realloc + to (VOID *) to suppress bogus warnings on some nonstandard hosts. + + * sdiff.c (xmalloc, xrealloc): Cast malloc, realloc + to (VOID *) to suppress bogus warnings on some nonstandard hosts. + (lf_copy, lf_skip, lf_snarf): Cast memchr to (char *) + to suppress bogus warnings on some nonstandard hosts. + (memchr): Make first arg char const * + to match standard. + +Mon Sep 27 00:23:37 1993 Paul Eggert (eggert@twinsun.com) + + * Version 2.5 released. + + * analyze.c (diff_2_files): Work around memcmp bug with size=0. + + * cmp.c (main, usage, version_string): Add --version option. + * Makefile.in (cmp): Add version.o. + + * diff.c (add_exclude_file): Cast memchr to (char *) + to suppress bogus warnings on some nonstandard hosts. + * sdiff.c (lf_copy, lf_skip, lf_snarf): Likewise. + + * diff3.c, sdiff.c, util.c (xmalloc, xrealloc): Cast malloc, realloc + to (VOID *) to suppress bogus warnings on some nonstandard hosts. + + * sdiff.c, util.c (memchr): Make first arg char const * + to match standard. + + * system.h (malloc, realloc): Declare only if !HAVE_STDLIB_H. + (memchr): Declare only if !HAVE_MEMCHR. These changes are + needed to keep some nonstandard hosts happy. + + * xmalloc.c: Include <sys/types.h> always; some nonstandard hosts + need it for size_t even if STDC_HEADERS. + +Sat Sep 18 01:33:07 1993 Paul Eggert (eggert@twinsun.com) + + * configure.in (AC_STAT_MACROS_BROKEN): Add. + * system.h (S_IS{BLK,CHR,DIR,FIFO,REG,SOCK}): Fix defns if + STAT_MACROS_BROKEN. + + * Makefile.in (diff3, sdiff, cmp): Do not link $(ALLOCA). + + * analyze.c (discard_confusing_lines): Make defn static, like decl. + * sdiff.c (xmalloc): Likewise. + + * ifdef.c (format_group): Ensure isdigit argument isn't < 0. + + * side.c (print_half_line): Use isprint, since some hosts lack isgraph. + * util.c (output_1_line): Likewise. Ensure its argument isn't < 0. + (xmalloc, xrealloc): Remove needless casts. + + * system.h (volatile, const): + Define these before including any system headers, + so that they're used consistently in all system includes. + (getenv, malloc, realloc): Declare even if HAVE_STDLIB_H, since some + <stdlib.h>s don't declare them. + (memchr): Likewise for <string.h>. + + * cmp.c, diff3.c, diff.h, sdiff.c: Include "system.h" first. + * diff.c: Remove redundant "system.h" inclusion. + + * diff3.c (xmalloc): Now static. + (xmalloc, realloc): Remove needless casts. + (READNUM): Ensure isdigit argument isn't negative. + +Wed Sep 14 07:14:15 1993 Paul Eggert (eggert@twinsun.com) + + * Version 2.4 released. + + * ifdef.c (scan_char_literal): New function, for new %c'x' and + %c'\ooo' format specs. + (format_group, print_ifdef_lines): Use it. Remove %0 format spec. + + * cmp.c (cmp): Don't try to read past end of file; this doesn't + work on ttys. + + * system.h, version.c: #include <config.h>, not "config.h", to allow + configuring in a separate directory when the source directory has + already been configured. + * Makefile.in (COMPILE): New defn, with proper -I options so that + `#include <config.h>' works. + (.c.o, diff3.o, sdiff.o): Use it. + +Mon Sep 13 06:45:43 1993 Paul Eggert (eggert@twinsun.com) + + * diff.c (main, longopts): Add --line-format=FORMAT option. + (specify_format): Args no longer const pointers. All callers changed. + + * ifdef.c: Add support for %?c, %(A=B?T:E), PRINTF_SPECn formats. + (struct group): New struct. + (print_ifdef_lines): Use it to simplify argument passing. + Remove the convention that last arg -1 signifies that the lines + from file 2 are the same as the lines from file 1; this + convention no longer works, now that line numbers might be + printed out, since the line numbers may differ. + Add first FILE * argument to output to. All callers changed. + Use a faster test for the single-fwrite optimization. + (format_group, scan_printf_spec, groups_letter_value): New functions. + + * diff.h (group_format, line_format): No longer const pointers. + (format_ifdef): 1st arg is no longer const pointer. + + * configure.in: Configure HAVE_LIMITS_H, HAVE_STDLIB_H. + * system.h <limits.h>, <stdlib.h>, <string.h>: + Include only if HAVE_LIMITS_H etc. + + * system.h (memcmp, memcpy, strchr, strrchr, struct dirent): Prefer + these standard names to the traditional names (bcmp, bcpy, index, + rindex, struct direct). All callers changed. + + * system.h (PARAMS, VOID): + Define earlier so that malloc decl can use VOID. + (STAT_BLOCKSIZE): Simplify ersatz defn; just use 8K. + +Fri Sep 3 00:21:02 1993 Paul Eggert (eggert@twinsun.com) + + * diff.c (compare_files): Two files with the same name must be + the same file; avoid a needless `stat' in that case. + +Fri Aug 27 06:59:03 1993 Paul Eggert (eggert@twinsun.com) + + * Pervasive changes for portability to 64-bit hosts: + Add prototypes to function declarations. + Use size_t, not int, when needed. + + * Other pervasive changes: + Use `const' more often. + Use STD{IN,OUT,ERR}_FILENO instead of [012]. + Use 0, not NULL, for portability to broken hosts. + + * Makefile.in: (srcs, objs, distfiles, cmp): New files cmpbuf.[ch]. + (distfiles): New files config.h.in, mkinstalldirs. + (.c.o): Add -DHAVE_CONFIG_H. + + * analyze.c: (diag): Pacify `gcc -Wall' with a useless assignment. + (diff_2_files): Use l.c.m., not max, of files' buffer sizes. + + * cmp.c: Make globals static when possible. + + (file): Now a 2-element array; replaces `file1' and `file2'. + (file_desc, buffer): Likewise, for file[12]_desc and buf[12]. + (main): Likewise, for stat_buf[12]. Index these variables with `i'. + + (ignore_initial): New var. + (long_options): Now const. Add `--ignore-initial'. + (usage): Sort options and add `--ignore-initial'. + (main, cmp): Add `--ignore-initial' support. + + (main): `cmp - -' now succeeds. + When comparing standard input to a file, and using a shortcut (e.g. + looking at file sizes or inode numbers), take the lseek offset into + account before deciding whether the files are identical. + Avoid mentioning `dev_t', `ino_t' for portability to nonstandard hosts. + Use l.c.m. of files' buffer sizes, not 8 * 1024. + ferror (stdout) does not imply errno has a useful value. + If 2nd file is "-", treat it first, in case stdin is closed. + + (cmp): Always compute `char_number', `smaller' for speed and simplicity. + Say `cmp: EOF on input', not `/usr/gnu/bin/cmp: EOF on input', + as per Posix.2. + + (block_compare_and_count): Increment line_number argument. + Remove end_char argument; it's always '\n'. All callers changed. + Do not assume sizeof(long) == 4; this isn't true on some 64-bit hosts. + (block_compare): Minimize differences with block_compare_and_count. + + (block_read): Coalesce `bp += nread's. + + (printc): Remove `FILE *' arg; output to stdout. All callers changed. + + * configure.in: Configure HAVE_SIGACTION, RETSIGTYPE, HAVE_VPRINTF. + Configure into config.h. + + * context.c (print_context_label): + Standard input's st_mtime is no longer a special case + here, since `compare_files' now sets it to the current time. + + * diff.c (usage): Sort options. + (filetype): New function. + (compare_files): Set stdin's st_mtime to be the current time. + Leave its name "-" instead of changing it to "Standard Input"; + to test whether a file is stdin, we must compare its name to "-" instead + of its desc to 0, since if it's closed other file descs may be 0. + When comparing standard input to a file, and using a shortcut (e.g. + looking at file sizes or inode numbers), take the lseek offset into + account before deciding whether the files are identical. + Pretend that nonexistent files have the same filetype as existing files. + Rename `errorcount' to `failed', since it's boolean. + In directory comparisons, if a file is neither a regular file nor a + directory, just print its type and the other file's type. + + * diff.h (Is_space, textchar): Remove. + (struct msg, msg_chain, msg_chain_end): Move to util.c. + (VOID): Move to system.h. + (line_cmp, version_string, change_letter, print_number_range, + find_change): New decls. + + * diff.texi: + whitespace -> white space. It now stands for whatever isspace yields. + Add --ignore-initial. + + * diff3.c (VOID): Move to system.h. + (version_string): Now char[]. + (usage): Sort options. + (process_diff): Pacify `gcc -Wall' with a useless assignment. + (read_diff): pid is of type pid_t, not int. Use waitpid if available. + (output_diff3): Simplify test for `\ No newline at end of file' message. + + * dir.c (struct dirdata): Rename `files' to `names' to avoid confusion + with external struct file_data `files'. + + * io.c (line_cmp): Move declaration to diff.h. + (textchar): Remove. + (find_and_hash_each_line): Use locale's definition of white space + instead of using one hardwired defn for -b and another for -w. + + * normal.c (change_letter, print_number_range, find_change): + Move decls to diff.h. + (print_normal_hunk): Now static. + + * sdiff.c (SEEK_SET): Move to system.h. + (version_string): Now char[], not char*. + (private_tempnam): Remove hardcoded limit on temporary file names. + (exiterr, perror_fatal, main): When exiting because of a signal, + exit with that signal's status. + (lf_refill, main, skip_white, edit, interact): Check for signal. + (ignore_SIGINT): Renamed from `ignore_signals'. + (NUM_SIGS, initial_handler): New macros. + (initial_action, signal_received, sigs_trapped): New vars. + (catchsig, trapsigs): Use sigaction if possible, since this closes the + windows of vulnerability that `signal' has. Use RETSIGTYPE not void. + When a signal comes in, just set a global variable; this is safer. + (checksigs, untrapsig): New functions. + (edit): Pacify `gcc -Wall' with a useless assignment. + Respond to each empty line with help, not to every other empty line. + (private_tempnam): Remove hardcoded limit on temporary file name length. + Don't assume sizeof (pid_t) <= sizeof (int). + + * system.h: (S_IXOTH, S_IXGRP, S_IXUSR, + SEEK_SET, SEEK_CUR, + STDIN_FILENO, STDOUT_FILENO, STDERR_FILENO): + New macros, if system doesn't define them. + (volatile): Don't define if already defined. + (PARAMS): New macro. + (VOID): Move here from diff.h. + + * util.c (struct msg, msg_chain, msg_chain_end): Moved here from diff.h. + (message5): New function. + (pr_pid): New var. + (begin_output): Allocate `name' more precisely. + Put child pid into pr_pid, so that we can wait for it later. + Don't check execl's return value, since any return must be an error. + (finish_output): Detect and report output errors. + Use waitpid if available. Check pr exit status. + (line_cmp): Use locale's definition of white space + instead of using one hardwired defn for -b and another for -w. + (analyze_cmp): Avoid double negation with `! nontrivial'. + Pacify `gcc -Wall' be rewriting for-loop into do-while-loop. + (dir_file_pathname): New function. + + * version.c (version_string): Now char[], not char*. + +Thu Jul 29 20:44:30 1993 David J. MacKenzie (djm@wookumz.gnu.ai.mit.edu) + + * Makefile.in (config.status): Run config.status --recheck, not + configure, to get the right args passed. + +Thu Jul 22 10:46:30 1993 Paul Eggert (eggert@twinsun.com) + + * Makefile.in (dist): Replace `if [ ! TEST ]; then ACTION; fi' + with `[ TEST ] || ACTION || exit' so that the containing for-loop exits + with proper status for `make'. + +Thu Jul 8 19:47:22 1993 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) + + * Makefile.in (installdirs): New target. + (install): Use it. + (Makefile, config.status, configure): New targets. + +Sat Jun 5 23:10:40 1993 Paul Eggert (eggert@twinsun.com) + + * Makefile.in (dist): Switch from .z to .gz. + +Wed May 26 17:16:02 1993 Paul Eggert (eggert@twinsun.com) + + * diff.c (main): Cast args to compare_files, for traditional C. + * side.c (print_sdiff_common_lines_print_sdiff_hunk): Likewise. + * analyze.c, diff3.c, sdiff.c, util.c: Don't assume NULL is defined + properly. + +Tue May 25 14:54:05 1993 Paul Eggert (eggert@twinsun.com) + + * analyze.c (diff_2_files): With -q, do not report that files differ + if all their differences are ignored. + (briefly_report): New function. + * diff.h (ignore_some_changes): New variable. + * diff.c (compare_files): Don't use the file size shortcut if + ignore_some_changes is nonzero, since the file size may differ + merely due to ignored changes. + (main): Set ignore_some_changes if we might ignore some changes. + Remove unsystematic assignment of 0 to static vars. + * io.c (read_files): New argument PRETEND_BINARY says whether to + pretend the files are binary. + + * diff3.c (tab_align_flag): New variable, for new -T option. + (main, usage, output_diff3): Add support for -T. + +Sun May 23 15:25:29 1993 Richard Stallman (rms@mole.gnu.ai.mit.edu) + + * dir.c (dir_sort): Always init `data' to avoid GCC warning. + +Sat May 22 15:35:02 1993 Paul Eggert (eggert@twinsun.com) + + * Makefile.in (dist): Change name of package from diff to diffutils. + Don't bother to build .Z dist; .z suffices. + +Fri May 21 16:35:22 1993 Paul Eggert (eggert@twinsun.com) + + * diff.c: Include "system.h" to get memchr declaration. + * system.h (memchr): Declare if !HAVE_MEMCHR, not if + !HAVE_MEMCHR && !STDC_HEADERS. + +Wed May 19 17:43:55 1993 Paul Eggert (eggert@twinsun.com) + + * Version 2.3 released. + +Fri Apr 23 17:18:44 1993 Paul Eggert (eggert@twinsun.com) + + * io.c (find_identical_ends): Do not discard the last HORIZON_LINES + lines of the prefix, or the first HORIZON_LINES lines of the suffix. + * diff.c (main, longopts, usage): Add --horizon-lines option. + * diff3.c (main, process_diff, read_diff): Invoke second diff + with --horizon-lines determined by the first diff. + * diff.h, diff3.c (horizon_lines): New variable. + +Mon Mar 22 16:16:00 1993 Roland McGrath (roland@churchy.gnu.ai.mit.edu) + + * system.h [HAVE_STRING_H || STDC_HEADERS] (bcopy, bcmp, bzero): + Don't define if already defined. + +Fri Mar 5 00:20:16 1993 Richard Stallman (rms@mole.gnu.ai.mit.edu) + + * diff.c (main): Use NULL in arg to compare_files. + +Thu Feb 25 15:26:01 1993 Roland McGrath (roland@churchy.gnu.ai.mit.edu) + + * system.h: Declare memchr #if !HAVE_MEMCHR && !STDC_HEADERS, + not #if !HAVE_MEMCHR || !STDC_HEADERS. + +Mon Feb 22 15:04:46 1993 Richard Stallman (rms@geech.gnu.ai.mit.edu) + + * io.c (find_identical_ends): Move complicated arg outside GUESS_LINES. + +Mon Feb 22 12:56:12 1993 Roland McGrath (roland@churchy.gnu.ai.mit.edu) + + * Makefile.in (.c.o): Add -I$(srcdir); put $(CFLAGS) last before $<. + +Sat Feb 20 19:18:56 1993 Richard Stallman (rms@mole.gnu.ai.mit.edu) + + * io.c (binary_file_p): Return zero if file size is zero. + +Fri Feb 19 17:31:32 1993 Roland McGrath (roland@geech.gnu.ai.mit.edu) + + * Version 2.2 released. + + * system.h [HAVE_STRING_H || STDC_HEADERS] (index, rindex): Don't + define if already defined. + +Wed Feb 17 17:08:00 1993 Roland McGrath (roland@churchy.gnu.ai.mit.edu) + + * Makefile.in (srcs): Remove limits.h. + +Thu Feb 11 03:36:00 1993 Richard Stallman (rms@mole.gnu.ai.mit.edu) + + * diff3.c (xmalloc): No longer static. + + * sdiff.c (edit): Allocate buf dynamically. + + * dir.c (dir_sort): Handle VOID_CLOSEDIR. + +Wed Feb 10 00:15:54 1993 Richard Stallman (rms@mole.gnu.ai.mit.edu) + + * limits.h: File deleted (should never have been there). + +Tue Feb 9 03:53:22 1993 Richard Stallman (rms@mole.gnu.ai.mit.edu) + + * Makefile.in (.c.o, diff3.o, sdiff.o): Put $(CFLAGS) last. + +Wed Feb 3 15:42:10 1993 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) + + * system.h: Don't #define const; let configure do it. + +Mon Feb 1 02:13:23 1993 Paul Eggert (eggert@hal.gnu.ai.mit.edu) + + * Version 2.1 released. + + * Makefile.in (dist): Survive ln failures. Create .tar.z + (gzipped tar) file as well as .tar.Z (compressed tar) file. + +Fri Jan 8 22:31:41 1993 Paul Eggert (eggert@twinsun.com) + + * side.c (print_half_line): When the input position falls + outside the column, do not output a tab even if the output + position still falls within the column. + +Mon Dec 21 13:54:36 1992 David J. MacKenzie (djm@kropotkin.gnu.ai.mit.edu) + + * Makefile.in (.c.o): Add -I. + +Fri Dec 18 14:08:20 1992 Paul Eggert (eggert@twinsun.com) + + * configure.in: Add HAVE_FCNTL_H, since system.h uses it. + +Tue Nov 24 10:06:48 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) + + * Makefile.in: Note change from USG to HAVE_STRING_H. + +Mon Nov 23 18:44:00 1992 Paul Eggert (eggert@twinsun.com) + + * io.c (find_and_hash_each_line): When running out of lines, + double the number of allocated lines, instead of just doubling + that number minus the prefix lines. This is more likely to + avoid the need for further memory allocation. + +Wed Nov 18 20:40:28 1992 Paul Eggert (eggert@twinsun.com) + + * dir.c (dir_sort): Plug memory leak: space holding directory + contents was not being reclaimed. Get directory size from + struct file_data for initial guess at memory needed. + Detect errors when reading and closing directory. + (diff_dirs): Pass struct file_data to dir_sort. Finish plugging leak. + * diff.c (compare_files): Pass struct file_data to diff_dirs. + + * io.c (find_and_hash_each_line): Don't assume alloc_lines is + nonzero when allocating more lines. + +Thu Nov 12 16:02:18 1992 Paul Eggert (eggert@twinsun.com) + + * diff.c (main): Add `-U lines' as an alias for `--unified=lines'. + + * diff3.c (usage): Add third --label option in example. + + * util.c (analyze_hunk): Fix test for ignoring blank lines. + + * configure.in, system.h: Avoid USG; use HAVE_TIME_H etc. instead. + +Mon Nov 9 05:13:25 1992 Paul Eggert (eggert@twinsun.com) + + * diff3.c (main, usage): Add -A or --show-all. + -m now defaults to -A, not -E. Allow up to three -L options. + (output_diff3_edscript, output_diff3_merge): + Remove spurious differences between these two functions. + Output ||||||| for -A. Distinguish between conflicts and overlaps. + (dotlines, undotlines): New functions that output `Ns', not `N,Ns'. + (output_diff3_edscript, output_diff3_merge): Use them. + + * io.c (find_identical_ends): shift_boundaries needs an extra + identical line at the end, not at the beginning. + + * sdiff.c (edit): execvp wants char **, not const char **. + +Mon Oct 19 04:39:32 1992 Paul Eggert (eggert@twinsun.com) + + * context.c (print_context_script, find_function): Context + line numbers start with - file->prefix_lines, not 0. + + * io.c (binary_file_p): Undo last change; it was a library bug. + +Sun Oct 18 00:17:29 1992 Richard Stallman (rms@mole.gnu.ai.mit.edu) + + * io.c (binary_file_p): Consider empty file as non-binary. + +Mon Oct 5 05:18:46 1992 Paul Eggert (eggert@twinsun.com) + + * diff3.c (main, make_3way_diff, using_to_diff3_block): Don't + report bogus differences (for one of -mexEX3) just because the + file0-file1 diffs don't line up with the file0-file2 diffs. + (This is entirely possible since we don't use diff's -n + option.) Always compare file1 to file2, so that diff3 sees + those changes directly. Typically, file2 is now the common + file, not file0. + (output_diff3_merge): The input file is file 0, not the common file. + + (FC, FO): New macros; they replace FILE1, FILE0 for two-way diffs, + to distinguish them from three-way diffs. + + * diff3.c (using_to_diff3_block): Fold repeated code into loops. + + * diff3.c (make_3way_diff, process_diff): Have the *_end + variable point to the next field to be changed, not to the last + object allocated; this saves an if-then-else. + + * diff3.c (process_diff): Use D_NUMLINES instead of its definiens. + + * diff3.c: Make fns and vars static unless they must be external. + +Wed Sep 30 09:21:59 1992 Paul Eggert (eggert@twinsun.com) + + * analyze.c (diff_2_files): OUTPUT_IFDEF is now robust. + * diff.h (ROBUST_OUTPUT_STYLE): Likewise. + (default_line_format): Remove. All refs removed. + + * ifdef.c (print_ifdef_lines): Add %L. Optimize %l\n even if user + specified it, as opposed to its being the default. + +Tue Sep 29 19:01:28 1992 Paul Eggert (eggert@twinsun.com) + + * diff.c (longopts, main): --{old,new,unchanged,changed}--group-format + are new options, so that -D is no longer overloaded. Set + no_diff_means_no_output if --unchanged-{line,group}-format allows it. + * diff.h (enum line_class): New type. + (group_format, line_format): Use it to regularize option flags. + All refs changed. + + * ifdef.c (format_ifdef, print_ifdef_lines): %n is no longer a format. + +Mon Sep 28 04:51:42 1992 Paul Eggert (eggert@twinsun.com) + + * diff.c (main, usage): Replace --line-prefix with the more general + --{old,new,unchanged}-line-format options. + * ifdef.c (format_ifdef, print_ifdef_lines): Likewise. + * diff.h (line_format): Renamed from line_prefix. All refs changed. + * diff.h, ifdef.c (default_line_format): New variable. + * util.c (output_1_line): New function. + (print_1_line): Use it. + + * ifdef.c: (format_ifdef, print_ifdef_lines): Add %0 format. + +Sun Sep 27 05:38:13 1992 Paul Eggert (eggert@twinsun.com) + + * diff.c (main): Add -E or --line-prefix option. Add -D'=xxx' + for common lines. Change default -D< format from copy of -D> + format to to -D<%<; similarly for default -D> format. + * diff.h (common_format, line_prefix): New variables. + * ifdef.c (format_ifdef): New function. + (print_ifdef_script, print_ifdef_hunk, print_ifdef_lines): + Use it for -D'=xxx', -E. + + * context.c (find_hunk): Glue together two non-ignorable changes that + are exactly CONTEXT * 2 lines apart. This shortens output, removes + a behavioral discontinuity at CONTEXT = 0, and is more compatible + with traditional diff. + + * io.c (find_identical_ends): Slurp stdin at most once. + + * util.c (print_line_line): line_flag is const char *. + +Thu Sep 24 15:18:07 1992 Paul Eggert (eggert@twinsun.com) + + * ifdef.c (print_ifdef_lines): New function, which fwrites a sequence + of lines all at once for speed. + (print_ifdef_script, print_ifdef_hunk): Use it. + +Thu Sep 24 05:54:14 1992 Paul Eggert (eggert@twinsun.com) + + * diff.c (main): Support new -D options for if-then-else formats. + (specify_format): New function. + * diff.h (ifndef_format, ifdef_format, ifnelse_format): New variables. + * ifdef.c (print_ifdef_hunk): Use the new variables instead of + a hardwired format. + + * side.c (print_1sdiff_line): Represent incomplete lines on output. + (print_sdiff_script): Likewise. Don't print 'q' at end, + since that doesn't work with incomplete lines. + * sdiff.c (interact): Don't assume diff output ends with 'q' line. + * diff.h (ROBUST_OUTPUT_STYLE): OUTPUT_SDIFF is now robust. + + * sdiff.c (lf_copy, lf_snarf): Use memchr instead of index, + to avoid dumping core when files contain null characters. + (memchr): New function (if memchr is missing). + + * io.c (sip): New arg SKIP_TEST to skip test for binary file. + (read_files): Don't bother testing second file if first is binary. + +Thu Sep 17 21:17:49 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * system.h [!USG && !_POSIX_VERSION]: Protect from conflicting + prototype for wait in sys/wait.h. + +Wed Sep 16 12:32:18 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * Makefile.in: Include binprefix in -DDIFF_PROGRAM. + +Tue Sep 15 14:27:25 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * Version 2.0. + +Sat Sep 12 01:31:19 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * util.c, diff.h, system.h [!HAVE_MEMCHR]: Don't use void * + and const when declaring memchr replacement. Declare memchr + if !STDC_HEADERS && !USG. + +Thu Sep 10 15:17:32 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * Makefile.in (uninstall): New target. + + * diff.c (excluded_filename): Use fnmatch, not wildmat. + (usage): Document -x, -X, --exclude, --exclude-from. + Makefile.in: Use fnmatch.c, not wildmat.c. + +Sun Sep 6 23:46:25 1992 Paul Eggert (eggert@twinsun.com) + + * configure.in: Add HAVE_MEMCHR. + * diff.h, util.c: Use it instead of MEMCHR_MISSING. + +Sun Sep 6 07:25:49 1992 Paul Eggert (eggert@twinsun.com) + + * diff.h: (struct line_def): Replace this 3-word struct with char *. + This uses less memory, particularly for large files with short lines. + (struct file_data): New member linbuf_base counts number of lines + in common prefix that are not recorded in linbuf; + this uses less memory if files are identical or differ only at end. + New member buffered_lines counts possibly differing lines. + New member valid_lines counts valid data. + New member alloc_lines - linbuf_base replaces old linbufsize. + linbuf[0] now always points at first differing line. + Remove unused members ltran, suffix_lines. + Add const where appropriate. + (Is_space): New macro, for consistent definition of `white space'. + (excluded_filename, memchr, sip, slurp): New declarations. + * ed.c (print_ed_hunk): Adjust to diff.h's struct changes. + * context.c (pr_context_hunk): Likewise. + * ifdef.c (print_ifdef_script): Likewise. + * side.c (print_sdiff_script, print_half_line): Likewise. + * util.c (analyze_hunk, line_cmp, print_1_line): Likewise. + + * analyze.c (shift_boundaries): Remove unneeded variable `end' and + unnecessary comparisons of `preceding' and `other_preceding' against 0. + (diff_2_files): When comparing files byte-by-byte for equality, + don't slurp them all in at once; just compare them a buffer at a time. + This can win big if they differ early on. + Move some code to compare_files to enable this change. + Use only one buffer for stdin with `diff - -'. + (discard_confusing_lines, diff_2_files): Coalesce malloc/free calls. + (build_script): Remove obsolete OUTPUT_RCS code. + + * diff.c (add_exclude, add_exclude_file, excluded_filename): New fns. + (main): Use them for the new --exclude and --exclude-from options. + (compare_files): Don't open a file unless it must be read. + Treat `diff file file' and `diff file dir' similarly. + Move some code here from diff_2_files to enable this. + Simplify file vs dir warning. + + * dir.c (dir_sort): Support new --exclude* options. + + * io.c (struct equivclass): Put hash code and line length here instead + of struct line_def, so that they can be shared. + (find_and_hash_each_line): Compute equivalence class as we go, + instead of doing it in a separate pass; this thrashes memory less. + Make buckets realloc-able, since we can't preallocate them. + Record one more line start than there are lines, so that we can compute + any line's length by subtracting its start from the next line's, + instead of storing the length explicitly. This saves memory. + Move prefix-handling code to find_identical_ends; + this wins with large prefixes. + Use Is_space, not is_space, for consistent treatment of white space. + (prepare_text_end): New function. + (find_identical_ends): Move slurping here, so it's only done when + needed. Work even if the buffers are the same (because of `diff - -'). + Compare prefixes a word at a time for speed. + (find_equiv_class): Delete; now done by find_and_hash_each_line. + (read_files): Don't slurp unless needed. + find_equiv_class's work is now folded into find_and_hash_each_line. + Don't copy stdin buffer if `diff - -'. + Check for running out of primes. + (sip, slurp): Split first part of `slurp' into another function `sip'. + `sip' sets things up and perhaps reads the first ST_BLKSIZE buffer to + see whether the file is binary; `slurp' now just finishes the job. + This lets diff_2_files compare binary files lazily. + Allocate a one-word sentinel to allow word-at-a-time prefix comparison. + Count prefix lines only if needed, only count the first file's prefix. + Don't bother to count suffix lines; it's never needed. + Set up linbuf[0] to point at first differing line. + (binary_file_p): Change test for binary files: + if it has a null byte in its first buffer, it's binary. + (primes): Add more primes. + + * util.c (line_cmp): Use bcmp for speed. + Use Is_space, not is_space, for consistent treatment of white space. + (translate_line_number): Internal line numbers now count from 0 + starting after the prefix. + (memchr): New function (if memchr is missing). + + * Makefile.in: Document HAVE_ST_BLKSIZE. Link with wildmat.o. + * system.h (STAT_BLOCKSIZE): New macro based on HAVE_ST_BLKSIZE. + * configure.in: Add AC_ST_BLKSIZE. + * wildmat.c: New file. + +Fri Sep 4 01:28:51 1992 Richard Stallman (rms@mole.gnu.ai.mit.edu) + + * sdiff.c (xmalloc): Renamed from ck_malloc. Callers changed. + +Thu Sep 3 15:28:59 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * diff.h: Don't declare free, index, rindex. + +Tue Aug 11 22:18:06 1992 John Gilmore (gnu at cygnus.com) + + * io.c (binary_file_p): Use heuristic to avoid declaring info + files as binary files. Allow about 1.5% non-printing + characters (in info's case, ^_). + +Tue Jul 7 01:09:26 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * diff.h: Replace function_regexp and ignore_regexp with lists + of compiled regexps. + * analyze.c, context.c, util.c: Test whether the lists, not + the old variables, are empty. + * util.c (analyze_hunk), context.c (find_function): Compare + lines with the lists of regexps. + * diff.c (add_regexp): New function. + (main): Use it. + + * diff3: Add -v --version option. + * Makefile.in: Link with version.o. + + * system.h: New file. + * diff.h, cmp.c, diff3.c, sdiff.c: Use it. + + * diff.h, diff3.c: Include string.h or strings.h, as appropriate. + Declare malloc and realloc. + + * diff3.c (perror_with_exit): Include program name in message. + + * diff3.c: Lowercase error messages for GNU standards. + + * sdiff.c [USG || STDC_HEADERS]: Define bcopy in terms of memcpy. + + * sdiff.c: Use the version number from version.c. + * Makefile.in: Link with version.o. + + * cmp.c error.c xmalloc.c: New files from textutils. + * Makefile.in: Add rules for them. + + * diff.c (longopts): --unidirectional-new-file is like -P, not -N. + Rename --file-label to --label (leave old name, but undocumented). + + * sdiff.c, diff.c (usage): Condense messages and fix some errors. + + * diff3.c (main, usage): Add long-named options. + +Fri Jul 3 14:31:18 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * diff.h, diff3.c, sdiff.c: Change FOO_MISSING macros to HAVE_FOO. + +Thu Jun 25 16:59:47 1992 David J. MacKenzie (djm@apple-gunkies.gnu.ai.mit.edu) + + * diff.c: --reversed-ed -> --forward-ed. + +Wed Feb 26 12:17:32 1992 Paul Eggert (eggert@yata.uucp) + + * analyze.c, diff.c, diff.h, io.c: For -y, compare even if same file. + +Fri Feb 14 22:46:38 1992 Richard Stallman (rms@mole.gnu.ai.mit.edu) + + * io.c, diff3.c, analyze.c: Add extra parentheses. + +Sun Feb 9 00:22:42 1992 Richard Stallman (rms@mole.gnu.ai.mit.edu) + + * diff.h (unidirectional_new_file_flag): New variable. + * diff.c (main): Set that for -P. + (compare_files): Support -P, somewhat like -N. + (longopts): Support long name for -P. + +Sat Jan 4 20:10:34 1992 Paul Eggert (eggert at yata.uucp) + + * Makefile.in: Distribute diff.info-* too. + + * README, sdiff.c: version number now matches version.c. + + * configure: Fix and document vfork test. + + * ifdef.c: Don't dump core if `diff -Dx f f'. + +Mon Dec 23 23:36:08 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * diff.h, diff3.c, sdiff.c: Change POSIX ifdefs to + HAVE_UNISTD_H and _POSIX_VERSION. + +Wed Dec 18 17:00:31 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * Makefile.in (srcs): Add sdiff.c. + (tapefiles): Add diff.texi and diff.info. + + * diff.h, diff3.c, sdiff.c: Use HAVE_VFORK_H instead of + VFORK_HEADER and VFORK_WORKS. + +Tue Dec 17 00:02:59 1991 Paul Eggert (eggert at yata.uucp) + + * Makefile.in (all): Add diff.info, sdiff. + + * configure, diff.c, sdiff.c: + Prefix long options with `--', not `+'. + * diff.c: Regularize option names. + + * configure: Fix check for vfork. + * configure, diff.c, diff.h, diff3.c, sdiff.c: + Use Posix definitions when possible. + + * context.c: Align context with tab if -T is given. Tune. + * diff.c, diff.h, side.c: Calculate column widths so that tabs line up. + * io.c: Add distinction between white space and printing chars. + * side.c: Don't expand tabs unless -t is given. + * side.c, util.c: Tab expansion now knows about '\b', '\f', '\r', '\v'. + * util.c: -w skips all white space. Remove lint. Tune. + + * sdiff.c: Support many more diff options, e.g. `-', `sdiff file dir'. + Ignore interrupts while the subsidiary editor is in control. + Clean up temporary file and kill subsidiary diff if interrupted. + Ensure subsidiary diff doesn't ignore SIGPIPE. + Don't get confused while waiting for two subprocesses. + Don't let buffers overflow. Check for I/O errors. + Convert to GNU style. Tune. + + * sdiff.c, util.c: Don't lose errno. + Don't confuse sdiff with messages like `Binary files differ'. + * sdiff.c, side.c: Don't assume that common lines are identical. + Simplify --sdiff-merge-assist format. + +Mon Sep 16 16:42:01 1991 Tom Lord (lord at churchy.gnu.ai.mit.edu) + + * Makefile.in, sdiff.c: introduced sdiff front end to diff. + + * Makefile.in, analyze.c, diff.c, diff.h, io.c, side.c: Added + sdiff-style output format to diff. + +Mon Aug 26 16:44:55 1991 David J. MacKenzie (djm at pogo.gnu.ai.mit.edu) + + * Makefile.in, configure: Only put $< in Makefile if using VPATH, + because older makes don't understand it. + +Fri Aug 2 12:22:30 1991 David J. MacKenzie (djm at apple-gunkies) + + * configure: Create config.status. Remove it and Makefile if + interrupted while creating them. + +Thu Aug 1 22:24:31 1991 David J. MacKenzie (djm at apple-gunkies) + + * configure: Check for +srcdir etc. arg and look for + Makefile.in in that directory. Set VPATH if srcdir is not `.'. + * Makefile.in: Get rid of $(archpfx). + +Tue Jul 30 21:28:44 1991 Richard Stallman (rms at mole.gnu.ai.mit.edu) + + * Makefile.in (prefix): Renamed from DESTDIR. + +Wed Jul 24 23:08:56 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * diff.h, diff3.c: Rearrange ifdefs to use POSIX, + STDC_HEADERS, VFORK_MISSING, DIRENT. This way it works on + more systems that aren't pure USG or BSD. + Don't not define const if __GNUC__ is defined -- that would + break with -traditional. + * configure: Check for those features. + +Wed Jul 10 01:39:23 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * configure, Makefile.in: $(INSTALLPROG) -> $(INSTALL). + +Sat Jul 6 16:39:04 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) + + * Replace Makefile with configure and Makefile.in. + Update README with current compilation instructions. + +Sat Jul 6 14:03:29 1991 Richard Stallman (rms at mole.gnu.ai.mit.edu) + + * util.c (setup_output): Just save the args for later use. + (begin_output): Do the real work, with the values that were saved. + It's safe to call begin_output more than once. + Print the special headers for context format here. + * analyze.c (diff_2_files): Don't print special headers here. + * context.c (pr_context_hunk, pr_unidiff_hunk): Call begin_output. + * ed.c (print_ed_hunk, print_forward_ed_hunk, print_rcs_hunk): + * normal.c (print_normal_hunk): Likewise. + * ifdef.c (print_ifdef_hunk): Likewise. + * util.c (finish_output): Don't die if begin_output was not called. + +Thu Jun 20 23:10:01 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) + + * Makefile: Add TAGS, distclean, and realclean targets. + Set SHELL. + +Tue Apr 30 13:54:36 1991 Richard Stallman (rms at mole.gnu.ai.mit.edu) + + * diff.h (TRUE, FALSE): Undefine these before defining. + +Thu Mar 14 18:27:27 1991 Richard Stallman (rms@mole.ai.mit.edu) + + * Makefile (objs): Include $(ALLOCA). + +Sat Mar 9 22:34:03 1991 Richard Stallman (rms at mole.ai.mit.edu) + + * diff.h: Include regex.h. + +Thu Feb 28 18:59:53 1991 Richard Stallman (rms at mole.ai.mit.edu) + + * Makefile (diff3): Link with GNU getopt. + +Sat Feb 23 12:49:43 1991 Richard Stallman (rms at mole.ai.mit.edu) + + * io.c (find_equiv_class): Make hash code unsigned before mod. + + * diff.h (files): Add EXTERN. + +Sun Jan 13 21:33:01 1991 Richard Stallman (rms at mole.ai.mit.edu) + + * diff.c: +print option renamed +paginate. Remove +all-text. + +Mon Jan 7 06:18:01 1991 David J. MacKenzie (djm at geech.ai.mit.edu) + + * Makefile (dist): New target, replacing diff.tar and + diff.tar.Z, to encode version number in distribution directory + and tar file names. + +Sun Jan 6 18:42:23 1991 Michael I Bushnell (mib at geech.ai.mit.edu) + + * Version 1.15 released. + + * version.c: Updated from 1.15 alpha to 1.15 + + * context.c (print_context_number_range, + print_unidiff_number_range): Don't print N,M when N=M, print + just N instead. + + * README: Updated for version 1.15. + Makefile: Updated for version 1.15. + + * diff3.c (main): Don't get confused if one of the arguments + is a directory. + + * diff.c (compare_files): Don't get confused if comparing + standard input to a directory; print error instead. + + * analyze.c (diff_2_files), context.c (print_context_header, + print_context_script), diff.c (main), diff.h (enum + output_style): Tread unidiff as an output style in its own + right. This also generates an error when both -u and -c are + given. + + * diff.c (main): Better error messages when regexps are bad. + + * diff.c (compare_files): Don't assume stdin is opened. + + * diff3.c (read_diff): Don't assume things about the order of + descriptor assignment and closes. + + * util.c (setup_output): Don't assume things about the order + of descriptor assignment and closes. + + * diff.c (compare_files): Set a flag so that closes don't + happen more than once. + + * diff.c (main): Don't just flush stdout, do a close. That + way on broken systems we can still get errors. + +Mon Dec 24 16:24:17 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * diff.c (usage): Use = for args of long options. + +Mon Dec 17 18:19:20 1990 Michael I Bushnell (mib at geech.ai.mit.edu) + + * context.c (print_context_label): Labels were interchanged badly. + + * context.c (pr_unidiff_hunk): Changes to deal with files + ending in incomplete lines. + * util.c (print_1_line): Other half of the changes. + +Mon Dec 3 14:23:55 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * diff.c (longopts, usage): unidiff => unified. + +Wed Nov 7 17:13:08 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * analyze.c (diff_2_files): No warnings about newlines for -D. + + * diff.c (pr_unidiff_hunk): Remove ref to output_patch_flag. + +Tue Oct 23 23:19:18 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * diff.c (compare_files): For -D, compare even args are same file. + * analyze.c (diff_2_files): Likewise. + Also, output even if files have no differences. + + * analyze.c (diff_2_files): Print missing newline messages last. + Return 2 if a newline is missing. + Print them even if files end with identical text. + +Mon Oct 22 19:40:09 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * diff.c (usage): Return 2. + +Wed Oct 10 20:54:04 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * diff.c (longopts): Add +new-files. + +Sun Sep 23 22:49:29 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * context.c (print_context_script): Handle unidiff_flag. + (print_context_header): Likewise. + (print_unidiff_number_range, pr_unidiff_hunk): New functions. + * diff.c (longopts): Add element for +unidiff. + (main): Handle +unidiff and -u. + (usage): Mention them. + +Wed Sep 5 16:33:22 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * io.c (find_and_hash_each_line): Deal with missing final newline + after buffering necessary context lines. + +Sat Sep 1 16:32:32 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * io.c (find_identical_ends): ROBUST_OUTPUT_FORMAT test was backward. + +Thu Aug 23 17:17:20 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * diff3.c (WIFEXITED): Undef it if WEXITSTATUS is not defined. + * context.c (find_function): Don't try to return values. + +Wed Aug 22 11:54:39 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * diff.h (O_RDONLY): Define if not defined. + +Tue Aug 21 13:49:26 1990 Richard Stallman (rms at mole.ai.mit.edu) + + * Handle -L option. + * context.c (print_context_label): New function. + (print_context_header): Use that. + * diff.c (main): Recognize the option. + (usage): Updated. + * diff.h (file_label): New variable. + * diff3.c (main): Recognize -L instead of -t. + + * diff3.c (main): Support -m without other option. + + * diff3.c (WEXITSTATUS, WIFEXITED): Define whenever not defined. + + * diff3.c (bcopy, index, rindex): Delete definitions; not used. + (D_LINENUM, D_LINELEN): Likewise. + (struct diff_block): lengths includes newlines. + (struct diff3_block): Likewise. + (always_text, merge): New variables. + (read_diff): Return address of end, not size read. Calls changed. + Pass -a to diff if given to diff3. + current_chunk_size now an int. Detect error in `pipe'. + Check for incomplete line of output here. + (scan_diff_line): Don't make scan_ptr + 2 before knowing it is valid. + No need to check validity of diff output here. + Include newline in length of line. + (main): Compute rev_mapping here. Handle -a and -m. + Error message if excess -t operands. Error for incompatible options. + Error if `-' given more than once. + Fix error storing in tag_strings. + (output_diff3): REV_MAPPING is now an arg. Call changed. + Change syntax of "missing newline" message. + Expect length of line to include newline. + (output_diff3_edscript): Return just 0 or 1. + REV_MAPPING is now an arg. Call changed. + (output_diff3_merge): New function. + (process_diff): Better error message for bad diff format. + (fatal, perror_with_exit): Return status 2. + + * analyze.c (diff_2_files): Report missing newline in either + or both files, if not robust output style. + + * util.c (setup_output): Detect error from pipe. + No need to close stdin. + + * util.c (print_1_line): Change format of missing-newline msg. + Change if statements to switch. + + * io.c (slurp): Don't mention differences in final newline if -B. + + * io.c (binary_file_p): Use ISO char set as criterion, not ASCII. + + * io.c (find_identical_ends): Increase value of BEG0 by 1. + Other changes in backwards scan to avoid decrementing pointers + before start of array, and set LINES properly. + + * diff.h (ROBUST_OUTPUT_STYLE): New macro. + * io.c (find_identical_ends, find_and_hash_each_line): Use that macro. + + * diff.h (dup2): Don't define if XENIX. + + * diff.c (main): Check for write error at end. + + * context.c (find_function): Don't return a value. + Use argument FILE rather than global files. + + * analyze.c: Add external function declarations. + * analyze.c (build_script): Turn off explicit check for final newline. + + * analyze.c (discard_confusing_lines): Make integers unsigned. + +Tue Jul 31 21:37:16 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * io.c (find_and_hash_each_line): Correct the criterion + for leaving out the newline from the end of the line. + +Tue May 29 21:28:16 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * dir.c (diff_dirs): Free things only if nonzero. + +Mon Apr 16 18:31:05 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff.h (NDIR_IN_SYS): New macro controls location of ndir.h. + + * diff3.c (xmalloc, xrealloc): Don't die if size == 0 returns 0. + +Sun Mar 25 15:58:42 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * analyze.c (discard_confusing_lines): + `many' wasn't being used; use it. + Cancelling provisionals near start of run must handle already + cancelled provisionals. + Cancelling subruns of provisionals was cancelling last nonprovisional. + +Sat Mar 24 14:02:51 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * analyze.c (discard_confusing_lines): + Threshold for line occurring many times scales by square root + of total lines. + Within each run, cancel any long subrun of provisionals. + Don't update `provisional' while cancelling provisionals. + In big outer loop, handle provisional and nonprovisional separately. + +Thu Mar 22 16:35:33 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * analyze.c (discard_confusing_lines): + The first loops to discard provisionals from ends failed to step. + In second such loops, keep discarding all consecutive provisionals. + Increase threshold for stopping discarding, and also check for + consecutive nondiscardables as separate threshold. + +Fri Mar 16 00:33:08 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff3.c (read_diff): Pass -- as first arg to diff. + + * diff3.c: Include wait.h or define equivalent macros. + (read_diff): Don't use stdio printing error in the inferior. + Remember the pid and wait for it. Report failing status. + Report failure of vfork. + +Sun Mar 11 17:10:32 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff3.c (main): Accept -t options and pass to output_diff3_edscript. + (usage): Mention -t. + (read_diff): Use vfork. + (vfork): Don't use it on Sparc. + + * diff.h (vfork): Don't use it on Sparc. + +Tue Mar 6 22:37:20 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff3.c (dup2): Don't define on Xenix. + + * Makefile: Comments for Xenix. + +Thu Mar 1 17:19:23 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * analyze.c (diff_2_files): `message' requires three args. + +Fri Feb 23 10:56:50 1990 David J. MacKenzie (djm at albert.ai.mit.edu) + + * diff.h, util.c, diff3.c: Change 'void *' to 'VOID *', with + VOID defined as void if __STDC__, char if not. + +Sun Feb 18 20:31:58 1990 David J. MacKenzie (djm at albert.ai.mit.edu) + + * Makefile: Add rules for getopt.c, getopt1.c, getopt.h. + + * getopt.c, getopt.h, getopt1.c: New files. + + * main.c (main, usage): Add long options. + + * analyze.c (shift_boundaries): Remove unused var 'j_end'. + +Thu Feb 8 02:43:16 1990 Jim Kingdon (kingdon at pogo.ai.mit.edu) + + * GNUmakefile: include ../Makerules before Makefile. + +Fri Feb 2 23:21:38 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * analyze.c (diif_2_files): If -B or -I, don't return 1 + if all changes were ignored. + +Wed Jan 24 20:43:57 1990 Richard Stallman (rms at albert.ai.mit.edu) + + * diff3.c (fatal): Output to stderr. + +Thu Jan 11 00:25:56 1990 David J. MacKenzie (djm at hobbes.ai.mit.edu) + + * diff.c (usage): Mention -v. + +Wed Jan 10 16:06:38 1990 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff3.c (output_diff3_edscript): Return number of overlaps. + (main): If have overlaps, exit with status 1. + +Sun Dec 24 10:29:20 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * io.c (find_equiv_class): Fix typo that came from changing init of B + to an assigment. + + * version.c: New file. + * diff.c (main): -v prints version number. + + * io.c (binary_file_p): Null char implies binary file. + +Fri Nov 17 23:44:55 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * util.c (print_1_line): Fix off by 1 error. + +Thu Nov 16 13:51:10 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * util.c (xcalloc): Function deleted. + + * io.c (slurp): Null-terminate the buffer. + + * io.c (read_files): Delete unused vars. + + * io.c (find_equiv_class): Don't index by N if too low. + + * dir.c (dir_sort): Delete the extra declaration of compare_names. + + * diff.h: Don't declare xcalloc. Declare some other functions. + + * analyze.c (shift_boundaries): + Test for END at end of range before indexing by it. + Fix typo `preceeding' in var names. + +Sat Nov 11 14:04:16 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff3.c (using_to_diff3_block): Delete unused vars. + (make_3way_diff, process_diff_control, read_diff, output_diff3): Likewise. + +Mon Nov 6 18:15:50 EST 1989 Jay Fenlason (hack@ai.mit.edu) + + * README Fix typo. + +Fri Nov 3 15:27:47 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff.c (usage): Mention -D. + + * ifdef.c (print_ifdef_hunk): Write comments on #else and #endif. + +Sun Oct 29 16:41:07 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff.c (compare_files): Don't fflush for identical files. + +Wed Oct 25 17:57:12 1989 Randy Smith (randy at apple-gunkies.ai.mit.edu) + + * diff3.c (using_to_diff3_block): When defaulting lines from + FILE0, only copy up to just under the *lowest* line mentioned + in the next diff. + + * diff3.c (fatal): Add \n to error messages. + +Wed Oct 25 15:05:49 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * Makefile (tapefiles): Add ChangeLog. + +Tue Oct 3 00:51:17 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff3.c (process_diff, create_diff3_block): Init ->next field. + +Fri Sep 29 08:16:45 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * util.c (line_cmp): Alter end char of line 2, not line 1. + +Wed Sep 20 00:12:37 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * Makefile (diff.tar): Expect ln to fail on some files; + copy them with cp. + +Mon Sep 18 02:54:29 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * Handle -D option: + * io.c (find_and_hash_each_line): Keep all lines of 1st file. + * diff.c (main): Handle -D option. + (compare_files): Reject -D if files spec'd are directories. + * analyze.c (diff_2_files): Handle OUTPUT_IFDEF case. + +Fri Sep 1 20:15:50 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff.c (option_list): Rename arg VECTOR as OPTIONVEC. + +Mon Aug 28 17:58:27 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff.c (compare_files): Clear entire inf[i].stat. + +Wed Aug 23 17:48:47 1989 Richard Stallman (rms at apple-gunkies.ai.mit.edu) + + * io.c (find_identical_ends): Sign was backward + determining where to bound the scan for the suffix. + +Wed Aug 16 12:49:16 1989 Richard Stallman (rms at hobbes.ai.mit.edu) + + * analyze.c (diff_2_files): If -q, treat all files as binary. + * diff.c (main): Detect -q, record in no_details_flag. + +Sun Jul 30 23:12:00 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff.c (usage): New function. + (main): Call it. + +Wed Jul 26 02:02:19 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff.c (main): Make -C imply -c. + +Thu Jul 20 17:57:51 1989 Chris Hanson (cph at kleph) + + * io.c (find_and_hash_each_line): Bug fix in context handling, + introduced by last change. + +Fri Jul 14 17:39:20 1989 Chris Hanson (cph at kleph) + + * analyze.c: To make RCS work correctly on files that don't + necessarily end in newline, introduce some changes that cause + diffs to be sensitive to missing final newline. Because + non-RCS modes don't want to be affected by these changes, they + are conditional on `output_style == OUTPUT_RCS'. + (diff_2_files) [OUTPUT_RCS]: Suppress the "File X missing + newline" message. + (build_script) [OUTPUT_RCS]: Cause the last line to compare as + different if exactly one of the files is missing its final + newline. + + * io.c (find_and_hash_each_line): Bug fix in + ignore_space_change mode. Change line's length to include the + newline. For OUTPUT_RCS, decrement last line's length if + there is no final newline. + (find_identical_ends) [OUTPUT_RCS]: If one of the files is + missing a final newline, make sure it's not included in either + the prefix or suffix. + + * util.c (print_1_line): Change line output routine to account + for line length including the newline. + +Tue Jun 27 02:35:28 1989 Roland McGrath (roland at hobbes.ai.mit.edu) + + * Makefile: Inserted $(archpfx) where appropriate. + +Wed May 17 20:18:43 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff3.c [USG]: Include fcntl.h. + + * diff.h [USG]: New compilation flags HAVE_NDIR, HAVE_DIRECT. + +Wed Apr 26 15:35:57 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * dir.c (diff_dirs): Two new args, NONEX1 and NONEX2, say to pretend + nonex dirs are empty. + (dir_sort): New arg NONEX, likewise. + * diff.c (compare_files): Pass those args. + Sometimes call diff_dirs if subdir exists in just one place. + +Wed Apr 12 01:10:27 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * io.c (find_identical_ends): Set END0 *after* last char + during backward scan for suffix. + +Sat Apr 8 15:49:49 1989 Randall Smith (randy at apple-gunkies.ai.mit.edu) + + * diff3.c (using_to_diff3_block): Now find high marks in files 1 + and 2 through mapping off of the last difference instead of the + first. + + * diff3.c: Many trivial changes to spelling inside comments. + +Fri Feb 24 12:38:03 1989 Randall Smith (randy at gluteus.ai.mit.edu) + + * util.c, normal.c, io.c, ed.c, dir.c, diff.h, diff.c, context.c, + analyze.c, Makefile: Changed copyright header to conform with new + GNU General Public license. + * diff3.c: Changed copyright header to conform with new GNU + General Public license. + * COPYING: Made a hard link to /gp/rms/COPYING. + +Fri Feb 24 10:01:58 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * io.c (slurp): Leave 2 chars space at end of buffer, not one. + (find_identical_ends): Special case if either file is empty; + don't try to make a sentinel since could crash. + +Wed Feb 15 14:24:48 1989 Jay Fenlason (hack at apple-gunkies.ai.mit.edu) + + * diff3.c (message) Re-wrote routine to avoid using alloca() + +Wed Feb 15 06:19:14 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * io.c (find_identical_ends): Delete the variable `bytes'. + +Sun Feb 12 11:50:36 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * io.c (slurp): ->bufsize is nominal amount we have room for; + add room for sentinel when calling xmalloc or xrealloc. + + * io.c (find_identical_ends): Do need overrun check in finding suffix. + +Fri Feb 10 01:28:15 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff.c (main): -C now takes arg to specify context length. + Now -p to show C function name--Damned IEEE! + Fatal error if context length spec'd twice. + + * ed.c (print_ed_hunk): Now special treatment only for lines containing + precisely a dot and nothing else. Output `..', end the insert, + substitute that one line, then resume the insert if nec. + + * io.c (find_and_hash_lines): When backing up over starting context, + don't move past buffer-beg. + + * io.c (find_identical_ends): Use sentinels to make the loops faster. + If files are identical, skip the 2nd loop and return quickly. + (slurp): Leave 1 char extra space after each buffer. + + * analyze.c (diff_2_files): Mention difference in final newlines. + +Wed Jan 25 22:44:44 1989 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * dir.c (diff_dirs): Use * when calling fcn ptr variable. + +Sat Dec 17 14:12:06 1988 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * Makefile: New vars INSTALL and LIBS used in some rules; + provide default defns plus commented-put defns for sysV. + +Thu Nov 17 16:42:53 1988 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * dir.c (dir_sort): Open-trouble not fatal; just say # files is -1. + (diff_dirs): If dir_sort does that, give up and return 2. + + * diff.c (compare_files): Don't open directories. + Don't close them specially either. + Cross-propagate inf[i].dir_p sooner. + +Sun Nov 13 11:19:36 1988 Richard Stallman (rms at sugar-bombs.ai.mit.edu) + + * diff.h: Declare index, rindex. + + * diff.c (compare_files): If comparing foodir with b/f, + use foodir/f, not foodir/b/f. + + * diff.c (compare_files): Don't print "are identical" msg for 2 dirs. + Status now 1 if one file is a dir and the other isn't, etc. + +Thu Nov 3 16:30:24 1988 Randall Smith (randy at gluteus.ai.mit.edu) + + * Makefile: Added a define for diff3 to define DIFF_PROGRAM. + + * util.c: Added hack to make sure that perror was not called with + a null pointer. + + * diff.c: Changed S_IFDIR to S_IFMT in masking type of file bits + out. + + * diff3.c: Included USG compatibility defines. + + * diff.h: Moved sys/file.h into #else USG section (not needed or + wanted on System V). + + * ed.c, analyze.c, context.c: Shortened names to 12 characters for + the sake of System V (too simple not to do). + +Local Variables: +mode: indented-text +left-margin: 8 +version-control: never +End: diff --git a/sys/src/ape/cmd/diff/FREEBSD-upgrade b/sys/src/ape/cmd/diff/FREEBSD-upgrade new file mode 100755 index 000000000..bc07d131d --- /dev/null +++ b/sys/src/ape/cmd/diff/FREEBSD-upgrade @@ -0,0 +1,24 @@ +Import of GNU diff 2.7 + +Original source available as ftp://prep.ai.mit.edu/pub/gnu/diffutils-2.7.tar.gz + +The following files and directories were removed for this import: + +Makefile.in +INSTALL +alloca.c +cmp.c +diff.info +diff.info-1 +diff.info-2 +diff.info-3 +diff.info-4 +error.c +fnmatch.c +fnmatch.h +memchr.c +mkinstalldirs +regex.c +regex.h +texinfo.tex +waitpid.c diff --git a/sys/src/ape/cmd/diff/NEWS b/sys/src/ape/cmd/diff/NEWS new file mode 100755 index 000000000..dcde1227d --- /dev/null +++ b/sys/src/ape/cmd/diff/NEWS @@ -0,0 +1,126 @@ +User-visible changes in version 2.7: + +* New diff option: --binary (useful only on non-Posix hosts) +* diff -b and -w now ignore line incompleteness; -B no longer does this. +* cmp -c now uses locale to decide which output characters to quote. +* Help and version messages are reorganized. + + +User-visible changes in version 2.6: + +* New cmp, diff, diff3, sdiff option: --help +* A new heuristic for diff greatly reduces the time needed to compare + large input files that contain many differences. +* Partly as a result, GNU diff's output is not exactly the same as before. + Usually it is a bit smaller, but sometimes it is a bit larger. + + +User-visible changes in version 2.5: + +* New cmp option: -v --version + + +User-visible changes in version 2.4: + +* New cmp option: --ignore-initial=BYTES +* New diff3 option: -T --initial-tab +* New diff option: --line-format=FORMAT +* New diff group format specifications: + <PRINTF_SPEC>[eflmnEFLMN] + A printf spec followed by one of the following letters + causes the integer corresponding to that letter to be + printed according to the printf specification. + E.g. `%5df' prints the number of the first line in the + group in the old file using the "%5d" format. + e: line number just before the group in old file; equals f - 1 + f: first line number in group in the old file + l: last line number in group in the old file + m: line number just after the group in old file; equals l + 1 + n: number of lines in group in the old file; equals l - f + 1 + E, F, L, M, N: likewise, for lines in the new file + %(A=B?T:E) + If A equals B then T else E. A and B are each either a decimal + constant or a single letter interpreted as above. T and E are + arbitrary format strings. This format spec is equivalent to T if + A's value equals B's; otherwise it is equivalent to E. For + example, `%(N=0?no:%dN) line%(N=1?:s)' is equivalent to `no lines' + if N (the number of lines in the group in the the new file) is 0, + to `1 line' if N is 1, and to `%dN lines' otherwise. + %c'C' + where C is a single character, stands for the character C. C may not + be a backslash or an apostrophe. E.g. %c':' stands for a colon. + %c'\O' + where O is a string of 1, 2, or 3 octal digits, stands for the + character with octal code O. E.g. %c'\0' stands for a null character. +* New diff line format specifications: + <PRINTF_SPEC>n + The line number, printed with <PRINTF_SPEC>. + E.g. `%5dn' prints the line number with a "%5d" format. + %c'C' + %c'\O' + The character C, or with octal code O, as above. +* Supported <PRINTF_SPEC>s have the same meaning as with printf, but must + match the extended regular expression %-*[0-9]*(\.[0-9]*)?[doxX]. +* The format spec %0 introduced in version 2.1 has been removed, since it + is incompatible with printf specs like %02d. To represent a null char, + use %c'\0' instead. +* cmp and diff now conform to Posix.2 (ISO/IEC 9945-2:1993) + if the underlying system conforms to Posix: + - Some messages' wordings are changed in minor ways. + - ``White space'' is now whatever C's `isspace' says it is. + - When comparing directories, if `diff' finds a file that is not a regular + file or a directory, it reports the file's type instead of diffing it. + (As usual, it follows symbolic links first.) + - When signaled, sdiff exits with the signal's status, not with status 2. +* Now portable to hosts where int, long, pointer, etc. are not all the same + size. +* `cmp - -' now works like `diff - -'. + + +User-visible changes in version 2.3: + +* New diff option: --horizon-lines=lines + + +User-visible changes in version 2.1: + +* New diff options: + --{old,new,unchanged}-line-format='format' + --{old,new,unchanged,changed}-group-format='format' + -U +* New diff3 option: + -A --show-all +* diff3 -m now defaults to -A, not -E. +* diff3 now takes up to three -L or --label options, not just two. + If just two options are given, they refer to the first two input files, + not the first and third input files. +* sdiff and diff -y handle incomplete lines. + + +User-visible changes in version 2.0: + +* Add sdiff and cmp programs. +* Add Texinfo documentation. +* Add configure script. +* Improve diff performance. +* New diff options: +-x --exclude +-X --exclude-from +-P --unidirectional-new-file +-W --width +-y --side-by-side +--left-column +--sdiff-merge-assist +--suppress-common-lines +* diff options renamed: +--label renamed from --file-label +--forward-ed renamed from --reversed-ed +--paginate renamed from --print +--entire-new-file renamed from --entire-new-files +--new-file renamed from --new-files +--all-text removed +* New diff3 options: +-v --version +* Add long-named equivalents for other diff3 options. +* diff options -F (--show-function-line) and -I (--ignore-matching-lines) + can now be given more than once. diff --git a/sys/src/ape/cmd/diff/README b/sys/src/ape/cmd/diff/README new file mode 100755 index 000000000..bfb99656a --- /dev/null +++ b/sys/src/ape/cmd/diff/README @@ -0,0 +1,9 @@ +This directory contains the GNU diff, diff3, sdiff, and cmp utilities. +Their features are a superset of the Unix features and they are +significantly faster. cmp has been moved here from the GNU textutils. + +See the file COPYING for copying conditions. +See the file diff.texi (or diff.info*) for documentation. +See the file INSTALL for compilation and installation instructions. + +Report bugs to bug-gnu-utils@prep.ai.mit.edu diff --git a/sys/src/ape/cmd/diff/analyze.c b/sys/src/ape/cmd/diff/analyze.c new file mode 100755 index 000000000..556d388c4 --- /dev/null +++ b/sys/src/ape/cmd/diff/analyze.c @@ -0,0 +1,1084 @@ +/* Analyze file differences for GNU DIFF. + Copyright (C) 1988, 1989, 1992, 1993 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* The basic algorithm is described in: + "An O(ND) Difference Algorithm and its Variations", Eugene Myers, + Algorithmica Vol. 1 No. 2, 1986, pp. 251-266; + see especially section 4.2, which describes the variation used below. + Unless the --minimal option is specified, this code uses the TOO_EXPENSIVE + heuristic, by Paul Eggert, to limit the cost to O(N**1.5 log N) + at the price of producing suboptimal output for large inputs with + many differences. + + The basic algorithm was independently discovered as described in: + "Algorithms for Approximate String Matching", E. Ukkonen, + Information and Control Vol. 64, 1985, pp. 100-118. */ + +#include "diff.h" +#include "cmpbuf.h" + +extern int no_discards; + +static int *xvec, *yvec; /* Vectors being compared. */ +static int *fdiag; /* Vector, indexed by diagonal, containing + 1 + the X coordinate of the point furthest + along the given diagonal in the forward + search of the edit matrix. */ +static int *bdiag; /* Vector, indexed by diagonal, containing + the X coordinate of the point furthest + along the given diagonal in the backward + search of the edit matrix. */ +static int too_expensive; /* Edit scripts longer than this are too + expensive to compute. */ + +#define SNAKE_LIMIT 20 /* Snakes bigger than this are considered `big'. */ + +struct partition +{ + int xmid, ymid; /* Midpoints of this partition. */ + int lo_minimal; /* Nonzero if low half will be analyzed minimally. */ + int hi_minimal; /* Likewise for high half. */ +}; + +static int diag PARAMS((int, int, int, int, int, struct partition *)); +static struct change *add_change PARAMS((int, int, int, int, struct change *)); +static struct change *build_reverse_script PARAMS((struct file_data const[])); +static struct change *build_script PARAMS((struct file_data const[])); +static void briefly_report PARAMS((int, struct file_data const[])); +static void compareseq PARAMS((int, int, int, int, int)); +static void discard_confusing_lines PARAMS((struct file_data[])); +static void shift_boundaries PARAMS((struct file_data[])); + +/* Find the midpoint of the shortest edit script for a specified + portion of the two files. + + Scan from the beginnings of the files, and simultaneously from the ends, + doing a breadth-first search through the space of edit-sequence. + When the two searches meet, we have found the midpoint of the shortest + edit sequence. + + If MINIMAL is nonzero, find the minimal edit script regardless + of expense. Otherwise, if the search is too expensive, use + heuristics to stop the search and report a suboptimal answer. + + Set PART->(XMID,YMID) to the midpoint (XMID,YMID). The diagonal number + XMID - YMID equals the number of inserted lines minus the number + of deleted lines (counting only lines before the midpoint). + Return the approximate edit cost; this is the total number of + lines inserted or deleted (counting only lines before the midpoint), + unless a heuristic is used to terminate the search prematurely. + + Set PART->LEFT_MINIMAL to nonzero iff the minimal edit script for the + left half of the partition is known; similarly for PART->RIGHT_MINIMAL. + + This function assumes that the first lines of the specified portions + of the two files do not match, and likewise that the last lines do not + match. The caller must trim matching lines from the beginning and end + of the portions it is going to specify. + + If we return the "wrong" partitions, + the worst this can do is cause suboptimal diff output. + It cannot cause incorrect diff output. */ + +static int +diag (xoff, xlim, yoff, ylim, minimal, part) + int xoff, xlim, yoff, ylim, minimal; + struct partition *part; +{ + int *const fd = fdiag; /* Give the compiler a chance. */ + int *const bd = bdiag; /* Additional help for the compiler. */ + int const *const xv = xvec; /* Still more help for the compiler. */ + int const *const yv = yvec; /* And more and more . . . */ + int const dmin = xoff - ylim; /* Minimum valid diagonal. */ + int const dmax = xlim - yoff; /* Maximum valid diagonal. */ + int const fmid = xoff - yoff; /* Center diagonal of top-down search. */ + int const bmid = xlim - ylim; /* Center diagonal of bottom-up search. */ + int fmin = fmid, fmax = fmid; /* Limits of top-down search. */ + int bmin = bmid, bmax = bmid; /* Limits of bottom-up search. */ + int c; /* Cost. */ + int odd = (fmid - bmid) & 1; /* True if southeast corner is on an odd + diagonal with respect to the northwest. */ + + fd[fmid] = xoff; + bd[bmid] = xlim; + + for (c = 1;; ++c) + { + int d; /* Active diagonal. */ + int big_snake = 0; + + /* Extend the top-down search by an edit step in each diagonal. */ + fmin > dmin ? fd[--fmin - 1] = -1 : ++fmin; + fmax < dmax ? fd[++fmax + 1] = -1 : --fmax; + for (d = fmax; d >= fmin; d -= 2) + { + int x, y, oldx, tlo = fd[d - 1], thi = fd[d + 1]; + + if (tlo >= thi) + x = tlo + 1; + else + x = thi; + oldx = x; + y = x - d; + while (x < xlim && y < ylim && xv[x] == yv[y]) + ++x, ++y; + if (x - oldx > SNAKE_LIMIT) + big_snake = 1; + fd[d] = x; + if (odd && bmin <= d && d <= bmax && bd[d] <= x) + { + part->xmid = x; + part->ymid = y; + part->lo_minimal = part->hi_minimal = 1; + return 2 * c - 1; + } + } + + /* Similarly extend the bottom-up search. */ + bmin > dmin ? bd[--bmin - 1] = INT_MAX : ++bmin; + bmax < dmax ? bd[++bmax + 1] = INT_MAX : --bmax; + for (d = bmax; d >= bmin; d -= 2) + { + int x, y, oldx, tlo = bd[d - 1], thi = bd[d + 1]; + + if (tlo < thi) + x = tlo; + else + x = thi - 1; + oldx = x; + y = x - d; + while (x > xoff && y > yoff && xv[x - 1] == yv[y - 1]) + --x, --y; + if (oldx - x > SNAKE_LIMIT) + big_snake = 1; + bd[d] = x; + if (!odd && fmin <= d && d <= fmax && x <= fd[d]) + { + part->xmid = x; + part->ymid = y; + part->lo_minimal = part->hi_minimal = 1; + return 2 * c; + } + } + + if (minimal) + continue; + + /* Heuristic: check occasionally for a diagonal that has made + lots of progress compared with the edit distance. + If we have any such, find the one that has made the most + progress and return it as if it had succeeded. + + With this heuristic, for files with a constant small density + of changes, the algorithm is linear in the file size. */ + + if (c > 200 && big_snake && heuristic) + { + int best; + + best = 0; + for (d = fmax; d >= fmin; d -= 2) + { + int dd = d - fmid; + int x = fd[d]; + int y = x - d; + int v = (x - xoff) * 2 - dd; + if (v > 12 * (c + (dd < 0 ? -dd : dd))) + { + if (v > best + && xoff + SNAKE_LIMIT <= x && x < xlim + && yoff + SNAKE_LIMIT <= y && y < ylim) + { + /* We have a good enough best diagonal; + now insist that it end with a significant snake. */ + int k; + + for (k = 1; xv[x - k] == yv[y - k]; k++) + if (k == SNAKE_LIMIT) + { + best = v; + part->xmid = x; + part->ymid = y; + break; + } + } + } + } + if (best > 0) + { + part->lo_minimal = 1; + part->hi_minimal = 0; + return 2 * c - 1; + } + + best = 0; + for (d = bmax; d >= bmin; d -= 2) + { + int dd = d - bmid; + int x = bd[d]; + int y = x - d; + int v = (xlim - x) * 2 + dd; + if (v > 12 * (c + (dd < 0 ? -dd : dd))) + { + if (v > best + && xoff < x && x <= xlim - SNAKE_LIMIT + && yoff < y && y <= ylim - SNAKE_LIMIT) + { + /* We have a good enough best diagonal; + now insist that it end with a significant snake. */ + int k; + + for (k = 0; xv[x + k] == yv[y + k]; k++) + if (k == SNAKE_LIMIT - 1) + { + best = v; + part->xmid = x; + part->ymid = y; + break; + } + } + } + } + if (best > 0) + { + part->lo_minimal = 0; + part->hi_minimal = 1; + return 2 * c - 1; + } + } + + /* Heuristic: if we've gone well beyond the call of duty, + give up and report halfway between our best results so far. */ + if (c >= too_expensive) + { + int fxybest, fxbest; + int bxybest, bxbest; + + fxbest = bxbest = 0; /* Pacify `gcc -Wall'. */ + + /* Find forward diagonal that maximizes X + Y. */ + fxybest = -1; + for (d = fmax; d >= fmin; d -= 2) + { + int x = min (fd[d], xlim); + int y = x - d; + if (ylim < y) + x = ylim + d, y = ylim; + if (fxybest < x + y) + { + fxybest = x + y; + fxbest = x; + } + } + + /* Find backward diagonal that minimizes X + Y. */ + bxybest = INT_MAX; + for (d = bmax; d >= bmin; d -= 2) + { + int x = max (xoff, bd[d]); + int y = x - d; + if (y < yoff) + x = yoff + d, y = yoff; + if (x + y < bxybest) + { + bxybest = x + y; + bxbest = x; + } + } + + /* Use the better of the two diagonals. */ + if ((xlim + ylim) - bxybest < fxybest - (xoff + yoff)) + { + part->xmid = fxbest; + part->ymid = fxybest - fxbest; + part->lo_minimal = 1; + part->hi_minimal = 0; + } + else + { + part->xmid = bxbest; + part->ymid = bxybest - bxbest; + part->lo_minimal = 0; + part->hi_minimal = 1; + } + return 2 * c - 1; + } + } +} + +/* Compare in detail contiguous subsequences of the two files + which are known, as a whole, to match each other. + + The results are recorded in the vectors files[N].changed_flag, by + storing a 1 in the element for each line that is an insertion or deletion. + + The subsequence of file 0 is [XOFF, XLIM) and likewise for file 1. + + Note that XLIM, YLIM are exclusive bounds. + All line numbers are origin-0 and discarded lines are not counted. + + If MINIMAL is nonzero, find a minimal difference no matter how + expensive it is. */ + +static void +compareseq (xoff, xlim, yoff, ylim, minimal) + int xoff, xlim, yoff, ylim, minimal; +{ + int * const xv = xvec; /* Help the compiler. */ + int * const yv = yvec; + + /* Slide down the bottom initial diagonal. */ + while (xoff < xlim && yoff < ylim && xv[xoff] == yv[yoff]) + ++xoff, ++yoff; + /* Slide up the top initial diagonal. */ + while (xlim > xoff && ylim > yoff && xv[xlim - 1] == yv[ylim - 1]) + --xlim, --ylim; + + /* Handle simple cases. */ + if (xoff == xlim) + while (yoff < ylim) + files[1].changed_flag[files[1].realindexes[yoff++]] = 1; + else if (yoff == ylim) + while (xoff < xlim) + files[0].changed_flag[files[0].realindexes[xoff++]] = 1; + else + { + int c; + struct partition part; + + /* Find a point of correspondence in the middle of the files. */ + + c = diag (xoff, xlim, yoff, ylim, minimal, &part); + + if (c == 1) + { + /* This should be impossible, because it implies that + one of the two subsequences is empty, + and that case was handled above without calling `diag'. + Let's verify that this is true. */ + abort (); +#if 0 + /* The two subsequences differ by a single insert or delete; + record it and we are done. */ + if (part.xmid - part.ymid < xoff - yoff) + files[1].changed_flag[files[1].realindexes[part.ymid - 1]] = 1; + else + files[0].changed_flag[files[0].realindexes[part.xmid]] = 1; +#endif + } + else + { + /* Use the partitions to split this problem into subproblems. */ + compareseq (xoff, part.xmid, yoff, part.ymid, part.lo_minimal); + compareseq (part.xmid, xlim, part.ymid, ylim, part.hi_minimal); + } + } +} + +/* Discard lines from one file that have no matches in the other file. + + A line which is discarded will not be considered by the actual + comparison algorithm; it will be as if that line were not in the file. + The file's `realindexes' table maps virtual line numbers + (which don't count the discarded lines) into real line numbers; + this is how the actual comparison algorithm produces results + that are comprehensible when the discarded lines are counted. + + When we discard a line, we also mark it as a deletion or insertion + so that it will be printed in the output. */ + +static void +discard_confusing_lines (filevec) + struct file_data filevec[]; +{ + unsigned int f, i; + char *discarded[2]; + int *equiv_count[2]; + int *p; + + /* Allocate our results. */ + p = (int *) xmalloc ((filevec[0].buffered_lines + filevec[1].buffered_lines) + * (2 * sizeof (int))); + for (f = 0; f < 2; f++) + { + filevec[f].undiscarded = p; p += filevec[f].buffered_lines; + filevec[f].realindexes = p; p += filevec[f].buffered_lines; + } + + /* Set up equiv_count[F][I] as the number of lines in file F + that fall in equivalence class I. */ + + p = (int *) xmalloc (filevec[0].equiv_max * (2 * sizeof (int))); + equiv_count[0] = p; + equiv_count[1] = p + filevec[0].equiv_max; + bzero (p, filevec[0].equiv_max * (2 * sizeof (int))); + + for (i = 0; i < filevec[0].buffered_lines; ++i) + ++equiv_count[0][filevec[0].equivs[i]]; + for (i = 0; i < filevec[1].buffered_lines; ++i) + ++equiv_count[1][filevec[1].equivs[i]]; + + /* Set up tables of which lines are going to be discarded. */ + + discarded[0] = xmalloc (sizeof (char) + * (filevec[0].buffered_lines + + filevec[1].buffered_lines)); + discarded[1] = discarded[0] + filevec[0].buffered_lines; + bzero (discarded[0], sizeof (char) * (filevec[0].buffered_lines + + filevec[1].buffered_lines)); + + /* Mark to be discarded each line that matches no line of the other file. + If a line matches many lines, mark it as provisionally discardable. */ + + for (f = 0; f < 2; f++) + { + unsigned int end = filevec[f].buffered_lines; + char *discards = discarded[f]; + int *counts = equiv_count[1 - f]; + int *equivs = filevec[f].equivs; + unsigned int many = 5; + unsigned int tem = end / 64; + + /* Multiply MANY by approximate square root of number of lines. + That is the threshold for provisionally discardable lines. */ + while ((tem = tem >> 2) > 0) + many *= 2; + + for (i = 0; i < end; i++) + { + int nmatch; + if (equivs[i] == 0) + continue; + nmatch = counts[equivs[i]]; + if (nmatch == 0) + discards[i] = 1; + else if (nmatch > many) + discards[i] = 2; + } + } + + /* Don't really discard the provisional lines except when they occur + in a run of discardables, with nonprovisionals at the beginning + and end. */ + + for (f = 0; f < 2; f++) + { + unsigned int end = filevec[f].buffered_lines; + register char *discards = discarded[f]; + + for (i = 0; i < end; i++) + { + /* Cancel provisional discards not in middle of run of discards. */ + if (discards[i] == 2) + discards[i] = 0; + else if (discards[i] != 0) + { + /* We have found a nonprovisional discard. */ + register int j; + unsigned int length; + unsigned int provisional = 0; + + /* Find end of this run of discardable lines. + Count how many are provisionally discardable. */ + for (j = i; j < end; j++) + { + if (discards[j] == 0) + break; + if (discards[j] == 2) + ++provisional; + } + + /* Cancel provisional discards at end, and shrink the run. */ + while (j > i && discards[j - 1] == 2) + discards[--j] = 0, --provisional; + + /* Now we have the length of a run of discardable lines + whose first and last are not provisional. */ + length = j - i; + + /* If 1/4 of the lines in the run are provisional, + cancel discarding of all provisional lines in the run. */ + if (provisional * 4 > length) + { + while (j > i) + if (discards[--j] == 2) + discards[j] = 0; + } + else + { + register unsigned int consec; + unsigned int minimum = 1; + unsigned int tem = length / 4; + + /* MINIMUM is approximate square root of LENGTH/4. + A subrun of two or more provisionals can stand + when LENGTH is at least 16. + A subrun of 4 or more can stand when LENGTH >= 64. */ + while ((tem = tem >> 2) > 0) + minimum *= 2; + minimum++; + + /* Cancel any subrun of MINIMUM or more provisionals + within the larger run. */ + for (j = 0, consec = 0; j < length; j++) + if (discards[i + j] != 2) + consec = 0; + else if (minimum == ++consec) + /* Back up to start of subrun, to cancel it all. */ + j -= consec; + else if (minimum < consec) + discards[i + j] = 0; + + /* Scan from beginning of run + until we find 3 or more nonprovisionals in a row + or until the first nonprovisional at least 8 lines in. + Until that point, cancel any provisionals. */ + for (j = 0, consec = 0; j < length; j++) + { + if (j >= 8 && discards[i + j] == 1) + break; + if (discards[i + j] == 2) + consec = 0, discards[i + j] = 0; + else if (discards[i + j] == 0) + consec = 0; + else + consec++; + if (consec == 3) + break; + } + + /* I advances to the last line of the run. */ + i += length - 1; + + /* Same thing, from end. */ + for (j = 0, consec = 0; j < length; j++) + { + if (j >= 8 && discards[i - j] == 1) + break; + if (discards[i - j] == 2) + consec = 0, discards[i - j] = 0; + else if (discards[i - j] == 0) + consec = 0; + else + consec++; + if (consec == 3) + break; + } + } + } + } + } + + /* Actually discard the lines. */ + for (f = 0; f < 2; f++) + { + char *discards = discarded[f]; + unsigned int end = filevec[f].buffered_lines; + unsigned int j = 0; + for (i = 0; i < end; ++i) + if (no_discards || discards[i] == 0) + { + filevec[f].undiscarded[j] = filevec[f].equivs[i]; + filevec[f].realindexes[j++] = i; + } + else + filevec[f].changed_flag[i] = 1; + filevec[f].nondiscarded_lines = j; + } + + free (discarded[0]); + free (equiv_count[0]); +} + +/* Adjust inserts/deletes of identical lines to join changes + as much as possible. + + We do something when a run of changed lines include a + line at one end and have an excluded, identical line at the other. + We are free to choose which identical line is included. + `compareseq' usually chooses the one at the beginning, + but usually it is cleaner to consider the following identical line + to be the "change". */ + +int inhibit; + +static void +shift_boundaries (filevec) + struct file_data filevec[]; +{ + int f; + + if (inhibit) + return; + + for (f = 0; f < 2; f++) + { + char *changed = filevec[f].changed_flag; + char const *other_changed = filevec[1-f].changed_flag; + int const *equivs = filevec[f].equivs; + int i = 0; + int j = 0; + int i_end = filevec[f].buffered_lines; + + while (1) + { + int runlength, start, corresponding; + + /* Scan forwards to find beginning of another run of changes. + Also keep track of the corresponding point in the other file. */ + + while (i < i_end && changed[i] == 0) + { + while (other_changed[j++]) + continue; + i++; + } + + if (i == i_end) + break; + + start = i; + + /* Find the end of this run of changes. */ + + while (changed[++i]) + continue; + while (other_changed[j]) + j++; + + do + { + /* Record the length of this run of changes, so that + we can later determine whether the run has grown. */ + runlength = i - start; + + /* Move the changed region back, so long as the + previous unchanged line matches the last changed one. + This merges with previous changed regions. */ + + while (start && equivs[start - 1] == equivs[i - 1]) + { + changed[--start] = 1; + changed[--i] = 0; + while (changed[start - 1]) + start--; + while (other_changed[--j]) + continue; + } + + /* Set CORRESPONDING to the end of the changed run, at the last + point where it corresponds to a changed run in the other file. + CORRESPONDING == I_END means no such point has been found. */ + corresponding = other_changed[j - 1] ? i : i_end; + + /* Move the changed region forward, so long as the + first changed line matches the following unchanged one. + This merges with following changed regions. + Do this second, so that if there are no merges, + the changed region is moved forward as far as possible. */ + + while (i != i_end && equivs[start] == equivs[i]) + { + changed[start++] = 0; + changed[i++] = 1; + while (changed[i]) + i++; + while (other_changed[++j]) + corresponding = i; + } + } + while (runlength != i - start); + + /* If possible, move the fully-merged run of changes + back to a corresponding run in the other file. */ + + while (corresponding < i) + { + changed[--start] = 1; + changed[--i] = 0; + while (other_changed[--j]) + continue; + } + } + } +} + +/* Cons an additional entry onto the front of an edit script OLD. + LINE0 and LINE1 are the first affected lines in the two files (origin 0). + DELETED is the number of lines deleted here from file 0. + INSERTED is the number of lines inserted here in file 1. + + If DELETED is 0 then LINE0 is the number of the line before + which the insertion was done; vice versa for INSERTED and LINE1. */ + +static struct change * +add_change (line0, line1, deleted, inserted, old) + int line0, line1, deleted, inserted; + struct change *old; +{ + struct change *new = (struct change *) xmalloc (sizeof (struct change)); + + new->line0 = line0; + new->line1 = line1; + new->inserted = inserted; + new->deleted = deleted; + new->link = old; + return new; +} + +/* Scan the tables of which lines are inserted and deleted, + producing an edit script in reverse order. */ + +static struct change * +build_reverse_script (filevec) + struct file_data const filevec[]; +{ + struct change *script = 0; + char *changed0 = filevec[0].changed_flag; + char *changed1 = filevec[1].changed_flag; + int len0 = filevec[0].buffered_lines; + int len1 = filevec[1].buffered_lines; + + /* Note that changedN[len0] does exist, and contains 0. */ + + int i0 = 0, i1 = 0; + + while (i0 < len0 || i1 < len1) + { + if (changed0[i0] || changed1[i1]) + { + int line0 = i0, line1 = i1; + + /* Find # lines changed here in each file. */ + while (changed0[i0]) ++i0; + while (changed1[i1]) ++i1; + + /* Record this change. */ + script = add_change (line0, line1, i0 - line0, i1 - line1, script); + } + + /* We have reached lines in the two files that match each other. */ + i0++, i1++; + } + + return script; +} + +/* Scan the tables of which lines are inserted and deleted, + producing an edit script in forward order. */ + +static struct change * +build_script (filevec) + struct file_data const filevec[]; +{ + struct change *script = 0; + char *changed0 = filevec[0].changed_flag; + char *changed1 = filevec[1].changed_flag; + int i0 = filevec[0].buffered_lines, i1 = filevec[1].buffered_lines; + + /* Note that changedN[-1] does exist, and contains 0. */ + + while (i0 >= 0 || i1 >= 0) + { + if (changed0[i0 - 1] || changed1[i1 - 1]) + { + int line0 = i0, line1 = i1; + + /* Find # lines changed here in each file. */ + while (changed0[i0 - 1]) --i0; + while (changed1[i1 - 1]) --i1; + + /* Record this change. */ + script = add_change (i0, i1, line0 - i0, line1 - i1, script); + } + + /* We have reached lines in the two files that match each other. */ + i0--, i1--; + } + + return script; +} + +/* If CHANGES, briefly report that two files differed. */ +static void +briefly_report (changes, filevec) + int changes; + struct file_data const filevec[]; +{ + if (changes) + message (no_details_flag ? "Files %s and %s differ\n" + : "Binary files %s and %s differ\n", + filevec[0].name, filevec[1].name); +} + +/* Report the differences of two files. DEPTH is the current directory + depth. */ +int +diff_2_files (filevec, depth) + struct file_data filevec[]; + int depth; +{ + int diags; + int i; + struct change *e, *p; + struct change *script; + int changes; + + + /* If we have detected that either file is binary, + compare the two files as binary. This can happen + only when the first chunk is read. + Also, --brief without any --ignore-* options means + we can speed things up by treating the files as binary. */ + + if (read_files (filevec, no_details_flag & ~ignore_some_changes)) + { + /* Files with different lengths must be different. */ + if (filevec[0].stat.st_size != filevec[1].stat.st_size + && (filevec[0].desc < 0 || S_ISREG (filevec[0].stat.st_mode)) + && (filevec[1].desc < 0 || S_ISREG (filevec[1].stat.st_mode))) + changes = 1; + + /* Standard input equals itself. */ + else if (filevec[0].desc == filevec[1].desc) + changes = 0; + + else + /* Scan both files, a buffer at a time, looking for a difference. */ + { + /* Allocate same-sized buffers for both files. */ + size_t buffer_size = buffer_lcm (STAT_BLOCKSIZE (filevec[0].stat), + STAT_BLOCKSIZE (filevec[1].stat)); + for (i = 0; i < 2; i++) + filevec[i].buffer = xrealloc (filevec[i].buffer, buffer_size); + + for (;; filevec[0].buffered_chars = filevec[1].buffered_chars = 0) + { + /* Read a buffer's worth from both files. */ + for (i = 0; i < 2; i++) + if (0 <= filevec[i].desc) + while (filevec[i].buffered_chars != buffer_size) + { + int r = read (filevec[i].desc, + filevec[i].buffer + + filevec[i].buffered_chars, + buffer_size - filevec[i].buffered_chars); + if (r == 0) + break; + if (r < 0) + pfatal_with_name (filevec[i].name); + filevec[i].buffered_chars += r; + } + + /* If the buffers differ, the files differ. */ + if (filevec[0].buffered_chars != filevec[1].buffered_chars + || (filevec[0].buffered_chars != 0 + && memcmp (filevec[0].buffer, + filevec[1].buffer, + filevec[0].buffered_chars) != 0)) + { + changes = 1; + break; + } + + /* If we reach end of file, the files are the same. */ + if (filevec[0].buffered_chars != buffer_size) + { + changes = 0; + break; + } + } + } + + briefly_report (changes, filevec); + } + else + { + /* Allocate vectors for the results of comparison: + a flag for each line of each file, saying whether that line + is an insertion or deletion. + Allocate an extra element, always zero, at each end of each vector. */ + + size_t s = filevec[0].buffered_lines + filevec[1].buffered_lines + 4; + filevec[0].changed_flag = xmalloc (s); + bzero (filevec[0].changed_flag, s); + filevec[0].changed_flag++; + filevec[1].changed_flag = filevec[0].changed_flag + + filevec[0].buffered_lines + 2; + + /* Some lines are obviously insertions or deletions + because they don't match anything. Detect them now, and + avoid even thinking about them in the main comparison algorithm. */ + + discard_confusing_lines (filevec); + + /* Now do the main comparison algorithm, considering just the + undiscarded lines. */ + + xvec = filevec[0].undiscarded; + yvec = filevec[1].undiscarded; + diags = filevec[0].nondiscarded_lines + filevec[1].nondiscarded_lines + 3; + fdiag = (int *) xmalloc (diags * (2 * sizeof (int))); + bdiag = fdiag + diags; + fdiag += filevec[1].nondiscarded_lines + 1; + bdiag += filevec[1].nondiscarded_lines + 1; + + /* Set TOO_EXPENSIVE to be approximate square root of input size, + bounded below by 256. */ + too_expensive = 1; + for (i = filevec[0].nondiscarded_lines + filevec[1].nondiscarded_lines; + i != 0; i >>= 2) + too_expensive <<= 1; + too_expensive = max (256, too_expensive); + + files[0] = filevec[0]; + files[1] = filevec[1]; + + compareseq (0, filevec[0].nondiscarded_lines, + 0, filevec[1].nondiscarded_lines, no_discards); + + free (fdiag - (filevec[1].nondiscarded_lines + 1)); + + /* Modify the results slightly to make them prettier + in cases where that can validly be done. */ + + shift_boundaries (filevec); + + /* Get the results of comparison in the form of a chain + of `struct change's -- an edit script. */ + + if (output_style == OUTPUT_ED) + script = build_reverse_script (filevec); + else + script = build_script (filevec); + + /* Set CHANGES if we had any diffs. + If some changes are ignored, we must scan the script to decide. */ + if (ignore_blank_lines_flag || ignore_regexp_list) + { + struct change *next = script; + changes = 0; + + while (next && changes == 0) + { + struct change *this, *end; + int first0, last0, first1, last1, deletes, inserts; + + /* Find a set of changes that belong together. */ + this = next; + end = find_change (next); + + /* Disconnect them from the rest of the changes, making them + a hunk, and remember the rest for next iteration. */ + next = end->link; + end->link = 0; + + /* Determine whether this hunk is really a difference. */ + analyze_hunk (this, &first0, &last0, &first1, &last1, + &deletes, &inserts); + + /* Reconnect the script so it will all be freed properly. */ + end->link = next; + + if (deletes || inserts) + changes = 1; + } + } + else + changes = (script != 0); + + if (no_details_flag) + briefly_report (changes, filevec); + else + { + if (changes || ! no_diff_means_no_output) + { + /* Record info for starting up output, + to be used if and when we have some output to print. */ + setup_output (files[0].name, files[1].name, depth); + + switch (output_style) + { + case OUTPUT_CONTEXT: + print_context_script (script, 0); + break; + + case OUTPUT_UNIFIED: + print_context_script (script, 1); + break; + + case OUTPUT_ED: + print_ed_script (script); + break; + + case OUTPUT_FORWARD_ED: + pr_forward_ed_script (script); + break; + + case OUTPUT_RCS: + print_rcs_script (script); + break; + + case OUTPUT_NORMAL: + print_normal_script (script); + break; + + case OUTPUT_IFDEF: + print_ifdef_script (script); + break; + + case OUTPUT_SDIFF: + print_sdiff_script (script); + } + + finish_output (); + } + } + + free (filevec[0].undiscarded); + + free (filevec[0].changed_flag - 1); + + for (i = 1; i >= 0; --i) + free (filevec[i].equivs); + + for (i = 0; i < 2; ++i) + free (filevec[i].linbuf + filevec[i].linbuf_base); + + for (e = script; e; e = p) + { + p = e->link; + free (e); + } + + if (! ROBUST_OUTPUT_STYLE (output_style)) + for (i = 0; i < 2; ++i) + if (filevec[i].missing_newline) + { + error ("No newline at end of file %s", filevec[i].name, ""); + changes = 2; + } + } + + if (filevec[0].buffer != filevec[1].buffer) + free (filevec[0].buffer); + free (filevec[1].buffer); + + return changes; +} diff --git a/sys/src/ape/cmd/diff/cmpbuf.c b/sys/src/ape/cmd/diff/cmpbuf.c new file mode 100755 index 000000000..e95a8f98e --- /dev/null +++ b/sys/src/ape/cmd/diff/cmpbuf.c @@ -0,0 +1,40 @@ +/* Buffer primitives for comparison operations. + Copyright (C) 1993 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#include "system.h" +#include "cmpbuf.h" + +/* Least common multiple of two buffer sizes A and B. */ + +size_t +buffer_lcm (a, b) + size_t a, b; +{ + size_t m, n, r; + + /* Yield reasonable values if buffer sizes are zero. */ + if (!a) + return b ? b : 8 * 1024; + if (!b) + return a; + + /* n = gcd (a, b) */ + for (m = a, n = b; (r = m % n) != 0; m = n, n = r) + continue; + + return a/n * b; +} diff --git a/sys/src/ape/cmd/diff/cmpbuf.h b/sys/src/ape/cmd/diff/cmpbuf.h new file mode 100755 index 000000000..e3852b7bd --- /dev/null +++ b/sys/src/ape/cmd/diff/cmpbuf.h @@ -0,0 +1,20 @@ +/* Buffer primitives for comparison operations. + Copyright (C) 1993 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +size_t buffer_lcm PARAMS((size_t, size_t)); diff --git a/sys/src/ape/cmd/diff/config.h b/sys/src/ape/cmd/diff/config.h new file mode 100755 index 000000000..f426774f2 --- /dev/null +++ b/sys/src/ape/cmd/diff/config.h @@ -0,0 +1,118 @@ +/* config.h. Generated automatically by configure. */ +/* config.hin. Generated automatically from configure.in by autoheader. */ + +/* Define if using alloca.c. */ +/* #undef C_ALLOCA */ + +/* Define if the closedir function returns void instead of int. */ +/* #undef CLOSEDIR_VOID */ + +/* Define to empty if the keyword does not work. */ +/* #undef const */ + +/* Define to one of _getb67, GETB67, getb67 for Cray-2 and Cray-YMP systems. + This function is required for alloca.c support on those systems. */ +/* #undef CRAY_STACKSEG_END */ + +/* Define if you have <alloca.h> and it should be used (not on Ultrix). */ +/* #undef HAVE_ALLOCA_H */ + +/* Define if you don't have vprintf but do have _doprnt. */ +/* #undef HAVE_DOPRNT */ + +/* Define if your struct stat has st_blksize. */ +/* #define HAVE_ST_BLKSIZE 1 */ + +/* Define if you have <vfork.h>. */ +/* #undef HAVE_VFORK_H */ + +/* Define if you have the vprintf function. */ +#define HAVE_VPRINTF 1 + +/* Define if on MINIX. */ +/* #undef _MINIX */ + +/* Define to `int' if <sys/types.h> doesn't define. */ +/* #undef pid_t */ + +/* Define if the system does not provide POSIX.1 features except + with this defined. */ +/* #undef _POSIX_1_SOURCE */ + +/* Define if you need to in order for stat and other things to work. */ +/* #undef _POSIX_SOURCE */ + +/* Define as the return type of signal handlers (int or void). */ +#define RETSIGTYPE void + +/* If using the C implementation of alloca, define if you know the + direction of stack growth for your system; otherwise it will be + automatically deduced at run-time. + STACK_DIRECTION > 0 => grows toward higher addresses + STACK_DIRECTION < 0 => grows toward lower addresses + STACK_DIRECTION = 0 => direction of growth unknown + */ +/* #undef STACK_DIRECTION */ + +/* Define if the `S_IS*' macros in <sys/stat.h> do not work properly. */ +/* #undef STAT_MACROS_BROKEN */ + +/* Define if you have the ANSI C header files. */ +#define STDC_HEADERS 1 + +/* Define if <sys/wait.h> is compatible with Posix applications. */ +#define HAVE_SYS_WAIT_H 1 + +/* Define vfork as fork if vfork does not work. */ +/* #undef vfork */ + +/* Define if you have the dup2 function. */ +#define HAVE_DUP2 1 + +/* Define if you have the memchr function. */ +#define HAVE_MEMCHR 1 + +/* Define if you have the sigaction function. */ +#define HAVE_SIGACTION 1 + +/* Define if you have the strchr function. */ +#define HAVE_STRCHR 1 + +/* Define if you have the strerror function. */ +#define HAVE_STRERROR 1 + +/* Define if you have the tmpnam function. */ +#define HAVE_TMPNAM 1 + +/* Define if you have the <dirent.h> header file. */ +#define HAVE_DIRENT_H 1 + +/* Define if you have the <fcntl.h> header file. */ +#define HAVE_FCNTL_H 1 + +/* Define if you have the <limits.h> header file. */ +#define HAVE_LIMITS_H 1 + +/* Define if you have the <ndir.h> header file. */ +/* #undef HAVE_NDIR_H */ + +/* Define if you have the <stdlib.h> header file. */ +#define HAVE_STDLIB_H 1 + +/* Define if you have the <string.h> header file. */ +#define HAVE_STRING_H 1 + +/* Define if you have the <sys/dir.h> header file. */ +/* #undef HAVE_SYS_DIR_H */ + +/* Define if you have the <sys/file.h> header file. */ +#define HAVE_SYS_FILE_H 1 + +/* Define if you have the <sys/ndir.h> header file. */ +/* #undef HAVE_SYS_NDIR_H */ + +/* Define if you have the <time.h> header file. */ +#define HAVE_TIME_H 1 + +/* Define if you have the <unistd.h> header file. */ +#define HAVE_UNISTD_H 1 diff --git a/sys/src/ape/cmd/diff/context.c b/sys/src/ape/cmd/diff/context.c new file mode 100755 index 000000000..14f950c52 --- /dev/null +++ b/sys/src/ape/cmd/diff/context.c @@ -0,0 +1,468 @@ +/* Context-format output routines for GNU DIFF. + Copyright (C) 1988,1989,1991,1992,1993,1994 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#include "diff.h" + +static struct change *find_hunk PARAMS((struct change *)); +static void find_function PARAMS((struct file_data const *, int, char const **, size_t *)); +static void mark_ignorable PARAMS((struct change *)); +static void pr_context_hunk PARAMS((struct change *)); +static void pr_unidiff_hunk PARAMS((struct change *)); +static void print_context_label PARAMS ((char const *, struct file_data *, char const *)); +static void print_context_number_range PARAMS((struct file_data const *, int, int)); +static void print_unidiff_number_range PARAMS((struct file_data const *, int, int)); + +/* Last place find_function started searching from. */ +static int find_function_last_search; + +/* The value find_function returned when it started searching there. */ +static int find_function_last_match; + +/* Print a label for a context diff, with a file name and date or a label. */ + +static void +print_context_label (mark, inf, label) + char const *mark; + struct file_data *inf; + char const *label; +{ + if (label) + fprintf (outfile, "%s %s\n", mark, label); + else + { + char const *ct = ctime (&inf->stat.st_mtime); + if (!ct) + ct = "?\n"; + /* See Posix.2 section 4.17.6.1.4 for this format. */ + fprintf (outfile, "%s %s\t%s", mark, inf->name, ct); + } +} + +/* Print a header for a context diff, with the file names and dates. */ + +void +print_context_header (inf, unidiff_flag) + struct file_data inf[]; + int unidiff_flag; +{ + if (unidiff_flag) + { + print_context_label ("---", &inf[0], file_label[0]); + print_context_label ("+++", &inf[1], file_label[1]); + } + else + { + print_context_label ("***", &inf[0], file_label[0]); + print_context_label ("---", &inf[1], file_label[1]); + } +} + +/* Print an edit script in context format. */ + +void +print_context_script (script, unidiff_flag) + struct change *script; + int unidiff_flag; +{ + if (ignore_blank_lines_flag || ignore_regexp_list) + mark_ignorable (script); + else + { + struct change *e; + for (e = script; e; e = e->link) + e->ignore = 0; + } + + find_function_last_search = - files[0].prefix_lines; + find_function_last_match = find_function_last_search - 1; + + if (unidiff_flag) + print_script (script, find_hunk, pr_unidiff_hunk); + else + print_script (script, find_hunk, pr_context_hunk); +} + +/* Print a pair of line numbers with a comma, translated for file FILE. + If the second number is not greater, use the first in place of it. + + Args A and B are internal line numbers. + We print the translated (real) line numbers. */ + +static void +print_context_number_range (file, a, b) + struct file_data const *file; + int a, b; +{ + int trans_a, trans_b; + translate_range (file, a, b, &trans_a, &trans_b); + + /* Note: we can have B < A in the case of a range of no lines. + In this case, we should print the line number before the range, + which is B. */ + if (trans_b > trans_a) + fprintf (outfile, "%d,%d", trans_a, trans_b); + else + fprintf (outfile, "%d", trans_b); +} + +/* Print a portion of an edit script in context format. + HUNK is the beginning of the portion to be printed. + The end is marked by a `link' that has been nulled out. + + Prints out lines from both files, and precedes each + line with the appropriate flag-character. */ + +static void +pr_context_hunk (hunk) + struct change *hunk; +{ + int first0, last0, first1, last1, show_from, show_to, i; + struct change *next; + char const *prefix; + char const *function; + size_t function_length; + FILE *out; + + /* Determine range of line numbers involved in each file. */ + + analyze_hunk (hunk, &first0, &last0, &first1, &last1, &show_from, &show_to); + + if (!show_from && !show_to) + return; + + /* Include a context's width before and after. */ + + i = - files[0].prefix_lines; + first0 = max (first0 - context, i); + first1 = max (first1 - context, i); + last0 = min (last0 + context, files[0].valid_lines - 1); + last1 = min (last1 + context, files[1].valid_lines - 1); + + /* If desired, find the preceding function definition line in file 0. */ + function = 0; + if (function_regexp_list) + find_function (&files[0], first0, &function, &function_length); + + begin_output (); + out = outfile; + + /* If we looked for and found a function this is part of, + include its name in the header of the diff section. */ + fprintf (out, "***************"); + + if (function) + { + fprintf (out, " "); + fwrite (function, 1, min (function_length - 1, 40), out); + } + + fprintf (out, "\n*** "); + print_context_number_range (&files[0], first0, last0); + fprintf (out, " ****\n"); + + if (show_from) + { + next = hunk; + + for (i = first0; i <= last0; i++) + { + /* Skip past changes that apply (in file 0) + only to lines before line I. */ + + while (next && next->line0 + next->deleted <= i) + next = next->link; + + /* Compute the marking for line I. */ + + prefix = " "; + if (next && next->line0 <= i) + /* The change NEXT covers this line. + If lines were inserted here in file 1, this is "changed". + Otherwise it is "deleted". */ + prefix = (next->inserted > 0 ? "!" : "-"); + + print_1_line (prefix, &files[0].linbuf[i]); + } + } + + fprintf (out, "--- "); + print_context_number_range (&files[1], first1, last1); + fprintf (out, " ----\n"); + + if (show_to) + { + next = hunk; + + for (i = first1; i <= last1; i++) + { + /* Skip past changes that apply (in file 1) + only to lines before line I. */ + + while (next && next->line1 + next->inserted <= i) + next = next->link; + + /* Compute the marking for line I. */ + + prefix = " "; + if (next && next->line1 <= i) + /* The change NEXT covers this line. + If lines were deleted here in file 0, this is "changed". + Otherwise it is "inserted". */ + prefix = (next->deleted > 0 ? "!" : "+"); + + print_1_line (prefix, &files[1].linbuf[i]); + } + } +} + +/* Print a pair of line numbers with a comma, translated for file FILE. + If the second number is smaller, use the first in place of it. + If the numbers are equal, print just one number. + + Args A and B are internal line numbers. + We print the translated (real) line numbers. */ + +static void +print_unidiff_number_range (file, a, b) + struct file_data const *file; + int a, b; +{ + int trans_a, trans_b; + translate_range (file, a, b, &trans_a, &trans_b); + + /* Note: we can have B < A in the case of a range of no lines. + In this case, we should print the line number before the range, + which is B. */ + if (trans_b <= trans_a) + fprintf (outfile, trans_b == trans_a ? "%d" : "%d,0", trans_b); + else + fprintf (outfile, "%d,%d", trans_a, trans_b - trans_a + 1); +} + +/* Print a portion of an edit script in unidiff format. + HUNK is the beginning of the portion to be printed. + The end is marked by a `link' that has been nulled out. + + Prints out lines from both files, and precedes each + line with the appropriate flag-character. */ + +static void +pr_unidiff_hunk (hunk) + struct change *hunk; +{ + int first0, last0, first1, last1, show_from, show_to, i, j, k; + struct change *next; + char const *function; + size_t function_length; + FILE *out; + + /* Determine range of line numbers involved in each file. */ + + analyze_hunk (hunk, &first0, &last0, &first1, &last1, &show_from, &show_to); + + if (!show_from && !show_to) + return; + + /* Include a context's width before and after. */ + + i = - files[0].prefix_lines; + first0 = max (first0 - context, i); + first1 = max (first1 - context, i); + last0 = min (last0 + context, files[0].valid_lines - 1); + last1 = min (last1 + context, files[1].valid_lines - 1); + + /* If desired, find the preceding function definition line in file 0. */ + function = 0; + if (function_regexp_list) + find_function (&files[0], first0, &function, &function_length); + + begin_output (); + out = outfile; + + fprintf (out, "@@ -"); + print_unidiff_number_range (&files[0], first0, last0); + fprintf (out, " +"); + print_unidiff_number_range (&files[1], first1, last1); + fprintf (out, " @@"); + + /* If we looked for and found a function this is part of, + include its name in the header of the diff section. */ + + if (function) + { + putc (' ', out); + fwrite (function, 1, min (function_length - 1, 40), out); + } + putc ('\n', out); + + next = hunk; + i = first0; + j = first1; + + while (i <= last0 || j <= last1) + { + + /* If the line isn't a difference, output the context from file 0. */ + + if (!next || i < next->line0) + { + putc (tab_align_flag ? '\t' : ' ', out); + print_1_line (0, &files[0].linbuf[i++]); + j++; + } + else + { + /* For each difference, first output the deleted part. */ + + k = next->deleted; + while (k--) + { + putc ('-', out); + if (tab_align_flag) + putc ('\t', out); + print_1_line (0, &files[0].linbuf[i++]); + } + + /* Then output the inserted part. */ + + k = next->inserted; + while (k--) + { + putc ('+', out); + if (tab_align_flag) + putc ('\t', out); + print_1_line (0, &files[1].linbuf[j++]); + } + + /* We're done with this hunk, so on to the next! */ + + next = next->link; + } + } +} + +/* Scan a (forward-ordered) edit script for the first place that more than + 2*CONTEXT unchanged lines appear, and return a pointer + to the `struct change' for the last change before those lines. */ + +static struct change * +find_hunk (start) + struct change *start; +{ + struct change *prev; + int top0, top1; + int thresh; + + do + { + /* Compute number of first line in each file beyond this changed. */ + top0 = start->line0 + start->deleted; + top1 = start->line1 + start->inserted; + prev = start; + start = start->link; + /* Threshold distance is 2*CONTEXT between two non-ignorable changes, + but only CONTEXT if one is ignorable. */ + thresh = ((prev->ignore || (start && start->ignore)) + ? context + : 2 * context + 1); + /* It is not supposed to matter which file we check in the end-test. + If it would matter, crash. */ + if (start && start->line0 - top0 != start->line1 - top1) + abort (); + } while (start + /* Keep going if less than THRESH lines + elapse before the affected line. */ + && start->line0 < top0 + thresh); + + return prev; +} + +/* Set the `ignore' flag properly in each change in SCRIPT. + It should be 1 if all the lines inserted or deleted in that change + are ignorable lines. */ + +static void +mark_ignorable (script) + struct change *script; +{ + while (script) + { + struct change *next = script->link; + int first0, last0, first1, last1, deletes, inserts; + + /* Turn this change into a hunk: detach it from the others. */ + script->link = 0; + + /* Determine whether this change is ignorable. */ + analyze_hunk (script, &first0, &last0, &first1, &last1, &deletes, &inserts); + /* Reconnect the chain as before. */ + script->link = next; + + /* If the change is ignorable, mark it. */ + script->ignore = (!deletes && !inserts); + + /* Advance to the following change. */ + script = next; + } +} + +/* Find the last function-header line in FILE prior to line number LINENUM. + This is a line containing a match for the regexp in `function_regexp'. + Store the address of the line text into LINEP and the length of the + line into LENP. + Do not store anything if no function-header is found. */ + +static void +find_function (file, linenum, linep, lenp) + struct file_data const *file; + int linenum; + char const **linep; + size_t *lenp; +{ + int i = linenum; + int last = find_function_last_search; + find_function_last_search = i; + + while (--i >= last) + { + /* See if this line is what we want. */ + struct regexp_list *r; + char const *line = file->linbuf[i]; + size_t len = file->linbuf[i + 1] - line; + + for (r = function_regexp_list; r; r = r->next) + if (0 <= re_search (&r->buf, line, len, 0, len, 0)) + { + *linep = line; + *lenp = len; + find_function_last_match = i; + return; + } + } + /* If we search back to where we started searching the previous time, + find the line we found last time. */ + if (find_function_last_match >= - file->prefix_lines) + { + i = find_function_last_match; + *linep = file->linbuf[i]; + *lenp = file->linbuf[i + 1] - *linep; + return; + } + return; +} diff --git a/sys/src/ape/cmd/diff/diagmeet.note b/sys/src/ape/cmd/diff/diagmeet.note new file mode 100755 index 000000000..8f7242c73 --- /dev/null +++ b/sys/src/ape/cmd/diff/diagmeet.note @@ -0,0 +1,71 @@ +Here is a comparison matrix which shows a case in which +it is possible for the forward and backward scan in `diag' +to meet along a nonzero length of diagonal simultaneous +(so that bdiag[d] and fdiag[d] are not equal) +even though there is no snake on that diagonal at the meeting point. + + + 85 1 1 1 159 1 1 17 + 1 2 3 4 +60 + 1 2 +1 + 2 2 3 4 +71 + 3 3 4 5 +85 + 4 3 4 5 +17 + 5 4 5 +1 + 6 4 5 6 +183 + 7 5 6 7 +10 + 8 6 7 +1 + 9 6 7 8 +12 + 7 8 9 10 +13 + 10 8 9 10 +14 + 10 9 10 +17 + 10 10 +1 + 10 9 10 +1 + 8 10 10 10 +183 + 8 7 9 9 9 +10 + 7 6 8 9 8 8 +1 + 6 5 7 7 +1 + 5 6 6 +1 + 5 5 5 +50 + 5 4 4 4 +1 + 4 3 3 +85 + 5 4 3 2 2 +1 + 2 1 +17 + 5 4 3 2 1 1 +1 + 1 0 + 85 1 1 1 159 1 1 17 + + + + + + + + + diff --git a/sys/src/ape/cmd/diff/diff.c b/sys/src/ape/cmd/diff/diff.c new file mode 100755 index 000000000..f9e49aa7c --- /dev/null +++ b/sys/src/ape/cmd/diff/diff.c @@ -0,0 +1,1124 @@ +/* GNU DIFF main routine. + Copyright (C) 1988, 1989, 1992, 1993, 1994 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* GNU DIFF was written by Mike Haertel, David Hayes, + Richard Stallman, Len Tower, and Paul Eggert. */ + +/* $FreeBSD: src/contrib/diff/diff.c,v 1.3 1999/11/26 02:51:44 obrien Exp $ */ + +#define GDIFF_MAIN +#include "diff.h" +#include <signal.h> +#include "getopt.h" +#ifdef __FreeBSD__ +#include <locale.h> +#include <fnmatch.h> +#else +#include "fnmatch.h" +#endif +#include "prepend_args.h" + +#ifndef DEFAULT_WIDTH +#define DEFAULT_WIDTH 130 +#endif + +#ifndef GUTTER_WIDTH_MINIMUM +#define GUTTER_WIDTH_MINIMUM 3 +#endif + +static char const *filetype PARAMS((struct stat const *)); +static char *option_list PARAMS((char **, int)); +static int add_exclude_file PARAMS((char const *)); +static int ck_atoi PARAMS((char const *, int *)); +static int compare_files PARAMS((char const *, char const *, char const *, char const *, int)); +static int specify_format PARAMS((char **, char *)); +static void add_exclude PARAMS((char const *)); +static void add_regexp PARAMS((struct regexp_list **, char const *)); +static void specify_style PARAMS((enum output_style)); +static void try_help PARAMS((char const *)); +static void check_stdout PARAMS((void)); +static void usage PARAMS((void)); + +/* Nonzero for -r: if comparing two directories, + compare their common subdirectories recursively. */ + +static int recursive; + +/* For debugging: don't do discard_confusing_lines. */ + +int no_discards; + +#if HAVE_SETMODE +/* I/O mode: nonzero only if using binary input/output. */ +static int binary_I_O; +#endif + +/* Return a string containing the command options with which diff was invoked. + Spaces appear between what were separate ARGV-elements. + There is a space at the beginning but none at the end. + If there were no options, the result is an empty string. + + Arguments: OPTIONVEC, a vector containing separate ARGV-elements, and COUNT, + the length of that vector. */ + +static char * +option_list (optionvec, count) + char **optionvec; /* Was `vector', but that collides on Alliant. */ + int count; +{ + int i; + size_t length = 0; + char *result; + + for (i = 0; i < count; i++) + length += strlen (optionvec[i]) + 1; + + result = xmalloc (length + 1); + result[0] = 0; + + for (i = 0; i < count; i++) + { + strcat (result, " "); + strcat (result, optionvec[i]); + } + + return result; +} + +/* Convert STR to a positive integer, storing the result in *OUT. + If STR is not a valid integer, return -1 (otherwise 0). */ +static int +ck_atoi (str, out) + char const *str; + int *out; +{ + char const *p; + for (p = str; *p; p++) + if (*p < '0' || *p > '9') + return -1; + + *out = atoi (optarg); + return 0; +} + +/* Keep track of excluded file name patterns. */ + +static char const **exclude; +static int exclude_alloc, exclude_count; + +int +excluded_filename (f) + char const *f; +{ + int i; + for (i = 0; i < exclude_count; i++) + if (fnmatch (exclude[i], f, 0) == 0) + return 1; + return 0; +} + +static void +add_exclude (pattern) + char const *pattern; +{ + if (exclude_alloc <= exclude_count) + exclude = (char const **) + (exclude_alloc == 0 + ? xmalloc ((exclude_alloc = 64) * sizeof (*exclude)) + : xrealloc (exclude, (exclude_alloc *= 2) * sizeof (*exclude))); + + exclude[exclude_count++] = pattern; +} + +static int +add_exclude_file (name) + char const *name; +{ + struct file_data f; + char *p, *q, *lim; + + f.name = optarg; + f.desc = (strcmp (name, "-") == 0 + ? STDIN_FILENO + : open (name, O_RDONLY, 0)); + if (f.desc < 0 || fstat (f.desc, &f.stat) != 0) + return -1; + + sip (&f, 1); + slurp (&f); + + for (p = f.buffer, lim = p + f.buffered_chars; p < lim; p = q) + { + q = (char *) memchr (p, '\n', lim - p); + if (!q) + q = lim; + *q++ = 0; + add_exclude (p); + } + + return close (f.desc); +} + +/* The numbers 129- that appear in the fourth element of some entries + tell the big switch in `main' how to process those options. */ + +static struct option const longopts[] = +{ + {"ignore-blank-lines", 0, 0, 'B'}, + {"context", 2, 0, 'C'}, + {"ifdef", 1, 0, 'D'}, + {"show-function-line", 1, 0, 'F'}, + {"speed-large-files", 0, 0, 'H'}, + {"ignore-matching-lines", 1, 0, 'I'}, + {"label", 1, 0, 'L'}, + {"file-label", 1, 0, 'L'}, /* An alias, no longer recommended */ + {"new-file", 0, 0, 'N'}, + {"entire-new-file", 0, 0, 'N'}, /* An alias, no longer recommended */ + {"unidirectional-new-file", 0, 0, 'P'}, + {"starting-file", 1, 0, 'S'}, + {"initial-tab", 0, 0, 'T'}, + {"width", 1, 0, 'W'}, + {"text", 0, 0, 'a'}, + {"ascii", 0, 0, 'a'}, /* An alias, no longer recommended */ + {"ignore-space-change", 0, 0, 'b'}, + {"minimal", 0, 0, 'd'}, + {"ed", 0, 0, 'e'}, + {"forward-ed", 0, 0, 'f'}, + {"ignore-case", 0, 0, 'i'}, + {"paginate", 0, 0, 'l'}, + {"print", 0, 0, 'l'}, /* An alias, no longer recommended */ + {"rcs", 0, 0, 'n'}, + {"show-c-function", 0, 0, 'p'}, + {"brief", 0, 0, 'q'}, + {"recursive", 0, 0, 'r'}, + {"report-identical-files", 0, 0, 's'}, + {"expand-tabs", 0, 0, 't'}, + {"version", 0, 0, 'v'}, + {"ignore-all-space", 0, 0, 'w'}, + {"exclude", 1, 0, 'x'}, + {"exclude-from", 1, 0, 'X'}, + {"side-by-side", 0, 0, 'y'}, + {"unified", 2, 0, 'U'}, + {"left-column", 0, 0, 129}, + {"suppress-common-lines", 0, 0, 130}, + {"sdiff-merge-assist", 0, 0, 131}, + {"old-line-format", 1, 0, 132}, + {"new-line-format", 1, 0, 133}, + {"unchanged-line-format", 1, 0, 134}, + {"line-format", 1, 0, 135}, + {"old-group-format", 1, 0, 136}, + {"new-group-format", 1, 0, 137}, + {"unchanged-group-format", 1, 0, 138}, + {"changed-group-format", 1, 0, 139}, + {"horizon-lines", 1, 0, 140}, + {"help", 0, 0, 141}, + {"binary", 0, 0, 142}, + {0, 0, 0, 0} +}; + +int +main (argc, argv) + int argc; + char *argv[]; +{ + int val; + int c; + int prev = -1; + int width = DEFAULT_WIDTH; + int show_c_function = 0; + +#ifdef __FreeBSD__ + setlocale(LC_ALL, ""); +#endif + /* Do our initializations. */ + initialize_main (&argc, &argv); + program_name = argv[0]; + output_style = OUTPUT_NORMAL; + context = -1; + + prepend_default_options (getenv ("DIFF_OPTIONS"), &argc, &argv); + + /* Decode the options. */ + + while ((c = getopt_long (argc, argv, + "0123456789abBcC:dD:efF:hHiI:lL:nNopPqrsS:tTuU:vwW:x:X:y", + longopts, 0)) != EOF) + { + switch (c) + { + /* All digits combine in decimal to specify the context-size. */ + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + case '0': + if (context == -1) + context = 0; + /* If a context length has already been specified, + more digits allowed only if they follow right after the others. + Reject two separate runs of digits, or digits after -C. */ + else if (prev < '0' || prev > '9') + fatal ("context length specified twice"); + + context = context * 10 + c - '0'; + break; + + case 'a': + /* Treat all files as text files; never treat as binary. */ + always_text_flag = 1; + break; + + case 'b': + /* Ignore changes in amount of white space. */ + ignore_space_change_flag = 1; + ignore_some_changes = 1; + ignore_some_line_changes = 1; + break; + + case 'B': + /* Ignore changes affecting only blank lines. */ + ignore_blank_lines_flag = 1; + ignore_some_changes = 1; + break; + + case 'C': /* +context[=lines] */ + case 'U': /* +unified[=lines] */ + if (optarg) + { + if (context >= 0) + fatal ("context length specified twice"); + + if (ck_atoi (optarg, &context)) + fatal ("invalid context length argument"); + } + + /* Falls through. */ + case 'c': + /* Make context-style output. */ + specify_style (c == 'U' ? OUTPUT_UNIFIED : OUTPUT_CONTEXT); + break; + + case 'd': + /* Don't discard lines. This makes things slower (sometimes much + slower) but will find a guaranteed minimal set of changes. */ + no_discards = 1; + break; + + case 'D': + /* Make merged #ifdef output. */ + specify_style (OUTPUT_IFDEF); + { + int i, err = 0; + static char const C_ifdef_group_formats[] = + "#ifndef %s\n%%<#endif /* not %s */\n%c#ifdef %s\n%%>#endif /* %s */\n%c%%=%c#ifndef %s\n%%<#else /* %s */\n%%>#endif /* %s */\n"; + char *b = xmalloc (sizeof (C_ifdef_group_formats) + + 7 * strlen(optarg) - 14 /* 7*"%s" */ + - 8 /* 5*"%%" + 3*"%c" */); + sprintf (b, C_ifdef_group_formats, + optarg, optarg, 0, + optarg, optarg, 0, 0, + optarg, optarg, optarg); + for (i = 0; i < 4; i++) + { + err |= specify_format (&group_format[i], b); + b += strlen (b) + 1; + } + if (err) + error ("conflicting #ifdef formats", 0, 0); + } + break; + + case 'e': + /* Make output that is a valid `ed' script. */ + specify_style (OUTPUT_ED); + break; + + case 'f': + /* Make output that looks vaguely like an `ed' script + but has changes in the order they appear in the file. */ + specify_style (OUTPUT_FORWARD_ED); + break; + + case 'F': + /* Show, for each set of changes, the previous line that + matches the specified regexp. Currently affects only + context-style output. */ + add_regexp (&function_regexp_list, optarg); + break; + + case 'h': + /* Split the files into chunks of around 1500 lines + for faster processing. Usually does not change the result. + + This currently has no effect. */ + break; + + case 'H': + /* Turn on heuristics that speed processing of large files + with a small density of changes. */ + heuristic = 1; + break; + + case 'i': + /* Ignore changes in case. */ + ignore_case_flag = 1; + ignore_some_changes = 1; + ignore_some_line_changes = 1; + break; + + case 'I': + /* Ignore changes affecting only lines that match the + specified regexp. */ + add_regexp (&ignore_regexp_list, optarg); + ignore_some_changes = 1; + break; + + case 'l': + /* Pass the output through `pr' to paginate it. */ + paginate_flag = 1; +#if !defined(SIGCHLD) && defined(SIGCLD) +#define SIGCHLD SIGCLD +#endif +#ifdef SIGCHLD + /* Pagination requires forking and waiting, and + System V fork+wait does not work if SIGCHLD is ignored. */ + signal (SIGCHLD, SIG_DFL); +#endif + break; + + case 'L': + /* Specify file labels for `-c' output headers. */ + if (!file_label[0]) + file_label[0] = optarg; + else if (!file_label[1]) + file_label[1] = optarg; + else + fatal ("too many file label options"); + break; + + case 'n': + /* Output RCS-style diffs, like `-f' except that each command + specifies the number of lines affected. */ + specify_style (OUTPUT_RCS); + break; + + case 'N': + /* When comparing directories, if a file appears only in one + directory, treat it as present but empty in the other. */ + entire_new_file_flag = 1; + break; + + case 'o': + /* Output in the old tradition style. */ + specify_style (OUTPUT_NORMAL); + break; + + case 'p': + /* Make context-style output and show name of last C function. */ + show_c_function = 1; + add_regexp (&function_regexp_list, "^[_a-zA-Z$]"); + break; + + case 'P': + /* When comparing directories, if a file appears only in + the second directory of the two, + treat it as present but empty in the other. */ + unidirectional_new_file_flag = 1; + break; + + case 'q': + no_details_flag = 1; + break; + + case 'r': + /* When comparing directories, + recursively compare any subdirectories found. */ + recursive = 1; + break; + + case 's': + /* Print a message if the files are the same. */ + print_file_same_flag = 1; + break; + + case 'S': + /* When comparing directories, start with the specified + file name. This is used for resuming an aborted comparison. */ + dir_start_file = optarg; + break; + + case 't': + /* Expand tabs to spaces in the output so that it preserves + the alignment of the input files. */ + tab_expand_flag = 1; + break; + + case 'T': + /* Use a tab in the output, rather than a space, before the + text of an input line, so as to keep the proper alignment + in the input line without changing the characters in it. */ + tab_align_flag = 1; + break; + + case 'u': + /* Output the context diff in unidiff format. */ + specify_style (OUTPUT_UNIFIED); + break; + + case 'v': + printf ("diff - GNU diffutils version %s\n", version_string); + exit (0); + + case 'w': + /* Ignore horizontal white space when comparing lines. */ + ignore_all_space_flag = 1; + ignore_some_changes = 1; + ignore_some_line_changes = 1; + break; + + case 'x': + add_exclude (optarg); + break; + + case 'X': + if (add_exclude_file (optarg) != 0) + pfatal_with_name (optarg); + break; + + case 'y': + /* Use side-by-side (sdiff-style) columnar output. */ + specify_style (OUTPUT_SDIFF); + break; + + case 'W': + /* Set the line width for OUTPUT_SDIFF. */ + if (ck_atoi (optarg, &width) || width <= 0) + fatal ("column width must be a positive integer"); + break; + + case 129: + sdiff_left_only = 1; + break; + + case 130: + sdiff_skip_common_lines = 1; + break; + + case 131: + /* sdiff-style columns output. */ + specify_style (OUTPUT_SDIFF); + sdiff_help_sdiff = 1; + break; + + case 132: + case 133: + case 134: + specify_style (OUTPUT_IFDEF); + if (specify_format (&line_format[c - 132], optarg) != 0) + error ("conflicting line format", 0, 0); + break; + + case 135: + specify_style (OUTPUT_IFDEF); + { + int i, err = 0; + for (i = 0; i < sizeof (line_format) / sizeof (*line_format); i++) + err |= specify_format (&line_format[i], optarg); + if (err) + error ("conflicting line format", 0, 0); + } + break; + + case 136: + case 137: + case 138: + case 139: + specify_style (OUTPUT_IFDEF); + if (specify_format (&group_format[c - 136], optarg) != 0) + error ("conflicting group format", 0, 0); + break; + + case 140: + if (ck_atoi (optarg, &horizon_lines) || horizon_lines < 0) + fatal ("horizon must be a nonnegative integer"); + break; + + case 141: + usage (); + check_stdout (); + exit (0); + + case 142: + /* Use binary I/O when reading and writing data. + On Posix hosts, this has no effect. */ +#if HAVE_SETMODE + binary_I_O = 1; + setmode (STDOUT_FILENO, O_BINARY); +#endif + break; + + default: + try_help (0); + } + prev = c; + } + + if (argc - optind != 2) + try_help (argc - optind < 2 ? "missing operand" : "extra operand"); + + + { + /* + * We maximize first the half line width, and then the gutter width, + * according to the following constraints: + * 1. Two half lines plus a gutter must fit in a line. + * 2. If the half line width is nonzero: + * a. The gutter width is at least GUTTER_WIDTH_MINIMUM. + * b. If tabs are not expanded to spaces, + * a half line plus a gutter is an integral number of tabs, + * so that tabs in the right column line up. + */ + int t = tab_expand_flag ? 1 : TAB_WIDTH; + int off = (width + t + GUTTER_WIDTH_MINIMUM) / (2*t) * t; + sdiff_half_width = max (0, min (off - GUTTER_WIDTH_MINIMUM, width - off)), + sdiff_column2_offset = sdiff_half_width ? off : width; + } + + if (show_c_function && output_style != OUTPUT_UNIFIED) + specify_style (OUTPUT_CONTEXT); + + if (output_style != OUTPUT_CONTEXT && output_style != OUTPUT_UNIFIED) + context = 0; + else if (context == -1) + /* Default amount of context for -c. */ + context = 3; + + if (output_style == OUTPUT_IFDEF) + { + /* Format arrays are char *, not char const *, + because integer formats are temporarily modified. + But it is safe to assign a constant like "%=" to a format array, + since "%=" does not format any integers. */ + int i; + for (i = 0; i < sizeof (line_format) / sizeof (*line_format); i++) + if (!line_format[i]) + line_format[i] = "%l\n"; + if (!group_format[OLD]) + group_format[OLD] + = group_format[UNCHANGED] ? group_format[UNCHANGED] : "%<"; + if (!group_format[NEW]) + group_format[NEW] + = group_format[UNCHANGED] ? group_format[UNCHANGED] : "%>"; + if (!group_format[UNCHANGED]) + group_format[UNCHANGED] = "%="; + if (!group_format[CHANGED]) + group_format[CHANGED] = concat (group_format[OLD], + group_format[NEW], ""); + } + + no_diff_means_no_output = + (output_style == OUTPUT_IFDEF ? + (!*group_format[UNCHANGED] + || (strcmp (group_format[UNCHANGED], "%=") == 0 + && !*line_format[UNCHANGED])) + : output_style == OUTPUT_SDIFF ? sdiff_skip_common_lines : 1); + + switch_string = option_list (argv + 1, optind - 1); + + val = compare_files (0, argv[optind], 0, argv[optind + 1], 0); + + /* Print any messages that were saved up for last. */ + print_message_queue (); + + check_stdout (); + exit (val); + return val; +} + +/* Add the compiled form of regexp PATTERN to REGLIST. */ + +static void +add_regexp (reglist, pattern) + struct regexp_list **reglist; + char const *pattern; +{ + struct regexp_list *r; + char const *m; + + r = (struct regexp_list *) xmalloc (sizeof (*r)); + bzero (r, sizeof (*r)); + r->buf.fastmap = xmalloc (256); + m = re_compile_pattern (pattern, strlen (pattern), &r->buf); + if (m != 0) + error ("%s: %s", pattern, m); + + /* Add to the start of the list, since it's easier than the end. */ + r->next = *reglist; + *reglist = r; +} + +static void +try_help (reason) + char const *reason; +{ + if (reason) + error ("%s", reason, 0); + error ("Try `%s --help' for more information.", program_name, 0); + exit (2); +} + +static void +check_stdout () +{ + if (ferror (stdout) || fclose (stdout) != 0) + fatal ("write error"); +} + +static char const * const option_help[] = { +"-i --ignore-case Consider upper- and lower-case to be the same.", +"-w --ignore-all-space Ignore all white space.", +"-b --ignore-space-change Ignore changes in the amount of white space.", +"-B --ignore-blank-lines Ignore changes whose lines are all blank.", +"-I RE --ignore-matching-lines=RE Ignore changes whose lines all match RE.", +#if HAVE_SETMODE +"--binary Read and write data in binary mode.", +#endif +"-a --text Treat all files as text.\n", +"-c -C NUM --context[=NUM] Output NUM (default 2) lines of copied context.", +"-u -U NUM --unified[=NUM] Output NUM (default 2) lines of unified context.", +" -NUM Use NUM context lines.", +" -L LABEL --label LABEL Use LABEL instead of file name.", +" -p --show-c-function Show which C function each change is in.", +" -F RE --show-function-line=RE Show the most recent line matching RE.", +"-q --brief Output only whether files differ.", +"-e --ed Output an ed script.", +"-n --rcs Output an RCS format diff.", +"-y --side-by-side Output in two columns.", +" -w NUM --width=NUM Output at most NUM (default 130) characters per line.", +" --left-column Output only the left column of common lines.", +" --suppress-common-lines Do not output common lines.", +"-DNAME --ifdef=NAME Output merged file to show `#ifdef NAME' diffs.", +"--GTYPE-group-format=GFMT Similar, but format GTYPE input groups with GFMT.", +"--line-format=LFMT Similar, but format all input lines with LFMT.", +"--LTYPE-line-format=LFMT Similar, but format LTYPE input lines with LFMT.", +" LTYPE is `old', `new', or `unchanged'. GTYPE is LTYPE or `changed'.", +" GFMT may contain:", +" %< lines from FILE1", +" %> lines from FILE2", +" %= lines common to FILE1 and FILE2", +" %[-][WIDTH][.[PREC]]{doxX}LETTER printf-style spec for LETTER", +" LETTERs are as follows for new group, lower case for old group:", +" F first line number", +" L last line number", +" N number of lines = L-F+1", +" E F-1", +" M L+1", +" LFMT may contain:", +" %L contents of line", +" %l contents of line, excluding any trailing newline", +" %[-][WIDTH][.[PREC]]{doxX}n printf-style spec for input line number", +" Either GFMT or LFMT may contain:", +" %% %", +" %c'C' the single character C", +" %c'\\OOO' the character with octal code OOO\n", +"-l --paginate Pass the output through `pr' to paginate it.", +"-t --expand-tabs Expand tabs to spaces in output.", +"-T --initial-tab Make tabs line up by prepending a tab.\n", +"-r --recursive Recursively compare any subdirectories found.", +"-N --new-file Treat absent files as empty.", +"-P --unidirectional-new-file Treat absent first files as empty.", +"-s --report-identical-files Report when two files are the same.", +"-x PAT --exclude=PAT Exclude files that match PAT.", +"-X FILE --exclude-from=FILE Exclude files that match any pattern in FILE.", +"-S FILE --starting-file=FILE Start with FILE when comparing directories.\n", +"--horizon-lines=NUM Keep NUM lines of the common prefix and suffix.", +"-d --minimal Try hard to find a smaller set of changes.", +"-H --speed-large-files Assume large files and many scattered small changes.\n", +"-v --version Output version info.", +"--help Output this help.", +0 +}; + +static void +usage () +{ + char const * const *p; + + printf ("Usage: %s [OPTION]... FILE1 FILE2\n\n", program_name); + for (p = option_help; *p; p++) + printf (" %s\n", *p); + printf ("\nIf FILE1 or FILE2 is `-', read standard input.\n"); +} + +static int +specify_format (var, value) + char **var; + char *value; +{ + int err = *var ? strcmp (*var, value) : 0; + *var = value; + return err; +} + +static void +specify_style (style) + enum output_style style; +{ + if (output_style != OUTPUT_NORMAL + && output_style != style) + error ("conflicting specifications of output style", 0, 0); + output_style = style; +} + +static char const * +filetype (st) + struct stat const *st; +{ + /* See Posix.2 section 4.17.6.1.1 and Table 5-1 for these formats. + To keep diagnostics grammatical, the returned string must start + with a consonant. */ + + if (S_ISREG (st->st_mode)) + { + if (st->st_size == 0) + return "regular empty file"; + /* Posix.2 section 5.14.2 seems to suggest that we must read the file + and guess whether it's C, Fortran, etc., but this is somewhat useless + and doesn't reflect historical practice. We're allowed to guess + wrong, so we don't bother to read the file. */ + return "regular file"; + } + if (S_ISDIR (st->st_mode)) return "directory"; + + /* other Posix.1 file types */ +#ifdef S_ISBLK + if (S_ISBLK (st->st_mode)) return "block special file"; +#endif +#ifdef S_ISCHR + if (S_ISCHR (st->st_mode)) return "character special file"; +#endif +#ifdef S_ISFIFO + if (S_ISFIFO (st->st_mode)) return "fifo"; +#endif + + /* other Posix.1b file types */ +#ifdef S_TYPEISMQ + if (S_TYPEISMQ (st)) return "message queue"; +#endif +#ifdef S_TYPEISSEM + if (S_TYPEISSEM (st)) return "semaphore"; +#endif +#ifdef S_TYPEISSHM + if (S_TYPEISSHM (st)) return "shared memory object"; +#endif + + /* other popular file types */ + /* S_ISLNK is impossible with `fstat' and `stat'. */ +#ifdef S_ISSOCK + if (S_ISSOCK (st->st_mode)) return "socket"; +#endif + + return "weird file"; +} + +/* Compare two files (or dirs) with specified names + DIR0/NAME0 and DIR1/NAME1, at level DEPTH in directory recursion. + (if DIR0 is 0, then the name is just NAME0, etc.) + This is self-contained; it opens the files and closes them. + + Value is 0 if files are the same, 1 if different, + 2 if there is a problem opening them. */ + +static int +compare_files (dir0, name0, dir1, name1, depth) + char const *dir0, *dir1; + char const *name0, *name1; + int depth; +{ + struct file_data inf[2]; + register int i; + int val; + int same_files; + int failed = 0; + char *free0 = 0, *free1 = 0; + + /* If this is directory comparison, perhaps we have a file + that exists only in one of the directories. + If so, just print a message to that effect. */ + + if (! ((name0 != 0 && name1 != 0) + || (unidirectional_new_file_flag && name1 != 0) + || entire_new_file_flag)) + { + char const *name = name0 == 0 ? name1 : name0; + char const *dir = name0 == 0 ? dir1 : dir0; + message ("Only in %s: %s\n", dir, name); + /* Return 1 so that diff_dirs will return 1 ("some files differ"). */ + return 1; + } + + bzero (inf, sizeof (inf)); + + /* Mark any nonexistent file with -1 in the desc field. */ + /* Mark unopened files (e.g. directories) with -2. */ + + inf[0].desc = name0 == 0 ? -1 : -2; + inf[1].desc = name1 == 0 ? -1 : -2; + + /* Now record the full name of each file, including nonexistent ones. */ + + if (name0 == 0) + name0 = name1; + if (name1 == 0) + name1 = name0; + + inf[0].name = dir0 == 0 ? name0 : (free0 = dir_file_pathname (dir0, name0)); + inf[1].name = dir1 == 0 ? name1 : (free1 = dir_file_pathname (dir1, name1)); + + /* Stat the files. Record whether they are directories. */ + + for (i = 0; i <= 1; i++) + { + if (inf[i].desc != -1) + { + int stat_result; + + if (i && filename_cmp (inf[i].name, inf[0].name) == 0) + { + inf[i].stat = inf[0].stat; + stat_result = 0; + } + else if (strcmp (inf[i].name, "-") == 0) + { + inf[i].desc = STDIN_FILENO; + stat_result = fstat (STDIN_FILENO, &inf[i].stat); + if (stat_result == 0 && S_ISREG (inf[i].stat.st_mode)) + { + off_t pos = lseek (STDIN_FILENO, (off_t) 0, SEEK_CUR); + if (pos == -1) + stat_result = -1; + else + { + if (pos <= inf[i].stat.st_size) + inf[i].stat.st_size -= pos; + else + inf[i].stat.st_size = 0; + /* Posix.2 4.17.6.1.4 requires current time for stdin. */ + time (&inf[i].stat.st_mtime); + } + } + } + else + stat_result = stat (inf[i].name, &inf[i].stat); + + if (stat_result != 0) + { + perror_with_name (inf[i].name); + failed = 1; + } + else + { + inf[i].dir_p = S_ISDIR (inf[i].stat.st_mode) && inf[i].desc != 0; + if (inf[1 - i].desc == -1) + { + inf[1 - i].dir_p = inf[i].dir_p; + inf[1 - i].stat.st_mode = inf[i].stat.st_mode; + } + } + } + } + + if (! failed && depth == 0 && inf[0].dir_p != inf[1].dir_p) + { + /* If one is a directory, and it was specified in the command line, + use the file in that dir with the other file's basename. */ + + int fnm_arg = inf[0].dir_p; + int dir_arg = 1 - fnm_arg; + char const *fnm = inf[fnm_arg].name; + char const *dir = inf[dir_arg].name; + char const *p = filename_lastdirchar (fnm); + char const *filename = inf[dir_arg].name + = dir_file_pathname (dir, p ? p + 1 : fnm); + + if (strcmp (fnm, "-") == 0) + fatal ("can't compare - to a directory"); + + if (stat (filename, &inf[dir_arg].stat) != 0) + { + perror_with_name (filename); + failed = 1; + } + else + inf[dir_arg].dir_p = S_ISDIR (inf[dir_arg].stat.st_mode); + } + + if (failed) + { + + /* If either file should exist but does not, return 2. */ + + val = 2; + + } + else if ((same_files = inf[0].desc != -1 && inf[1].desc != -1 + && 0 < same_file (&inf[0].stat, &inf[1].stat)) + && no_diff_means_no_output) + { + /* The two named files are actually the same physical file. + We know they are identical without actually reading them. */ + + val = 0; + } + else if (inf[0].dir_p & inf[1].dir_p) + { + if (output_style == OUTPUT_IFDEF) + fatal ("-D option not supported with directories"); + + /* If both are directories, compare the files in them. */ + + if (depth > 0 && !recursive) + { + /* But don't compare dir contents one level down + unless -r was specified. */ + message ("Common subdirectories: %s and %s\n", + inf[0].name, inf[1].name); + val = 0; + } + else + { + val = diff_dirs (inf, compare_files, depth); + } + + } + else if ((inf[0].dir_p | inf[1].dir_p) + || (depth > 0 + && (! S_ISREG (inf[0].stat.st_mode) + || ! S_ISREG (inf[1].stat.st_mode)))) + { + /* Perhaps we have a subdirectory that exists only in one directory. + If so, just print a message to that effect. */ + + if (inf[0].desc == -1 || inf[1].desc == -1) + { + if ((inf[0].dir_p | inf[1].dir_p) + && recursive + && (entire_new_file_flag + || (unidirectional_new_file_flag && inf[0].desc == -1))) + val = diff_dirs (inf, compare_files, depth); + else + { + char const *dir = (inf[0].desc == -1) ? dir1 : dir0; + /* See Posix.2 section 4.17.6.1.1 for this format. */ + message ("Only in %s: %s\n", dir, name0); + val = 1; + } + } + else + { + /* We have two files that are not to be compared. */ + + /* See Posix.2 section 4.17.6.1.1 for this format. */ + message5 ("File %s is a %s while file %s is a %s\n", + inf[0].name, filetype (&inf[0].stat), + inf[1].name, filetype (&inf[1].stat)); + + /* This is a difference. */ + val = 1; + } + } + else if ((no_details_flag & ~ignore_some_changes) + && inf[0].stat.st_size != inf[1].stat.st_size + && (inf[0].desc == -1 || S_ISREG (inf[0].stat.st_mode)) + && (inf[1].desc == -1 || S_ISREG (inf[1].stat.st_mode))) + { + message ("Files %s and %s differ\n", inf[0].name, inf[1].name); + val = 1; + } + else + { + /* Both exist and neither is a directory. */ + + /* Open the files and record their descriptors. */ + + if (inf[0].desc == -2) + if ((inf[0].desc = open (inf[0].name, O_RDONLY, 0)) < 0) + { + perror_with_name (inf[0].name); + failed = 1; + } + if (inf[1].desc == -2) + if (same_files) + inf[1].desc = inf[0].desc; + else if ((inf[1].desc = open (inf[1].name, O_RDONLY, 0)) < 0) + { + perror_with_name (inf[1].name); + failed = 1; + } + +#if HAVE_SETMODE + if (binary_I_O) + for (i = 0; i <= 1; i++) + if (0 <= inf[i].desc) + setmode (inf[i].desc, O_BINARY); +#endif + + /* Compare the files, if no error was found. */ + + val = failed ? 2 : diff_2_files (inf, depth); + + /* Close the file descriptors. */ + + if (inf[0].desc >= 0 && close (inf[0].desc) != 0) + { + perror_with_name (inf[0].name); + val = 2; + } + if (inf[1].desc >= 0 && inf[0].desc != inf[1].desc + && close (inf[1].desc) != 0) + { + perror_with_name (inf[1].name); + val = 2; + } + } + + /* Now the comparison has been done, if no error prevented it, + and VAL is the value this function will return. */ + + if (val == 0 && !inf[0].dir_p) + { + if (print_file_same_flag) + message ("Files %s and %s are identical\n", + inf[0].name, inf[1].name); + } + else + fflush (stdout); + + if (free0) + free (free0); + if (free1) + free (free1); + + return val; +} diff --git a/sys/src/ape/cmd/diff/diff.h b/sys/src/ape/cmd/diff/diff.h new file mode 100755 index 000000000..24f088598 --- /dev/null +++ b/sys/src/ape/cmd/diff/diff.h @@ -0,0 +1,344 @@ +/* Shared definitions for GNU DIFF + Copyright (C) 1988, 89, 91, 92, 93 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#include "system.h" +#include <stdio.h> +#ifdef __FreeBSD__ +#include <gnuregex.h> +#else +#include "regex.h" +#endif + +#define TAB_WIDTH 8 + +/* Variables for command line options */ + +#ifndef GDIFF_MAIN +#define EXTERN extern +#else +#define EXTERN +#endif + +enum output_style { + /* Default output style. */ + OUTPUT_NORMAL, + /* Output the differences with lines of context before and after (-c). */ + OUTPUT_CONTEXT, + /* Output the differences in a unified context diff format (-u). */ + OUTPUT_UNIFIED, + /* Output the differences as commands suitable for `ed' (-e). */ + OUTPUT_ED, + /* Output the diff as a forward ed script (-f). */ + OUTPUT_FORWARD_ED, + /* Like -f, but output a count of changed lines in each "command" (-n). */ + OUTPUT_RCS, + /* Output merged #ifdef'd file (-D). */ + OUTPUT_IFDEF, + /* Output sdiff style (-y). */ + OUTPUT_SDIFF +}; + +/* True for output styles that are robust, + i.e. can handle a file that ends in a non-newline. */ +#define ROBUST_OUTPUT_STYLE(S) ((S) != OUTPUT_ED && (S) != OUTPUT_FORWARD_ED) + +EXTERN enum output_style output_style; + +/* Nonzero if output cannot be generated for identical files. */ +EXTERN int no_diff_means_no_output; + +/* Number of lines of context to show in each set of diffs. + This is zero when context is not to be shown. */ +EXTERN int context; + +/* Consider all files as text files (-a). + Don't interpret codes over 0177 as implying a "binary file". */ +EXTERN int always_text_flag; + +/* Number of lines to keep in identical prefix and suffix. */ +EXTERN int horizon_lines; + +/* Ignore changes in horizontal white space (-b). */ +EXTERN int ignore_space_change_flag; + +/* Ignore all horizontal white space (-w). */ +EXTERN int ignore_all_space_flag; + +/* Ignore changes that affect only blank lines (-B). */ +EXTERN int ignore_blank_lines_flag; + +/* 1 if lines may match even if their contents do not match exactly. + This depends on various options. */ +EXTERN int ignore_some_line_changes; + +/* 1 if files may match even if their contents are not byte-for-byte identical. + This depends on various options. */ +EXTERN int ignore_some_changes; + +/* Ignore differences in case of letters (-i). */ +EXTERN int ignore_case_flag; + +/* File labels for `-c' output headers (-L). */ +EXTERN char *file_label[2]; + +struct regexp_list +{ + struct re_pattern_buffer buf; + struct regexp_list *next; +}; + +/* Regexp to identify function-header lines (-F). */ +EXTERN struct regexp_list *function_regexp_list; + +/* Ignore changes that affect only lines matching this regexp (-I). */ +EXTERN struct regexp_list *ignore_regexp_list; + +/* Say only whether files differ, not how (-q). */ +EXTERN int no_details_flag; + +/* Report files compared that match (-s). + Normally nothing is output when that happens. */ +EXTERN int print_file_same_flag; + +/* Output the differences with exactly 8 columns added to each line + so that any tabs in the text line up properly (-T). */ +EXTERN int tab_align_flag; + +/* Expand tabs in the output so the text lines up properly + despite the characters added to the front of each line (-t). */ +EXTERN int tab_expand_flag; + +/* In directory comparison, specify file to start with (-S). + All file names less than this name are ignored. */ +EXTERN char *dir_start_file; + +/* If a file is new (appears in only one dir) + include its entire contents (-N). + Then `patch' would create the file with appropriate contents. */ +EXTERN int entire_new_file_flag; + +/* If a file is new (appears in only the second dir) + include its entire contents (-P). + Then `patch' would create the file with appropriate contents. */ +EXTERN int unidirectional_new_file_flag; + +/* Pipe each file's output through pr (-l). */ +EXTERN int paginate_flag; + +enum line_class { + /* Lines taken from just the first file. */ + OLD, + /* Lines taken from just the second file. */ + NEW, + /* Lines common to both files. */ + UNCHANGED, + /* A hunk containing both old and new lines (line groups only). */ + CHANGED +}; + +/* Line group formats for old, new, unchanged, and changed groups. */ +EXTERN char *group_format[CHANGED + 1]; + +/* Line formats for old, new, and unchanged lines. */ +EXTERN char *line_format[UNCHANGED + 1]; + +/* If using OUTPUT_SDIFF print extra information to help the sdiff filter. */ +EXTERN int sdiff_help_sdiff; + +/* Tell OUTPUT_SDIFF to show only the left version of common lines. */ +EXTERN int sdiff_left_only; + +/* Tell OUTPUT_SDIFF to not show common lines. */ +EXTERN int sdiff_skip_common_lines; + +/* The half line width and column 2 offset for OUTPUT_SDIFF. */ +EXTERN unsigned sdiff_half_width; +EXTERN unsigned sdiff_column2_offset; + +/* String containing all the command options diff received, + with spaces between and at the beginning but none at the end. + If there were no options given, this string is empty. */ +EXTERN char * switch_string; + +/* Nonzero means use heuristics for better speed. */ +EXTERN int heuristic; + +/* Name of program the user invoked (for error messages). */ +EXTERN char *program_name; + +/* The result of comparison is an "edit script": a chain of `struct change'. + Each `struct change' represents one place where some lines are deleted + and some are inserted. + + LINE0 and LINE1 are the first affected lines in the two files (origin 0). + DELETED is the number of lines deleted here from file 0. + INSERTED is the number of lines inserted here in file 1. + + If DELETED is 0 then LINE0 is the number of the line before + which the insertion was done; vice versa for INSERTED and LINE1. */ + +struct change +{ + struct change *link; /* Previous or next edit command */ + int inserted; /* # lines of file 1 changed here. */ + int deleted; /* # lines of file 0 changed here. */ + int line0; /* Line number of 1st deleted line. */ + int line1; /* Line number of 1st inserted line. */ + char ignore; /* Flag used in context.c */ +}; + +/* Structures that describe the input files. */ + +/* Data on one input file being compared. */ + +struct file_data { + int desc; /* File descriptor */ + char const *name; /* File name */ + struct stat stat; /* File status from fstat() */ + int dir_p; /* nonzero if file is a directory */ + + /* Buffer in which text of file is read. */ + char * buffer; + /* Allocated size of buffer. */ + size_t bufsize; + /* Number of valid characters now in the buffer. */ + size_t buffered_chars; + + /* Array of pointers to lines in the file. */ + char const **linbuf; + + /* linbuf_base <= buffered_lines <= valid_lines <= alloc_lines. + linebuf[linbuf_base ... buffered_lines - 1] are possibly differing. + linebuf[linbuf_base ... valid_lines - 1] contain valid data. + linebuf[linbuf_base ... alloc_lines - 1] are allocated. */ + int linbuf_base, buffered_lines, valid_lines, alloc_lines; + + /* Pointer to end of prefix of this file to ignore when hashing. */ + char const *prefix_end; + + /* Count of lines in the prefix. + There are this many lines in the file before linbuf[0]. */ + int prefix_lines; + + /* Pointer to start of suffix of this file to ignore when hashing. */ + char const *suffix_begin; + + /* Vector, indexed by line number, containing an equivalence code for + each line. It is this vector that is actually compared with that + of another file to generate differences. */ + int *equivs; + + /* Vector, like the previous one except that + the elements for discarded lines have been squeezed out. */ + int *undiscarded; + + /* Vector mapping virtual line numbers (not counting discarded lines) + to real ones (counting those lines). Both are origin-0. */ + int *realindexes; + + /* Total number of nondiscarded lines. */ + int nondiscarded_lines; + + /* Vector, indexed by real origin-0 line number, + containing 1 for a line that is an insertion or a deletion. + The results of comparison are stored here. */ + char *changed_flag; + + /* 1 if file ends in a line with no final newline. */ + int missing_newline; + + /* 1 more than the maximum equivalence value used for this or its + sibling file. */ + int equiv_max; +}; + +/* Describe the two files currently being compared. */ + +EXTERN struct file_data files[2]; + +/* Stdio stream to output diffs to. */ + +EXTERN FILE *outfile; + +/* Declare various functions. */ + +/* analyze.c */ +int diff_2_files PARAMS((struct file_data[], int)); + +/* context.c */ +void print_context_header PARAMS((struct file_data[], int)); +void print_context_script PARAMS((struct change *, int)); + +/* diff.c */ +int excluded_filename PARAMS((char const *)); + +/* dir.c */ +int diff_dirs PARAMS((struct file_data const[], int (*) PARAMS((char const *, char const *, char const *, char const *, int)), int)); + +/* ed.c */ +void print_ed_script PARAMS((struct change *)); +void pr_forward_ed_script PARAMS((struct change *)); + +/* ifdef.c */ +void print_ifdef_script PARAMS((struct change *)); + +/* io.c */ +int read_files PARAMS((struct file_data[], int)); +int sip PARAMS((struct file_data *, int)); +void slurp PARAMS((struct file_data *)); + +/* normal.c */ +void print_normal_script PARAMS((struct change *)); + +/* rcs.c */ +void print_rcs_script PARAMS((struct change *)); + +/* side.c */ +void print_sdiff_script PARAMS((struct change *)); + +/* util.c */ +VOID *xmalloc PARAMS((size_t)); +VOID *xrealloc PARAMS((VOID *, size_t)); +char *concat PARAMS((char const *, char const *, char const *)); +char *dir_file_pathname PARAMS((char const *, char const *)); +int change_letter PARAMS((int, int)); +int line_cmp PARAMS((char const *, char const *)); +int translate_line_number PARAMS((struct file_data const *, int)); +struct change *find_change PARAMS((struct change *)); +struct change *find_reverse_change PARAMS((struct change *)); +void analyze_hunk PARAMS((struct change *, int *, int *, int *, int *, int *, int *)); +void begin_output PARAMS((void)); +void debug_script PARAMS((struct change *)); +void error PARAMS((char const *, char const *, char const *)); +void fatal PARAMS((char const *)); +void finish_output PARAMS((void)); +void message PARAMS((char const *, char const *, char const *)); +void message5 PARAMS((char const *, char const *, char const *, char const *, char const *)); +void output_1_line PARAMS((char const *, char const *, char const *, char const *)); +void perror_with_name PARAMS((char const *)); +void pfatal_with_name PARAMS((char const *)); +void print_1_line PARAMS((char const *, char const * const *)); +void print_message_queue PARAMS((void)); +void print_number_range PARAMS((int, struct file_data *, int, int)); +void print_script PARAMS((struct change *, struct change * (*) PARAMS((struct change *)), void (*) PARAMS((struct change *)))); +void setup_output PARAMS((char const *, char const *, int)); +void translate_range PARAMS((struct file_data const *, int, int, int *, int *)); + +/* version.c */ +extern char const version_string[]; diff --git a/sys/src/ape/cmd/diff/diff.texi b/sys/src/ape/cmd/diff/diff.texi new file mode 100755 index 000000000..7551acba8 --- /dev/null +++ b/sys/src/ape/cmd/diff/diff.texi @@ -0,0 +1,3916 @@ +\input texinfo @c -*-texinfo-*- +@c %**start of header +@setfilename diff.info +@settitle Comparing and Merging Files +@setchapternewpage odd +@c %**end of header + +@ifinfo +This file documents the the GNU @code{diff}, @code{diff3}, @code{sdiff}, +and @code{cmp} commands for showing the differences between text files +and the @code{patch} command for using their output to update files. + +Copyright (C) 1992, 1993, 1994 Free Software Foundation, Inc. + +Permission is granted to make and distribute verbatim copies of +this manual provided the copyright notice and this permission notice +are preserved on all copies. + +@ignore +Permission is granted to process this file through TeX and print the +results, provided the printed document carries copying permission +notice identical to this one except for the removal of this paragraph +(this paragraph not being relevant to the printed manual). + +@end ignore +Permission is granted to copy and distribute modified versions of this +manual under the conditions for verbatim copying, provided that the entire +resulting derived work is distributed under the terms of a permission +notice identical to this one. + +Permission is granted to copy and distribute translations of this manual +into another language, under the above conditions for modified versions, +except that this permission notice may be stated in a translation approved +by the Foundation. +@end ifinfo + +@titlepage +@title Comparing and Merging Files +@subtitle @code{diff}, @code{diff3}, @code{sdiff}, @code{cmp}, and @code{patch} +@subtitle Edition 1.3, for @code{diff} 2.5 and @code{patch} 2.1 +@subtitle September 1993 +@author by David MacKenzie, Paul Eggert, and Richard Stallman + +@page +@vskip 0pt plus 1filll +Copyright @copyright{} 1992, 1993, 1994 Free Software Foundation, Inc. + +Permission is granted to make and distribute verbatim copies of +this manual provided the copyright notice and this permission notice +are preserved on all copies. + +Permission is granted to copy and distribute modified versions of this +manual under the conditions for verbatim copying, provided that the entire +resulting derived work is distributed under the terms of a permission +notice identical to this one. + +Permission is granted to copy and distribute translations of this manual +into another language, under the above conditions for modified versions, +except that this permission notice may be stated in a translation approved +by the Foundation. +@end titlepage + +@node Top, , , (dir) + +@ifinfo +This file documents the the GNU @code{diff}, @code{diff3}, @code{sdiff}, +and @code{cmp} commands for showing the differences between text files +and the @code{patch} command for using their output to update files. + +This is Edition 1.2, for @code{diff} 2.4 and @code{patch} 2.1. +@end ifinfo + +@menu +* Overview:: Preliminary information. + +* Comparison:: What file comparison means. +* Output Formats:: Formats for difference reports. +* Comparing Directories:: Comparing files and directories. +* Adjusting Output:: Making @code{diff} output prettier. +* diff Performance:: Making @code{diff} smarter or faster. +* Comparing Three Files:: Formats for three-way difference reports. + +* diff3 Merging:: Merging from a common ancestor. +* Interactive Merging:: Interactive merging with @code{sdiff}. +* Merging with patch:: Using @code{patch} to change old files into new ones. +* Making Patches:: Tips for making patch distributions. + +* Invoking cmp:: How to run @code{cmp} and a summary of its options. +* Invoking diff:: How to run @code{diff} and a summary of its options. +* Invoking diff3:: How to run @code{diff3} and a summary of its options. +* Invoking patch:: How to run @code{patch} and a summary of its options. +* Invoking sdiff:: How to run @code{sdiff} and a summary of its options. + +* Incomplete Lines:: Lines that lack trailing newlines. +* Projects:: If you think you've found a bug or other shortcoming. + +* Concept Index:: Index of concepts. +@end menu + +@node Overview, Comparison, , Top +@unnumbered Overview +@cindex overview of @code{diff} and @code{patch} + +Computer users often find occasion to ask how two files differ. Perhaps +one file is a newer version of the other file. Or maybe the two files +started out as identical copies but were changed by different people. + +You can use the @code{diff} command to show differences between two +files, or each corresponding file in two directories. @code{diff} +outputs differences between files line by line in any of several +formats, selectable by command line options. This set of differences is +often called a @dfn{diff} or @dfn{patch}. For files that are identical, +@code{diff} normally produces no output; for binary (non-text) files, +@code{diff} normally reports only that they are different. + +You can use the @code{cmp} command to show the offsets and line numbers +where two files differ. @code{cmp} can also show all the characters +that differ between the two files, side by side. Another way to compare +two files character by character is the Emacs command @kbd{M-x +compare-windows}. @xref{Other Window, , Other Window, emacs, The GNU +Emacs Manual}, for more information on that command. + +You can use the @code{diff3} command to show differences among three +files. When two people have made independent changes to a common +original, @code{diff3} can report the differences between the original +and the two changed versions, and can produce a merged file that +contains both persons' changes together with warnings about conflicts. + +You can use the @code{sdiff} command to merge two files interactively. + +You can use the set of differences produced by @code{diff} to distribute +updates to text files (such as program source code) to other people. +This method is especially useful when the differences are small compared +to the complete files. Given @code{diff} output, you can use the +@code{patch} program to update, or @dfn{patch}, a copy of the file. If you +think of @code{diff} as subtracting one file from another to produce +their difference, you can think of @code{patch} as adding the difference +to one file to reproduce the other. + +This manual first concentrates on making diffs, and later shows how to +use diffs to update files. + +GNU @code{diff} was written by Mike Haertel, David Hayes, Richard +Stallman, Len Tower, and Paul Eggert. Wayne Davison designed and +implemented the unified output format. The basic algorithm is described +in ``An O(ND) Difference Algorithm and its Variations'', Eugene W. Myers, +@cite{Algorithmica} Vol.@: 1 No.@: 2, 1986, pp.@: 251--266; and in ``A File +Comparison Program'', Webb Miller and Eugene W. Myers, +@cite{Software---Practice and Experience} Vol.@: 15 No.@: 11, 1985, +pp.@: 1025--1040. +@c From: "Gene Myers" <gene@cs.arizona.edu> +@c They are about the same basic algorithm; the Algorithmica +@c paper gives a rigorous treatment and the sub-algorithm for +@c delivering scripts and should be the primary reference, but +@c both should be mentioned. +The algorithm was independently discovered as described in +``Algorithms for Approximate String Matching'', +E. Ukkonen, @cite{Information and Control} Vol.@: 64, 1985, pp.@: 100--118. +@c From: "Gene Myers" <gene@cs.arizona.edu> +@c Date: Wed, 29 Sep 1993 08:27:55 MST +@c Ukkonen should be given credit for also discovering the algorithm used +@c in GNU diff. + +GNU @code{diff3} was written by Randy Smith. GNU @code{sdiff} was +written by Thomas Lord. GNU @code{cmp} was written by Torbjorn Granlund +and David MacKenzie. + +@code{patch} was written mainly by Larry Wall; the GNU enhancements were +written mainly by Wayne Davison and David MacKenzie. Parts of this +manual are adapted from a manual page written by Larry Wall, with his +permission. + +@node Comparison, Output Formats, Overview, Top +@chapter What Comparison Means +@cindex introduction + +There are several ways to think about the differences between two files. +One way to think of the differences is as a series of lines that were +deleted from, inserted in, or changed in one file to produce the other +file. @code{diff} compares two files line by line, finds groups of +lines that differ, and reports each group of differing lines. It can +report the differing lines in several formats, which have different +purposes. + +GNU @code{diff} can show whether files are different without detailing +the differences. It also provides ways to suppress certain kinds of +differences that are not important to you. Most commonly, such +differences are changes in the amount of white space between words or +lines. @code{diff} also provides ways to suppress differences in +alphabetic case or in lines that match a regular expression that you +provide. These options can accumulate; for example, you can ignore +changes in both white space and alphabetic case. + +Another way to think of the differences between two files is as a +sequence of pairs of characters that can be either identical or +different. @code{cmp} reports the differences between two files +character by character, instead of line by line. As a result, it is +more useful than @code{diff} for comparing binary files. For text +files, @code{cmp} is useful mainly when you want to know only whether +two files are identical. + +To illustrate the effect that considering changes character by character +can have compared with considering them line by line, think of what +happens if a single newline character is added to the beginning of a +file. If that file is then compared with an otherwise identical file +that lacks the newline at the beginning, @code{diff} will report that a +blank line has been added to the file, while @code{cmp} will report that +almost every character of the two files differs. + +@code{diff3} normally compares three input files line by line, finds +groups of lines that differ, and reports each group of differing lines. +Its output is designed to make it easy to inspect two different sets of +changes to the same file. + +@menu +* Hunks:: Groups of differing lines. +* White Space:: Suppressing differences in white space. +* Blank Lines:: Suppressing differences in blank lines. +* Case Folding:: Suppressing differences in alphabetic case. +* Specified Folding:: Suppressing differences that match regular expressions. +* Brief:: Summarizing which files are different. +* Binary:: Comparing binary files or forcing text comparisons. +@end menu + +@node Hunks, White Space, , Comparison +@section Hunks +@cindex hunks + +When comparing two files, @code{diff} finds sequences of lines common to +both files, interspersed with groups of differing lines called +@dfn{hunks}. Comparing two identical files yields one sequence of +common lines and no hunks, because no lines differ. Comparing two +entirely different files yields no common lines and one large hunk that +contains all lines of both files. In general, there are many ways to +match up lines between two given files. @code{diff} tries to minimize +the total hunk size by finding large sequences of common lines +interspersed with small hunks of differing lines. + +For example, suppose the file @file{F} contains the three lines +@samp{a}, @samp{b}, @samp{c}, and the file @file{G} contains the same +three lines in reverse order @samp{c}, @samp{b}, @samp{a}. If +@code{diff} finds the line @samp{c} as common, then the command +@samp{diff F G} produces this output: + +@example +1,2d0 +< a +< b +3a2,3 +> b +> a +@end example + +@noindent +But if @code{diff} notices the common line @samp{b} instead, it produces +this output: + +@example +1c1 +< a +--- +> c +3c3 +< c +--- +> a +@end example + +@noindent +It is also possible to find @samp{a} as the common line. @code{diff} +does not always find an optimal matching between the files; it takes +shortcuts to run faster. But its output is usually close to the +shortest possible. You can adjust this tradeoff with the +@samp{--minimal} option (@pxref{diff Performance}). + +@node White Space, Blank Lines, Hunks, Comparison +@section Suppressing Differences in Blank and Tab Spacing +@cindex blank and tab difference suppression +@cindex tab and blank difference suppression + +The @samp{-b} and @samp{--ignore-space-change} options ignore white space +at line end, and considers all other sequences of one or more +white space characters to be equivalent. With these options, +@code{diff} considers the following two lines to be equivalent, where +@samp{$} denotes the line end: + +@example +Here lyeth muche rychnesse in lytell space. -- John Heywood$ +Here lyeth muche rychnesse in lytell space. -- John Heywood $ +@end example + +The @samp{-w} and @samp{--ignore-all-space} options are stronger than +@samp{-b}. They ignore difference even if one file has white space where +the other file has none. @dfn{White space} characters include +tab, newline, vertical tab, form feed, carriage return, and space; +some locales may define additional characters to be white space. +With these options, @code{diff} considers the +following two lines to be equivalent, where @samp{$} denotes the line +end and @samp{^M} denotes a carriage return: + +@example +Here lyeth muche rychnesse in lytell space.-- John Heywood$ + He relyeth much erychnes seinly tells pace. --John Heywood ^M$ +@end example + +@node Blank Lines, Case Folding, White Space, Comparison +@section Suppressing Differences in Blank Lines +@cindex blank line difference suppression + +The @samp{-B} and @samp{--ignore-blank-lines} options ignore insertions +or deletions of blank lines. These options normally affect only lines +that are completely empty; they do not affect lines that look empty but +contain space or tab characters. With these options, for example, a +file containing +@example +1. A point is that which has no part. + +2. A line is breadthless length. +-- Euclid, The Elements, I +@end example +@noindent +is considered identical to a file containing +@example +1. A point is that which has no part. +2. A line is breadthless length. + + +-- Euclid, The Elements, I +@end example + +@node Case Folding, Specified Folding, Blank Lines, Comparison +@section Suppressing Case Differences +@cindex case difference suppression + +GNU @code{diff} can treat lowercase letters as equivalent to their +uppercase counterparts, so that, for example, it considers @samp{Funky +Stuff}, @samp{funky STUFF}, and @samp{fUNKy stuFf} to all be the same. +To request this, use the @samp{-i} or @samp{--ignore-case} option. + +@node Specified Folding, Brief, Case Folding, Comparison +@section Suppressing Lines Matching a Regular Expression +@cindex regular expression suppression + +To ignore insertions and deletions of lines that match a regular +expression, use the @samp{-I @var{regexp}} or +@samp{--ignore-matching-lines=@var{regexp}} option. You should escape +regular expressions that contain shell metacharacters to prevent the +shell from expanding them. For example, @samp{diff -I '^[0-9]'} ignores +all changes to lines beginning with a digit. + +However, @samp{-I} only ignores the insertion or deletion of lines that +contain the regular expression if every changed line in the hunk---every +insertion and every deletion---matches the regular expression. In other +words, for each nonignorable change, @code{diff} prints the complete set +of changes in its vicinity, including the ignorable ones. + +You can specify more than one regular expression for lines to ignore by +using more than one @samp{-I} option. @code{diff} tries to match each +line against each regular expression, starting with the last one given. + +@node Brief, Binary, Specified Folding, Comparison +@section Summarizing Which Files Differ +@cindex summarizing which files differ +@cindex brief difference reports + +When you only want to find out whether files are different, and you +don't care what the differences are, you can use the summary output +format. In this format, instead of showing the differences between the +files, @code{diff} simply reports whether files differ. The @samp{-q} +and @samp{--brief} options select this output format. + +This format is especially useful when comparing the contents of two +directories. It is also much faster than doing the normal line by line +comparisons, because @code{diff} can stop analyzing the files as soon as +it knows that there are any differences. + +You can also get a brief indication of whether two files differ by using +@code{cmp}. For files that are identical, @code{cmp} produces no +output. When the files differ, by default, @code{cmp} outputs the byte +offset and line number where the first difference occurs. You can use +the @samp{-s} option to suppress that information, so that @code{cmp} +produces no output and reports whether the files differ using only its +exit status (@pxref{Invoking cmp}). + +@c Fix this. +Unlike @code{diff}, @code{cmp} cannot compare directories; it can only +compare two files. + +@node Binary, , Brief, Comparison +@section Binary Files and Forcing Text Comparisons +@cindex binary file diff +@cindex text versus binary diff + +If @code{diff} thinks that either of the two files it is comparing is +binary (a non-text file), it normally treats that pair of files much as +if the summary output format had been selected (@pxref{Brief}), and +reports only that the binary files are different. This is because line +by line comparisons are usually not meaningful for binary files. + +@code{diff} determines whether a file is text or binary by checking the +first few bytes in the file; the exact number of bytes is system +dependent, but it is typically several thousand. If every character in +that part of the file is non-null, @code{diff} considers the file to be +text; otherwise it considers the file to be binary. + +Sometimes you might want to force @code{diff} to consider files to be +text. For example, you might be comparing text files that contain +null characters; @code{diff} would erroneously decide that those are +non-text files. Or you might be comparing documents that are in a +format used by a word processing system that uses null characters to +indicate special formatting. You can force @code{diff} to consider all +files to be text files, and compare them line by line, by using the +@samp{-a} or @samp{--text} option. If the files you compare using this +option do not in fact contain text, they will probably contain few +newline characters, and the @code{diff} output will consist of hunks +showing differences between long lines of whatever characters the files +contain. + +You can also force @code{diff} to consider all files to be binary files, +and report only whether they differ (but not how). Use the +@samp{--brief} option for this. + +In operating systems that distinguish between text and binary files, +@code{diff} normally reads and writes all data as text. Use the +@samp{--binary} option to force @code{diff} to read and write binary +data instead. This option has no effect on a Posix-compliant system +like GNU or traditional Unix. However, many personal computer +operating systems represent the end of a line with a carriage return +followed by a newline. On such systems, @code{diff} normally ignores +these carriage returns on input and generates them at the end of each +output line, but with the @samp{--binary} option @code{diff} treats +each carriage return as just another input character, and does not +generate a carriage return at the end of each output line. This can be +useful when dealing with non-text files that are meant to be +interchanged with Posix-compliant systems. + +If you want to compare two files byte by byte, you can use the +@code{cmp} program with the @samp{-l} option to show the values of each +differing byte in the two files. With GNU @code{cmp}, you can also use +the @samp{-c} option to show the ASCII representation of those bytes. +@xref{Invoking cmp}, for more information. + +If @code{diff3} thinks that any of the files it is comparing is binary +(a non-text file), it normally reports an error, because such +comparisons are usually not useful. @code{diff3} uses the same test as +@code{diff} to decide whether a file is binary. As with @code{diff}, if +the input files contain a few non-text characters but otherwise are like +text files, you can force @code{diff3} to consider all files to be text +files and compare them line by line by using the @samp{-a} or +@samp{--text} options. + +@node Output Formats, Comparing Directories, Comparison, Top +@chapter @code{diff} Output Formats +@cindex output formats +@cindex format of @code{diff} output + +@code{diff} has several mutually exclusive options for output format. +The following sections describe each format, illustrating how +@code{diff} reports the differences between two sample input files. + +@menu +* Sample diff Input:: Sample @code{diff} input files for examples. +* Normal:: Showing differences without surrounding text. +* Context:: Showing differences with the surrounding text. +* Side by Side:: Showing differences in two columns. +* Scripts:: Generating scripts for other programs. +* If-then-else:: Merging files with if-then-else. +@end menu + +@node Sample diff Input, Normal, , Output Formats +@section Two Sample Input Files +@cindex @code{diff} sample input +@cindex sample input for @code{diff} + +Here are two sample files that we will use in numerous examples to +illustrate the output of @code{diff} and how various options can change +it. + +This is the file @file{lao}: + +@example +The Way that can be told of is not the eternal Way; +The name that can be named is not the eternal name. +The Nameless is the origin of Heaven and Earth; +The Named is the mother of all things. +Therefore let there always be non-being, + so we may see their subtlety, +And let there always be being, + so we may see their outcome. +The two are the same, +But after they are produced, + they have different names. +@end example + +This is the file @file{tzu}: + +@example +The Nameless is the origin of Heaven and Earth; +The named is the mother of all things. + +Therefore let there always be non-being, + so we may see their subtlety, +And let there always be being, + so we may see their outcome. +The two are the same, +But after they are produced, + they have different names. +They both may be called deep and profound. +Deeper and more profound, +The door of all subtleties! +@end example + +In this example, the first hunk contains just the first two lines of +@file{lao}, the second hunk contains the fourth line of @file{lao} +opposing the second and third lines of @file{tzu}, and the last hunk +contains just the last three lines of @file{tzu}. + +@node Normal, Context, Sample diff Input, Output Formats +@section Showing Differences Without Context +@cindex normal output format +@cindex @samp{<} output format + +The ``normal'' @code{diff} output format shows each hunk of differences +without any surrounding context. Sometimes such output is the clearest +way to see how lines have changed, without the clutter of nearby +unchanged lines (although you can get similar results with the context +or unified formats by using 0 lines of context). However, this format +is no longer widely used for sending out patches; for that purpose, the +context format (@pxref{Context Format}) and the unified format +(@pxref{Unified Format}) are superior. Normal format is the default for +compatibility with older versions of @code{diff} and the Posix standard. + +@menu +* Detailed Normal:: A detailed description of normal output format. +* Example Normal:: Sample output in the normal format. +@end menu + +@node Detailed Normal, Example Normal, , Normal +@subsection Detailed Description of Normal Format + +The normal output format consists of one or more hunks of differences; +each hunk shows one area where the files differ. Normal format hunks +look like this: + +@example +@var{change-command} +< @var{from-file-line} +< @var{from-file-line}@dots{} +--- +> @var{to-file-line} +> @var{to-file-line}@dots{} +@end example + +There are three types of change commands. Each consists of a line +number or comma-separated range of lines in the first file, a single +character indicating the kind of change to make, and a line number or +comma-separated range of lines in the second file. All line numbers are +the original line numbers in each file. The types of change commands +are: + +@table @samp +@item @var{l}a@var{r} +Add the lines in range @var{r} of the second file after line @var{l} of +the first file. For example, @samp{8a12,15} means append lines 12--15 +of file 2 after line 8 of file 1; or, if changing file 2 into file 1, +delete lines 12--15 of file 2. + +@item @var{f}c@var{t} +Replace the lines in range @var{f} of the first file with lines in range +@var{t} of the second file. This is like a combined add and delete, but +more compact. For example, @samp{5,7c8,10} means change lines 5--7 of +file 1 to read as lines 8--10 of file 2; or, if changing file 2 into +file 1, change lines 8--10 of file 2 to read as lines 5--7 of file 1. + +@item @var{r}d@var{l} +Delete the lines in range @var{r} from the first file; line @var{l} is where +they would have appeared in the second file had they not been deleted. +For example, @samp{5,7d3} means delete lines 5--7 of file 1; or, if +changing file 2 into file 1, append lines 5--7 of file 1 after line 3 of +file 2. +@end table + +@node Example Normal, , Detailed Normal, Normal +@subsection An Example of Normal Format + +Here is the output of the command @samp{diff lao tzu} +(@pxref{Sample diff Input}, for the complete contents of the two files). +Notice that it shows only the lines that are different between the two +files. + +@example +1,2d0 +< The Way that can be told of is not the eternal Way; +< The name that can be named is not the eternal name. +4c2,3 +< The Named is the mother of all things. +--- +> The named is the mother of all things. +> +11a11,13 +> They both may be called deep and profound. +> Deeper and more profound, +> The door of all subtleties! +@end example + +@node Context, Side by Side, Normal, Output Formats +@section Showing Differences in Their Context +@cindex context output format +@cindex @samp{!} output format + +Usually, when you are looking at the differences between files, you will +also want to see the parts of the files near the lines that differ, to +help you understand exactly what has changed. These nearby parts of the +files are called the @dfn{context}. + +GNU @code{diff} provides two output formats that show context around the +differing lines: @dfn{context format} and @dfn{unified format}. It can +optionally show in which function or section of the file the differing +lines are found. + +If you are distributing new versions of files to other people in the +form of @code{diff} output, you should use one of the output formats +that show context so that they can apply the diffs even if they have +made small changes of their own to the files. @code{patch} can apply +the diffs in this case by searching in the files for the lines of +context around the differing lines; if those lines are actually a few +lines away from where the diff says they are, @code{patch} can adjust +the line numbers accordingly and still apply the diff correctly. +@xref{Imperfect}, for more information on using @code{patch} to apply +imperfect diffs. + +@menu +* Context Format:: An output format that shows surrounding lines. +* Unified Format:: A more compact output format that shows context. +* Sections:: Showing which sections of the files differences are in. +* Alternate Names:: Showing alternate file names in context headers. +@end menu + +@node Context Format, Unified Format, , Context +@subsection Context Format + +The context output format shows several lines of context around the +lines that differ. It is the standard format for distributing updates +to source code. + +To select this output format, use the @samp{-C @var{lines}}, +@samp{--context@r{[}=@var{lines}@r{]}}, or @samp{-c} option. The +argument @var{lines} that some of these options take is the number of +lines of context to show. If you do not specify @var{lines}, it +defaults to three. For proper operation, @code{patch} typically needs +at least two lines of context. + +@menu +* Detailed Context:: A detailed description of the context output format. +* Example Context:: Sample output in context format. +* Less Context:: Another sample with less context. +@end menu + +@node Detailed Context, Example Context, , Context Format +@subsubsection Detailed Description of Context Format + +The context output format starts with a two-line header, which looks +like this: + +@example +*** @var{from-file} @var{from-file-modification-time} +--- @var{to-file} @var{to-file-modification time} +@end example + +@noindent +You can change the header's content with the @samp{-L @var{label}} or +@samp{--label=@var{label}} option; see @ref{Alternate Names}. + +Next come one or more hunks of differences; each hunk shows one area +where the files differ. Context format hunks look like this: + +@example +*************** +*** @var{from-file-line-range} **** + @var{from-file-line} + @var{from-file-line}@dots{} +--- @var{to-file-line-range} ---- + @var{to-file-line} + @var{to-file-line}@dots{} +@end example + +The lines of context around the lines that differ start with two space +characters. The lines that differ between the two files start with one +of the following indicator characters, followed by a space character: + +@table @samp +@item ! +A line that is part of a group of one or more lines that changed between +the two files. There is a corresponding group of lines marked with +@samp{!} in the part of this hunk for the other file. + +@item + +An ``inserted'' line in the second file that corresponds to nothing in +the first file. + +@item - +A ``deleted'' line in the first file that corresponds to nothing in the +second file. +@end table + +If all of the changes in a hunk are insertions, the lines of +@var{from-file} are omitted. If all of the changes are deletions, the +lines of @var{to-file} are omitted. + +@node Example Context, Less Context, Detailed Context, Context Format +@subsubsection An Example of Context Format + +Here is the output of @samp{diff -c lao tzu} (@pxref{Sample diff Input}, +for the complete contents of the two files). Notice that up to three +lines that are not different are shown around each line that is +different; they are the context lines. Also notice that the first two +hunks have run together, because their contents overlap. + +@example +*** lao Sat Jan 26 23:30:39 1991 +--- tzu Sat Jan 26 23:30:50 1991 +*************** +*** 1,7 **** +- The Way that can be told of is not the eternal Way; +- The name that can be named is not the eternal name. + The Nameless is the origin of Heaven and Earth; +! The Named is the mother of all things. + Therefore let there always be non-being, + so we may see their subtlety, + And let there always be being, +--- 1,6 ---- + The Nameless is the origin of Heaven and Earth; +! The named is the mother of all things. +! + Therefore let there always be non-being, + so we may see their subtlety, + And let there always be being, +*************** +*** 9,11 **** +--- 8,13 ---- + The two are the same, + But after they are produced, + they have different names. ++ They both may be called deep and profound. ++ Deeper and more profound, ++ The door of all subtleties! +@end example + +@node Less Context, , Example Context, Context Format +@subsubsection An Example of Context Format with Less Context + +Here is the output of @samp{diff --context=1 lao tzu} (@pxref{Sample +diff Input}, for the complete contents of the two files). Notice that +at most one context line is reported here. + +@example +*** lao Sat Jan 26 23:30:39 1991 +--- tzu Sat Jan 26 23:30:50 1991 +*************** +*** 1,5 **** +- The Way that can be told of is not the eternal Way; +- The name that can be named is not the eternal name. + The Nameless is the origin of Heaven and Earth; +! The Named is the mother of all things. + Therefore let there always be non-being, +--- 1,4 ---- + The Nameless is the origin of Heaven and Earth; +! The named is the mother of all things. +! + Therefore let there always be non-being, +*************** +*** 11 **** +--- 10,13 ---- + they have different names. ++ They both may be called deep and profound. ++ Deeper and more profound, ++ The door of all subtleties! +@end example + +@node Unified Format, Sections, Context Format, Context +@subsection Unified Format +@cindex unified output format +@cindex @samp{+-} output format + +The unified output format is a variation on the context format that is +more compact because it omits redundant context lines. To select this +output format, use the @samp{-U @var{lines}}, +@samp{--unified@r{[}=@var{lines}@r{]}}, or @samp{-u} +option. The argument @var{lines} is the number of lines of context to +show. When it is not given, it defaults to three. + +At present, only GNU @code{diff} can produce this format and only GNU +@code{patch} can automatically apply diffs in this format. For proper +operation, @code{patch} typically needs at least two lines of context. + +@menu +* Detailed Unified:: A detailed description of unified format. +* Example Unified:: Sample output in unified format. +@end menu + +@node Detailed Unified, Example Unified, , Unified Format +@subsubsection Detailed Description of Unified Format + +The unified output format starts with a two-line header, which looks +like this: + +@example +--- @var{from-file} @var{from-file-modification-time} ++++ @var{to-file} @var{to-file-modification-time} +@end example + +@noindent +You can change the header's content with the @samp{-L @var{label}} or +@samp{--label=@var{label}} option; see @xref{Alternate Names}. + +Next come one or more hunks of differences; each hunk shows one area +where the files differ. Unified format hunks look like this: + +@example +@@@@ @var{from-file-range} @var{to-file-range} @@@@ + @var{line-from-either-file} + @var{line-from-either-file}@dots{} +@end example + +The lines common to both files begin with a space character. The lines +that actually differ between the two files have one of the following +indicator characters in the left column: + +@table @samp +@item + +A line was added here to the first file. + +@item - +A line was removed here from the first file. +@end table + +@node Example Unified, , Detailed Unified, Unified Format +@subsubsection An Example of Unified Format + +Here is the output of the command @samp{diff -u lao tzu} +(@pxref{Sample diff Input}, for the complete contents of the two files): + +@example +--- lao Sat Jan 26 23:30:39 1991 ++++ tzu Sat Jan 26 23:30:50 1991 +@@@@ -1,7 +1,6 @@@@ +-The Way that can be told of is not the eternal Way; +-The name that can be named is not the eternal name. + The Nameless is the origin of Heaven and Earth; +-The Named is the mother of all things. ++The named is the mother of all things. ++ + Therefore let there always be non-being, + so we may see their subtlety, + And let there always be being, +@@@@ -9,3 +8,6 @@@@ + The two are the same, + But after they are produced, + they have different names. ++They both may be called deep and profound. ++Deeper and more profound, ++The door of all subtleties! +@end example + +@node Sections, Alternate Names, Unified Format, Context +@subsection Showing Which Sections Differences Are in +@cindex headings +@cindex section headings + +Sometimes you might want to know which part of the files each change +falls in. If the files are source code, this could mean which function +was changed. If the files are documents, it could mean which chapter or +appendix was changed. GNU @code{diff} can show this by displaying the +nearest section heading line that precedes the differing lines. Which +lines are ``section headings'' is determined by a regular expression. + +@menu +* Specified Headings:: Showing headings that match regular expressions. +* C Function Headings:: Showing headings of C functions. +@end menu + +@node Specified Headings, C Function Headings, , Sections +@subsubsection Showing Lines That Match Regular Expressions +@cindex specified headings +@cindex regular expression matching headings + +To show in which sections differences occur for files that are not +source code for C or similar languages, use the @samp{-F @var{regexp}} +or @samp{--show-function-line=@var{regexp}} option. @code{diff} +considers lines that match the argument @var{regexp} to be the beginning +of a section of the file. Here are suggested regular expressions for +some common languages: + +@c Please add to this list, e.g. Fortran, Pascal. +@table @samp +@item ^[A-Za-z_] +C, C++, Prolog +@item ^( +Lisp +@item ^@@\(chapter\|appendix\|unnumbered\|chapheading\) +Texinfo +@end table + +This option does not automatically select an output format; in order to +use it, you must select the context format (@pxref{Context Format}) or +unified format (@pxref{Unified Format}). In other output formats it +has no effect. + +The @samp{-F} and @samp{--show-function-line} options find the nearest +unchanged line that precedes each hunk of differences and matches the +given regular expression. Then they add that line to the end of the +line of asterisks in the context format, or to the @samp{@@@@} line in +unified format. If no matching line exists, they leave the output for +that hunk unchanged. If that line is more than 40 characters long, they +output only the first 40 characters. You can specify more than one +regular expression for such lines; @code{diff} tries to match each line +against each regular expression, starting with the last one given. This +means that you can use @samp{-p} and @samp{-F} together, if you wish. + +@node C Function Headings, , Specified Headings, Sections +@subsubsection Showing C Function Headings +@cindex C function headings +@cindex function headings, C + +To show in which functions differences occur for C and similar +languages, you can use the @samp{-p} or @samp{--show-c-function} option. +This option automatically defaults to the context output format +(@pxref{Context Format}), with the default number of lines of context. +You can override that number with @samp{-C @var{lines}} elsewhere in the +command line. You can override both the format and the number with +@samp{-U @var{lines}} elsewhere in the command line. + +The @samp{-p} and @samp{--show-c-function} options are equivalent to +@samp{-F'^[_a-zA-Z$]'} if the unified format is specified, otherwise +@samp{-c -F'^[_a-zA-Z$]'} (@pxref{Specified Headings}). GNU @code{diff} +provides them for the sake of convenience. + +@node Alternate Names, , Sections, Context +@subsection Showing Alternate File Names +@cindex alternate file names +@cindex file name alternates + +If you are comparing two files that have meaningless or uninformative +names, you might want @code{diff} to show alternate names in the header +of the context and unified output formats. To do this, use the @samp{-L +@var{label}} or @samp{--label=@var{label}} option. The first time +you give this option, its argument replaces the name and date of the +first file in the header; the second time, its argument replaces the +name and date of the second file. If you give this option more than +twice, @code{diff} reports an error. The @samp{-L} option does not +affect the file names in the @code{pr} header when the @samp{-l} or +@samp{--paginate} option is used (@pxref{Pagination}). + +Here are the first two lines of the output from @samp{diff -C2 +-Loriginal -Lmodified lao tzu}: + +@example +*** original +--- modified +@end example + +@node Side by Side, Scripts, Context, Output Formats +@section Showing Differences Side by Side +@cindex side by side +@cindex two-column output +@cindex columnar output + +@code{diff} can produce a side by side difference listing of two files. +The files are listed in two columns with a gutter between them. The +gutter contains one of the following markers: + +@table @asis +@item white space +The corresponding lines are in common. That is, either the lines are +identical, or the difference is ignored because of one of the +@samp{--ignore} options (@pxref{White Space}). + +@item @samp{|} +The corresponding lines differ, and they are either both complete +or both incomplete. + +@item @samp{<} +The files differ and only the first file contains the line. + +@item @samp{>} +The files differ and only the second file contains the line. + +@item @samp{(} +Only the first file contains the line, but the difference is ignored. + +@item @samp{)} +Only the second file contains the line, but the difference is ignored. + +@item @samp{\} +The corresponding lines differ, and only the first line is incomplete. + +@item @samp{/} +The corresponding lines differ, and only the second line is incomplete. +@end table + +Normally, an output line is incomplete if and only if the lines that it +contains are incomplete; @xref{Incomplete Lines}. However, when an +output line represents two differing lines, one might be incomplete +while the other is not. In this case, the output line is complete, +but its the gutter is marked @samp{\} if the first line is incomplete, +@samp{/} if the second line is. + +Side by side format is sometimes easiest to read, but it has limitations. +It generates much wider output than usual, and truncates lines that are +too long to fit. Also, it relies on lining up output more heavily than +usual, so its output looks particularly bad if you use varying +width fonts, nonstandard tab stops, or nonprinting characters. + +You can use the @code{sdiff} command to interactively merge side by side +differences. @xref{Interactive Merging}, for more information on merging files. + +@menu +* Side by Side Format:: Controlling side by side output format. +* Example Side by Side:: Sample side by side output. +@end menu + +@node Side by Side Format, Example Side by Side, , Side by Side +@section Controlling Side by Side Format +@cindex side by side format + +The @samp{-y} or @samp{--side-by-side} option selects side by side +format. Because side by side output lines contain two input lines, they +are wider than usual. They are normally 130 columns, which can fit onto +a traditional printer line. You can set the length of output lines with +the @samp{-W @var{columns}} or @samp{--width=@var{columns}} option. The +output line is split into two halves of equal length, separated by a +small gutter to mark differences; the right half is aligned to a tab +stop so that tabs line up. Input lines that are too long to fit in half +of an output line are truncated for output. + +The @samp{--left-column} option prints only the left column of two +common lines. The @samp{--suppress-common-lines} option suppresses +common lines entirely. + +@node Example Side by Side, , Side by Side Format, Side by Side +@subsection An Example of Side by Side Format + +Here is the output of the command @samp{diff -y -W 72 lao tzu} +(@pxref{Sample diff Input}, for the complete contents of the two files). + +@example +The Way that can be told of is n < +The name that can be named is no < +The Nameless is the origin of He The Nameless is the origin of He +The Named is the mother of all t | The named is the mother of all t + > +Therefore let there always be no Therefore let there always be no + so we may see their subtlety, so we may see their subtlety, +And let there always be being, And let there always be being, + so we may see their outcome. so we may see their outcome. +The two are the same, The two are the same, +But after they are produced, But after they are produced, + they have different names. they have different names. + > They both may be called deep and + > Deeper and more profound, + > The door of all subtleties! +@end example + +@node Scripts, If-then-else, Side by Side, Output Formats +@section Making Edit Scripts +@cindex script output formats + +Several output modes produce command scripts for editing @var{from-file} +to produce @var{to-file}. + +@menu +* ed Scripts:: Using @code{diff} to produce commands for @code{ed}. +* Forward ed:: Making forward @code{ed} scripts. +* RCS:: A special @code{diff} output format used by RCS. +@end menu + +@node ed Scripts, Forward ed, , Scripts +@subsection @code{ed} Scripts +@cindex @code{ed} script output format + +@code{diff} can produce commands that direct the @code{ed} text editor +to change the first file into the second file. Long ago, this was the +only output mode that was suitable for editing one file into another +automatically; today, with @code{patch}, it is almost obsolete. Use the +@samp{-e} or @samp{--ed} option to select this output format. + +Like the normal format (@pxref{Normal}), this output format does not +show any context; unlike the normal format, it does not include the +information necessary to apply the diff in reverse (to produce the first +file if all you have is the second file and the diff). + +If the file @file{d} contains the output of @samp{diff -e old new}, then +the command @samp{(cat d && echo w) | ed - old} edits @file{old} to make +it a copy of @file{new}. More generally, if @file{d1}, @file{d2}, +@dots{}, @file{dN} contain the outputs of @samp{diff -e old new1}, +@samp{diff -e new1 new2}, @dots{}, @samp{diff -e newN-1 newN}, +respectively, then the command @samp{(cat d1 d2 @dots{} dN && echo w) | +ed - old} edits @file{old} to make it a copy of @file{newN}. + +@menu +* Detailed ed:: A detailed description of @code{ed} format. +* Example ed:: A sample @code{ed} script. +@end menu + +@node Detailed ed, Example ed, , ed Scripts +@subsubsection Detailed Description of @code{ed} Format + +The @code{ed} output format consists of one or more hunks of +differences. The changes closest to the ends of the files come first so +that commands that change the number of lines do not affect how +@code{ed} interprets line numbers in succeeding commands. @code{ed} +format hunks look like this: + +@example +@var{change-command} +@var{to-file-line} +@var{to-file-line}@dots{} +. +@end example + +Because @code{ed} uses a single period on a line to indicate the end of +input, GNU @code{diff} protects lines of changes that contain a single +period on a line by writing two periods instead, then writing a +subsequent @code{ed} command to change the two periods into one. The +@code{ed} format cannot represent an incomplete line, so if the second +file ends in a changed incomplete line, @code{diff} reports an error and +then pretends that a newline was appended. + +There are three types of change commands. Each consists of a line +number or comma-separated range of lines in the first file and a single +character indicating the kind of change to make. All line numbers are +the original line numbers in the file. The types of change commands +are: + +@table @samp +@item @var{l}a +Add text from the second file after line @var{l} in the first file. For +example, @samp{8a} means to add the following lines after line 8 of file +1. + +@item @var{r}c +Replace the lines in range @var{r} in the first file with the following +lines. Like a combined add and delete, but more compact. For example, +@samp{5,7c} means change lines 5--7 of file 1 to read as the text file +2. + +@item @var{r}d +Delete the lines in range @var{r} from the first file. For example, +@samp{5,7d} means delete lines 5--7 of file 1. +@end table + +@node Example ed, , Detailed ed, ed Scripts +@subsubsection Example @code{ed} Script + +Here is the output of @samp{diff -e lao tzu} (@pxref{Sample +diff Input}, for the complete contents of the two files): + +@example +11a +They both may be called deep and profound. +Deeper and more profound, +The door of all subtleties! +. +4c +The named is the mother of all things. + +. +1,2d +@end example + +@node Forward ed, RCS, ed Scripts, Scripts +@subsection Forward @code{ed} Scripts +@cindex forward @code{ed} script output format + +@code{diff} can produce output that is like an @code{ed} script, but +with hunks in forward (front to back) order. The format of the commands +is also changed slightly: command characters precede the lines they +modify, spaces separate line numbers in ranges, and no attempt is made +to disambiguate hunk lines consisting of a single period. Like +@code{ed} format, forward @code{ed} format cannot represent incomplete +lines. + +Forward @code{ed} format is not very useful, because neither @code{ed} +nor @code{patch} can apply diffs in this format. It exists mainly for +compatibility with older versions of @code{diff}. Use the @samp{-f} or +@samp{--forward-ed} option to select it. + +@node RCS, , Forward ed, Scripts +@subsection RCS Scripts +@cindex RCS script output format + +The RCS output format is designed specifically for use by the Revision +Control System, which is a set of free programs used for organizing +different versions and systems of files. Use the @samp{-n} or +@samp{--rcs} option to select this output format. It is like the +forward @code{ed} format (@pxref{Forward ed}), but it can represent +arbitrary changes to the contents of a file because it avoids the +forward @code{ed} format's problems with lines consisting of a single +period and with incomplete lines. Instead of ending text sections with +a line consisting of a single period, each command specifies the number +of lines it affects; a combination of the @samp{a} and @samp{d} +commands are used instead of @samp{c}. Also, if the second file ends +in a changed incomplete line, then the output also ends in an +incomplete line. + +Here is the output of @samp{diff -n lao tzu} (@pxref{Sample +diff Input}, for the complete contents of the two files): + +@example +d1 2 +d4 1 +a4 2 +The named is the mother of all things. + +a11 3 +They both may be called deep and profound. +Deeper and more profound, +The door of all subtleties! +@end example + +@node If-then-else, , Scripts, Output Formats +@section Merging Files with If-then-else +@cindex merged output format +@cindex if-then-else output format +@cindex C if-then-else output format +@cindex @code{ifdef} output format + +You can use @code{diff} to merge two files of C source code. The output +of @code{diff} in this format contains all the lines of both files. +Lines common to both files are output just once; the differing parts are +separated by the C preprocessor directives @code{#ifdef @var{name}} or +@code{#ifndef @var{name}}, @code{#else}, and @code{#endif}. When +compiling the output, you select which version to use by either defining +or leaving undefined the macro @var{name}. + +To merge two files, use @code{diff} with the @samp{-D @var{name}} or +@samp{--ifdef=@var{name}} option. The argument @var{name} is the C +preprocessor identifier to use in the @code{#ifdef} and @code{#ifndef} +directives. + +For example, if you change an instance of @code{wait (&s)} to +@code{waitpid (-1, &s, 0)} and then merge the old and new files with +the @samp{--ifdef=HAVE_WAITPID} option, then the affected part of your code +might look like this: + +@example + do @{ +#ifndef HAVE_WAITPID + if ((w = wait (&s)) < 0 && errno != EINTR) +#else /* HAVE_WAITPID */ + if ((w = waitpid (-1, &s, 0)) < 0 && errno != EINTR) +#endif /* HAVE_WAITPID */ + return w; + @} while (w != child); +@end example + +You can specify formats for languages other than C by using line group +formats and line formats, as described in the next sections. + +@menu +* Line Group Formats:: Formats for general if-then-else line groups. +* Line Formats:: Formats for each line in a line group. +* Detailed If-then-else:: A detailed description of if-then-else format. +* Example If-then-else:: Sample if-then-else format output. +@end menu + +@node Line Group Formats, Line Formats, , If-then-else +@subsection Line Group Formats +@cindex line group formats +@cindex formats for if-then-else line groups + +Line group formats let you specify formats suitable for many +applications that allow if-then-else input, including programming +languages and text formatting languages. A line group format specifies +the output format for a contiguous group of similar lines. + +For example, the following command compares the TeX files @file{old} +and @file{new}, and outputs a merged file in which old regions are +surrounded by @samp{\begin@{em@}}-@samp{\end@{em@}} lines, and new +regions are surrounded by @samp{\begin@{bf@}}-@samp{\end@{bf@}} lines. + +@example +diff \ + --old-group-format='\begin@{em@} +%<\end@{em@} +' \ + --new-group-format='\begin@{bf@} +%>\end@{bf@} +' \ + old new +@end example + +The following command is equivalent to the above example, but it is a +little more verbose, because it spells out the default line group formats. + +@example +diff \ + --old-group-format='\begin@{em@} +%<\end@{em@} +' \ + --new-group-format='\begin@{bf@} +%>\end@{bf@} +' \ + --unchanged-group-format='%=' \ + --changed-group-format='\begin@{em@} +%<\end@{em@} +\begin@{bf@} +%>\end@{bf@} +' \ + old new +@end example + +Here is a more advanced example, which outputs a diff listing with +headers containing line numbers in a ``plain English'' style. + +@example +diff \ + --unchanged-group-format='' \ + --old-group-format='-------- %dn line%(n=1?:s) deleted at %df: +%<' \ + --new-group-format='-------- %dN line%(N=1?:s) added after %de: +%>' \ + --changed-group-format='-------- %dn line%(n=1?:s) changed at %df: +%<-------- to: +%>' \ + old new +@end example + +To specify a line group format, use @code{diff} with one of the options +listed below. You can specify up to four line group formats, one for +each kind of line group. You should quote @var{format}, because it +typically contains shell metacharacters. + +@table @samp +@item --old-group-format=@var{format} +These line groups are hunks containing only lines from the first file. +The default old group format is the same as the changed group format if +it is specified; otherwise it is a format that outputs the line group as-is. + +@item --new-group-format=@var{format} +These line groups are hunks containing only lines from the second +file. The default new group format is same as the the changed group +format if it is specified; otherwise it is a format that outputs the +line group as-is. + +@item --changed-group-format=@var{format} +These line groups are hunks containing lines from both files. The +default changed group format is the concatenation of the old and new +group formats. + +@item --unchanged-group-format=@var{format} +These line groups contain lines common to both files. The default +unchanged group format is a format that outputs the line group as-is. +@end table + +In a line group format, ordinary characters represent themselves; +conversion specifications start with @samp{%} and have one of the +following forms. + +@table @samp +@item %< +stands for the lines from the first file, including the trailing newline. +Each line is formatted according to the old line format (@pxref{Line Formats}). + +@item %> +stands for the lines from the second file, including the trailing newline. +Each line is formatted according to the new line format. + +@item %= +stands for the lines common to both files, including the trailing newline. +Each line is formatted according to the unchanged line format. + +@item %% +stands for @samp{%}. + +@item %c'@var{C}' +where @var{C} is a single character, stands for @var{C}. +@var{C} may not be a backslash or an apostrophe. +For example, @samp{%c':'} stands for a colon, even inside +the then-part of an if-then-else format, which a colon would +normally terminate. + +@item %c'\@var{O}' +where @var{O} is a string of 1, 2, or 3 octal digits, +stands for the character with octal code @var{O}. +For example, @samp{%c'\0'} stands for a null character. + +@item @var{F}@var{n} +where @var{F} is a @code{printf} conversion specification and @var{n} is one +of the following letters, stands for @var{n}'s value formatted with @var{F}. + +@table @samp +@item e +The line number of the line just before the group in the old file. + +@item f +The line number of the first line in the group in the old file; +equals @var{e} + 1. + +@item l +The line number of the last line in the group in the old file. + +@item m +The line number of the line just after the group in the old file; +equals @var{l} + 1. + +@item n +The number of lines in the group in the old file; equals @var{l} - @var{f} + 1. + +@item E, F, L, M, N +Likewise, for lines in the new file. + +@end table + +The @code{printf} conversion specification can be @samp{%d}, +@samp{%o}, @samp{%x}, or @samp{%X}, specifying decimal, octal, +lower case hexadecimal, or upper case hexadecimal output +respectively. After the @samp{%} the following options can appear in +sequence: a @samp{-} specifying left-justification; an integer +specifying the minimum field width; and a period followed by an +optional integer specifying the minimum number of digits. +For example, @samp{%5dN} prints the number of new lines in the group +in a field of width 5 characters, using the @code{printf} format @code{"%5d"}. + +@item (@var{A}=@var{B}?@var{T}:@var{E}) +If @var{A} equals @var{B} then @var{T} else @var{E}. +@var{A} and @var{B} are each either a decimal constant +or a single letter interpreted as above. +This format spec is equivalent to @var{T} if +@var{A}'s value equals @var{B}'s; otherwise it is equivalent to @var{E}. + +For example, @samp{%(N=0?no:%dN) line%(N=1?:s)} is equivalent to +@samp{no lines} if @var{N} (the number of lines in the group in the the +new file) is 0, to @samp{1 line} if @var{N} is 1, and to @samp{%dN lines} +otherwise. +@end table + +@node Line Formats, Detailed If-then-else, Line Group Formats, If-then-else +@subsection Line Formats +@cindex line formats + +Line formats control how each line taken from an input file is +output as part of a line group in if-then-else format. + +For example, the following command outputs text with a one-column +change indicator to the left of the text. The first column of output +is @samp{-} for deleted lines, @samp{|} for added lines, and a space +for unchanged lines. The formats contain newline characters where +newlines are desired on output. + +@example +diff \ + --old-line-format='-%l +' \ + --new-line-format='|%l +' \ + --unchanged-line-format=' %l +' \ + old new +@end example + +To specify a line format, use one of the following options. You should +quote @var{format}, since it often contains shell metacharacters. + +@table @samp +@item --old-line-format=@var{format} +formats lines just from the first file. + +@item --new-line-format=@var{format} +formats lines just from the second file. + +@item --unchanged-line-format=@var{format} +formats lines common to both files. + +@item --line-format=@var{format} +formats all lines; in effect, it sets all three above options simultaneously. +@end table + +In a line format, ordinary characters represent themselves; +conversion specifications start with @samp{%} and have one of the +following forms. + +@table @samp +@item %l +stands for the the contents of the line, not counting its trailing +newline (if any). This format ignores whether the line is incomplete; +@xref{Incomplete Lines}. + +@item %L +stands for the the contents of the line, including its trailing newline +(if any). If a line is incomplete, this format preserves its +incompleteness. + +@item %% +stands for @samp{%}. + +@item %c'@var{C}' +where @var{C} is a single character, stands for @var{C}. +@var{C} may not be a backslash or an apostrophe. +For example, @samp{%c':'} stands for a colon. + +@item %c'\@var{O}' +where @var{O} is a string of 1, 2, or 3 octal digits, +stands for the character with octal code @var{O}. +For example, @samp{%c'\0'} stands for a null character. + +@item @var{F}n +where @var{F} is a @code{printf} conversion specification, +stands for the line number formatted with @var{F}. +For example, @samp{%.5dn} prints the line number using the +@code{printf} format @code{"%.5d"}. @xref{Line Group Formats}, for +more about printf conversion specifications. + +@end table + +The default line format is @samp{%l} followed by a newline character. + +If the input contains tab characters and it is important that they line +up on output, you should ensure that @samp{%l} or @samp{%L} in a line +format is just after a tab stop (e.g.@: by preceding @samp{%l} or +@samp{%L} with a tab character), or you should use the @samp{-t} or +@samp{--expand-tabs} option. + +Taken together, the line and line group formats let you specify many +different formats. For example, the following command uses a format +similar to @code{diff}'s normal format. You can tailor this command +to get fine control over @code{diff}'s output. + +@example +diff \ + --old-line-format='< %l +' \ + --new-line-format='> %l +' \ + --old-group-format='%df%(f=l?:,%dl)d%dE +%<' \ + --new-group-format='%dea%dF%(F=L?:,%dL) +%>' \ + --changed-group-format='%df%(f=l?:,%dl)c%dF%(F=L?:,%dL) +%<--- +%>' \ + --unchanged-group-format='' \ + old new +@end example + +@node Detailed If-then-else, Example If-then-else, Line Formats, If-then-else +@subsection Detailed Description of If-then-else Format + +For lines common to both files, @code{diff} uses the unchanged line +group format. For each hunk of differences in the merged output +format, if the hunk contains only lines from the first file, +@code{diff} uses the old line group format; if the hunk contains only +lines from the second file, @code{diff} uses the new group format; +otherwise, @code{diff} uses the changed group format. + +The old, new, and unchanged line formats specify the output format of +lines from the first file, lines from the second file, and lines common +to both files, respectively. + +The option @samp{--ifdef=@var{name}} is equivalent to +the following sequence of options using shell syntax: + +@example +--old-group-format='#ifndef @var{name} +%<#endif /* not @var{name} */ +' \ +--new-group-format='#ifdef @var{name} +%>#endif /* @var{name} */ +' \ +--unchanged-group-format='%=' \ +--changed-group-format='#ifndef @var{name} +%<#else /* @var{name} */ +%>#endif /* @var{name} */ +' +@end example + +You should carefully check the @code{diff} output for proper nesting. +For example, when using the the @samp{-D @var{name}} or +@samp{--ifdef=@var{name}} option, you should check that if the +differing lines contain any of the C preprocessor directives +@samp{#ifdef}, @samp{#ifndef}, @samp{#else}, @samp{#elif}, or +@samp{#endif}, they are nested properly and match. If they don't, you +must make corrections manually. It is a good idea to carefully check +the resulting code anyway to make sure that it really does what you +want it to; depending on how the input files were produced, the output +might contain duplicate or otherwise incorrect code. + +The @code{patch} @samp{-D @var{name}} option behaves just like +the @code{diff} @samp{-D @var{name}} option, except it operates on +a file and a diff to produce a merged file; @xref{patch Options}. + +@node Example If-then-else, , Detailed If-then-else, If-then-else +@subsection An Example of If-then-else Format + +Here is the output of @samp{diff -DTWO lao tzu} (@pxref{Sample +diff Input}, for the complete contents of the two files): + +@example +#ifndef TWO +The Way that can be told of is not the eternal Way; +The name that can be named is not the eternal name. +#endif /* not TWO */ +The Nameless is the origin of Heaven and Earth; +#ifndef TWO +The Named is the mother of all things. +#else /* TWO */ +The named is the mother of all things. + +#endif /* TWO */ +Therefore let there always be non-being, + so we may see their subtlety, +And let there always be being, + so we may see their outcome. +The two are the same, +But after they are produced, + they have different names. +#ifdef TWO +They both may be called deep and profound. +Deeper and more profound, +The door of all subtleties! +#endif /* TWO */ +@end example + +@node Comparing Directories, Adjusting Output, Output Formats, Top +@chapter Comparing Directories + +You can use @code{diff} to compare some or all of the files in two +directory trees. When both file name arguments to @code{diff} are +directories, it compares each file that is contained in both +directories, examining file names in alphabetical order. Normally +@code{diff} is silent about pairs of files that contain no differences, +but if you use the @samp{-s} or @samp{--report-identical-files} option, +it reports pairs of identical files. Normally @code{diff} reports +subdirectories common to both directories without comparing +subdirectories' files, but if you use the @samp{-r} or +@samp{--recursive} option, it compares every corresponding pair of files +in the directory trees, as many levels deep as they go. + +For file names that are in only one of the directories, @code{diff} +normally does not show the contents of the file that exists; it reports +only that the file exists in that directory and not in the other. You +can make @code{diff} act as though the file existed but was empty in the +other directory, so that it outputs the entire contents of the file that +actually exists. (It is output as either an insertion or a +deletion, depending on whether it is in the first or the second +directory given.) To do this, use the @samp{-N} or @samp{--new-file} +option. + +If the older directory contains one or more large files that are not in +the newer directory, you can make the patch smaller by using the +@samp{-P} or @samp{--unidirectional-new-file} option instead of @samp{-N}. +This option is like @samp{-N} except that it only inserts the contents +of files that appear in the second directory but not the first (that is, +files that were added). At the top of the patch, write instructions for +the user applying the patch to remove the files that were deleted before +applying the patch. @xref{Making Patches}, for more discussion of +making patches for distribution. + +To ignore some files while comparing directories, use the @samp{-x +@var{pattern}} or @samp{--exclude=@var{pattern}} option. This option +ignores any files or subdirectories whose base names match the shell +pattern @var{pattern}. Unlike in the shell, a period at the start of +the base of a file name matches a wildcard at the start of a pattern. +You should enclose @var{pattern} in quotes so that the shell does not +expand it. For example, the option @samp{-x '*.[ao]'} ignores any file +whose name ends with @samp{.a} or @samp{.o}. + +This option accumulates if you specify it more than once. For example, +using the options @samp{-x 'RCS' -x '*,v'} ignores any file or +subdirectory whose base name is @samp{RCS} or ends with @samp{,v}. + +If you need to give this option many times, you can instead put the +patterns in a file, one pattern per line, and use the @samp{-X +@var{file}} or @samp{--exclude-from=@var{file}} option. + +If you have been comparing two directories and stopped partway through, +later you might want to continue where you left off. You can do this by +using the @samp{-S @var{file}} or @samp{--starting-file=@var{file}} +option. This compares only the file @var{file} and all alphabetically +later files in the topmost directory level. + +@node Adjusting Output, diff Performance, Comparing Directories, Top +@chapter Making @code{diff} Output Prettier + +@code{diff} provides several ways to adjust the appearance of its output. +These adjustments can be applied to any output format. + +@menu +* Tabs:: Preserving the alignment of tabstops. +* Pagination:: Page numbering and timestamping @code{diff} output. +@end menu + +@node Tabs, Pagination, , Adjusting Output +@section Preserving Tabstop Alignment +@cindex tabstop alignment +@cindex aligning tabstops + +The lines of text in some of the @code{diff} output formats are preceded +by one or two characters that indicate whether the text is inserted, +deleted, or changed. The addition of those characters can cause tabs to +move to the next tabstop, throwing off the alignment of columns in the +line. GNU @code{diff} provides two ways to make tab-aligned columns +line up correctly. + +The first way is to have @code{diff} convert all tabs into the correct +number of spaces before outputting them; select this method with the +@samp{-t} or @samp{--expand-tabs} option. @code{diff} assumes that +tabstops are set every 8 columns. To use this form of output with +@code{patch}, you must give @code{patch} the @samp{-l} or +@samp{--ignore-white-space} option (@pxref{Changed White Space}, for more +information). + +The other method for making tabs line up correctly is to add a tab +character instead of a space after the indicator character at the +beginning of the line. This ensures that all following tab characters +are in the same position relative to tabstops that they were in the +original files, so that the output is aligned correctly. Its +disadvantage is that it can make long lines too long to fit on one line +of the screen or the paper. It also does not work with the unified +output format, which does not have a space character after the change +type indicator character. Select this method with the @samp{-T} or +@samp{--initial-tab} option. + +@node Pagination, , Tabs, Adjusting Output +@section Paginating @code{diff} Output +@cindex paginating @code{diff} output + +It can be convenient to have long output page-numbered and time-stamped. +The @samp{-l} and @samp{--paginate} options do this by sending the +@code{diff} output through the @code{pr} program. Here is what the page +header might look like for @samp{diff -lc lao tzu}: + +@example +Mar 11 13:37 1991 diff -lc lao tzu Page 1 +@end example + +@node diff Performance, Comparing Three Files, Adjusting Output, Top +@chapter @code{diff} Performance Tradeoffs +@cindex performance of @code{diff} + +GNU @code{diff} runs quite efficiently; however, in some circumstances +you can cause it to run faster or produce a more compact set of changes. +There are two ways that you can affect the performance of GNU +@code{diff} by changing the way it compares files. + +Performance has more than one dimension. These options improve one +aspect of performance at the cost of another, or they improve +performance in some cases while hurting it in others. + +The way that GNU @code{diff} determines which lines have changed always +comes up with a near-minimal set of differences. Usually it is good +enough for practical purposes. If the @code{diff} output is large, you +might want @code{diff} to use a modified algorithm that sometimes +produces a smaller set of differences. The @samp{-d} or +@samp{--minimal} option does this; however, it can also cause +@code{diff} to run more slowly than usual, so it is not the default +behavior. + +When the files you are comparing are large and have small groups of +changes scattered throughout them, you can use the @samp{-H} or +@samp{--speed-large-files} option to make a different modification to +the algorithm that @code{diff} uses. If the input files have a constant +small density of changes, this option speeds up the comparisons without +changing the output. If not, @code{diff} might produce a larger set of +differences; however, the output will still be correct. + +Normally @code{diff} discards the prefix and suffix that is common to +both files before it attempts to find a minimal set of differences. +This makes @code{diff} run faster, but occasionally it may produce +non-minimal output. The @samp{--horizon-lines=@var{lines}} option +prevents @code{diff} from discarding the last @var{lines} lines of the +prefix and the first @var{lines} lines of the suffix. This gives +@code{diff} further opportunities to find a minimal output. + +@node Comparing Three Files, diff3 Merging, diff Performance, Top +@chapter Comparing Three Files +@cindex comparing three files +@cindex format of @code{diff3} output + +Use the program @code{diff3} to compare three files and show any +differences among them. (@code{diff3} can also merge files; see +@ref{diff3 Merging}). + +The ``normal'' @code{diff3} output format shows each hunk of +differences without surrounding context. Hunks are labeled depending +on whether they are two-way or three-way, and lines are annotated by +their location in the input files. + +@xref{Invoking diff3}, for more information on how to run @code{diff3}. + +@menu +* Sample diff3 Input:: Sample @code{diff3} input for examples. +* Detailed diff3 Normal:: A detailed description of normal output format. +* diff3 Hunks:: The format of normal output format. +* Example diff3 Normal:: Sample output in the normal format. +@end menu + +@node Sample diff3 Input, Detailed diff3 Normal, , Comparing Three Files +@section A Third Sample Input File +@cindex @code{diff3} sample input +@cindex sample input for @code{diff3} + +Here is a third sample file that will be used in examples to illustrate +the output of @code{diff3} and how various options can change it. The +first two files are the same that we used for @code{diff} (@pxref{Sample +diff Input}). This is the third sample file, called @file{tao}: + +@example +The Way that can be told of is not the eternal Way; +The name that can be named is not the eternal name. +The Nameless is the origin of Heaven and Earth; +The named is the mother of all things. + +Therefore let there always be non-being, + so we may see their subtlety, +And let there always be being, + so we may see their result. +The two are the same, +But after they are produced, + they have different names. + + -- The Way of Lao-Tzu, tr. Wing-tsit Chan +@end example + +@node Detailed diff3 Normal, diff3 Hunks, Sample diff3 Input, Comparing Three Files +@section Detailed Description of @code{diff3} Normal Format + +Each hunk begins with a line marked @samp{====}. Three-way hunks have +plain @samp{====} lines, and two-way hunks have @samp{1}, @samp{2}, or +@samp{3} appended to specify which of the three input files differ in +that hunk. The hunks contain copies of two or three sets of input +lines each preceded by one or two commands identifying where the lines +came from. + +Normally, two spaces precede each copy of an input line to distinguish +it from the commands. But with the @samp{-T} or @samp{--initial-tab} +option, @code{diff3} uses a tab instead of two spaces; this lines up +tabs correctly. @xref{Tabs}, for more information. + +Commands take the following forms: + +@table @samp +@item @var{file}:@var{l}a +This hunk appears after line @var{l} of file @var{file}, and +contains no lines in that file. To edit this file to yield the other +files, one must append hunk lines taken from the other files. For +example, @samp{1:11a} means that the hunk follows line 11 in the first +file and contains no lines from that file. + +@item @var{file}:@var{r}c +This hunk contains the lines in the range @var{r} of file @var{file}. +The range @var{r} is a comma-separated pair of line numbers, or just one +number if the range is a singleton. To edit this file to yield the +other files, one must change the specified lines to be the lines taken +from the other files. For example, @samp{2:11,13c} means that the hunk +contains lines 11 through 13 from the second file. +@end table + +If the last line in a set of input lines is incomplete +(@pxref{Incomplete Lines}), it is distinguished on output from a full +line by a following line that starts with @samp{\}. + +@node diff3 Hunks, Example diff3 Normal, Detailed diff3 Normal, Comparing Three Files +@section @code{diff3} Hunks +@cindex hunks for @code{diff3} +@cindex @code{diff3} hunks + +Groups of lines that differ in two or three of the input files are +called @dfn{diff3 hunks}, by analogy with @code{diff} hunks +(@pxref{Hunks}). If all three input files differ in a @code{diff3} +hunk, the hunk is called a @dfn{three-way hunk}; if just two input files +differ, it is a @dfn{two-way hunk}. + +As with @code{diff}, several solutions are possible. When comparing the +files @samp{A}, @samp{B}, and @samp{C}, @code{diff3} normally finds +@code{diff3} hunks by merging the two-way hunks output by the two +commands @samp{diff A B} and @samp{diff A C}. This does not necessarily +minimize the size of the output, but exceptions should be rare. + +For example, suppose @file{F} contains the three lines @samp{a}, +@samp{b}, @samp{f}, @file{G} contains the lines @samp{g}, @samp{b}, +@samp{g}, and @file{H} contains the lines @samp{a}, @samp{b}, +@samp{h}. @samp{diff3 F G H} might output the following: + +@example +====2 +1:1c +3:1c + a +2:1c + g +==== +1:3c + f +2:3c + g +3:3c + h +@end example + +@noindent +because it found a two-way hunk containing @samp{a} in the first and +third files and @samp{g} in the second file, then the single line +@samp{b} common to all three files, then a three-way hunk containing +the last line of each file. + +@node Example diff3 Normal, , diff3 Hunks, Comparing Three Files +@section An Example of @code{diff3} Normal Format + +Here is the output of the command @samp{diff3 lao tzu tao} +(@pxref{Sample diff3 Input}, for the complete contents of the files). +Notice that it shows only the lines that are different among the three +files. + +@example +====2 +1:1,2c +3:1,2c + The Way that can be told of is not the eternal Way; + The name that can be named is not the eternal name. +2:0a +====1 +1:4c + The Named is the mother of all things. +2:2,3c +3:4,5c + The named is the mother of all things. + +====3 +1:8c +2:7c + so we may see their outcome. +3:9c + so we may see their result. +==== +1:11a +2:11,13c + They both may be called deep and profound. + Deeper and more profound, + The door of all subtleties! +3:13,14c + + -- The Way of Lao-Tzu, tr. Wing-tsit Chan +@end example + +@node diff3 Merging, Interactive Merging, Comparing Three Files, Top +@chapter Merging From a Common Ancestor +@cindex merging from a common ancestor + +When two people have made changes to copies of the same file, +@code{diff3} can produce a merged output that contains both sets of +changes together with warnings about conflicts. + +One might imagine programs with names like @code{diff4} and @code{diff5} +to compare more than three files simultaneously, but in practice the +need rarely arises. You can use @code{diff3} to merge three or more +sets of changes to a file by merging two change sets at a time. + +@code{diff3} can incorporate changes from two modified versions into a +common preceding version. This lets you merge the sets of changes +represented by the two newer files. Specify the common ancestor version +as the second argument and the two newer versions as the first and third +arguments, like this: + +@example +diff3 @var{mine} @var{older} @var{yours} +@end example + +@noindent +You can remember the order of the arguments by noting that they are in +alphabetical order. + +@cindex conflict +@cindex overlap +You can think of this as subtracting @var{older} from @var{yours} and +adding the result to @var{mine}, or as merging into @var{mine} the +changes that would turn @var{older} into @var{yours}. This merging is +well-defined as long as @var{mine} and @var{older} match in the +neighborhood of each such change. This fails to be true when all three +input files differ or when only @var{older} differs; we call this +a @dfn{conflict}. When all three input files differ, we call the +conflict an @dfn{overlap}. + +@code{diff3} gives you several ways to handle overlaps and conflicts. +You can omit overlaps or conflicts, or select only overlaps, +or mark conflicts with special @samp{<<<<<<<} and @samp{>>>>>>>} lines. + +@code{diff3} can output the merge results as an @code{ed} script that +that can be applied to the first file to yield the merged output. +However, it is usually better to have @code{diff3} generate the merged +output directly; this bypasses some problems with @code{ed}. + +@menu +* Which Changes:: Selecting changes to incorporate. +* Marking Conflicts:: Marking conflicts. +* Bypassing ed:: Generating merged output directly. +* Merging Incomplete Lines:: How @code{diff3} merges incomplete lines. +* Saving the Changed File:: Emulating System V behavior. +@end menu + +@node Which Changes, Marking Conflicts, , diff3 Merging +@section Selecting Which Changes to Incorporate +@cindex overlapping change, selection of +@cindex unmerged change + +You can select all unmerged changes from @var{older} to @var{yours} for merging +into @var{mine} with the @samp{-e} or @samp{--ed} option. You can +select only the nonoverlapping unmerged changes with @samp{-3} or +@samp{--easy-only}, and you can select only the overlapping changes with +@samp{-x} or @samp{--overlap-only}. + +The @samp{-e}, @samp{-3} and @samp{-x} options select only +@dfn{unmerged changes}, i.e.@: changes where @var{mine} and @var{yours} +differ; they ignore changes from @var{older} to @var{yours} where +@var{mine} and @var{yours} are identical, because they assume that such +changes have already been merged. If this assumption is not a safe +one, you can use the @samp{-A} or @samp{--show-all} option +(@pxref{Marking Conflicts}). + +Here is the output of the command @code{diff3} with each of these three +options (@pxref{Sample diff3 Input}, for the complete contents of the files). +Notice that @samp{-e} outputs the union of the disjoint sets of changes +output by @samp{-3} and @samp{-x}. + +Output of @samp{diff3 -e lao tzu tao}: +@example +11a + + -- The Way of Lao-Tzu, tr. Wing-tsit Chan +. +8c + so we may see their result. +. +@end example + +Output of @samp{diff3 -3 lao tzu tao}: +@example +8c + so we may see their result. +. +@end example + +Output of @samp{diff3 -x lao tzu tao}: +@example +11a + + -- The Way of Lao-Tzu, tr. Wing-tsit Chan +. +@end example + +@node Marking Conflicts, Bypassing ed, Which Changes, diff3 Merging +@section Marking Conflicts +@cindex conflict marking +@cindex @samp{<<<<<<<} for marking conflicts + +@code{diff3} can mark conflicts in the merged output by +bracketing them with special marker lines. A conflict +that comes from two files @var{A} and @var{B} is marked as follows: + +@example +<<<<<<< @var{A} +@r{lines from @var{A}} +======= +@r{lines from @var{B}} +>>>>>>> @var{B} +@end example + +A conflict that comes from three files @var{A}, @var{B} and @var{C} is +marked as follows: + +@example +<<<<<<< @var{A} +@r{lines from @var{A}} +||||||| @var{B} +@r{lines from @var{B}} +======= +@r{lines from @var{C}} +>>>>>>> @var{C} +@end example + +The @samp{-A} or @samp{--show-all} option acts like the @samp{-e} +option, except that it brackets conflicts, and it outputs all changes +from @var{older} to @var{yours}, not just the unmerged changes. Thus, +given the sample input files (@pxref{Sample diff3 Input}), @samp{diff3 +-A lao tzu tao} puts brackets around the conflict where only @file{tzu} +differs: + +@example +<<<<<<< tzu +======= +The Way that can be told of is not the eternal Way; +The name that can be named is not the eternal name. +>>>>>>> tao +@end example + +And it outputs the three-way conflict as follows: + +@example +<<<<<<< lao +||||||| tzu +They both may be called deep and profound. +Deeper and more profound, +The door of all subtleties! +======= + + -- The Way of Lao-Tzu, tr. Wing-tsit Chan +>>>>>>> tao +@end example + +The @samp{-E} or @samp{--show-overlap} option outputs less information +than the @samp{-A} or @samp{--show-all} option, because it outputs only +unmerged changes, and it never outputs the contents of the second +file. Thus the @samp{-E} option acts like the @samp{-e} option, +except that it brackets the first and third files from three-way +overlapping changes. Similarly, @samp{-X} acts like @samp{-x}, except +it brackets all its (necessarily overlapping) changes. For example, +for the three-way overlapping change above, the @samp{-E} and @samp{-X} +options output the following: + +@example +<<<<<<< lao +======= + + -- The Way of Lao-Tzu, tr. Wing-tsit Chan +>>>>>>> tao +@end example + +If you are comparing files that have meaningless or uninformative names, +you can use the @samp{-L @var{label}} or @samp{--label=@var{label}} +option to show alternate names in the @samp{<<<<<<<}, @samp{|||||||} +and @samp{>>>>>>>} brackets. This option can be given up to three +times, once for each input file. Thus @samp{diff3 -A -L X -L Y -L Z A +B C} acts like @samp{diff3 -A A B C}, except that the output looks like +it came from files named @samp{X}, @samp{Y} and @samp{Z} rather than +from files named @samp{A}, @samp{B} and @samp{C}. + +@node Bypassing ed, Merging Incomplete Lines, Marking Conflicts, diff3 Merging +@section Generating the Merged Output Directly +@cindex merged @code{diff3} format + +With the @samp{-m} or @samp{--merge} option, @code{diff3} outputs the +merged file directly. This is more efficient than using @code{ed} to +generate it, and works even with non-text files that @code{ed} would +reject. If you specify @samp{-m} without an @code{ed} script option, +@samp{-A} (@samp{--show-all}) is assumed. + +For example, the command @samp{diff3 -m lao tzu tao} +(@pxref{Sample diff3 Input} for a copy of the input files) would output +the following: + +@example +<<<<<<< tzu +======= +The Way that can be told of is not the eternal Way; +The name that can be named is not the eternal name. +>>>>>>> tao +The Nameless is the origin of Heaven and Earth; +The Named is the mother of all things. +Therefore let there always be non-being, + so we may see their subtlety, +And let there always be being, + so we may see their result. +The two are the same, +But after they are produced, + they have different names. +<<<<<<< lao +||||||| tzu +They both may be called deep and profound. +Deeper and more profound, +The door of all subtleties! +======= + + -- The Way of Lao-Tzu, tr. Wing-tsit Chan +>>>>>>> tao +@end example + +@node Merging Incomplete Lines, Saving the Changed File, Bypassing ed, diff3 Merging +@section How @code{diff3} Merges Incomplete Lines +@cindex incomplete line merging + +With @samp{-m}, incomplete lines (@pxref{Incomplete Lines}) are simply +copied to the output as they are found; if the merged output ends in an +conflict and one of the input files ends in an incomplete +line, succeeding @samp{|||||||}, @samp{=======} or @samp{>>>>>>>} +brackets appear somewhere other than the start of a line because +they are appended to the incomplete line. + +Without @samp{-m}, if an @code{ed} script option is specified and an +incomplete line is found, @code{diff3} generates a warning and acts as +if a newline had been present. + +@node Saving the Changed File, , Merging Incomplete Lines, diff3 Merging +@section Saving the Changed File +@cindex System V @code{diff3} compatibility + +Traditional Unix @code{diff3} generates an @code{ed} script without the +trailing @samp{w} and and @samp{q} commands that save the changes. +System V @code{diff3} generates these extra commands. GNU @code{diff3} +normally behaves like traditional Unix @code{diff3}, but with the +@samp{-i} option it behaves like System V @code{diff3} and appends the +@samp{w} and @samp{q} commands. + +The @samp{-i} option requires one of the @code{ed} script options +@samp{-AeExX3}, and is incompatible with the merged output option +@samp{-m}. + +@node Interactive Merging, Merging with patch, diff3 Merging, Top +@chapter Interactive Merging with @code{sdiff} +@cindex diff merging +@cindex interactive merging + +With @code{sdiff}, you can merge two files interactively based on a +side-by-side @samp{-y} format comparison (@pxref{Side by Side}). Use +@samp{-o @var{file}} or @samp{--output=@var{file}} to specify where to +put the merged text. @xref{Invoking sdiff}, for more details on the +options to @code{sdiff}. + +Another way to merge files interactively is to use the Emacs Lisp +package @code{emerge}. @xref{emerge, , emerge, emacs, The GNU Emacs +Manual}, for more information. + +@menu +* sdiff Option Summary::Summary of @code{sdiff} options. +* Merge Commands:: Merging two files interactively. +@end menu + +@node sdiff Option Summary, Merge Commands, , Interactive Merging +@section Specifying @code{diff} Options to @code{sdiff} +@cindex @code{sdiff} output format + +The following @code{sdiff} options have the same meaning as for +@code{diff}. @xref{diff Options}, for the use of these options. + +@example +-a -b -d -i -t -v +-B -H -I @var{regexp} + +--ignore-blank-lines --ignore-case +--ignore-matching-lines=@var{regexp} --ignore-space-change +--left-column --minimal --speed-large-files +--suppress-common-lines --expand-tabs +--text --version --width=@var{columns} +@end example + +For historical reasons, @code{sdiff} has alternate names for some +options. The @samp{-l} option is equivalent to the @samp{--left-column} +option, and similarly @samp{-s} is equivalent to +@samp{--suppress-common-lines}. The meaning of the @code{sdiff} +@samp{-w} and @samp{-W} options is interchanged from that of +@code{diff}: with @code{sdiff}, @samp{-w @var{columns}} is equivalent to +@samp{--width=@var{columns}}, and @samp{-W} is equivalent to +@samp{--ignore-all-space}. @code{sdiff} without the @samp{-o} option is +equivalent to @code{diff} with the @samp{-y} or @samp{--side-by-side} +option (@pxref{Side by Side}). + +@node Merge Commands, , sdiff Option Summary, Interactive Merging +@section Merge Commands +@cindex merge commands +@cindex merging interactively + +Groups of common lines, with a blank gutter, are copied from the first +file to the output. After each group of differing lines, @code{sdiff} +prompts with @samp{%} and pauses, waiting for one of the following +commands. Follow each command with @key{RET}. + +@table @samp +@item e +Discard both versions. +Invoke a text editor on an empty temporary file, +then copy the resulting file to the output. + +@item eb +Concatenate the two versions, edit the result in a temporary file, +then copy the edited result to the output. + +@item el +Edit a copy of the left version, then copy the result to the output. + +@item er +Edit a copy of the right version, then copy the result to the output. + +@item l +Copy the left version to the output. + +@item q +Quit. + +@item r +Copy the right version to the output. + +@item s +Silently copy common lines. + +@item v +Verbosely copy common lines. This is the default. +@end table + +The text editor invoked is specified by the @code{EDITOR} environment +variable if it is set. The default is system-dependent. + +@node Merging with patch, Making Patches, Interactive Merging, Top +@chapter Merging with @code{patch} + +@code{patch} takes comparison output produced by @code{diff} and applies +the differences to a copy of the original file, producing a patched +version. With @code{patch}, you can distribute just the changes to a +set of files instead of distributing the entire file set; your +correspondents can apply @code{patch} to update their copy of the files +with your changes. @code{patch} automatically determines the diff +format, skips any leading or trailing headers, and uses the headers to +determine which file to patch. This lets your correspondents feed an +article or message containing a difference listing directly to +@code{patch}. + +@code{patch} detects and warns about common problems like forward +patches. It saves the original version of the files it patches, and +saves any patches that it could not apply. It can also maintain a +@code{patchlevel.h} file to ensures that your correspondents apply +diffs in the proper order. + +@code{patch} accepts a series of diffs in its standard input, usually +separated by headers that specify which file to patch. It applies +@code{diff} hunks (@pxref{Hunks}) one by one. If a hunk does not +exactly match the original file, @code{patch} uses heuristics to try to +patch the file as well as it can. If no approximate match can be found, +@code{patch} rejects the hunk and skips to the next hunk. @code{patch} +normally replaces each file @var{f} with its new version, saving the +original file in @samp{@var{f}.orig}, and putting reject hunks (if any) +into @samp{@var{f}.rej}. + +@xref{Invoking patch}, for detailed information on the options to +@code{patch}. @xref{Backups}, for more information on how +@code{patch} names backup files. @xref{Rejects}, for more information +on where @code{patch} puts reject hunks. + +@menu +* patch Input:: Selecting the type of @code{patch} input. +* Imperfect:: Dealing with imperfect patches. +* Empty Files:: Removing empty files after patching. +* Multiple Patches:: Handling multiple patches in a file specially. +* patch Messages:: Messages and questions @code{patch} can produce. +@end menu + +@node patch Input, Imperfect, , Merging with patch +@section Selecting the @code{patch} Input Format +@cindex @code{patch} input format + +@code{patch} normally determines which @code{diff} format the patch +file uses by examining its contents. For patch files that contain +particularly confusing leading text, you might need to use one of the +following options to force @code{patch} to interpret the patch file as a +certain format of diff. The output formats listed here are the only +ones that @code{patch} can understand. + +@table @samp +@item -c +@itemx --context +context diff. + +@item -e +@itemx --ed +@code{ed} script. + +@item -n +@itemx --normal +normal diff. + +@item -u +@itemx --unified +unified diff. +@end table + +@node Imperfect, Empty Files, patch Input, Merging with patch +@section Applying Imperfect Patches +@cindex imperfect patch application + +@code{patch} tries to skip any leading text in the patch file, apply the +diff, and then skip any trailing text. Thus you can feed a news article +or mail message directly to @code{patch}, and it should work. If the +entire diff is indented by a constant amount of white space, @code{patch} +automatically ignores the indentation. + +However, certain other types of imperfect input require user +intervention. + +@menu +* Changed White Space:: When tabs and spaces don't match exactly. +* Reversed Patches:: Applying reversed patches correctly. +* Inexact:: Helping @code{patch} find close matches. +@end menu + +@node Changed White Space, Reversed Patches, , Imperfect +@subsection Applying Patches with Changed White Space +@cindex white space in patches + +Sometimes mailers, editors, or other programs change spaces into tabs, +or vice versa. If this happens to a patch file or an input file, the +files might look the same, but @code{patch} will not be able to match +them properly. If this problem occurs, use the @samp{-l} or +@samp{--ignore-white-space} option, which makes @code{patch} compare +white space loosely so that any sequence of white space in the patch file +matches any sequence of white space in the input files. Non-white-space +characters must still match exactly. Each line of the context must +still match a line in the input file. + +@node Reversed Patches, Inexact, Changed White Space, Imperfect +@subsection Applying Reversed Patches +@cindex reversed patches + +Sometimes people run @code{diff} with the new file first instead of +second. This creates a diff that is ``reversed''. To apply such +patches, give @code{patch} the @samp{-R} or @samp{--reverse} option. +@code{patch} then attempts to swap each hunk around before applying it. +Rejects come out in the swapped format. The @samp{-R} option does not +work with @code{ed} scripts because there is too little information in +them to reconstruct the reverse operation. + +Often @code{patch} can guess that the patch is reversed. If the first +hunk of a patch fails, @code{patch} reverses the hunk to see if it can +apply it that way. If it can, @code{patch} asks you if you want to have +the @samp{-R} option set; if it can't, @code{patch} continues to apply +the patch normally. This method cannot detect a reversed patch if it is +a normal diff and the first command is an append (which should have been +a delete) since appends always succeed, because a null context matches +anywhere. But most patches add or change lines rather than delete them, +so most reversed normal diffs begin with a delete, which fails, and +@code{patch} notices. + +If you apply a patch that you have already applied, @code{patch} thinks +it is a reversed patch and offers to un-apply the patch. This could be +construed as a feature. If you did this inadvertently and you don't +want to un-apply the patch, just answer @samp{n} to this offer and to +the subsequent ``apply anyway'' question---or type @kbd{C-c} to kill the +@code{patch} process. + +@node Inexact, , Reversed Patches, Imperfect +@subsection Helping @code{patch} Find Inexact Matches +@cindex inexact patches +@cindex fuzz factor when patching + +For context diffs, and to a lesser extent normal diffs, @code{patch} can +detect when the line numbers mentioned in the patch are incorrect, and +it attempts to find the correct place to apply each hunk of the patch. +As a first guess, it takes the line number mentioned in the hunk, plus +or minus any offset used in applying the previous hunk. If that is not +the correct place, @code{patch} scans both forward and backward for a +set of lines matching the context given in the hunk. + +First @code{patch} looks for a place where all lines of the context +match. If it cannot find such a place, and it is reading a context or +unified diff, and the maximum fuzz factor is set to 1 or more, then +@code{patch} makes another scan, ignoring the first and last line of +context. If that fails, and the maximum fuzz factor is set to 2 or +more, it makes another scan, ignoring the first two and last two lines +of context are ignored. It continues similarly if the maximum fuzz +factor is larger. + +The @samp{-F @var{lines}} or @samp{--fuzz=@var{lines}} option sets the +maximum fuzz factor to @var{lines}. This option only applies to context +and unified diffs; it ignores up to @var{lines} lines while looking for +the place to install a hunk. Note that a larger fuzz factor increases +the odds of making a faulty patch. The default fuzz factor is 2; it may +not be set to more than the number of lines of context in the diff, +ordinarily 3. + +If @code{patch} cannot find a place to install a hunk of the patch, it +writes the hunk out to a reject file (@pxref{Rejects}, for information +on how reject files are named). It writes out rejected hunks in context +format no matter what form the input patch is in. If the input is a +normal or @code{ed} diff, many of the contexts are simply null. The +line numbers on the hunks in the reject file may be different from those +in the patch file: they show the approximate location where @code{patch} +thinks the failed hunks belong in the new file rather than in the old +one. + +As it completes each hunk, @code{patch} tells you whether the hunk +succeeded or failed, and if it failed, on which line (in the new file) +@code{patch} thinks the hunk should go. If this is different from the +line number specified in the diff, it tells you the offset. A single +large offset @emph{may} indicate that @code{patch} installed a hunk in +the wrong place. @code{patch} also tells you if it used a fuzz factor +to make the match, in which case you should also be slightly suspicious. + +@code{patch} cannot tell if the line numbers are off in an @code{ed} +script, and can only detect wrong line numbers in a normal diff when it +finds a change or delete command. It may have the same problem with a +context diff using a fuzz factor equal to or greater than the number of +lines of context shown in the diff (typically 3). In these cases, you +should probably look at a context diff between your original and patched +input files to see if the changes make sense. Compiling without errors +is a pretty good indication that the patch worked, but not a guarantee. + +@code{patch} usually produces the correct results, even when it must +make many guesses. However, the results are guaranteed only when +the patch is applied to an exact copy of the file that the patch was +generated from. + +@node Empty Files, Multiple Patches, Imperfect, Merging with patch +@section Removing Empty Files +@cindex empty files, removing +@cindex removing empty files + +Sometimes when comparing two directories, the first directory contains a +file that the second directory does not. If you give @code{diff} the +@samp{-N} or @samp{--new-file} option, it outputs a diff that deletes +the contents of this file. By default, @code{patch} leaves an empty +file after applying such a diff. The @samp{-E} or +@samp{--remove-empty-files} option to @code{patch} deletes output files +that are empty after applying the diff. + +@node Multiple Patches, patch Messages, Empty Files, Merging with patch +@section Multiple Patches in a File +@cindex multiple patches + +If the patch file contains more than one patch, @code{patch} tries to +apply each of them as if they came from separate patch files. This +means that it determines the name of the file to patch for each patch, +and that it examines the leading text before each patch for file names +and prerequisite revision level (@pxref{Making Patches}, for more on +that topic). + +For the second and subsequent patches in the patch file, you can give +options and another original file name by separating their argument +lists with a @samp{+}. However, the argument list for a second or +subsequent patch may not specify a new patch file, since that does not +make sense. + +For example, to tell @code{patch} to strip the first three slashes from +the name of the first patch in the patch file and none from subsequent +patches, and to use @file{code.c} as the first input file, you can use: + +@example +patch -p3 code.c + -p0 < patchfile +@end example + +The @samp{-S} or @samp{--skip} option ignores the current patch from the +patch file, but continue looking for the next patch in the file. Thus, +to ignore the first and third patches in the patch file, you can use: + +@example +patch -S + + -S + < patch file +@end example + +@node patch Messages, , Multiple Patches, Merging with patch +@section Messages and Questions from @code{patch} +@cindex @code{patch} messages and questions +@cindex diagnostics from @code{patch} +@cindex messages from @code{patch} + +@code{patch} can produce a variety of messages, especially if it has +trouble decoding its input. In a few situations where it's not sure how +to proceed, @code{patch} normally prompts you for more information from +the keyboard. There are options to suppress printing non-fatal messages +and stopping for keyboard input. + +The message @samp{Hmm...} indicates that @code{patch} is reading text in +the patch file, attempting to determine whether there is a patch in that +text, and if so, what kind of patch it is. + +You can inhibit all terminal output from @code{patch}, unless an error +occurs, by using the @samp{-s}, @samp{--quiet}, or @samp{--silent} +option. + +There are two ways you can prevent @code{patch} from asking you any +questions. The @samp{-f} or @samp{--force} option assumes that you know +what you are doing. It assumes the following: + +@itemize @bullet +@item +skip patches that do not contain file names in their headers; + +@item +patch files even though they have the wrong version for the +@samp{Prereq:} line in the patch; + +@item +assume that patches are not reversed even if they look like they are. +@end itemize + +The @samp{-t} or @samp{--batch} option is similar to @samp{-f}, in that +it suppresses questions, but it makes somewhat different assumptions: + +@itemize @bullet +@item +skip patches that do not contain file names in their headers +(the same as @samp{-f}); + +@item +skip patches for which the file has the wrong version for the +@samp{Prereq:} line in the patch; + +@item +assume that patches are reversed if they look like they are. +@end itemize + +@code{patch} exits with a non-zero status if it creates any reject +files. When applying a set of patches in a loop, you should check the +exit status, so you don't apply a later patch to a partially patched +file. + +@node Making Patches, Invoking cmp, Merging with patch, Top +@chapter Tips for Making Patch Distributions +@cindex patch making tips +@cindex tips for patch making + +Here are some things you should keep in mind if you are going to +distribute patches for updating a software package. + +Make sure you have specified the file names correctly, either in a +context diff header or with an @samp{Index:} line. If you are patching +files in a subdirectory, be sure to tell the patch user to specify a +@samp{-p} or @samp{--strip} option as needed. Take care to not send out +reversed patches, since these make people wonder whether they have +already applied the patch. + +To save people from partially applying a patch before other patches that +should have gone before it, you can make the first patch in the patch +file update a file with a name like @file{patchlevel.h} or +@file{version.c}, which contains a patch level or version number. If +the input file contains the wrong version number, @code{patch} will +complain immediately. + +An even clearer way to prevent this problem is to put a @samp{Prereq:} +line before the patch. If the leading text in the patch file contains a +line that starts with @samp{Prereq:}, @code{patch} takes the next word +from that line (normally a version number) and checks whether the next +input file contains that word, preceded and followed by either +white space or a newline. If not, @code{patch} prompts you for +confirmation before proceeding. This makes it difficult to accidentally +apply patches in the wrong order. + +Since @code{patch} does not handle incomplete lines properly, make sure +that all the source files in your program end with a newline whenever +you release a version. + +To create a patch that changes an older version of a package into a +newer version, first make a copy of the older version in a scratch +directory. Typically you do that by unpacking a @code{tar} or +@code{shar} archive of the older version. + +You might be able to reduce the size of the patch by renaming or +removing some files before making the patch. If the older version of +the package contains any files that the newer version does not, or if +any files have been renamed between the two versions, make a list of +@code{rm} and @code{mv} commands for the user to execute in the old +version directory before applying the patch. Then run those commands +yourself in the scratch directory. + +If there are any files that you don't need to include in the patch +because they can easily be rebuilt from other files (for example, +@file{TAGS} and output from @code{yacc} and @code{makeinfo}), replace +the versions in the scratch directory with the newer versions, using +@code{rm} and @code{ln} or @code{cp}. + +Now you can create the patch. The de-facto standard @code{diff} format +for patch distributions is context format with two lines of context, +produced by giving @code{diff} the @samp{-C 2} option. Do not use less +than two lines of context, because @code{patch} typically needs at +least two lines for proper operation. Give @code{diff} the @samp{-P} +option in case the newer version of the package contains any files that +the older one does not. Make sure to specify the scratch directory +first and the newer directory second. + +Add to the top of the patch a note telling the user any @code{rm} and +@code{mv} commands to run before applying the patch. Then you can +remove the scratch directory. + +@node Invoking cmp, Invoking diff, Making Patches, Top +@chapter Invoking @code{cmp} +@cindex invoking @code{cmp} +@cindex @code{cmp} invocation + +The @code{cmp} command compares two files, and if they differ, tells the +first byte and line number where they differ. Its arguments are as +follows: + +@example +cmp @var{options}@dots{} @var{from-file} @r{[}@var{to-file}@var{]} +@end example + +The file name @samp{-} is always the standard input. @code{cmp} also +uses the standard input if one file name is omitted. + +An exit status of 0 means no differences were found, 1 means some +differences were found, and 2 means trouble. + +@menu +* cmp Options:: Summary of options to @code{cmp}. +@end menu + +@node cmp Options, , , Invoking cmp +@section Options to @code{cmp} +@cindex @code{cmp} options +@cindex options for @code{cmp} + +Below is a summary of all of the options that GNU @code{cmp} accepts. +Most options have two equivalent names, one of which is a single letter +preceded by @samp{-}, and the other of which is a long name preceded by +@samp{--}. Multiple single letter options (unless they take an +argument) can be combined into a single command line word: @samp{-cl} is +equivalent to @samp{-c -l}. + +@table @samp +@item -c +Print the differing characters. Display control characters as a +@samp{^} followed by a letter of the alphabet and precede characters +that have the high bit set with @samp{M-} (which stands for ``meta''). + +@item --ignore-initial=@var{bytes} +Ignore any differences in the the first @var{bytes} bytes of the input files. +Treat files with fewer than @var{bytes} bytes as if they are empty. + +@item -l +Print the (decimal) offsets and (octal) values of all differing bytes. + +@item --print-chars +Print the differing characters. Display control characters as a +@samp{^} followed by a letter of the alphabet and precede characters +that have the high bit set with @samp{M-} (which stands for ``meta''). + +@item --quiet +@itemx -s +@itemx --silent +Do not print anything; only return an exit status indicating whether +the files differ. + +@item --verbose +Print the (decimal) offsets and (octal) values of all differing bytes. + +@item -v +@item --version +Output the version number of @code{cmp}. +@end table + +@node Invoking diff, Invoking diff3, Invoking cmp, Top +@chapter Invoking @code{diff} +@cindex invoking @code{diff} +@cindex @code{diff} invocation + +The format for running the @code{diff} command is: + +@example +diff @var{options}@dots{} @var{from-file} @var{to-file} +@end example + +In the simplest case, @code{diff} compares the contents of the two files +@var{from-file} and @var{to-file}. A file name of @samp{-} stands for +text read from the standard input. As a special case, @samp{diff - -} +compares a copy of standard input to itself. + +If @var{from-file} is a directory and @var{to-file} is not, @code{diff} +compares the file in @var{from-file} whose file name is that of @var{to-file}, +and vice versa. The non-directory file must not be @samp{-}. + +If both @var{from-file} and @var{to-file} are directories, +@code{diff} compares corresponding files in both directories, in +alphabetical order; this comparison is not recursive unless the +@samp{-r} or @samp{--recursive} option is given. @code{diff} never +compares the actual contents of a directory as if it were a file. The +file that is fully specified may not be standard input, because standard +input is nameless and the notion of ``file with the same name'' does not +apply. + +@code{diff} options begin with @samp{-}, so normally @var{from-file} and +@var{to-file} may not begin with @samp{-}. However, @samp{--} as an +argument by itself treats the remaining arguments as file names even if +they begin with @samp{-}. + +An exit status of 0 means no differences were found, 1 means some +differences were found, and 2 means trouble. + +@menu +* diff Options:: Summary of options to @code{diff}. +@end menu + +@node diff Options, , , Invoking diff +@section Options to @code{diff} +@cindex @code{diff} options +@cindex options for @code{diff} + +Below is a summary of all of the options that GNU @code{diff} accepts. +Most options have two equivalent names, one of which is a single letter +preceded by @samp{-}, and the other of which is a long name preceded by +@samp{--}. Multiple single letter options (unless they take an +argument) can be combined into a single command line word: @samp{-ac} is +equivalent to @samp{-a -c}. Long named options can be abbreviated to +any unique prefix of their name. Brackets ([ and ]) indicate that an +option takes an optional argument. + +@table @samp +@item -@var{lines} +Show @var{lines} (an integer) lines of context. This option does not +specify an output format by itself; it has no effect unless it is +combined with @samp{-c} (@pxref{Context Format}) or @samp{-u} +(@pxref{Unified Format}). This option is obsolete. For proper +operation, @code{patch} typically needs at least two lines of context. + +@item -a +Treat all files as text and compare them line-by-line, even if they +do not seem to be text. @xref{Binary}. + +@item -b +Ignore changes in amount of white space. @xref{White Space}. + +@item -B +Ignore changes that just insert or delete blank lines. @xref{Blank +Lines}. + +@item --binary +Read and write data in binary mode. @xref{Binary}. + +@item --brief +Report only whether the files differ, not the details of the +differences. @xref{Brief}. + +@item -c +Use the context output format. @xref{Context Format}. + +@item -C @var{lines} +@itemx --context@r{[}=@var{lines}@r{]} +Use the context output format, showing @var{lines} (an integer) lines of +context, or three if @var{lines} is not given. @xref{Context Format}. +For proper operation, @code{patch} typically needs at least two lines of +context. + +@item --changed-group-format=@var{format} +Use @var{format} to output a line group containing differing lines from +both files in if-then-else format. @xref{Line Group Formats}. + +@item -d +Change the algorithm perhaps find a smaller set of changes. This makes +@code{diff} slower (sometimes much slower). @xref{diff Performance}. + +@item -D @var{name} +Make merged @samp{#ifdef} format output, conditional on the preprocessor +macro @var{name}. @xref{If-then-else}. + +@item -e +@itemx --ed +Make output that is a valid @code{ed} script. @xref{ed Scripts}. + +@item --exclude=@var{pattern} +When comparing directories, ignore files and subdirectories whose basenames +match @var{pattern}. @xref{Comparing Directories}. + +@item --exclude-from=@var{file} +When comparing directories, ignore files and subdirectories whose basenames +match any pattern contained in @var{file}. @xref{Comparing Directories}. + +@item --expand-tabs +Expand tabs to spaces in the output, to preserve the alignment of tabs +in the input files. @xref{Tabs}. + +@item -f +Make output that looks vaguely like an @code{ed} script but has changes +in the order they appear in the file. @xref{Forward ed}. + +@item -F @var{regexp} +In context and unified format, for each hunk of differences, show some +of the last preceding line that matches @var{regexp}. @xref{Specified +Headings}. + +@item --forward-ed +Make output that looks vaguely like an @code{ed} script but has changes +in the order they appear in the file. @xref{Forward ed}. + +@item -h +This option currently has no effect; it is present for Unix +compatibility. + +@item -H +Use heuristics to speed handling of large files that have numerous +scattered small changes. @xref{diff Performance}. + +@item --horizon-lines=@var{lines} +Do not discard the last @var{lines} lines of the common prefix +and the first @var{lines} lines of the common suffix. +@xref{diff Performance}. + +@item -i +Ignore changes in case; consider upper- and lower-case letters +equivalent. @xref{Case Folding}. + +@item -I @var{regexp} +Ignore changes that just insert or delete lines that match @var{regexp}. +@xref{Specified Folding}. + +@item --ifdef=@var{name} +Make merged if-then-else output using @var{name}. @xref{If-then-else}. + +@item --ignore-all-space +Ignore white space when comparing lines. @xref{White Space}. + +@item --ignore-blank-lines +Ignore changes that just insert or delete blank lines. @xref{Blank +Lines}. + +@item --ignore-case +Ignore changes in case; consider upper- and lower-case to be the same. +@xref{Case Folding}. + +@item --ignore-matching-lines=@var{regexp} +Ignore changes that just insert or delete lines that match @var{regexp}. +@xref{Specified Folding}. + +@item --ignore-space-change +Ignore changes in amount of white space. +@xref{White Space}. + +@item --initial-tab +Output a tab rather than a space before the text of a line in normal or +context format. This causes the alignment of tabs in the line to look +normal. @xref{Tabs}. + +@item -l +Pass the output through @code{pr} to paginate it. @xref{Pagination}. + +@item -L @var{label} +Use @var{label} instead of the file name in the context format +(@pxref{Context Format}) and unified format (@pxref{Unified Format}) +headers. @xref{RCS}. + +@item --label=@var{label} +Use @var{label} instead of the file name in the context format +(@pxref{Context Format}) and unified format (@pxref{Unified Format}) +headers. + +@item --left-column +Print only the left column of two common lines in side by side format. +@xref{Side by Side Format}. + +@item --line-format=@var{format} +Use @var{format} to output all input lines in if-then-else format. +@xref{Line Formats}. + +@item --minimal +Change the algorithm to perhaps find a smaller set of changes. This +makes @code{diff} slower (sometimes much slower). @xref{diff +Performance}. + +@item -n +Output RCS-format diffs; like @samp{-f} except that each command +specifies the number of lines affected. @xref{RCS}. + +@item -N +@itemx --new-file +In directory comparison, if a file is found in only one directory, +treat it as present but empty in the other directory. @xref{Comparing +Directories}. + +@item --new-group-format=@var{format} +Use @var{format} to output a group of lines taken from just the second +file in if-then-else format. @xref{Line Group Formats}. + +@item --new-line-format=@var{format} +Use @var{format} to output a line taken from just the second file in +if-then-else format. @xref{Line Formats}. + +@item --old-group-format=@var{format} +Use @var{format} to output a group of lines taken from just the first +file in if-then-else format. @xref{Line Group Formats}. + +@item --old-line-format=@var{format} +Use @var{format} to output a line taken from just the first file in +if-then-else format. @xref{Line Formats}. + +@item -p +Show which C function each change is in. @xref{C Function Headings}. + +@item -P +When comparing directories, if a file appears only in the second +directory of the two, treat it as present but empty in the other. +@xref{Comparing Directories}. + +@item --paginate +Pass the output through @code{pr} to paginate it. @xref{Pagination}. + +@item -q +Report only whether the files differ, not the details of the +differences. @xref{Brief}. + +@item -r +When comparing directories, recursively compare any subdirectories +found. @xref{Comparing Directories}. + +@item --rcs +Output RCS-format diffs; like @samp{-f} except that each command +specifies the number of lines affected. @xref{RCS}. + +@item --recursive +When comparing directories, recursively compare any subdirectories +found. @xref{Comparing Directories}. + +@item --report-identical-files +Report when two files are the same. @xref{Comparing Directories}. + +@item -s +Report when two files are the same. @xref{Comparing Directories}. + +@item -S @var{file} +When comparing directories, start with the file @var{file}. This is +used for resuming an aborted comparison. @xref{Comparing Directories}. + +@item --sdiff-merge-assist +Print extra information to help @code{sdiff}. @code{sdiff} uses this +option when it runs @code{diff}. This option is not intended for users +to use directly. + +@item --show-c-function +Show which C function each change is in. @xref{C Function Headings}. + +@item --show-function-line=@var{regexp} +In context and unified format, for each hunk of differences, show some +of the last preceding line that matches @var{regexp}. @xref{Specified +Headings}. + +@item --side-by-side +Use the side by side output format. @xref{Side by Side Format}. + +@item --speed-large-files +Use heuristics to speed handling of large files that have numerous +scattered small changes. @xref{diff Performance}. + +@item --starting-file=@var{file} +When comparing directories, start with the file @var{file}. This is +used for resuming an aborted comparison. @xref{Comparing Directories}. + +@item --suppress-common-lines +Do not print common lines in side by side format. +@xref{Side by Side Format}. + +@item -t +Expand tabs to spaces in the output, to preserve the alignment of tabs +in the input files. @xref{Tabs}. + +@item -T +Output a tab rather than a space before the text of a line in normal or +context format. This causes the alignment of tabs in the line to look +normal. @xref{Tabs}. + +@item --text +Treat all files as text and compare them line-by-line, even if they +do not appear to be text. @xref{Binary}. + +@item -u +Use the unified output format. @xref{Unified Format}. + +@item --unchanged-group-format=@var{format} +Use @var{format} to output a group of common lines taken from both files +in if-then-else format. @xref{Line Group Formats}. + +@item --unchanged-line-format=@var{format} +Use @var{format} to output a line common to both files in if-then-else +format. @xref{Line Formats}. + +@item --unidirectional-new-file +When comparing directories, if a file appears only in the second +directory of the two, treat it as present but empty in the other. +@xref{Comparing Directories}. + +@item -U @var{lines} +@itemx --unified@r{[}=@var{lines}@r{]} +Use the unified output format, showing @var{lines} (an integer) lines of +context, or three if @var{lines} is not given. @xref{Unified Format}. +For proper operation, @code{patch} typically needs at least two lines of +context. + +@item -v +@itemx --version +Output the version number of @code{diff}. + +@item -w +Ignore white space when comparing lines. @xref{White Space}. + +@item -W @var{columns} +@itemx --width=@var{columns} +Use an output width of @var{columns} in side by side format. +@xref{Side by Side Format}. + +@item -x @var{pattern} +When comparing directories, ignore files and subdirectories whose basenames +match @var{pattern}. @xref{Comparing Directories}. + +@item -X @var{file} +When comparing directories, ignore files and subdirectories whose basenames +match any pattern contained in @var{file}. @xref{Comparing Directories}. + +@item -y +Use the side by side output format. @xref{Side by Side Format}. +@end table + +@node Invoking diff3, Invoking patch, Invoking diff, Top +@chapter Invoking @code{diff3} +@cindex invoking @code{diff3} +@cindex @code{diff3} invocation + +The @code{diff3} command compares three files and outputs descriptions +of their differences. Its arguments are as follows: + +@example +diff3 @var{options}@dots{} @var{mine} @var{older} @var{yours} +@end example + +The files to compare are @var{mine}, @var{older}, and @var{yours}. +At most one of these three file names may be @samp{-}, +which tells @code{diff3} to read the standard input for that file. + +An exit status of 0 means @code{diff3} was successful, 1 means some +conflicts were found, and 2 means trouble. + +@menu +* diff3 Options:: Summary of options to @code{diff3}. +@end menu + +@node diff3 Options, , , Invoking diff3 +@section Options to @code{diff3} +@cindex @code{diff3} options +@cindex options for @code{diff3} + +Below is a summary of all of the options that GNU @code{diff3} +accepts. Multiple single letter options (unless they take an argument) +can be combined into a single command line argument. + +@table @samp +@item -a +Treat all files as text and compare them line-by-line, even if they +do not appear to be text. @xref{Binary}. + +@item -A +Incorporate all changes from @var{older} to @var{yours} into @var{mine}, +surrounding all conflicts with bracket lines. +@xref{Marking Conflicts}. + +@item -e +Generate an @code{ed} script that incorporates all the changes from +@var{older} to @var{yours} into @var{mine}. @xref{Which Changes}. + +@item -E +Like @samp{-e}, except bracket lines from overlapping changes' first +and third files. +@xref{Marking Conflicts}. +With @samp{-e}, an overlapping change looks like this: + +@example +<<<<<<< @var{mine} +@r{lines from @var{mine}} +======= +@r{lines from @var{yours}} +>>>>>>> @var{yours} +@end example + +@item --ed +Generate an @code{ed} script that incorporates all the changes from +@var{older} to @var{yours} into @var{mine}. @xref{Which Changes}. + +@item --easy-only +Like @samp{-e}, except output only the nonoverlapping changes. +@xref{Which Changes}. + +@item -i +Generate @samp{w} and @samp{q} commands at the end of the @code{ed} +script for System V compatibility. This option must be combined with +one of the @samp{-AeExX3} options, and may not be combined with @samp{-m}. +@xref{Saving the Changed File}. + +@item --initial-tab +Output a tab rather than two spaces before the text of a line in normal format. +This causes the alignment of tabs in the line to look normal. @xref{Tabs}. + +@item -L @var{label} +@itemx --label=@var{label} +Use the label @var{label} for the brackets output by the @samp{-A}, +@samp{-E} and @samp{-X} options. This option may be given up to three +times, one for each input file. The default labels are the names of +the input files. Thus @samp{diff3 -L X -L Y -L Z -m A B C} acts like +@samp{diff3 -m A B C}, except that the output looks like it came from +files named @samp{X}, @samp{Y} and @samp{Z} rather than from files +named @samp{A}, @samp{B} and @samp{C}. @xref{Marking Conflicts}. + +@item -m +@itemx --merge +Apply the edit script to the first file and send the result to standard +output. Unlike piping the output from @code{diff3} to @code{ed}, this +works even for binary files and incomplete lines. @samp{-A} is assumed +if no edit script option is specified. @xref{Bypassing ed}. + +@item --overlap-only +Like @samp{-e}, except output only the overlapping changes. +@xref{Which Changes}. + +@item --show-all +Incorporate all unmerged changes from @var{older} to @var{yours} into +@var{mine}, surrounding all overlapping changes with bracket lines. +@xref{Marking Conflicts}. + +@item --show-overlap +Like @samp{-e}, except bracket lines from overlapping changes' first +and third files. +@xref{Marking Conflicts}. + +@item -T +Output a tab rather than two spaces before the text of a line in normal format. +This causes the alignment of tabs in the line to look normal. @xref{Tabs}. + +@item --text +Treat all files as text and compare them line-by-line, even if they +do not appear to be text. @xref{Binary}. + +@item -v +@itemx --version +Output the version number of @code{diff3}. + +@item -x +Like @samp{-e}, except output only the overlapping changes. +@xref{Which Changes}. + +@item -X +Like @samp{-E}, except output only the overlapping changes. +In other words, like @samp{-x}, except bracket changes as in @samp{-E}. +@xref{Marking Conflicts}. + +@item -3 +Like @samp{-e}, except output only the nonoverlapping changes. +@xref{Which Changes}. +@end table + +@node Invoking patch, Invoking sdiff, Invoking diff3, Top +@chapter Invoking @code{patch} +@cindex invoking @code{patch} +@cindex @code{patch} invocation + +Normally @code{patch} is invoked like this: + +@example +patch <@var{patchfile} +@end example + +The full format for invoking @code{patch} is: + +@example +patch @var{options}@dots{} @r{[}@var{origfile} @r{[}@var{patchfile}@r{]}@r{]} @r{[}+ @var{options}@dots{} @r{[}@var{origfile}@r{]}@r{]}@dots{} +@end example + +If you do not specify @var{patchfile}, or if @var{patchfile} is +@samp{-}, @code{patch} reads the patch (that is, the @code{diff} output) +from the standard input. + +You can specify one or more of the original files as @var{orig} arguments; +each one and options for interpreting it is separated from the others with a +@samp{+}. @xref{Multiple Patches}, for more information. + +If you do not specify an input file on the command line, @code{patch} +tries to figure out from the @dfn{leading text} (any text in the patch +that comes before the @code{diff} output) which file to edit. In the +header of a context or unified diff, @code{patch} looks in lines +beginning with @samp{***}, @samp{---}, or @samp{+++}; among those, it +chooses the shortest name of an existing file. Otherwise, if there is +an @samp{Index:} line in the leading text, @code{patch} tries to use the +file name from that line. If @code{patch} cannot figure out the name of +an existing file from the leading text, it prompts you for the name of +the file to patch. + +If the input file does not exist or is read-only, and a suitable RCS or +SCCS file exists, @code{patch} attempts to check out or get the file +before proceeding. + +By default, @code{patch} replaces the original input file with the +patched version, after renaming the original file into a backup file +(@pxref{Backups}, for a description of how @code{patch} names backup +files). You can also specify where to put the output with the @samp{-o +@var{output-file}} or @samp{--output=@var{output-file}} option. + +@menu +* patch Directories:: Changing directory and stripping directories. +* Backups:: Backup file names. +* Rejects:: Reject file names. +* patch Options:: Summary table of options to @code{patch}. +@end menu + +@node patch Directories, Backups, , Invoking patch +@section Applying Patches in Other Directories +@cindex directories and patch +@cindex patching directories + +The @samp{-d @var{directory}} or @samp{--directory=@var{directory}} +option to @code{patch} makes directory @var{directory} the current +directory for interpreting both file names in the patch file, and file +names given as arguments to other options (such as @samp{-B} and +@samp{-o}). For example, while in a news reading program, you can patch +a file in the @file{/usr/src/emacs} directory directly from the article +containing the patch like this: + +@example +| patch -d /usr/src/emacs +@end example + +Sometimes the file names given in a patch contain leading directories, +but you keep your files in a directory different from the one given in +the patch. In those cases, you can use the +@samp{-p@r{[}@var{number}@r{]}} or @samp{--strip@r{[}=@var{number}@r{]}} +option to set the file name strip count to @var{number}. The strip +count tells @code{patch} how many slashes, along with the directory +names between them, to strip from the front of file names. @samp{-p} +with no @var{number} given is equivalent to @samp{-p0}. By default, +@code{patch} strips off all leading directories, leaving just the base file +names, except that when a file name given in the patch is a relative +file name and all of its leading directories already exist, @code{patch} does +not strip off the leading directory. (A @dfn{relative} file name is one +that does not start with a slash.) + +@code{patch} looks for each file (after any slashes have been stripped) +in the current directory, or if you used the @samp{-d @var{directory}} +option, in that directory. + +For example, suppose the file name in the patch file is +@file{/gnu/src/emacs/etc/NEWS}. Using @samp{-p} or @samp{-p0} gives the +entire file name unmodified, @samp{-p1} gives +@file{gnu/src/emacs/etc/NEWS} (no leading slash), @samp{-p4} gives +@file{etc/NEWS}, and not specifying @samp{-p} at all gives @file{NEWS}. + +@node Backups, Rejects, patch Directories, Invoking patch +@section Backup File Names +@cindex backup file names + +Normally, @code{patch} renames an original input file into a backup file +by appending to its name the extension @samp{.orig}, or @samp{~} on +systems that do not support long file names. The @samp{-b +@var{backup-suffix}} or @samp{--suffix=@var{backup-suffix}} option uses +@var{backup-suffix} as the backup extension instead. + +Alternately, you can specify the extension for backup files with the +@code{SIMPLE_BACKUP_SUFFIX} environment variable, which the options +override. + +@code{patch} can also create numbered backup files the way GNU Emacs +does. With this method, instead of having a single backup of each file, +@code{patch} makes a new backup file name each time it patches a file. +For example, the backups of a file named @file{sink} would be called, +successively, @file{sink.~1~}, @file{sink.~2~}, @file{sink.~3~}, etc. + +The @samp{-V @var{backup-style}} or +@samp{--version-control=@var{backup-style}} option takes as an argument +a method for creating backup file names. You can alternately control +the type of backups that @code{patch} makes with the +@code{VERSION_CONTROL} environment variable, which the @samp{-V} option +overrides. The value of the @code{VERSION_CONTROL} environment variable +and the argument to the @samp{-V} option are like the GNU Emacs +@code{version-control} variable (@pxref{Backups, +emacs, The GNU Emacs Manual}, for more information on backup versions in +Emacs). They also recognize synonyms that are more descriptive. The +valid values are listed below; unique abbreviations are acceptable. + +@table @samp +@item t +@itemx numbered +Always make numbered backups. + +@item nil +@itemx existing +Make numbered backups of files that already have them, simple backups of +the others. This is the default. + +@item never +@itemx simple +Always make simple backups. +@end table + +Alternately, you can tell @code{patch} to prepend a prefix, such as a +directory name, to produce backup file names. The @samp{-B +@var{backup-prefix}} or @samp{--prefix=@var{backup-prefix}} option makes +backup files by prepending @var{backup-prefix} to them. If you use this +option, @code{patch} ignores any @samp{-b} option that you give. + +If the backup file already exists, @code{patch} creates a new backup +file name by changing the first lowercase letter in the last component +of the file name into uppercase. If there are no more lowercase letters +in the name, it removes the first character from the name. It repeats +this process until it comes up with a backup file name that does not +already exist. + +If you specify the output file with the @samp{-o} option, that file is +the one that is backed up, not the input file. + +@node Rejects, patch Options, Backups, Invoking patch +@section Reject File Names +@cindex reject file names + +The names for reject files (files containing patches that @code{patch} +could not find a place to apply) are normally the name of the output +file with @samp{.rej} appended (or @samp{#} on systems that do not +support long file names). + +Alternatively, you can tell @code{patch} to place all of the rejected +patches in a single file. The @samp{-r @var{reject-file}} or +@samp{--reject-file=@var{reject-file}} option uses @var{reject-file} as +the reject file name. + +@node patch Options, , Rejects, Invoking patch +@section Options to @code{patch} +@cindex @code{patch} options +@cindex options for @code{patch} + +Here is a summary of all of the options that @code{patch} accepts. +Older versions of @code{patch} do not accept long-named options or the +@samp{-t}, @samp{-E}, or @samp{-V} options. + +Multiple single-letter options that do not take an argument can be +combined into a single command line argument (with only one dash). +Brackets ([ and ]) indicate that an option takes an optional argument. + +@table @samp +@item -b @var{backup-suffix} +Use @var{backup-suffix} as the backup extension instead of +@samp{.orig} or @samp{~}. @xref{Backups}. + +@item -B @var{backup-prefix} +Use @var{backup-prefix} as a prefix to the backup file name. If this +option is specified, any @samp{-b} option is ignored. @xref{Backups}. + +@item --batch +Do not ask any questions. @xref{patch Messages}. + +@item -c +@itemx --context +Interpret the patch file as a context diff. @xref{patch Input}. + +@item -d @var{directory} +@itemx --directory=@var{directory} +Makes directory @var{directory} the current directory for interpreting +both file names in the patch file, and file names given as arguments to +other options. @xref{patch Directories}. + +@item -D @var{name} +Make merged if-then-else output using @var{format}. @xref{If-then-else}. + +@item --debug=@var{number} +Set internal debugging flags. Of interest only to @code{patch} +patchers. + +@item -e +@itemx --ed +Interpret the patch file as an @code{ed} script. @xref{patch Input}. + +@item -E +Remove output files that are empty after the patches have been applied. +@xref{Empty Files}. + +@item -f +Assume that the user knows exactly what he or she is doing, and do not +ask any questions. @xref{patch Messages}. + +@item -F @var{lines} +Set the maximum fuzz factor to @var{lines}. @xref{Inexact}. + +@item --force +Assume that the user knows exactly what he or she is doing, and do not +ask any questions. @xref{patch Messages}. + +@item --forward +Ignore patches that @code{patch} thinks are reversed or already applied. +See also @samp{-R}. @xref{Reversed Patches}. + +@item --fuzz=@var{lines} +Set the maximum fuzz factor to @var{lines}. @xref{Inexact}. + +@item --help +Print a summary of the options that @code{patch} recognizes, then exit. + +@item --ifdef=@var{name} +Make merged if-then-else output using @var{format}. @xref{If-then-else}. + +@item --ignore-white-space +@itemx -l +Let any sequence of white space in the patch file match any sequence of +white space in the input file. @xref{Changed White Space}. + +@item -n +@itemx --normal +Interpret the patch file as a normal diff. @xref{patch Input}. + +@item -N +Ignore patches that @code{patch} thinks are reversed or already applied. +See also @samp{-R}. @xref{Reversed Patches}. + +@item -o @var{output-file} +@itemx --output=@var{output-file} +Use @var{output-file} as the output file name. @xref{patch Options}. + +@item -p@r{[}@var{number}@r{]} +Set the file name strip count to @var{number}. @xref{patch Directories}. + +@item --prefix=@var{backup-prefix} +Use @var{backup-prefix} as a prefix to the backup file name. If this +option is specified, any @samp{-b} option is ignored. @xref{Backups}. + +@item --quiet +Work silently unless an error occurs. @xref{patch Messages}. + +@item -r @var{reject-file} +Use @var{reject-file} as the reject file name. @xref{Rejects}. + +@item -R +Assume that this patch was created with the old and new files swapped. +@xref{Reversed Patches}. + +@item --reject-file=@var{reject-file} +Use @var{reject-file} as the reject file name. @xref{Rejects}. + +@item --remove-empty-files +Remove output files that are empty after the patches have been applied. +@xref{Empty Files}. + +@item --reverse +Assume that this patch was created with the old and new files swapped. +@xref{Reversed Patches}. + +@item -s +Work silently unless an error occurs. @xref{patch Messages}. + +@item -S +Ignore this patch from the patch file, but continue looking for the next +patch in the file. @xref{Multiple Patches}. + +@item --silent +Work silently unless an error occurs. @xref{patch Messages}. + +@item --skip +Ignore this patch from the patch file, but continue looking for the next +patch in the file. @xref{Multiple Patches}. + +@item --strip@r{[}=@var{number}@r{]} +Set the file name strip count to @var{number}. @xref{patch Directories}. + +@item --suffix=@var{backup-suffix} +Use @var{backup-suffix} as the backup extension instead of +@samp{.orig} or @samp{~}. @xref{Backups}. + +@item -t +Do not ask any questions. @xref{patch Messages}. + +@item -u +@itemx --unified +Interpret the patch file as a unified diff. @xref{patch Input}. + +@item -v +Output the revision header and patch level of @code{patch}. + +@item -V @var{backup-style} +Select the kind of backups to make. @xref{Backups}. + +@item --version +Output the revision header and patch level of @code{patch}, then exit. + +@item --version=control=@var{backup-style} +Select the kind of backups to make. @xref{Backups}. + +@item -x @var{number} +Set internal debugging flags. Of interest only to @code{patch} +patchers. +@end table + +@node Invoking sdiff, Incomplete Lines, Invoking patch, Top +@chapter Invoking @code{sdiff} +@cindex invoking @code{sdiff} +@cindex @code{sdiff} invocation + +The @code{sdiff} command merges two files and interactively outputs the +results. Its arguments are as follows: + +@example +sdiff -o @var{outfile} @var{options}@dots{} @var{from-file} @var{to-file} +@end example + +This merges @var{from-file} with @var{to-file}, with output to @var{outfile}. +If @var{from-file} is a directory and @var{to-file} is not, @code{sdiff} +compares the file in @var{from-file} whose file name is that of @var{to-file}, +and vice versa. @var{from-file} and @var{to-file} may not both be +directories. + +@code{sdiff} options begin with @samp{-}, so normally @var{from-file} +and @var{to-file} may not begin with @samp{-}. However, @samp{--} as an +argument by itself treats the remaining arguments as file names even if +they begin with @samp{-}. You may not use @samp{-} as an input file. + +An exit status of 0 means no differences were found, 1 means some +differences were found, and 2 means trouble. + +@code{sdiff} without @samp{-o} (or @samp{--output}) produces a +side-by-side difference. This usage is obsolete; use @samp{diff +--side-by-side} instead. + +@menu +* sdiff Options:: Summary of options to @code{diff}. +@end menu + +@node sdiff Options, , , Invoking sdiff +@section Options to @code{sdiff} +@cindex @code{sdiff} options +@cindex options for @code{sdiff} + +Below is a summary of all of the options that GNU @code{sdiff} accepts. +Each option has two equivalent names, one of which is a single +letter preceded by @samp{-}, and the other of which is a long name +preceded by @samp{--}. Multiple single letter options (unless they take +an argument) can be combined into a single command line argument. Long +named options can be abbreviated to any unique prefix of their name. + +@table @samp +@item -a +Treat all files as text and compare them line-by-line, even if they +do not appear to be text. @xref{Binary}. + +@item -b +Ignore changes in amount of white space. @xref{White Space}. + +@item -B +Ignore changes that just insert or delete blank lines. @xref{Blank +Lines}. + +@item -d +Change the algorithm to perhaps find a smaller set of changes. This +makes @code{sdiff} slower (sometimes much slower). @xref{diff +Performance}. + +@item -H +Use heuristics to speed handling of large files that have numerous +scattered small changes. @xref{diff Performance}. + +@item --expand-tabs +Expand tabs to spaces in the output, to preserve the alignment of tabs +in the input files. @xref{Tabs}. + +@item -i +Ignore changes in case; consider upper- and lower-case to be the same. +@xref{Case Folding}. + +@item -I @var{regexp} +Ignore changes that just insert or delete lines that match @var{regexp}. +@xref{Specified Folding}. + +@item --ignore-all-space +Ignore white space when comparing lines. @xref{White Space}. + +@item --ignore-blank-lines +Ignore changes that just insert or delete blank lines. @xref{Blank +Lines}. + +@item --ignore-case +Ignore changes in case; consider upper- and lower-case to be the same. +@xref{Case Folding}. + +@item --ignore-matching-lines=@var{regexp} +Ignore changes that just insert or delete lines that match @var{regexp}. +@xref{Specified Folding}. + +@item --ignore-space-change +Ignore changes in amount of white space. +@xref{White Space}. + +@item -l +@itemx --left-column +Print only the left column of two common lines. +@xref{Side by Side Format}. + +@item --minimal +Change the algorithm to perhaps find a smaller set of changes. This +makes @code{sdiff} slower (sometimes much slower). @xref{diff +Performance}. + +@item -o @var{file} +@itemx --output=@var{file} +Put merged output into @var{file}. This option is required for merging. + +@item -s +@itemx --suppress-common-lines +Do not print common lines. @xref{Side by Side Format}. + +@item --speed-large-files +Use heuristics to speed handling of large files that have numerous +scattered small changes. @xref{diff Performance}. + +@item -t +Expand tabs to spaces in the output, to preserve the alignment of tabs +in the input files. @xref{Tabs}. + +@item --text +Treat all files as text and compare them line-by-line, even if they +do not appear to be text. @xref{Binary}. + +@item -v +@itemx --version +Output the version number of @code{sdiff}. + +@item -w @var{columns} +@itemx --width=@var{columns} +Use an output width of @var{columns}. @xref{Side by Side Format}. +Note that for historical reasons, this option is @samp{-W} in @code{diff}, +@samp{-w} in @code{sdiff}. + +@item -W +Ignore horizontal white space when comparing lines. @xref{White Space}. +Note that for historical reasons, this option is @samp{-w} in @code{diff}, +@samp{-W} in @code{sdiff}. +@end table + +@node Incomplete Lines, Projects, Invoking sdiff, Top +@chapter Incomplete Lines +@cindex incomplete lines +@cindex full lines +@cindex newline treatment by @code{diff} + +When an input file ends in a non-newline character, its last line is +called an @dfn{incomplete line} because its last character is not a +newline. All other lines are called @dfn{full lines} and end in a +newline character. Incomplete lines do not match full lines unless +differences in white space are ignored (@pxref{White Space}). + +An incomplete line is normally distinguished on output from a full line +by a following line that starts with @samp{\}. However, the RCS format +(@pxref{RCS}) outputs the incomplete line as-is, without any trailing +newline or following line. The side by side format normally represents +incomplete lines as-is, but in some cases uses a @samp{\} or @samp{/} +gutter marker; @xref{Side by Side}. The if-then-else line format +preserves a line's incompleteness with @samp{%L}, and discards the +newline with @samp{%l}; @xref{Line Formats}. Finally, with the +@code{ed} and forward @code{ed} output formats (@pxref{Output Formats}) +@code{diff} cannot represent an incomplete line, so it pretends there +was a newline and reports an error. + +For example, suppose @file{F} and @file{G} are one-byte files that +contain just @samp{f} and @samp{g}, respectively. Then @samp{diff F G} +outputs + +@example +1c1 +< f +\ No newline at end of file +--- +> g +\ No newline at end of file +@end example + +@noindent +(The exact message may differ in non-English locales.) +@samp{diff -n F G} outputs the following without a trailing newline: + +@example +d1 1 +a1 1 +g +@end example + +@samp{diff -e F G} reports two errors and outputs the following: + +@example +1c +g +. +@end example + +@node Projects, Concept Index, Incomplete Lines, Top +@chapter Future Projects + +Here are some ideas for improving GNU @code{diff} and @code{patch}. The +GNU project has identified some improvements as potential programming +projects for volunteers. You can also help by reporting any bugs that +you find. + +If you are a programmer and would like to contribute something to the +GNU project, please consider volunteering for one of these projects. If +you are seriously contemplating work, please write to +@samp{gnu@@prep.ai.mit.edu} to coordinate with other volunteers. + +@menu +* Shortcomings:: Suggested projects for improvements. +* Bugs:: Reporting bugs. +@end menu + +@node Shortcomings, Bugs, , Projects +@section Suggested Projects for Improving GNU @code{diff} and @code{patch} +@cindex projects for directories + +One should be able to use GNU @code{diff} to generate a patch from any +pair of directory trees, and given the patch and a copy of one such +tree, use @code{patch} to generate a faithful copy of the other. +Unfortunately, some changes to directory trees cannot be expressed using +current patch formats; also, @code{patch} does not handle some of the +existing formats. These shortcomings motivate the following suggested +projects. + +@menu +* Changing Structure:: Handling changes to the directory structure. +* Special Files:: Handling symbolic links, device special files, etc. +* Unusual File Names:: Handling file names that contain unusual characters. +* Arbitrary Limits:: Patching non-text files. +* Large Files:: Handling files that do not fit in memory. +* Ignoring Changes:: Ignoring certain changes while showing others. +@end menu + +@node Changing Structure, Special Files, , Shortcomings +@subsection Handling Changes to the Directory Structure +@cindex directory structure changes + +@code{diff} and @code{patch} do not handle some changes to directory +structure. For example, suppose one directory tree contains a directory +named @samp{D} with some subsidiary files, and another contains a file +with the same name @samp{D}. @samp{diff -r} does not output enough +information for @code{patch} to transform the the directory subtree into +the file. + +There should be a way to specify that a file has been deleted without +having to include its entire contents in the patch file. There should +also be a way to tell @code{patch} that a file was renamed, even if +there is no way for @code{diff} to generate such information. + +These problems can be fixed by extending the @code{diff} output format +to represent changes in directory structure, and extending @code{patch} +to understand these extensions. + +@node Special Files, Unusual File Names, Changing Structure, Shortcomings +@subsection Files that are Neither Directories Nor Regular Files +@cindex special files + +Some files are neither directories nor regular files: they are unusual +files like symbolic links, device special files, named pipes, and +sockets. Currently, @code{diff} treats symbolic links like regular files; +it treats other special files like regular files if they are specified +at the top level, but simply reports their presence when comparing +directories. This means that @code{patch} cannot represent changes +to such files. For example, if you change which file a symbolic link +points to, @code{diff} outputs the difference between the two files, +instead of the change to the symbolic link. + +@c This might not be a good idea; is it wise for root to install devices +@c this way? +@code{diff} should optionally report changes to special files specially, +and @code{patch} should be extended to understand these extensions. + +@node Unusual File Names, Arbitrary Limits, Special Files, Shortcomings +@subsection File Names that Contain Unusual Characters +@cindex file names with unusual characters + +When a file name contains an unusual character like a newline or +white space, @samp{diff -r} generates a patch that @code{patch} cannot +parse. The problem is with format of @code{diff} output, not just with +@code{patch}, because with odd enough file names one can cause +@code{diff} to generate a patch that is syntactically correct but +patches the wrong files. The format of @code{diff} output should be +extended to handle all possible file names. + +@node Arbitrary Limits, Large Files, Unusual File Names, Shortcomings +@subsection Arbitrary Limits +@cindex binary file patching + +GNU @code{diff} can analyze files with arbitrarily long lines and files +that end in incomplete lines. However, @code{patch} cannot patch such +files. The @code{patch} internal limits on line lengths should be +removed, and @code{patch} should be extended to parse @code{diff} +reports of incomplete lines. + +@node Large Files, Ignoring Changes, Arbitrary Limits, Shortcomings +@subsection Handling Files that Do Not Fit in Memory +@cindex large files + +@code{diff} operates by reading both files into memory. This method +fails if the files are too large, and @code{diff} should have a fallback. + +One way to do this is to scan the files sequentially to compute hash +codes of the lines and put the lines in equivalence classes based only +on hash code. Then compare the files normally. This does produce some +false matches. + +Then scan the two files sequentially again, checking each match to see +whether it is real. When a match is not real, mark both the +``matching'' lines as changed. Then build an edit script as usual. + +The output routines would have to be changed to scan the files +sequentially looking for the text to print. + +@node Ignoring Changes,, Large Files, Shortcomings +@subsection Ignoring Certain Changes + +It would be nice to have a feature for specifying two strings, one in +@var{from-file} and one in @var{to-file}, which should be considered to +match. Thus, if the two strings are @samp{foo} and @samp{bar}, then if +two lines differ only in that @samp{foo} in file 1 corresponds to +@samp{bar} in file 2, the lines are treated as identical. + +It is not clear how general this feature can or should be, or +what syntax should be used for it. + +@node Bugs, , Shortcomings, Projects +@section Reporting Bugs +@cindex bug reports +@cindex reporting bugs + +If you think you have found a bug in GNU @code{cmp}, @code{diff}, +@code{diff3}, @code{sdiff}, or @code{patch}, please report it by +electronic mail to @samp{bug-gnu-utils@@prep.ai.mit.edu}. Send as +precise a description of the problem as you can, including sample input +files that produce the bug, if applicable. + +Because Larry Wall has not released a new version of @code{patch} since +mid 1988 and the GNU version of @code{patch} has been changed since +then, please send bug reports for @code{patch} by electronic mail to +both @samp{bug-gnu-utils@@prep.ai.mit.edu} and +@samp{lwall@@netlabs.com}. + +@node Concept Index, , Projects, Top +@unnumbered Concept Index + +@printindex cp + +@shortcontents +@contents +@bye diff --git a/sys/src/ape/cmd/diff/diff3.c b/sys/src/ape/cmd/diff/diff3.c new file mode 100755 index 000000000..c6e641399 --- /dev/null +++ b/sys/src/ape/cmd/diff/diff3.c @@ -0,0 +1,1778 @@ +/* Three way file comparison program (diff3) for Project GNU. + Copyright (C) 1988, 1989, 1992, 1993, 1994 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* Written by Randy Smith */ + +#include "system.h" +#include <stdio.h> +#include <signal.h> +#include "getopt.h" + +extern char const version_string[]; + +/* + * Internal data structures and macros for the diff3 program; includes + * data structures for both diff3 diffs and normal diffs. + */ + +/* Different files within a three way diff. */ +#define FILE0 0 +#define FILE1 1 +#define FILE2 2 + +/* + * A three way diff is built from two two-way diffs; the file which + * the two two-way diffs share is: + */ +#define FILEC FILE2 + +/* + * Different files within a two way diff. + * FC is the common file, FO the other file. + */ +#define FO 0 +#define FC 1 + +/* The ranges are indexed by */ +#define START 0 +#define END 1 + +enum diff_type { + ERROR, /* Should not be used */ + ADD, /* Two way diff add */ + CHANGE, /* Two way diff change */ + DELETE, /* Two way diff delete */ + DIFF_ALL, /* All three are different */ + DIFF_1ST, /* Only the first is different */ + DIFF_2ND, /* Only the second */ + DIFF_3RD /* Only the third */ +}; + +/* Two way diff */ +struct diff_block { + int ranges[2][2]; /* Ranges are inclusive */ + char **lines[2]; /* The actual lines (may contain nulls) */ + size_t *lengths[2]; /* Line lengths (including newlines, if any) */ + struct diff_block *next; +}; + +/* Three way diff */ + +struct diff3_block { + enum diff_type correspond; /* Type of diff */ + int ranges[3][2]; /* Ranges are inclusive */ + char **lines[3]; /* The actual lines (may contain nulls) */ + size_t *lengths[3]; /* Line lengths (including newlines, if any) */ + struct diff3_block *next; +}; + +/* + * Access the ranges on a diff block. + */ +#define D_LOWLINE(diff, filenum) \ + ((diff)->ranges[filenum][START]) +#define D_HIGHLINE(diff, filenum) \ + ((diff)->ranges[filenum][END]) +#define D_NUMLINES(diff, filenum) \ + (D_HIGHLINE (diff, filenum) - D_LOWLINE (diff, filenum) + 1) + +/* + * Access the line numbers in a file in a diff by relative line + * numbers (i.e. line number within the diff itself). Note that these + * are lvalues and can be used for assignment. + */ +#define D_RELNUM(diff, filenum, linenum) \ + ((diff)->lines[filenum][linenum]) +#define D_RELLEN(diff, filenum, linenum) \ + ((diff)->lengths[filenum][linenum]) + +/* + * And get at them directly, when that should be necessary. + */ +#define D_LINEARRAY(diff, filenum) \ + ((diff)->lines[filenum]) +#define D_LENARRAY(diff, filenum) \ + ((diff)->lengths[filenum]) + +/* + * Next block. + */ +#define D_NEXT(diff) ((diff)->next) + +/* + * Access the type of a diff3 block. + */ +#define D3_TYPE(diff) ((diff)->correspond) + +/* + * Line mappings based on diffs. The first maps off the top of the + * diff, the second off of the bottom. + */ +#define D_HIGH_MAPLINE(diff, fromfile, tofile, lineno) \ + ((lineno) \ + - D_HIGHLINE ((diff), (fromfile)) \ + + D_HIGHLINE ((diff), (tofile))) + +#define D_LOW_MAPLINE(diff, fromfile, tofile, lineno) \ + ((lineno) \ + - D_LOWLINE ((diff), (fromfile)) \ + + D_LOWLINE ((diff), (tofile))) + +/* + * General memory allocation function. + */ +#define ALLOCATE(number, type) \ + (type *) xmalloc ((number) * sizeof (type)) + +/* Options variables for flags set on command line. */ + +/* If nonzero, treat all files as text files, never as binary. */ +static int always_text; + +/* If nonzero, write out an ed script instead of the standard diff3 format. */ +static int edscript; + +/* If nonzero, in the case of overlapping diffs (type DIFF_ALL), + preserve the lines which would normally be deleted from + file 1 with a special flagging mechanism. */ +static int flagging; + +/* Number of lines to keep in identical prefix and suffix. */ +static int horizon_lines = 10; + +/* Use a tab to align output lines (-T). */ +static int tab_align_flag; + +/* If nonzero, do not output information for overlapping diffs. */ +static int simple_only; + +/* If nonzero, do not output information for non-overlapping diffs. */ +static int overlap_only; + +/* If nonzero, show information for DIFF_2ND diffs. */ +static int show_2nd; + +/* If nonzero, include `:wq' at the end of the script + to write out the file being edited. */ +static int finalwrite; + +/* If nonzero, output a merged file. */ +static int merge; + +static char *program_name; + +static VOID *xmalloc PARAMS((size_t)); +static VOID *xrealloc PARAMS((VOID *, size_t)); + +static char *read_diff PARAMS((char const *, char const *, char **)); +static char *scan_diff_line PARAMS((char *, char **, size_t *, char *, int)); +static enum diff_type process_diff_control PARAMS((char **, struct diff_block *)); +static int compare_line_list PARAMS((char * const[], size_t const[], char * const[], size_t const[], int)); +static int copy_stringlist PARAMS((char * const[], size_t const[], char *[], size_t[], int)); +static int dotlines PARAMS((FILE *, struct diff3_block *, int)); +static int output_diff3_edscript PARAMS((FILE *, struct diff3_block *, int const[3], int const[3], char const *, char const *, char const *)); +static int output_diff3_merge PARAMS((FILE *, FILE *, struct diff3_block *, int const[3], int const[3], char const *, char const *, char const *)); +static size_t myread PARAMS((int, char *, size_t)); +static struct diff3_block *create_diff3_block PARAMS((int, int, int, int, int, int)); +static struct diff3_block *make_3way_diff PARAMS((struct diff_block *, struct diff_block *)); +static struct diff3_block *reverse_diff3_blocklist PARAMS((struct diff3_block *)); +static struct diff3_block *using_to_diff3_block PARAMS((struct diff_block *[2], struct diff_block *[2], int, int, struct diff3_block const *)); +static struct diff_block *process_diff PARAMS((char const *, char const *, struct diff_block **)); +static void check_stdout PARAMS((void)); +static void fatal PARAMS((char const *)); +static void output_diff3 PARAMS((FILE *, struct diff3_block *, int const[3], int const[3])); +static void perror_with_exit PARAMS((char const *)); +static void try_help PARAMS((char const *)); +static void undotlines PARAMS((FILE *, int, int, int)); +static void usage PARAMS((void)); + +static char const diff_program[] = DIFF_PROGRAM; + +static struct option const longopts[] = +{ + {"text", 0, 0, 'a'}, + {"show-all", 0, 0, 'A'}, + {"ed", 0, 0, 'e'}, + {"show-overlap", 0, 0, 'E'}, + {"label", 1, 0, 'L'}, + {"merge", 0, 0, 'm'}, + {"initial-tab", 0, 0, 'T'}, + {"overlap-only", 0, 0, 'x'}, + {"easy-only", 0, 0, '3'}, + {"version", 0, 0, 'v'}, + {"help", 0, 0, 129}, + {0, 0, 0, 0} +}; + +/* + * Main program. Calls diff twice on two pairs of input files, + * combines the two diffs, and outputs them. + */ +int +main (argc, argv) + int argc; + char **argv; +{ + int c, i; + int mapping[3]; + int rev_mapping[3]; + int incompat = 0; + int conflicts_found; + struct diff_block *thread0, *thread1, *last_block; + struct diff3_block *diff3; + int tag_count = 0; + char *tag_strings[3]; + char *commonname; + char **file; + struct stat statb; + + initialize_main (&argc, &argv); + program_name = argv[0]; + + while ((c = getopt_long (argc, argv, "aeimvx3AEL:TX", longopts, 0)) != EOF) + { + switch (c) + { + case 'a': + always_text = 1; + break; + case 'A': + show_2nd = 1; + flagging = 1; + incompat++; + break; + case 'x': + overlap_only = 1; + incompat++; + break; + case '3': + simple_only = 1; + incompat++; + break; + case 'i': + finalwrite = 1; + break; + case 'm': + merge = 1; + break; + case 'X': + overlap_only = 1; + /* Falls through */ + case 'E': + flagging = 1; + /* Falls through */ + case 'e': + incompat++; + break; + case 'T': + tab_align_flag = 1; + break; + case 'v': + printf ("diff3 - GNU diffutils version %s\n", version_string); + exit (0); + case 129: + usage (); + check_stdout (); + exit (0); + case 'L': + /* Handle up to three -L options. */ + if (tag_count < 3) + { + tag_strings[tag_count++] = optarg; + break; + } + try_help ("Too many labels were given. The limit is 3."); + default: + try_help (0); + } + } + + edscript = incompat & ~merge; /* -AeExX3 without -m implies ed script. */ + show_2nd |= ~incompat & merge; /* -m without -AeExX3 implies -A. */ + flagging |= ~incompat & merge; + + if (incompat > 1 /* Ensure at most one of -AeExX3. */ + || finalwrite & merge /* -i -m would rewrite input file. */ + || (tag_count && ! flagging)) /* -L requires one of -AEX. */ + try_help ("incompatible options"); + + if (argc - optind != 3) + try_help (argc - optind < 3 ? "missing operand" : "extra operand"); + + file = &argv[optind]; + + for (i = tag_count; i < 3; i++) + tag_strings[i] = file[i]; + + /* Always compare file1 to file2, even if file2 is "-". + This is needed for -mAeExX3. Using the file0 as + the common file would produce wrong results, because if the + file0-file1 diffs didn't line up with the file0-file2 diffs + (which is entirely possible since we don't use diff's -n option), + diff3 might report phantom changes from file1 to file2. */ + + if (strcmp (file[2], "-") == 0) + { + /* Sigh. We've got standard input as the last arg. We can't + call diff twice on stdin. Use the middle arg as the common + file instead. */ + if (strcmp (file[0], "-") == 0 || strcmp (file[1], "-") == 0) + fatal ("`-' specified for more than one input file"); + mapping[0] = 0; + mapping[1] = 2; + mapping[2] = 1; + } + else + { + /* Normal, what you'd expect */ + mapping[0] = 0; + mapping[1] = 1; + mapping[2] = 2; + } + + for (i = 0; i < 3; i++) + rev_mapping[mapping[i]] = i; + + for (i = 0; i < 3; i++) + if (strcmp (file[i], "-") != 0) + { + if (stat (file[i], &statb) < 0) + perror_with_exit (file[i]); + else if (S_ISDIR(statb.st_mode)) + { + fprintf (stderr, "%s: %s: Is a directory\n", + program_name, file[i]); + exit (2); + } + } + +#if !defined(SIGCHLD) && defined(SIGCLD) +#define SIGCHLD SIGCLD +#endif +#ifdef SIGCHLD + /* System V fork+wait does not work if SIGCHLD is ignored. */ + signal (SIGCHLD, SIG_DFL); +#endif + + commonname = file[rev_mapping[FILEC]]; + thread1 = process_diff (file[rev_mapping[FILE1]], commonname, &last_block); + if (thread1) + for (i = 0; i < 2; i++) + { + horizon_lines = max (horizon_lines, D_NUMLINES (thread1, i)); + horizon_lines = max (horizon_lines, D_NUMLINES (last_block, i)); + } + thread0 = process_diff (file[rev_mapping[FILE0]], commonname, &last_block); + diff3 = make_3way_diff (thread0, thread1); + if (edscript) + conflicts_found + = output_diff3_edscript (stdout, diff3, mapping, rev_mapping, + tag_strings[0], tag_strings[1], tag_strings[2]); + else if (merge) + { + if (! freopen (file[rev_mapping[FILE0]], "r", stdin)) + perror_with_exit (file[rev_mapping[FILE0]]); + conflicts_found + = output_diff3_merge (stdin, stdout, diff3, mapping, rev_mapping, + tag_strings[0], tag_strings[1], tag_strings[2]); + if (ferror (stdin)) + fatal ("read error"); + } + else + { + output_diff3 (stdout, diff3, mapping, rev_mapping); + conflicts_found = 0; + } + + check_stdout (); + exit (conflicts_found); + return conflicts_found; +} + +static void +try_help (reason) + char const *reason; +{ + if (reason) + fprintf (stderr, "%s: %s\n", program_name, reason); + fprintf (stderr, "%s: Try `%s --help' for more information.\n", + program_name, program_name); + exit (2); +} + +static void +check_stdout () +{ + if (ferror (stdout) || fclose (stdout) != 0) + fatal ("write error"); +} + +/* + * Explain, patiently and kindly, how to use this program. + */ +static void +usage () +{ + printf ("Usage: %s [OPTION]... MYFILE OLDFILE YOURFILE\n\n", program_name); + + printf ("%s", "\ + -e --ed Output unmerged changes from OLDFILE to YOURFILE into MYFILE.\n\ + -E --show-overlap Output unmerged changes, bracketing conflicts.\n\ + -A --show-all Output all changes, bracketing conflicts.\n\ + -x --overlap-only Output overlapping changes.\n\ + -X Output overlapping changes, bracketing them.\n\ + -3 --easy-only Output unmerged nonoverlapping changes.\n\n"); + printf ("%s", "\ + -m --merge Output merged file instead of ed script (default -A).\n\ + -L LABEL --label=LABEL Use LABEL instead of file name.\n\ + -i Append `w' and `q' commands to ed scripts.\n\ + -a --text Treat all files as text.\n\ + -T --initial-tab Make tabs line up by prepending a tab.\n\n"); + printf ("%s", "\ + -v --version Output version info.\n\ + --help Output this help.\n\n"); + printf ("If a FILE is `-', read standard input.\n"); +} + +/* + * Routines that combine the two diffs together into one. The + * algorithm used follows: + * + * File2 is shared in common between the two diffs. + * Diff02 is the diff between 0 and 2. + * Diff12 is the diff between 1 and 2. + * + * 1) Find the range for the first block in File2. + * a) Take the lowest of the two ranges (in File2) in the two + * current blocks (one from each diff) as being the low + * water mark. Assign the upper end of this block as + * being the high water mark and move the current block up + * one. Mark the block just moved over as to be used. + * b) Check the next block in the diff that the high water + * mark is *not* from. + * + * *If* the high water mark is above + * the low end of the range in that block, + * + * mark that block as to be used and move the current + * block up. Set the high water mark to the max of + * the high end of this block and the current. Repeat b. + * + * 2) Find the corresponding ranges in File0 (from the blocks + * in diff02; line per line outside of diffs) and in File1. + * Create a diff3_block, reserving space as indicated by the ranges. + * + * 3) Copy all of the pointers for file2 in. At least for now, + * do memcmp's between corresponding strings in the two diffs. + * + * 4) Copy all of the pointers for file0 and 1 in. Get what you + * need from file2 (when there isn't a diff block, it's + * identical to file2 within the range between diff blocks). + * + * 5) If the diff blocks you used came from only one of the two + * strings of diffs, then that file (i.e. the one other than + * the common file in that diff) is the odd person out. If you used + * diff blocks from both sets, check to see if files 0 and 1 match: + * + * Same number of lines? If so, do a set of memcmp's (if a + * memcmp matches; copy the pointer over; it'll be easier later + * if you have to do any compares). If they match, 0 & 1 are + * the same. If not, all three different. + * + * Then you do it again, until you run out of blocks. + * + */ + +/* + * This routine makes a three way diff (chain of diff3_block's) from two + * two way diffs (chains of diff_block's). It is assumed that each of + * the two diffs passed are onto the same file (i.e. that each of the + * diffs were made "to" the same file). The three way diff pointer + * returned will have numbering FILE0--the other file in diff02, + * FILE1--the other file in diff12, and FILEC--the common file. + */ +static struct diff3_block * +make_3way_diff (thread0, thread1) + struct diff_block *thread0, *thread1; +{ +/* + * This routine works on the two diffs passed to it as threads. + * Thread number 0 is diff02, thread number 1 is diff12. The USING + * array is set to the base of the list of blocks to be used to + * construct each block of the three way diff; if no blocks from a + * particular thread are to be used, that element of the using array + * is set to 0. The elements LAST_USING array are set to the last + * elements on each of the using lists. + * + * The HIGH_WATER_MARK is set to the highest line number in the common file + * described in any of the diffs in either of the USING lists. The + * HIGH_WATER_THREAD names the thread. Similarly the BASE_WATER_MARK + * and BASE_WATER_THREAD describe the lowest line number in the common file + * described in any of the diffs in either of the USING lists. The + * HIGH_WATER_DIFF is the diff from which the HIGH_WATER_MARK was + * taken. + * + * The HIGH_WATER_DIFF should always be equal to LAST_USING + * [HIGH_WATER_THREAD]. The OTHER_DIFF is the next diff to check for + * higher water, and should always be equal to + * CURRENT[HIGH_WATER_THREAD ^ 0x1]. The OTHER_THREAD is the thread + * in which the OTHER_DIFF is, and hence should always be equal to + * HIGH_WATER_THREAD ^ 0x1. + * + * The variable LAST_DIFF is kept set to the last diff block produced + * by this routine, for line correspondence purposes between that diff + * and the one currently being worked on. It is initialized to + * ZERO_DIFF before any blocks have been created. + */ + + struct diff_block + *using[2], + *last_using[2], + *current[2]; + + int + high_water_mark; + + int + high_water_thread, + base_water_thread, + other_thread; + + struct diff_block + *high_water_diff, + *other_diff; + + struct diff3_block + *result, + *tmpblock, + **result_end; + + struct diff3_block const *last_diff3; + + static struct diff3_block const zero_diff3; + + /* Initialization */ + result = 0; + result_end = &result; + current[0] = thread0; current[1] = thread1; + last_diff3 = &zero_diff3; + + /* Sniff up the threads until we reach the end */ + + while (current[0] || current[1]) + { + using[0] = using[1] = last_using[0] = last_using[1] = 0; + + /* Setup low and high water threads, diffs, and marks. */ + if (!current[0]) + base_water_thread = 1; + else if (!current[1]) + base_water_thread = 0; + else + base_water_thread = + (D_LOWLINE (current[0], FC) > D_LOWLINE (current[1], FC)); + + high_water_thread = base_water_thread; + + high_water_diff = current[high_water_thread]; + +#if 0 + /* low and high waters start off same diff */ + base_water_mark = D_LOWLINE (high_water_diff, FC); +#endif + + high_water_mark = D_HIGHLINE (high_water_diff, FC); + + /* Make the diff you just got info from into the using class */ + using[high_water_thread] + = last_using[high_water_thread] + = high_water_diff; + current[high_water_thread] = high_water_diff->next; + last_using[high_water_thread]->next = 0; + + /* And mark the other diff */ + other_thread = high_water_thread ^ 0x1; + other_diff = current[other_thread]; + + /* Shuffle up the ladder, checking the other diff to see if it + needs to be incorporated. */ + while (other_diff + && D_LOWLINE (other_diff, FC) <= high_water_mark + 1) + { + + /* Incorporate this diff into the using list. Note that + this doesn't take it off the current list */ + if (using[other_thread]) + last_using[other_thread]->next = other_diff; + else + using[other_thread] = other_diff; + last_using[other_thread] = other_diff; + + /* Take it off the current list. Note that this following + code assumes that other_diff enters it equal to + current[high_water_thread ^ 0x1] */ + current[other_thread] = current[other_thread]->next; + other_diff->next = 0; + + /* Set the high_water stuff + If this comparison is equal, then this is the last pass + through this loop; since diff blocks within a given + thread cannot overlap, the high_water_mark will be + *below* the range_start of either of the next diffs. */ + + if (high_water_mark < D_HIGHLINE (other_diff, FC)) + { + high_water_thread ^= 1; + high_water_diff = other_diff; + high_water_mark = D_HIGHLINE (other_diff, FC); + } + + /* Set the other diff */ + other_thread = high_water_thread ^ 0x1; + other_diff = current[other_thread]; + } + + /* The using lists contain a list of all of the blocks to be + included in this diff3_block. Create it. */ + + tmpblock = using_to_diff3_block (using, last_using, + base_water_thread, high_water_thread, + last_diff3); + + if (!tmpblock) + fatal ("internal error: screwup in format of diff blocks"); + + /* Put it on the list. */ + *result_end = tmpblock; + result_end = &tmpblock->next; + + /* Set up corresponding lines correctly. */ + last_diff3 = tmpblock; + } + return result; +} + +/* + * using_to_diff3_block: + * This routine takes two lists of blocks (from two separate diff + * threads) and puts them together into one diff3 block. + * It then returns a pointer to this diff3 block or 0 for failure. + * + * All arguments besides using are for the convenience of the routine; + * they could be derived from the using array. + * LAST_USING is a pair of pointers to the last blocks in the using + * structure. + * LOW_THREAD and HIGH_THREAD tell which threads contain the lowest + * and highest line numbers for File0. + * last_diff3 contains the last diff produced in the calling routine. + * This is used for lines mappings which would still be identical to + * the state that diff ended in. + * + * A distinction should be made in this routine between the two diffs + * that are part of a normal two diff block, and the three diffs that + * are part of a diff3_block. + */ +static struct diff3_block * +using_to_diff3_block (using, last_using, low_thread, high_thread, last_diff3) + struct diff_block + *using[2], + *last_using[2]; + int low_thread, high_thread; + struct diff3_block const *last_diff3; +{ + int low[2], high[2]; + struct diff3_block *result; + struct diff_block *ptr; + int d, i; + + /* Find the range in the common file. */ + int lowc = D_LOWLINE (using[low_thread], FC); + int highc = D_HIGHLINE (last_using[high_thread], FC); + + /* Find the ranges in the other files. + If using[d] is null, that means that the file to which that diff + refers is equivalent to the common file over this range. */ + + for (d = 0; d < 2; d++) + if (using[d]) + { + low[d] = D_LOW_MAPLINE (using[d], FC, FO, lowc); + high[d] = D_HIGH_MAPLINE (last_using[d], FC, FO, highc); + } + else + { + low[d] = D_HIGH_MAPLINE (last_diff3, FILEC, FILE0 + d, lowc); + high[d] = D_HIGH_MAPLINE (last_diff3, FILEC, FILE0 + d, highc); + } + + /* Create a block with the appropriate sizes */ + result = create_diff3_block (low[0], high[0], low[1], high[1], lowc, highc); + + /* Copy information for the common file. + Return with a zero if any of the compares failed. */ + + for (d = 0; d < 2; d++) + for (ptr = using[d]; ptr; ptr = D_NEXT (ptr)) + { + int result_offset = D_LOWLINE (ptr, FC) - lowc; + + if (!copy_stringlist (D_LINEARRAY (ptr, FC), + D_LENARRAY (ptr, FC), + D_LINEARRAY (result, FILEC) + result_offset, + D_LENARRAY (result, FILEC) + result_offset, + D_NUMLINES (ptr, FC))) + return 0; + } + + /* Copy information for file d. First deal with anything that might be + before the first diff. */ + + for (d = 0; d < 2; d++) + { + struct diff_block *u = using[d]; + int lo = low[d], hi = high[d]; + + for (i = 0; + i + lo < (u ? D_LOWLINE (u, FO) : hi + 1); + i++) + { + D_RELNUM (result, FILE0 + d, i) = D_RELNUM (result, FILEC, i); + D_RELLEN (result, FILE0 + d, i) = D_RELLEN (result, FILEC, i); + } + + for (ptr = u; ptr; ptr = D_NEXT (ptr)) + { + int result_offset = D_LOWLINE (ptr, FO) - lo; + int linec; + + if (!copy_stringlist (D_LINEARRAY (ptr, FO), + D_LENARRAY (ptr, FO), + D_LINEARRAY (result, FILE0 + d) + result_offset, + D_LENARRAY (result, FILE0 + d) + result_offset, + D_NUMLINES (ptr, FO))) + return 0; + + /* Catch the lines between here and the next diff */ + linec = D_HIGHLINE (ptr, FC) + 1 - lowc; + for (i = D_HIGHLINE (ptr, FO) + 1 - lo; + i < (D_NEXT (ptr) ? D_LOWLINE (D_NEXT (ptr), FO) : hi + 1) - lo; + i++) + { + D_RELNUM (result, FILE0 + d, i) = D_RELNUM (result, FILEC, linec); + D_RELLEN (result, FILE0 + d, i) = D_RELLEN (result, FILEC, linec); + linec++; + } + } + } + + /* Set correspond */ + if (!using[0]) + D3_TYPE (result) = DIFF_2ND; + else if (!using[1]) + D3_TYPE (result) = DIFF_1ST; + else + { + int nl0 = D_NUMLINES (result, FILE0); + int nl1 = D_NUMLINES (result, FILE1); + + if (nl0 != nl1 + || !compare_line_list (D_LINEARRAY (result, FILE0), + D_LENARRAY (result, FILE0), + D_LINEARRAY (result, FILE1), + D_LENARRAY (result, FILE1), + nl0)) + D3_TYPE (result) = DIFF_ALL; + else + D3_TYPE (result) = DIFF_3RD; + } + + return result; +} + +/* + * This routine copies pointers from a list of strings to a different list + * of strings. If a spot in the second list is already filled, it + * makes sure that it is filled with the same string; if not it + * returns 0, the copy incomplete. + * Upon successful completion of the copy, it returns 1. + */ +static int +copy_stringlist (fromptrs, fromlengths, toptrs, tolengths, copynum) + char * const fromptrs[]; + char *toptrs[]; + size_t const fromlengths[]; + size_t tolengths[]; + int copynum; +{ + register char * const *f = fromptrs; + register char **t = toptrs; + register size_t const *fl = fromlengths; + register size_t *tl = tolengths; + + while (copynum--) + { + if (*t) + { if (*fl != *tl || memcmp (*f, *t, *fl)) return 0; } + else + { *t = *f ; *tl = *fl; } + + t++; f++; tl++; fl++; + } + return 1; +} + +/* + * Create a diff3_block, with ranges as specified in the arguments. + * Allocate the arrays for the various pointers (and zero them) based + * on the arguments passed. Return the block as a result. + */ +static struct diff3_block * +create_diff3_block (low0, high0, low1, high1, low2, high2) + register int low0, high0, low1, high1, low2, high2; +{ + struct diff3_block *result = ALLOCATE (1, struct diff3_block); + int numlines; + + D3_TYPE (result) = ERROR; + D_NEXT (result) = 0; + + /* Assign ranges */ + D_LOWLINE (result, FILE0) = low0; + D_HIGHLINE (result, FILE0) = high0; + D_LOWLINE (result, FILE1) = low1; + D_HIGHLINE (result, FILE1) = high1; + D_LOWLINE (result, FILE2) = low2; + D_HIGHLINE (result, FILE2) = high2; + + /* Allocate and zero space */ + numlines = D_NUMLINES (result, FILE0); + if (numlines) + { + D_LINEARRAY (result, FILE0) = ALLOCATE (numlines, char *); + D_LENARRAY (result, FILE0) = ALLOCATE (numlines, size_t); + bzero (D_LINEARRAY (result, FILE0), (numlines * sizeof (char *))); + bzero (D_LENARRAY (result, FILE0), (numlines * sizeof (size_t))); + } + else + { + D_LINEARRAY (result, FILE0) = 0; + D_LENARRAY (result, FILE0) = 0; + } + + numlines = D_NUMLINES (result, FILE1); + if (numlines) + { + D_LINEARRAY (result, FILE1) = ALLOCATE (numlines, char *); + D_LENARRAY (result, FILE1) = ALLOCATE (numlines, size_t); + bzero (D_LINEARRAY (result, FILE1), (numlines * sizeof (char *))); + bzero (D_LENARRAY (result, FILE1), (numlines * sizeof (size_t))); + } + else + { + D_LINEARRAY (result, FILE1) = 0; + D_LENARRAY (result, FILE1) = 0; + } + + numlines = D_NUMLINES (result, FILE2); + if (numlines) + { + D_LINEARRAY (result, FILE2) = ALLOCATE (numlines, char *); + D_LENARRAY (result, FILE2) = ALLOCATE (numlines, size_t); + bzero (D_LINEARRAY (result, FILE2), (numlines * sizeof (char *))); + bzero (D_LENARRAY (result, FILE2), (numlines * sizeof (size_t))); + } + else + { + D_LINEARRAY (result, FILE2) = 0; + D_LENARRAY (result, FILE2) = 0; + } + + /* Return */ + return result; +} + +/* + * Compare two lists of lines of text. + * Return 1 if they are equivalent, 0 if not. + */ +static int +compare_line_list (list1, lengths1, list2, lengths2, nl) + char * const list1[], * const list2[]; + size_t const lengths1[], lengths2[]; + int nl; +{ + char + * const *l1 = list1, + * const *l2 = list2; + size_t const + *lgths1 = lengths1, + *lgths2 = lengths2; + + while (nl--) + if (!*l1 || !*l2 || *lgths1 != *lgths2++ + || memcmp (*l1++, *l2++, *lgths1++)) + return 0; + return 1; +} + +/* + * Routines to input and parse two way diffs. + */ + +extern char **environ; + +static struct diff_block * +process_diff (filea, fileb, last_block) + char const *filea, *fileb; + struct diff_block **last_block; +{ + char *diff_contents; + char *diff_limit; + char *scan_diff; + enum diff_type dt; + int i; + struct diff_block *block_list, **block_list_end, *bptr; + + diff_limit = read_diff (filea, fileb, &diff_contents); + scan_diff = diff_contents; + block_list_end = &block_list; + bptr = 0; /* Pacify `gcc -W'. */ + + while (scan_diff < diff_limit) + { + bptr = ALLOCATE (1, struct diff_block); + bptr->lines[0] = bptr->lines[1] = 0; + bptr->lengths[0] = bptr->lengths[1] = 0; + + dt = process_diff_control (&scan_diff, bptr); + if (dt == ERROR || *scan_diff != '\n') + { + fprintf (stderr, "%s: diff error: ", program_name); + do + { + putc (*scan_diff, stderr); + } + while (*scan_diff++ != '\n'); + exit (2); + } + scan_diff++; + + /* Force appropriate ranges to be null, if necessary */ + switch (dt) + { + case ADD: + bptr->ranges[0][0]++; + break; + case DELETE: + bptr->ranges[1][0]++; + break; + case CHANGE: + break; + default: + fatal ("internal error: invalid diff type in process_diff"); + break; + } + + /* Allocate space for the pointers for the lines from filea, and + parcel them out among these pointers */ + if (dt != ADD) + { + int numlines = D_NUMLINES (bptr, 0); + bptr->lines[0] = ALLOCATE (numlines, char *); + bptr->lengths[0] = ALLOCATE (numlines, size_t); + for (i = 0; i < numlines; i++) + scan_diff = scan_diff_line (scan_diff, + &(bptr->lines[0][i]), + &(bptr->lengths[0][i]), + diff_limit, + '<'); + } + + /* Get past the separator for changes */ + if (dt == CHANGE) + { + if (strncmp (scan_diff, "---\n", 4)) + fatal ("invalid diff format; invalid change separator"); + scan_diff += 4; + } + + /* Allocate space for the pointers for the lines from fileb, and + parcel them out among these pointers */ + if (dt != DELETE) + { + int numlines = D_NUMLINES (bptr, 1); + bptr->lines[1] = ALLOCATE (numlines, char *); + bptr->lengths[1] = ALLOCATE (numlines, size_t); + for (i = 0; i < numlines; i++) + scan_diff = scan_diff_line (scan_diff, + &(bptr->lines[1][i]), + &(bptr->lengths[1][i]), + diff_limit, + '>'); + } + + /* Place this block on the blocklist. */ + *block_list_end = bptr; + block_list_end = &bptr->next; + } + + *block_list_end = 0; + *last_block = bptr; + return block_list; +} + +/* + * This routine will parse a normal format diff control string. It + * returns the type of the diff (ERROR if the format is bad). All of + * the other important information is filled into to the structure + * pointed to by db, and the string pointer (whose location is passed + * to this routine) is updated to point beyond the end of the string + * parsed. Note that only the ranges in the diff_block will be set by + * this routine. + * + * If some specific pair of numbers has been reduced to a single + * number, then both corresponding numbers in the diff block are set + * to that number. In general these numbers are interpetted as ranges + * inclusive, unless being used by the ADD or DELETE commands. It is + * assumed that these will be special cased in a superior routine. + */ + +static enum diff_type +process_diff_control (string, db) + char **string; + struct diff_block *db; +{ + char *s = *string; + int holdnum; + enum diff_type type; + +/* These macros are defined here because they can use variables + defined in this function. Don't try this at home kids, we're + trained professionals! + + Also note that SKIPWHITE only recognizes tabs and spaces, and + that READNUM can only read positive, integral numbers */ + +#define SKIPWHITE(s) { while (*s == ' ' || *s == '\t') s++; } +#define READNUM(s, num) \ + { unsigned char c = *s; if (!ISDIGIT (c)) return ERROR; holdnum = 0; \ + do { holdnum = (c - '0' + holdnum * 10); } \ + while (ISDIGIT (c = *++s)); (num) = holdnum; } + + /* Read first set of digits */ + SKIPWHITE (s); + READNUM (s, db->ranges[0][START]); + + /* Was that the only digit? */ + SKIPWHITE (s); + if (*s == ',') + { + /* Get the next digit */ + s++; + READNUM (s, db->ranges[0][END]); + } + else + db->ranges[0][END] = db->ranges[0][START]; + + /* Get the letter */ + SKIPWHITE (s); + switch (*s) + { + case 'a': + type = ADD; + break; + case 'c': + type = CHANGE; + break; + case 'd': + type = DELETE; + break; + default: + return ERROR; /* Bad format */ + } + s++; /* Past letter */ + + /* Read second set of digits */ + SKIPWHITE (s); + READNUM (s, db->ranges[1][START]); + + /* Was that the only digit? */ + SKIPWHITE (s); + if (*s == ',') + { + /* Get the next digit */ + s++; + READNUM (s, db->ranges[1][END]); + SKIPWHITE (s); /* To move to end */ + } + else + db->ranges[1][END] = db->ranges[1][START]; + + *string = s; + return type; +} + +static char * +read_diff (filea, fileb, output_placement) + char const *filea, *fileb; + char **output_placement; +{ + char *diff_result; + size_t bytes, current_chunk_size, total; + int fd, wstatus; + struct stat pipestat; + + /* 302 / 1000 is log10(2.0) rounded up. Subtract 1 for the sign bit; + add 1 for integer division truncation; add 1 more for a minus sign. */ +#define INT_STRLEN_BOUND(type) ((sizeof(type)*CHAR_BIT - 1) * 302 / 1000 + 2) + +#if HAVE_FORK + + char const *argv[7]; + char horizon_arg[17 + INT_STRLEN_BOUND (int)]; + char const **ap; + int fds[2]; + pid_t pid; + + ap = argv; + *ap++ = diff_program; + if (always_text) + *ap++ = "-a"; + sprintf (horizon_arg, "--horizon-lines=%d", horizon_lines); + *ap++ = horizon_arg; + *ap++ = "--"; + *ap++ = filea; + *ap++ = fileb; + *ap = 0; + + if (pipe (fds) != 0) + perror_with_exit ("pipe"); + + pid = fork (); + if (pid == 0) + { + /* Child */ + close (fds[0]); + if (fds[1] != STDOUT_FILENO) + { + dup2 (fds[1], STDOUT_FILENO); + close (fds[1]); + } + execve (diff_program, (char **) argv, environ); + /* Avoid stdio, because the parent process's buffers are inherited. */ + write (STDERR_FILENO, diff_program, strlen (diff_program)); + write (STDERR_FILENO, ": not found\n", 12); + _exit (2); + } + + if (pid == -1) + perror_with_exit ("fork failed"); + + close (fds[1]); /* Prevent erroneous lack of EOF */ + fd = fds[0]; + +#else /* ! HAVE_FORK */ + + FILE *fpipe; + char *command = xmalloc (sizeof (diff_program) + 30 + INT_STRLEN_BOUND (int) + + 4 * (strlen (filea) + strlen (fileb))); + char *p; + sprintf (command, "%s -a --horizon-lines=%d -- ", + diff_program, horizon_lines); + p = command + strlen (command); + SYSTEM_QUOTE_ARG (p, filea); + *p++ = ' '; + SYSTEM_QUOTE_ARG (p, fileb); + *p = '\0'; + fpipe = popen (command, "r"); + if (!fpipe) + perror_with_exit (command); + free (command); + fd = fileno (fpipe); + +#endif /* ! HAVE_FORK */ + + current_chunk_size = 8 * 1024; + if (fstat (fd, &pipestat) == 0) + current_chunk_size = max (current_chunk_size, STAT_BLOCKSIZE (pipestat)); + + diff_result = xmalloc (current_chunk_size); + total = 0; + do { + bytes = myread (fd, + diff_result + total, + current_chunk_size - total); + total += bytes; + if (total == current_chunk_size) + { + if (current_chunk_size < 2 * current_chunk_size) + current_chunk_size = 2 * current_chunk_size; + else if (current_chunk_size < (size_t) -1) + current_chunk_size = (size_t) -1; + else + fatal ("files are too large to fit into memory"); + diff_result = xrealloc (diff_result, (current_chunk_size *= 2)); + } + } while (bytes); + + if (total != 0 && diff_result[total-1] != '\n') + fatal ("invalid diff format; incomplete last line"); + + *output_placement = diff_result; + +#if ! HAVE_FORK + + wstatus = pclose (fpipe); + +#else /* HAVE_FORK */ + + if (close (fd) != 0) + perror_with_exit ("pipe close"); + if (waitpid (pid, &wstatus, 0) < 0) + perror_with_exit ("waitpid failed"); + +#endif /* HAVE_FORK */ + + if (! (WIFEXITED (wstatus) && WEXITSTATUS (wstatus) < 2)) + fatal ("subsidiary diff failed"); + + return diff_result + total; +} + + +/* + * Scan a regular diff line (consisting of > or <, followed by a + * space, followed by text (including nulls) up to a newline. + * + * This next routine began life as a macro and many parameters in it + * are used as call-by-reference values. + */ +static char * +scan_diff_line (scan_ptr, set_start, set_length, limit, leadingchar) + char *scan_ptr, **set_start; + size_t *set_length; + char *limit; + int leadingchar; +{ + char *line_ptr; + + if (!(scan_ptr[0] == leadingchar + && scan_ptr[1] == ' ')) + fatal ("invalid diff format; incorrect leading line chars"); + + *set_start = line_ptr = scan_ptr + 2; + while (*line_ptr++ != '\n') + ; + + /* Include newline if the original line ended in a newline, + or if an edit script is being generated. + Copy any missing newline message to stderr if an edit script is being + generated, because edit scripts cannot handle missing newlines. + Return the beginning of the next line. */ + *set_length = line_ptr - *set_start; + if (line_ptr < limit && *line_ptr == '\\') + { + if (edscript) + fprintf (stderr, "%s:", program_name); + else + --*set_length; + line_ptr++; + do + { + if (edscript) + putc (*line_ptr, stderr); + } + while (*line_ptr++ != '\n'); + } + + return line_ptr; +} + +/* + * This routine outputs a three way diff passed as a list of + * diff3_block's. + * The argument MAPPING is indexed by external file number (in the + * argument list) and contains the internal file number (from the + * diff passed). This is important because the user expects his + * outputs in terms of the argument list number, and the diff passed + * may have been done slightly differently (if the last argument + * was "-", for example). + * REV_MAPPING is the inverse of MAPPING. + */ +static void +output_diff3 (outputfile, diff, mapping, rev_mapping) + FILE *outputfile; + struct diff3_block *diff; + int const mapping[3], rev_mapping[3]; +{ + int i; + int oddoneout; + char *cp; + struct diff3_block *ptr; + int line; + size_t length; + int dontprint; + static int skew_increment[3] = { 2, 3, 1 }; /* 0==>2==>1==>3 */ + char const *line_prefix = tab_align_flag ? "\t" : " "; + + for (ptr = diff; ptr; ptr = D_NEXT (ptr)) + { + char x[2]; + + switch (ptr->correspond) + { + case DIFF_ALL: + x[0] = '\0'; + dontprint = 3; /* Print them all */ + oddoneout = 3; /* Nobody's odder than anyone else */ + break; + case DIFF_1ST: + case DIFF_2ND: + case DIFF_3RD: + oddoneout = rev_mapping[(int) ptr->correspond - (int) DIFF_1ST]; + + x[0] = oddoneout + '1'; + x[1] = '\0'; + dontprint = oddoneout==0; + break; + default: + fatal ("internal error: invalid diff type passed to output"); + } + fprintf (outputfile, "====%s\n", x); + + /* Go 0, 2, 1 if the first and third outputs are equivalent. */ + for (i = 0; i < 3; + i = (oddoneout == 1 ? skew_increment[i] : i + 1)) + { + int realfile = mapping[i]; + int + lowt = D_LOWLINE (ptr, realfile), + hight = D_HIGHLINE (ptr, realfile); + + fprintf (outputfile, "%d:", i + 1); + switch (lowt - hight) + { + case 1: + fprintf (outputfile, "%da\n", lowt - 1); + break; + case 0: + fprintf (outputfile, "%dc\n", lowt); + break; + default: + fprintf (outputfile, "%d,%dc\n", lowt, hight); + break; + } + + if (i == dontprint) continue; + + if (lowt <= hight) + { + line = 0; + do + { + fprintf (outputfile, line_prefix); + cp = D_RELNUM (ptr, realfile, line); + length = D_RELLEN (ptr, realfile, line); + fwrite (cp, sizeof (char), length, outputfile); + } + while (++line < hight - lowt + 1); + if (cp[length - 1] != '\n') + fprintf (outputfile, "\n\\ No newline at end of file\n"); + } + } + } +} + + +/* + * Output to OUTPUTFILE the lines of B taken from FILENUM. + * Double any initial '.'s; yield nonzero if any initial '.'s were doubled. + */ +static int +dotlines (outputfile, b, filenum) + FILE *outputfile; + struct diff3_block *b; + int filenum; +{ + int i; + int leading_dot = 0; + + for (i = 0; + i < D_NUMLINES (b, filenum); + i++) + { + char *line = D_RELNUM (b, filenum, i); + if (line[0] == '.') + { + leading_dot = 1; + fprintf (outputfile, "."); + } + fwrite (line, sizeof (char), + D_RELLEN (b, filenum, i), outputfile); + } + + return leading_dot; +} + +/* + * Output to OUTPUTFILE a '.' line. If LEADING_DOT is nonzero, + * also output a command that removes initial '.'s + * starting with line START and continuing for NUM lines. + */ +static void +undotlines (outputfile, leading_dot, start, num) + FILE *outputfile; + int leading_dot, start, num; +{ + fprintf (outputfile, ".\n"); + if (leading_dot) + if (num == 1) + fprintf (outputfile, "%ds/^\\.//\n", start); + else + fprintf (outputfile, "%d,%ds/^\\.//\n", start, start + num - 1); +} + +/* + * This routine outputs a diff3 set of blocks as an ed script. This + * script applies the changes between file's 2 & 3 to file 1. It + * takes the precise format of the ed script to be output from global + * variables set during options processing. Note that it does + * destructive things to the set of diff3 blocks it is passed; it + * reverses their order (this gets around the problems involved with + * changing line numbers in an ed script). + * + * Note that this routine has the same problem of mapping as the last + * one did; the variable MAPPING maps from file number according to + * the argument list to file number according to the diff passed. All + * files listed below are in terms of the argument list. + * REV_MAPPING is the inverse of MAPPING. + * + * The arguments FILE0, FILE1 and FILE2 are the strings to print + * as the names of the three files. These may be the actual names, + * or may be the arguments specified with -L. + * + * Returns 1 if conflicts were found. + */ + +static int +output_diff3_edscript (outputfile, diff, mapping, rev_mapping, + file0, file1, file2) + FILE *outputfile; + struct diff3_block *diff; + int const mapping[3], rev_mapping[3]; + char const *file0, *file1, *file2; +{ + int leading_dot; + int conflicts_found = 0, conflict; + struct diff3_block *b; + + for (b = reverse_diff3_blocklist (diff); b; b = b->next) + { + /* Must do mapping correctly. */ + enum diff_type type + = ((b->correspond == DIFF_ALL) ? + DIFF_ALL : + ((enum diff_type) + (((int) DIFF_1ST) + + rev_mapping[(int) b->correspond - (int) DIFF_1ST]))); + + /* If we aren't supposed to do this output block, skip it. */ + switch (type) + { + default: continue; + case DIFF_2ND: if (!show_2nd) continue; conflict = 1; break; + case DIFF_3RD: if (overlap_only) continue; conflict = 0; break; + case DIFF_ALL: if (simple_only) continue; conflict = flagging; break; + } + + if (conflict) + { + conflicts_found = 1; + + + /* Mark end of conflict. */ + + fprintf (outputfile, "%da\n", D_HIGHLINE (b, mapping[FILE0])); + leading_dot = 0; + if (type == DIFF_ALL) + { + if (show_2nd) + { + /* Append lines from FILE1. */ + fprintf (outputfile, "||||||| %s\n", file1); + leading_dot = dotlines (outputfile, b, mapping[FILE1]); + } + /* Append lines from FILE2. */ + fprintf (outputfile, "=======\n"); + leading_dot |= dotlines (outputfile, b, mapping[FILE2]); + } + fprintf (outputfile, ">>>>>>> %s\n", file2); + undotlines (outputfile, leading_dot, + D_HIGHLINE (b, mapping[FILE0]) + 2, + (D_NUMLINES (b, mapping[FILE1]) + + D_NUMLINES (b, mapping[FILE2]) + 1)); + + + /* Mark start of conflict. */ + + fprintf (outputfile, "%da\n<<<<<<< %s\n", + D_LOWLINE (b, mapping[FILE0]) - 1, + type == DIFF_ALL ? file0 : file1); + leading_dot = 0; + if (type == DIFF_2ND) + { + /* Prepend lines from FILE1. */ + leading_dot = dotlines (outputfile, b, mapping[FILE1]); + fprintf (outputfile, "=======\n"); + } + undotlines (outputfile, leading_dot, + D_LOWLINE (b, mapping[FILE0]) + 1, + D_NUMLINES (b, mapping[FILE1])); + } + else if (D_NUMLINES (b, mapping[FILE2]) == 0) + /* Write out a delete */ + { + if (D_NUMLINES (b, mapping[FILE0]) == 1) + fprintf (outputfile, "%dd\n", + D_LOWLINE (b, mapping[FILE0])); + else + fprintf (outputfile, "%d,%dd\n", + D_LOWLINE (b, mapping[FILE0]), + D_HIGHLINE (b, mapping[FILE0])); + } + else + /* Write out an add or change */ + { + switch (D_NUMLINES (b, mapping[FILE0])) + { + case 0: + fprintf (outputfile, "%da\n", + D_HIGHLINE (b, mapping[FILE0])); + break; + case 1: + fprintf (outputfile, "%dc\n", + D_HIGHLINE (b, mapping[FILE0])); + break; + default: + fprintf (outputfile, "%d,%dc\n", + D_LOWLINE (b, mapping[FILE0]), + D_HIGHLINE (b, mapping[FILE0])); + break; + } + + undotlines (outputfile, dotlines (outputfile, b, mapping[FILE2]), + D_LOWLINE (b, mapping[FILE0]), + D_NUMLINES (b, mapping[FILE2])); + } + } + if (finalwrite) fprintf (outputfile, "w\nq\n"); + return conflicts_found; +} + +/* + * Read from INFILE and output to OUTPUTFILE a set of diff3_ blocks DIFF + * as a merged file. This acts like 'ed file0 <[output_diff3_edscript]', + * except that it works even for binary data or incomplete lines. + * + * As before, MAPPING maps from arg list file number to diff file number, + * REV_MAPPING is its inverse, + * and FILE0, FILE1, and FILE2 are the names of the files. + * + * Returns 1 if conflicts were found. + */ + +static int +output_diff3_merge (infile, outputfile, diff, mapping, rev_mapping, + file0, file1, file2) + FILE *infile, *outputfile; + struct diff3_block *diff; + int const mapping[3], rev_mapping[3]; + char const *file0, *file1, *file2; +{ + int c, i; + int conflicts_found = 0, conflict; + struct diff3_block *b; + int linesread = 0; + + for (b = diff; b; b = b->next) + { + /* Must do mapping correctly. */ + enum diff_type type + = ((b->correspond == DIFF_ALL) ? + DIFF_ALL : + ((enum diff_type) + (((int) DIFF_1ST) + + rev_mapping[(int) b->correspond - (int) DIFF_1ST]))); + char const *format_2nd = "<<<<<<< %s\n"; + + /* If we aren't supposed to do this output block, skip it. */ + switch (type) + { + default: continue; + case DIFF_2ND: if (!show_2nd) continue; conflict = 1; break; + case DIFF_3RD: if (overlap_only) continue; conflict = 0; break; + case DIFF_ALL: if (simple_only) continue; conflict = flagging; + format_2nd = "||||||| %s\n"; + break; + } + + /* Copy I lines from file 0. */ + i = D_LOWLINE (b, FILE0) - linesread - 1; + linesread += i; + while (0 <= --i) + do + { + c = getc (infile); + if (c == EOF) + if (ferror (infile)) + perror_with_exit ("input file"); + else if (feof (infile)) + fatal ("input file shrank"); + putc (c, outputfile); + } + while (c != '\n'); + + if (conflict) + { + conflicts_found = 1; + + if (type == DIFF_ALL) + { + /* Put in lines from FILE0 with bracket. */ + fprintf (outputfile, "<<<<<<< %s\n", file0); + for (i = 0; + i < D_NUMLINES (b, mapping[FILE0]); + i++) + fwrite (D_RELNUM (b, mapping[FILE0], i), sizeof (char), + D_RELLEN (b, mapping[FILE0], i), outputfile); + } + + if (show_2nd) + { + /* Put in lines from FILE1 with bracket. */ + fprintf (outputfile, format_2nd, file1); + for (i = 0; + i < D_NUMLINES (b, mapping[FILE1]); + i++) + fwrite (D_RELNUM (b, mapping[FILE1], i), sizeof (char), + D_RELLEN (b, mapping[FILE1], i), outputfile); + } + + fprintf (outputfile, "=======\n"); + } + + /* Put in lines from FILE2. */ + for (i = 0; + i < D_NUMLINES (b, mapping[FILE2]); + i++) + fwrite (D_RELNUM (b, mapping[FILE2], i), sizeof (char), + D_RELLEN (b, mapping[FILE2], i), outputfile); + + if (conflict) + fprintf (outputfile, ">>>>>>> %s\n", file2); + + /* Skip I lines in file 0. */ + i = D_NUMLINES (b, FILE0); + linesread += i; + while (0 <= --i) + while ((c = getc (infile)) != '\n') + if (c == EOF) + if (ferror (infile)) + perror_with_exit ("input file"); + else if (feof (infile)) + { + if (i || b->next) + fatal ("input file shrank"); + return conflicts_found; + } + } + /* Copy rest of common file. */ + while ((c = getc (infile)) != EOF || !(ferror (infile) | feof (infile))) + putc (c, outputfile); + return conflicts_found; +} + +/* + * Reverse the order of the list of diff3 blocks. + */ +static struct diff3_block * +reverse_diff3_blocklist (diff) + struct diff3_block *diff; +{ + register struct diff3_block *tmp, *next, *prev; + + for (tmp = diff, prev = 0; tmp; tmp = next) + { + next = tmp->next; + tmp->next = prev; + prev = tmp; + } + + return prev; +} + +static size_t +myread (fd, ptr, size) + int fd; + char *ptr; + size_t size; +{ + size_t result = read (fd, ptr, size); + if (result == -1) + perror_with_exit ("read failed"); + return result; +} + +static VOID * +xmalloc (size) + size_t size; +{ + VOID *result = (VOID *) malloc (size ? size : 1); + if (!result) + fatal ("memory exhausted"); + return result; +} + +static VOID * +xrealloc (ptr, size) + VOID *ptr; + size_t size; +{ + VOID *result = (VOID *) realloc (ptr, size ? size : 1); + if (!result) + fatal ("memory exhausted"); + return result; +} + +static void +fatal (string) + char const *string; +{ + fprintf (stderr, "%s: %s\n", program_name, string); + exit (2); +} + +static void +perror_with_exit (string) + char const *string; +{ + int e = errno; + fprintf (stderr, "%s: ", program_name); + errno = e; + perror (string); + exit (2); +} diff --git a/sys/src/ape/cmd/diff/dir.c b/sys/src/ape/cmd/diff/dir.c new file mode 100755 index 000000000..036a86f11 --- /dev/null +++ b/sys/src/ape/cmd/diff/dir.c @@ -0,0 +1,216 @@ +/* Read, sort and compare two directories. Used for GNU DIFF. + Copyright (C) 1988, 1989, 1992, 1993, 1994 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#include "diff.h" + +/* Read the directory named by DIR and store into DIRDATA a sorted vector + of filenames for its contents. DIR->desc == -1 means this directory is + known to be nonexistent, so set DIRDATA to an empty vector. + Return -1 (setting errno) if error, 0 otherwise. */ + +struct dirdata +{ + char const **names; /* Sorted names of files in dir, 0-terminated. */ + char *data; /* Allocated storage for file names. */ +}; + +static int compare_names PARAMS((void const *, void const *)); +static int dir_sort PARAMS((struct file_data const *, struct dirdata *)); + +static int +dir_sort (dir, dirdata) + struct file_data const *dir; + struct dirdata *dirdata; +{ + register struct dirent *next; + register int i; + + /* Address of block containing the files that are described. */ + char const **names; + + /* Number of files in directory. */ + size_t nnames; + + /* Allocated and used storage for file name data. */ + char *data; + size_t data_alloc, data_used; + + dirdata->names = 0; + dirdata->data = 0; + nnames = 0; + data = 0; + + if (dir->desc != -1) + { + /* Open the directory and check for errors. */ + register DIR *reading = opendir (dir->name); + if (!reading) + return -1; + + /* Initialize the table of filenames. */ + + data_alloc = max (1, (size_t) dir->stat.st_size); + data_used = 0; + dirdata->data = data = xmalloc (data_alloc); + + /* Read the directory entries, and insert the subfiles + into the `data' table. */ + + while ((errno = 0, (next = readdir (reading)) != 0)) + { + char *d_name = next->d_name; + size_t d_size = NAMLEN (next) + 1; + + /* Ignore the files `.' and `..' */ + if (d_name[0] == '.' + && (d_name[1] == 0 || (d_name[1] == '.' && d_name[2] == 0))) + continue; + + if (excluded_filename (d_name)) + continue; + + while (data_alloc < data_used + d_size) + dirdata->data = data = xrealloc (data, data_alloc *= 2); + memcpy (data + data_used, d_name, d_size); + data_used += d_size; + nnames++; + } + if (errno) + { + int e = errno; + closedir (reading); + errno = e; + return -1; + } +#if CLOSEDIR_VOID + closedir (reading); +#else + if (closedir (reading) != 0) + return -1; +#endif + } + + /* Create the `names' table from the `data' table. */ + dirdata->names = names = (char const **) xmalloc (sizeof (char *) + * (nnames + 1)); + for (i = 0; i < nnames; i++) + { + names[i] = data; + data += strlen (data) + 1; + } + names[nnames] = 0; + + /* Sort the table. */ + qsort (names, nnames, sizeof (char *), compare_names); + + return 0; +} + +/* Sort the files now in the table. */ + +static int +compare_names (file1, file2) + void const *file1, *file2; +{ + return filename_cmp (* (char const *const *) file1, + * (char const *const *) file2); +} + +/* Compare the contents of two directories named in FILEVEC[0] and FILEVEC[1]. + This is a top-level routine; it does everything necessary for diff + on two directories. + + FILEVEC[0].desc == -1 says directory FILEVEC[0] doesn't exist, + but pretend it is empty. Likewise for FILEVEC[1]. + + HANDLE_FILE is a caller-provided subroutine called to handle each file. + It gets five operands: dir and name (rel to original working dir) of file + in dir 0, dir and name pathname of file in dir 1, and the recursion depth. + + For a file that appears in only one of the dirs, one of the name-args + to HANDLE_FILE is zero. + + DEPTH is the current depth in recursion, used for skipping top-level + files by the -S option. + + Returns the maximum of all the values returned by HANDLE_FILE, + or 2 if trouble is encountered in opening files. */ + +int +diff_dirs (filevec, handle_file, depth) + struct file_data const filevec[]; + int (*handle_file) PARAMS((char const *, char const *, char const *, char const *, int)); + int depth; +{ + struct dirdata dirdata[2]; + int val = 0; /* Return value. */ + int i; + + /* Get sorted contents of both dirs. */ + for (i = 0; i < 2; i++) + if (dir_sort (&filevec[i], &dirdata[i]) != 0) + { + perror_with_name (filevec[i].name); + val = 2; + } + + if (val == 0) + { + register char const * const *names0 = dirdata[0].names; + register char const * const *names1 = dirdata[1].names; + char const *name0 = filevec[0].name; + char const *name1 = filevec[1].name; + + /* If `-S name' was given, and this is the topmost level of comparison, + ignore all file names less than the specified starting name. */ + + if (dir_start_file && depth == 0) + { + while (*names0 && filename_cmp (*names0, dir_start_file) < 0) + names0++; + while (*names1 && filename_cmp (*names1, dir_start_file) < 0) + names1++; + } + + /* Loop while files remain in one or both dirs. */ + while (*names0 || *names1) + { + /* Compare next name in dir 0 with next name in dir 1. + At the end of a dir, + pretend the "next name" in that dir is very large. */ + int nameorder = (!*names0 ? 1 : !*names1 ? -1 + : filename_cmp (*names0, *names1)); + int v1 = (*handle_file) (name0, 0 < nameorder ? 0 : *names0++, + name1, nameorder < 0 ? 0 : *names1++, + depth + 1); + if (v1 > val) + val = v1; + } + } + + for (i = 0; i < 2; i++) + { + if (dirdata[i].names) + free (dirdata[i].names); + if (dirdata[i].data) + free (dirdata[i].data); + } + + return val; +} diff --git a/sys/src/ape/cmd/diff/ed.c b/sys/src/ape/cmd/diff/ed.c new file mode 100755 index 000000000..717ef358d --- /dev/null +++ b/sys/src/ape/cmd/diff/ed.c @@ -0,0 +1,200 @@ +/* Output routines for ed-script format. + Copyright (C) 1988, 89, 91, 92, 93 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#include "diff.h" + +static void print_ed_hunk PARAMS((struct change *)); +static void print_rcs_hunk PARAMS((struct change *)); +static void pr_forward_ed_hunk PARAMS((struct change *)); + +/* Print our script as ed commands. */ + +void +print_ed_script (script) + struct change *script; +{ + print_script (script, find_reverse_change, print_ed_hunk); +} + +/* Print a hunk of an ed diff */ + +static void +print_ed_hunk (hunk) + struct change *hunk; +{ + int f0, l0, f1, l1; + int deletes, inserts; + +#if 0 + hunk = flip_script (hunk); +#endif +#ifdef DEBUG + debug_script (hunk); +#endif + + /* Determine range of line numbers involved in each file. */ + analyze_hunk (hunk, &f0, &l0, &f1, &l1, &deletes, &inserts); + if (!deletes && !inserts) + return; + + begin_output (); + + /* Print out the line number header for this hunk */ + print_number_range (',', &files[0], f0, l0); + fprintf (outfile, "%c\n", change_letter (inserts, deletes)); + + /* Print new/changed lines from second file, if needed */ + if (inserts) + { + int i; + int inserting = 1; + for (i = f1; i <= l1; i++) + { + /* Resume the insert, if we stopped. */ + if (! inserting) + fprintf (outfile, "%da\n", + i - f1 + translate_line_number (&files[0], f0) - 1); + inserting = 1; + + /* If the file's line is just a dot, it would confuse `ed'. + So output it with a double dot, and set the flag LEADING_DOT + so that we will output another ed-command later + to change the double dot into a single dot. */ + + if (files[1].linbuf[i][0] == '.' + && files[1].linbuf[i][1] == '\n') + { + fprintf (outfile, "..\n"); + fprintf (outfile, ".\n"); + /* Now change that double dot to the desired single dot. */ + fprintf (outfile, "%ds/^\\.\\././\n", + i - f1 + translate_line_number (&files[0], f0)); + inserting = 0; + } + else + /* Line is not `.', so output it unmodified. */ + print_1_line ("", &files[1].linbuf[i]); + } + + /* End insert mode, if we are still in it. */ + if (inserting) + fprintf (outfile, ".\n"); + } +} + +/* Print change script in the style of ed commands, + but print the changes in the order they appear in the input files, + which means that the commands are not truly useful with ed. */ + +void +pr_forward_ed_script (script) + struct change *script; +{ + print_script (script, find_change, pr_forward_ed_hunk); +} + +static void +pr_forward_ed_hunk (hunk) + struct change *hunk; +{ + int i; + int f0, l0, f1, l1; + int deletes, inserts; + + /* Determine range of line numbers involved in each file. */ + analyze_hunk (hunk, &f0, &l0, &f1, &l1, &deletes, &inserts); + if (!deletes && !inserts) + return; + + begin_output (); + + fprintf (outfile, "%c", change_letter (inserts, deletes)); + print_number_range (' ', files, f0, l0); + fprintf (outfile, "\n"); + + /* If deletion only, print just the number range. */ + + if (!inserts) + return; + + /* For insertion (with or without deletion), print the number range + and the lines from file 2. */ + + for (i = f1; i <= l1; i++) + print_1_line ("", &files[1].linbuf[i]); + + fprintf (outfile, ".\n"); +} + +/* Print in a format somewhat like ed commands + except that each insert command states the number of lines it inserts. + This format is used for RCS. */ + +void +print_rcs_script (script) + struct change *script; +{ + print_script (script, find_change, print_rcs_hunk); +} + +/* Print a hunk of an RCS diff */ + +static void +print_rcs_hunk (hunk) + struct change *hunk; +{ + int i; + int f0, l0, f1, l1; + int deletes, inserts; + int tf0, tl0, tf1, tl1; + + /* Determine range of line numbers involved in each file. */ + analyze_hunk (hunk, &f0, &l0, &f1, &l1, &deletes, &inserts); + if (!deletes && !inserts) + return; + + begin_output (); + + translate_range (&files[0], f0, l0, &tf0, &tl0); + + if (deletes) + { + fprintf (outfile, "d"); + /* For deletion, print just the starting line number from file 0 + and the number of lines deleted. */ + fprintf (outfile, "%d %d\n", + tf0, + (tl0 >= tf0 ? tl0 - tf0 + 1 : 1)); + } + + if (inserts) + { + fprintf (outfile, "a"); + + /* Take last-line-number from file 0 and # lines from file 1. */ + translate_range (&files[1], f1, l1, &tf1, &tl1); + fprintf (outfile, "%d %d\n", + tl0, + (tl1 >= tf1 ? tl1 - tf1 + 1 : 1)); + + /* Print the inserted lines. */ + for (i = f1; i <= l1; i++) + print_1_line ("", &files[1].linbuf[i]); + } +} diff --git a/sys/src/ape/cmd/diff/fnmatch.c b/sys/src/ape/cmd/diff/fnmatch.c new file mode 100755 index 000000000..cf2d75190 --- /dev/null +++ b/sys/src/ape/cmd/diff/fnmatch.c @@ -0,0 +1,181 @@ +/* Copyright (C) 1992 Free Software Foundation, Inc. +This file is part of the GNU C Library. + +The GNU C Library is free software; you can redistribute it and/or +modify it under the terms of the GNU Library General Public License as +published by the Free Software Foundation; either version 2 of the +License, or (at your option) any later version. + +The GNU C Library is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +Library General Public License for more details. */ + +/* Modified slightly by Brian Berliner <berliner@sun.com> and + Jim Blandy <jimb@cyclic.com> for CVS use */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "system.h" + +/* IGNORE(@ */ +/* #include <ansidecl.h> */ +/* @) */ +#include <errno.h> +#include "fnmatch.h" + +#if !defined(__GNU_LIBRARY__) && !defined(STDC_HEADERS) +extern int errno; +#endif + +/* Match STRING against the filename pattern PATTERN, returning zero if + it matches, nonzero if not. */ +int +#if __STDC__ +fnmatch (const char *pattern, const char *string, int flags) +#else +fnmatch (pattern, string, flags) + char *pattern; + char *string; + int flags; +#endif +{ + register const char *p = pattern, *n = string; + register char c; + + if ((flags & ~__FNM_FLAGS) != 0) + { + errno = EINVAL; + return -1; + } + + while ((c = *p++) != '\0') + { + switch (c) + { + case '?': + if (*n == '\0') + return FNM_NOMATCH; + else if ((flags & FNM_PATHNAME) && *n == '/') + return FNM_NOMATCH; + else if ((flags & FNM_PERIOD) && *n == '.' && + (n == string || ((flags & FNM_PATHNAME) && n[-1] == '/'))) + return FNM_NOMATCH; + break; + + case '\\': + if (!(flags & FNM_NOESCAPE)) + c = *p++; + if (FOLD_FN_CHAR (*n) != FOLD_FN_CHAR (c)) + return FNM_NOMATCH; + break; + + case '*': + if ((flags & FNM_PERIOD) && *n == '.' && + (n == string || ((flags & FNM_PATHNAME) && n[-1] == '/'))) + return FNM_NOMATCH; + + for (c = *p++; c == '?' || c == '*'; c = *p++, ++n) + if (((flags & FNM_PATHNAME) && *n == '/') || + (c == '?' && *n == '\0')) + return FNM_NOMATCH; + + if (c == '\0') + return 0; + + { + char c1 = (!(flags & FNM_NOESCAPE) && c == '\\') ? *p : c; + for (--p; *n != '\0'; ++n) + if ((c == '[' || FOLD_FN_CHAR (*n) == FOLD_FN_CHAR (c1)) && + fnmatch(p, n, flags & ~FNM_PERIOD) == 0) + return 0; + return FNM_NOMATCH; + } + + case '[': + { + /* Nonzero if the sense of the character class is inverted. */ + register int not; + + if (*n == '\0') + return FNM_NOMATCH; + + if ((flags & FNM_PERIOD) && *n == '.' && + (n == string || ((flags & FNM_PATHNAME) && n[-1] == '/'))) + return FNM_NOMATCH; + + not = (*p == '!' || *p == '^'); + if (not) + ++p; + + c = *p++; + for (;;) + { + register char cstart = c, cend = c; + + if (!(flags & FNM_NOESCAPE) && c == '\\') + cstart = cend = *p++; + + if (c == '\0') + /* [ (unterminated) loses. */ + return FNM_NOMATCH; + + c = *p++; + + if ((flags & FNM_PATHNAME) && c == '/') + /* [/] can never match. */ + return FNM_NOMATCH; + + if (c == '-' && *p != ']') + { + cend = *p++; + if (!(flags & FNM_NOESCAPE) && cend == '\\') + cend = *p++; + if (cend == '\0') + return FNM_NOMATCH; + c = *p++; + } + + if (*n >= cstart && *n <= cend) + goto matched; + + if (c == ']') + break; + } + if (!not) + return FNM_NOMATCH; + break; + + matched:; + /* Skip the rest of the [...] that already matched. */ + while (c != ']') + { + if (c == '\0') + /* [... (unterminated) loses. */ + return FNM_NOMATCH; + + c = *p++; + if (!(flags & FNM_NOESCAPE) && c == '\\') + /* 1003.2d11 is unclear if this is right. %%% */ + ++p; + } + if (not) + return FNM_NOMATCH; + } + break; + + default: + if (FOLD_FN_CHAR (c) != FOLD_FN_CHAR (*n)) + return FNM_NOMATCH; + } + + ++n; + } + + if (*n == '\0') + return 0; + + return FNM_NOMATCH; +} diff --git a/sys/src/ape/cmd/diff/fnmatch.h b/sys/src/ape/cmd/diff/fnmatch.h new file mode 100755 index 000000000..b15734721 --- /dev/null +++ b/sys/src/ape/cmd/diff/fnmatch.h @@ -0,0 +1,40 @@ +/* Copyright (C) 1992 Free Software Foundation, Inc. +This file is part of the GNU C Library. + +The GNU C Library is free software; you can redistribute it and/or +modify it under the terms of the GNU Library General Public License as +published by the Free Software Foundation; either version 2 of the +License, or (at your option) any later version. + +The GNU C Library is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +Library General Public License for more details. */ + +#ifndef _FNMATCH_H + +#define _FNMATCH_H 1 + +/* Bits set in the FLAGS argument to `fnmatch'. */ +#undef FNM_PATHNAME +#define FNM_PATHNAME (1 << 0)/* No wildcard can ever match `/'. */ +#undef FNM_NOESCAPE +#define FNM_NOESCAPE (1 << 1)/* Backslashes don't quote special chars. */ +#undef FNM_PERIOD +#define FNM_PERIOD (1 << 2)/* Leading `.' is matched only explicitly. */ +#undef __FNM_FLAGS +#define __FNM_FLAGS (FNM_PATHNAME|FNM_NOESCAPE|FNM_PERIOD) + +/* Value returned by `fnmatch' if STRING does not match PATTERN. */ +#undef FNM_NOMATCH +#define FNM_NOMATCH 1 + +/* Match STRING against the filename pattern PATTERN, + returning zero if it matches, FNM_NOMATCH if not. */ +#if __STDC__ +extern int fnmatch (const char *pattern, const char *string, int flags); +#else +extern int fnmatch (); +#endif + +#endif /* fnmatch.h */ diff --git a/sys/src/ape/cmd/diff/getopt.c b/sys/src/ape/cmd/diff/getopt.c new file mode 100755 index 000000000..964189d2d --- /dev/null +++ b/sys/src/ape/cmd/diff/getopt.c @@ -0,0 +1,748 @@ +/* Getopt for GNU. + NOTE: getopt is now part of the C library, so if you don't know what + "Keep this file name-space clean" means, talk to roland@gnu.ai.mit.edu + before changing it! + + Copyright (C) 1987, 88, 89, 90, 91, 92, 93, 94 + Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* This tells Alpha OSF/1 not to define a getopt prototype in <stdio.h>. + Ditto for AIX 3.2 and <stdlib.h>. */ +#ifndef _NO_PROTO +#define _NO_PROTO +#endif + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#ifndef __STDC__ +/* This is a separate conditional since some stdc systems + reject `defined (const)'. */ +#ifndef const +#define const +#endif +#endif + +#include <stdio.h> + +/* Comment out all this code if we are using the GNU C Library, and are not + actually compiling the library itself. This code is part of the GNU C + Library, but also included in many other GNU distributions. Compiling + and linking in this code is a waste when using the GNU C library + (especially if it is a shared library). Rather than having every GNU + program understand `configure --with-gnu-libc' and omit the object files, + it is simpler to just do this in the source for each such file. */ + +#if defined (_LIBC) || !defined (__GNU_LIBRARY__) + + +/* This needs to come after some library #include + to get __GNU_LIBRARY__ defined. */ +#ifdef __GNU_LIBRARY__ +/* Don't include stdlib.h for non-GNU C libraries because some of them + contain conflicting prototypes for getopt. */ +#include <stdlib.h> +#endif /* GNU C library. */ + +/* This version of `getopt' appears to the caller like standard Unix `getopt' + but it behaves differently for the user, since it allows the user + to intersperse the options with the other arguments. + + As `getopt' works, it permutes the elements of ARGV so that, + when it is done, all the options precede everything else. Thus + all application programs are extended to handle flexible argument order. + + Setting the environment variable POSIXLY_CORRECT disables permutation. + Then the behavior is completely standard. + + GNU application programs can use a third alternative mode in which + they can distinguish the relative order of options and other arguments. */ + +#include "getopt.h" + +/* For communication from `getopt' to the caller. + When `getopt' finds an option that takes an argument, + the argument value is returned here. + Also, when `ordering' is RETURN_IN_ORDER, + each non-option ARGV-element is returned here. */ + +char *optarg = NULL; + +/* Index in ARGV of the next element to be scanned. + This is used for communication to and from the caller + and for communication between successive calls to `getopt'. + + On entry to `getopt', zero means this is the first call; initialize. + + When `getopt' returns EOF, this is the index of the first of the + non-option elements that the caller should itself scan. + + Otherwise, `optind' communicates from one call to the next + how much of ARGV has been scanned so far. */ + +/* XXX 1003.2 says this must be 1 before any call. */ +int optind = 0; + +/* The next char to be scanned in the option-element + in which the last option character we returned was found. + This allows us to pick up the scan where we left off. + + If this is zero, or a null string, it means resume the scan + by advancing to the next ARGV-element. */ + +static char *nextchar; + +/* Callers store zero here to inhibit the error message + for unrecognized options. */ + +int opterr = 1; + +/* Set to an option character which was unrecognized. + This must be initialized on some systems to avoid linking in the + system's own getopt implementation. */ + +int optopt = '?'; + +/* Describe how to deal with options that follow non-option ARGV-elements. + + If the caller did not specify anything, + the default is REQUIRE_ORDER if the environment variable + POSIXLY_CORRECT is defined, PERMUTE otherwise. + + REQUIRE_ORDER means don't recognize them as options; + stop option processing when the first non-option is seen. + This is what Unix does. + This mode of operation is selected by either setting the environment + variable POSIXLY_CORRECT, or using `+' as the first character + of the list of option characters. + + PERMUTE is the default. We permute the contents of ARGV as we scan, + so that eventually all the non-options are at the end. This allows options + to be given in any order, even with programs that were not written to + expect this. + + RETURN_IN_ORDER is an option available to programs that were written + to expect options and other ARGV-elements in any order and that care about + the ordering of the two. We describe each non-option ARGV-element + as if it were the argument of an option with character code 1. + Using `-' as the first character of the list of option characters + selects this mode of operation. + + The special argument `--' forces an end of option-scanning regardless + of the value of `ordering'. In the case of RETURN_IN_ORDER, only + `--' can cause `getopt' to return EOF with `optind' != ARGC. */ + +static enum +{ + REQUIRE_ORDER, PERMUTE, RETURN_IN_ORDER +} ordering; + +/* Value of POSIXLY_CORRECT environment variable. */ +static char *posixly_correct; + +#ifdef __GNU_LIBRARY__ +/* We want to avoid inclusion of string.h with non-GNU libraries + because there are many ways it can cause trouble. + On some systems, it contains special magic macros that don't work + in GCC. */ +#include <string.h> +#define my_index strchr +#else + +/* Avoid depending on library functions or files + whose names are inconsistent. */ + +char *getenv (); + +static char * +my_index (str, chr) + const char *str; + int chr; +{ + while (*str) + { + if (*str == chr) + return (char *) str; + str++; + } + return 0; +} + +/* If using GCC, we can safely declare strlen this way. + If not using GCC, it is ok not to declare it. */ +#ifdef __GNUC__ +/* Note that Motorola Delta 68k R3V7 comes with GCC but not stddef.h. + That was relevant to code that was here before. */ +#ifndef __STDC__ +/* gcc with -traditional declares the built-in strlen to return int, + and has done so at least since version 2.4.5. -- rms. */ +extern int strlen (const char *); +#endif /* not __STDC__ */ +#endif /* __GNUC__ */ + +#endif /* not __GNU_LIBRARY__ */ + +/* Handle permutation of arguments. */ + +/* Describe the part of ARGV that contains non-options that have + been skipped. `first_nonopt' is the index in ARGV of the first of them; + `last_nonopt' is the index after the last of them. */ + +static int first_nonopt; +static int last_nonopt; + +/* Exchange two adjacent subsequences of ARGV. + One subsequence is elements [first_nonopt,last_nonopt) + which contains all the non-options that have been skipped so far. + The other is elements [last_nonopt,optind), which contains all + the options processed since those non-options were skipped. + + `first_nonopt' and `last_nonopt' are relocated so that they describe + the new indices of the non-options in ARGV after they are moved. */ + +static void +exchange (argv) + char **argv; +{ + int bottom = first_nonopt; + int middle = last_nonopt; + int top = optind; + char *tem; + + /* Exchange the shorter segment with the far end of the longer segment. + That puts the shorter segment into the right place. + It leaves the longer segment in the right place overall, + but it consists of two parts that need to be swapped next. */ + + while (top > middle && middle > bottom) + { + if (top - middle > middle - bottom) + { + /* Bottom segment is the short one. */ + int len = middle - bottom; + register int i; + + /* Swap it with the top part of the top segment. */ + for (i = 0; i < len; i++) + { + tem = argv[bottom + i]; + argv[bottom + i] = argv[top - (middle - bottom) + i]; + argv[top - (middle - bottom) + i] = tem; + } + /* Exclude the moved bottom segment from further swapping. */ + top -= len; + } + else + { + /* Top segment is the short one. */ + int len = top - middle; + register int i; + + /* Swap it with the bottom part of the bottom segment. */ + for (i = 0; i < len; i++) + { + tem = argv[bottom + i]; + argv[bottom + i] = argv[middle + i]; + argv[middle + i] = tem; + } + /* Exclude the moved top segment from further swapping. */ + bottom += len; + } + } + + /* Update records for the slots the non-options now occupy. */ + + first_nonopt += (optind - last_nonopt); + last_nonopt = optind; +} + +/* Initialize the internal data when the first call is made. */ + +static const char * +_getopt_initialize (optstring) + const char *optstring; +{ + /* Start processing options with ARGV-element 1 (since ARGV-element 0 + is the program name); the sequence of previously skipped + non-option ARGV-elements is empty. */ + + first_nonopt = last_nonopt = optind = 1; + + nextchar = NULL; + + posixly_correct = getenv ("POSIXLY_CORRECT"); + + /* Determine how to handle the ordering of options and nonoptions. */ + + if (optstring[0] == '-') + { + ordering = RETURN_IN_ORDER; + ++optstring; + } + else if (optstring[0] == '+') + { + ordering = REQUIRE_ORDER; + ++optstring; + } + else if (posixly_correct != NULL) + ordering = REQUIRE_ORDER; + else + ordering = PERMUTE; + + return optstring; +} + +/* Scan elements of ARGV (whose length is ARGC) for option characters + given in OPTSTRING. + + If an element of ARGV starts with '-', and is not exactly "-" or "--", + then it is an option element. The characters of this element + (aside from the initial '-') are option characters. If `getopt' + is called repeatedly, it returns successively each of the option characters + from each of the option elements. + + If `getopt' finds another option character, it returns that character, + updating `optind' and `nextchar' so that the next call to `getopt' can + resume the scan with the following option character or ARGV-element. + + If there are no more option characters, `getopt' returns `EOF'. + Then `optind' is the index in ARGV of the first ARGV-element + that is not an option. (The ARGV-elements have been permuted + so that those that are not options now come last.) + + OPTSTRING is a string containing the legitimate option characters. + If an option character is seen that is not listed in OPTSTRING, + return '?' after printing an error message. If you set `opterr' to + zero, the error message is suppressed but we still return '?'. + + If a char in OPTSTRING is followed by a colon, that means it wants an arg, + so the following text in the same ARGV-element, or the text of the following + ARGV-element, is returned in `optarg'. Two colons mean an option that + wants an optional arg; if there is text in the current ARGV-element, + it is returned in `optarg', otherwise `optarg' is set to zero. + + If OPTSTRING starts with `-' or `+', it requests different methods of + handling the non-option ARGV-elements. + See the comments about RETURN_IN_ORDER and REQUIRE_ORDER, above. + + Long-named options begin with `--' instead of `-'. + Their names may be abbreviated as long as the abbreviation is unique + or is an exact match for some defined option. If they have an + argument, it follows the option name in the same ARGV-element, separated + from the option name by a `=', or else the in next ARGV-element. + When `getopt' finds a long-named option, it returns 0 if that option's + `flag' field is nonzero, the value of the option's `val' field + if the `flag' field is zero. + + The elements of ARGV aren't really const, because we permute them. + But we pretend they're const in the prototype to be compatible + with other systems. + + LONGOPTS is a vector of `struct option' terminated by an + element containing a name which is zero. + + LONGIND returns the index in LONGOPT of the long-named option found. + It is only valid when a long-named option has been found by the most + recent call. + + If LONG_ONLY is nonzero, '-' as well as '--' can introduce + long-named options. */ + +int +_getopt_internal (argc, argv, optstring, longopts, longind, long_only) + int argc; + char *const *argv; + const char *optstring; + const struct option *longopts; + int *longind; + int long_only; +{ + optarg = NULL; + + if (optind == 0) + optstring = _getopt_initialize (optstring); + + if (nextchar == NULL || *nextchar == '\0') + { + /* Advance to the next ARGV-element. */ + + if (ordering == PERMUTE) + { + /* If we have just processed some options following some non-options, + exchange them so that the options come first. */ + + if (first_nonopt != last_nonopt && last_nonopt != optind) + exchange ((char **) argv); + else if (last_nonopt != optind) + first_nonopt = optind; + + /* Skip any additional non-options + and extend the range of non-options previously skipped. */ + + while (optind < argc + && (argv[optind][0] != '-' || argv[optind][1] == '\0')) + optind++; + last_nonopt = optind; + } + + /* The special ARGV-element `--' means premature end of options. + Skip it like a null option, + then exchange with previous non-options as if it were an option, + then skip everything else like a non-option. */ + + if (optind != argc && !strcmp (argv[optind], "--")) + { + optind++; + + if (first_nonopt != last_nonopt && last_nonopt != optind) + exchange ((char **) argv); + else if (first_nonopt == last_nonopt) + first_nonopt = optind; + last_nonopt = argc; + + optind = argc; + } + + /* If we have done all the ARGV-elements, stop the scan + and back over any non-options that we skipped and permuted. */ + + if (optind == argc) + { + /* Set the next-arg-index to point at the non-options + that we previously skipped, so the caller will digest them. */ + if (first_nonopt != last_nonopt) + optind = first_nonopt; + return EOF; + } + + /* If we have come to a non-option and did not permute it, + either stop the scan or describe it to the caller and pass it by. */ + + if ((argv[optind][0] != '-' || argv[optind][1] == '\0')) + { + if (ordering == REQUIRE_ORDER) + return EOF; + optarg = argv[optind++]; + return 1; + } + + /* We have found another option-ARGV-element. + Skip the initial punctuation. */ + + nextchar = (argv[optind] + 1 + + (longopts != NULL && argv[optind][1] == '-')); + } + + /* Decode the current option-ARGV-element. */ + + /* Check whether the ARGV-element is a long option. + + If long_only and the ARGV-element has the form "-f", where f is + a valid short option, don't consider it an abbreviated form of + a long option that starts with f. Otherwise there would be no + way to give the -f short option. + + On the other hand, if there's a long option "fubar" and + the ARGV-element is "-fu", do consider that an abbreviation of + the long option, just like "--fu", and not "-f" with arg "u". + + This distinction seems to be the most useful approach. */ + + if (longopts != NULL + && (argv[optind][1] == '-' + || (long_only && (argv[optind][2] || !my_index (optstring, argv[optind][1]))))) + { + char *nameend; + const struct option *p; + const struct option *pfound = NULL; + int exact = 0; + int ambig = 0; + int indfound; + int option_index; + + for (nameend = nextchar; *nameend && *nameend != '='; nameend++) + /* Do nothing. */ ; + + /* Test all long options for either exact match + or abbreviated matches. */ + for (p = longopts, option_index = 0; p->name; p++, option_index++) + if (!strncmp (p->name, nextchar, nameend - nextchar)) + { + if (nameend - nextchar == strlen (p->name)) + { + /* Exact match found. */ + pfound = p; + indfound = option_index; + exact = 1; + break; + } + else if (pfound == NULL) + { + /* First nonexact match found. */ + pfound = p; + indfound = option_index; + } + else + /* Second or later nonexact match found. */ + ambig = 1; + } + + if (ambig && !exact) + { + if (opterr) + fprintf (stderr, "%s: option `%s' is ambiguous\n", + argv[0], argv[optind]); + nextchar += strlen (nextchar); + optind++; + return '?'; + } + + if (pfound != NULL) + { + option_index = indfound; + optind++; + if (*nameend) + { + /* Don't test has_arg with >, because some C compilers don't + allow it to be used on enums. */ + if (pfound->has_arg) + optarg = nameend + 1; + else + { + if (opterr) + { + if (argv[optind - 1][1] == '-') + /* --option */ + fprintf (stderr, + "%s: option `--%s' doesn't allow an argument\n", + argv[0], pfound->name); + else + /* +option or -option */ + fprintf (stderr, + "%s: option `%c%s' doesn't allow an argument\n", + argv[0], argv[optind - 1][0], pfound->name); + } + nextchar += strlen (nextchar); + return '?'; + } + } + else if (pfound->has_arg == 1) + { + if (optind < argc) + optarg = argv[optind++]; + else + { + if (opterr) + fprintf (stderr, "%s: option `%s' requires an argument\n", + argv[0], argv[optind - 1]); + nextchar += strlen (nextchar); + return optstring[0] == ':' ? ':' : '?'; + } + } + nextchar += strlen (nextchar); + if (longind != NULL) + *longind = option_index; + if (pfound->flag) + { + *(pfound->flag) = pfound->val; + return 0; + } + return pfound->val; + } + + /* Can't find it as a long option. If this is not getopt_long_only, + or the option starts with '--' or is not a valid short + option, then it's an error. + Otherwise interpret it as a short option. */ + if (!long_only || argv[optind][1] == '-' + || my_index (optstring, *nextchar) == NULL) + { + if (opterr) + { + if (argv[optind][1] == '-') + /* --option */ + fprintf (stderr, "%s: unrecognized option `--%s'\n", + argv[0], nextchar); + else + /* +option or -option */ + fprintf (stderr, "%s: unrecognized option `%c%s'\n", + argv[0], argv[optind][0], nextchar); + } + nextchar = (char *) ""; + optind++; + return '?'; + } + } + + /* Look at and handle the next short option-character. */ + + { + char c = *nextchar++; + char *temp = my_index (optstring, c); + + /* Increment `optind' when we start to process its last character. */ + if (*nextchar == '\0') + ++optind; + + if (temp == NULL || c == ':') + { + if (opterr) + { + if (posixly_correct) + /* 1003.2 specifies the format of this message. */ + fprintf (stderr, "%s: illegal option -- %c\n", argv[0], c); + else + fprintf (stderr, "%s: invalid option -- %c\n", argv[0], c); + } + optopt = c; + return '?'; + } + if (temp[1] == ':') + { + if (temp[2] == ':') + { + /* This is an option that accepts an argument optionally. */ + if (*nextchar != '\0') + { + optarg = nextchar; + optind++; + } + else + optarg = NULL; + nextchar = NULL; + } + else + { + /* This is an option that requires an argument. */ + if (*nextchar != '\0') + { + optarg = nextchar; + /* If we end this ARGV-element by taking the rest as an arg, + we must advance to the next element now. */ + optind++; + } + else if (optind == argc) + { + if (opterr) + { + /* 1003.2 specifies the format of this message. */ + fprintf (stderr, "%s: option requires an argument -- %c\n", + argv[0], c); + } + optopt = c; + if (optstring[0] == ':') + c = ':'; + else + c = '?'; + } + else + /* We already incremented `optind' once; + increment it again when taking next ARGV-elt as argument. */ + optarg = argv[optind++]; + nextchar = NULL; + } + } + return c; + } +} + +int +getopt (argc, argv, optstring) + int argc; + char *const *argv; + const char *optstring; +{ + return _getopt_internal (argc, argv, optstring, + (const struct option *) 0, + (int *) 0, + 0); +} + +#endif /* _LIBC or not __GNU_LIBRARY__. */ + +#ifdef TEST + +/* Compile with -DTEST to make an executable for use in testing + the above definition of `getopt'. */ + +int +main (argc, argv) + int argc; + char **argv; +{ + int c; + int digit_optind = 0; + + while (1) + { + int this_option_optind = optind ? optind : 1; + + c = getopt (argc, argv, "abc:d:0123456789"); + if (c == EOF) + break; + + switch (c) + { + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (digit_optind != 0 && digit_optind != this_option_optind) + printf ("digits occur in two different argv-elements.\n"); + digit_optind = this_option_optind; + printf ("option %c\n", c); + break; + + case 'a': + printf ("option a\n"); + break; + + case 'b': + printf ("option b\n"); + break; + + case 'c': + printf ("option c with value `%s'\n", optarg); + break; + + case '?': + break; + + default: + printf ("?? getopt returned character code 0%o ??\n", c); + } + } + + if (optind < argc) + { + printf ("non-option ARGV-elements: "); + while (optind < argc) + printf ("%s ", argv[optind++]); + printf ("\n"); + } + + exit (0); +} + +#endif /* TEST */ diff --git a/sys/src/ape/cmd/diff/getopt.h b/sys/src/ape/cmd/diff/getopt.h new file mode 100755 index 000000000..45541f5ac --- /dev/null +++ b/sys/src/ape/cmd/diff/getopt.h @@ -0,0 +1,129 @@ +/* Declarations for getopt. + Copyright (C) 1989, 1990, 1991, 1992, 1993 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#ifndef _GETOPT_H +#define _GETOPT_H 1 + +#ifdef __cplusplus +extern "C" { +#endif + +/* For communication from `getopt' to the caller. + When `getopt' finds an option that takes an argument, + the argument value is returned here. + Also, when `ordering' is RETURN_IN_ORDER, + each non-option ARGV-element is returned here. */ + +extern char *optarg; + +/* Index in ARGV of the next element to be scanned. + This is used for communication to and from the caller + and for communication between successive calls to `getopt'. + + On entry to `getopt', zero means this is the first call; initialize. + + When `getopt' returns EOF, this is the index of the first of the + non-option elements that the caller should itself scan. + + Otherwise, `optind' communicates from one call to the next + how much of ARGV has been scanned so far. */ + +extern int optind; + +/* Callers store zero here to inhibit the error message `getopt' prints + for unrecognized options. */ + +extern int opterr; + +/* Set to an option character which was unrecognized. */ + +extern int optopt; + +/* Describe the long-named options requested by the application. + The LONG_OPTIONS argument to getopt_long or getopt_long_only is a vector + of `struct option' terminated by an element containing a name which is + zero. + + The field `has_arg' is: + no_argument (or 0) if the option does not take an argument, + required_argument (or 1) if the option requires an argument, + optional_argument (or 2) if the option takes an optional argument. + + If the field `flag' is not NULL, it points to a variable that is set + to the value given in the field `val' when the option is found, but + left unchanged if the option is not found. + + To have a long-named option do something other than set an `int' to + a compiled-in constant, such as set a value from `optarg', set the + option's `flag' field to zero and its `val' field to a nonzero + value (the equivalent single-letter option character, if there is + one). For long options that have a zero `flag' field, `getopt' + returns the contents of the `val' field. */ + +struct option +{ +#if __STDC__ + const char *name; +#else + char *name; +#endif + /* has_arg can't be an enum because some compilers complain about + type mismatches in all the code that assumes it is an int. */ + int has_arg; + int *flag; + int val; +}; + +/* Names for the values of the `has_arg' field of `struct option'. */ + +#define no_argument 0 +#define required_argument 1 +#define optional_argument 2 + +#if __STDC__ +#if defined(__GNU_LIBRARY__) +/* Many other libraries have conflicting prototypes for getopt, with + differences in the consts, in stdlib.h. To avoid compilation + errors, only prototype getopt for the GNU C library. */ +extern int getopt (int argc, char *const *argv, const char *shortopts); +#else /* not __GNU_LIBRARY__ */ +extern int getopt (); +#endif /* not __GNU_LIBRARY__ */ +extern int getopt_long (int argc, char *const *argv, const char *shortopts, + const struct option *longopts, int *longind); +extern int getopt_long_only (int argc, char *const *argv, + const char *shortopts, + const struct option *longopts, int *longind); + +/* Internal only. Users should not call this directly. */ +extern int _getopt_internal (int argc, char *const *argv, + const char *shortopts, + const struct option *longopts, int *longind, + int long_only); +#else /* not __STDC__ */ +extern int getopt (); +extern int getopt_long (); +extern int getopt_long_only (); + +extern int _getopt_internal (); +#endif /* not __STDC__ */ + +#ifdef __cplusplus +} +#endif + +#endif /* _GETOPT_H */ diff --git a/sys/src/ape/cmd/diff/getopt1.c b/sys/src/ape/cmd/diff/getopt1.c new file mode 100755 index 000000000..725c653bb --- /dev/null +++ b/sys/src/ape/cmd/diff/getopt1.c @@ -0,0 +1,180 @@ +/* getopt_long and getopt_long_only entry points for GNU getopt. + Copyright (C) 1987, 88, 89, 90, 91, 92, 1993 + Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#include "getopt.h" + +#ifndef __STDC__ +/* This is a separate conditional since some stdc systems + reject `defined (const)'. */ +#ifndef const +#define const +#endif +#endif + +#include <stdio.h> + +/* Comment out all this code if we are using the GNU C Library, and are not + actually compiling the library itself. This code is part of the GNU C + Library, but also included in many other GNU distributions. Compiling + and linking in this code is a waste when using the GNU C library + (especially if it is a shared library). Rather than having every GNU + program understand `configure --with-gnu-libc' and omit the object files, + it is simpler to just do this in the source for each such file. */ + +#if defined (_LIBC) || !defined (__GNU_LIBRARY__) + + +/* This needs to come after some library #include + to get __GNU_LIBRARY__ defined. */ +#ifdef __GNU_LIBRARY__ +#include <stdlib.h> +#else +char *getenv (); +#endif + +#ifndef NULL +#define NULL 0 +#endif + +int +getopt_long (argc, argv, options, long_options, opt_index) + int argc; + char *const *argv; + const char *options; + const struct option *long_options; + int *opt_index; +{ + return _getopt_internal (argc, argv, options, long_options, opt_index, 0); +} + +/* Like getopt_long, but '-' as well as '--' can indicate a long option. + If an option that starts with '-' (not '--') doesn't match a long option, + but does match a short option, it is parsed as a short option + instead. */ + +int +getopt_long_only (argc, argv, options, long_options, opt_index) + int argc; + char *const *argv; + const char *options; + const struct option *long_options; + int *opt_index; +{ + return _getopt_internal (argc, argv, options, long_options, opt_index, 1); +} + + +#endif /* _LIBC or not __GNU_LIBRARY__. */ + +#ifdef TEST + +#include <stdio.h> + +int +main (argc, argv) + int argc; + char **argv; +{ + int c; + int digit_optind = 0; + + while (1) + { + int this_option_optind = optind ? optind : 1; + int option_index = 0; + static struct option long_options[] = + { + {"add", 1, 0, 0}, + {"append", 0, 0, 0}, + {"delete", 1, 0, 0}, + {"verbose", 0, 0, 0}, + {"create", 0, 0, 0}, + {"file", 1, 0, 0}, + {0, 0, 0, 0} + }; + + c = getopt_long (argc, argv, "abc:d:0123456789", + long_options, &option_index); + if (c == EOF) + break; + + switch (c) + { + case 0: + printf ("option %s", long_options[option_index].name); + if (optarg) + printf (" with arg %s", optarg); + printf ("\n"); + break; + + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (digit_optind != 0 && digit_optind != this_option_optind) + printf ("digits occur in two different argv-elements.\n"); + digit_optind = this_option_optind; + printf ("option %c\n", c); + break; + + case 'a': + printf ("option a\n"); + break; + + case 'b': + printf ("option b\n"); + break; + + case 'c': + printf ("option c with value `%s'\n", optarg); + break; + + case 'd': + printf ("option d with value `%s'\n", optarg); + break; + + case '?': + break; + + default: + printf ("?? getopt returned character code 0%o ??\n", c); + } + } + + if (optind < argc) + { + printf ("non-option ARGV-elements: "); + while (optind < argc) + printf ("%s ", argv[optind++]); + printf ("\n"); + } + + exit (0); +} + +#endif /* TEST */ diff --git a/sys/src/ape/cmd/diff/ifdef.c b/sys/src/ape/cmd/diff/ifdef.c new file mode 100755 index 000000000..2834cbdfa --- /dev/null +++ b/sys/src/ape/cmd/diff/ifdef.c @@ -0,0 +1,428 @@ +/* #ifdef-format output routines for GNU DIFF. + Copyright (C) 1989, 1991, 1992, 1993, 1994 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY. No author or distributor +accepts responsibility to anyone for the consequences of using it +or for whether it serves any particular purpose or works at all, +unless he says so in writing. Refer to the GNU DIFF General Public +License for full details. + +Everyone is granted permission to copy, modify and redistribute +GNU DIFF, but only under the conditions described in the +GNU DIFF General Public License. A copy of this license is +supposed to have been given to you along with GNU DIFF so you +can know your rights and responsibilities. It should be in a +file named COPYING. Among other things, the copyright notice +and this notice must be preserved on all copies. */ + + +#include "diff.h" + +struct group +{ + struct file_data const *file; + int from, upto; /* start and limit lines for this group of lines */ +}; + +static char *format_group PARAMS((FILE *, char *, int, struct group const *)); +static char *scan_char_literal PARAMS((char *, int *)); +static char *scan_printf_spec PARAMS((char *)); +static int groups_letter_value PARAMS((struct group const *, int)); +static void format_ifdef PARAMS((char *, int, int, int, int)); +static void print_ifdef_hunk PARAMS((struct change *)); +static void print_ifdef_lines PARAMS((FILE *, char *, struct group const *)); + +static int next_line; + +/* Print the edit-script SCRIPT as a merged #ifdef file. */ + +void +print_ifdef_script (script) + struct change *script; +{ + next_line = - files[0].prefix_lines; + print_script (script, find_change, print_ifdef_hunk); + if (next_line < files[0].valid_lines) + { + begin_output (); + format_ifdef (group_format[UNCHANGED], next_line, files[0].valid_lines, + next_line - files[0].valid_lines + files[1].valid_lines, + files[1].valid_lines); + } +} + +/* Print a hunk of an ifdef diff. + This is a contiguous portion of a complete edit script, + describing changes in consecutive lines. */ + +static void +print_ifdef_hunk (hunk) + struct change *hunk; +{ + int first0, last0, first1, last1, deletes, inserts; + char *format; + + /* Determine range of line numbers involved in each file. */ + analyze_hunk (hunk, &first0, &last0, &first1, &last1, &deletes, &inserts); + if (inserts) + format = deletes ? group_format[CHANGED] : group_format[NEW]; + else if (deletes) + format = group_format[OLD]; + else + return; + + begin_output (); + + /* Print lines up to this change. */ + if (next_line < first0) + format_ifdef (group_format[UNCHANGED], next_line, first0, + next_line - first0 + first1, first1); + + /* Print this change. */ + next_line = last0 + 1; + format_ifdef (format, first0, next_line, first1, last1 + 1); +} + +/* Print a set of lines according to FORMAT. + Lines BEG0 up to END0 are from the first file; + lines BEG1 up to END1 are from the second file. */ + +static void +format_ifdef (format, beg0, end0, beg1, end1) + char *format; + int beg0, end0, beg1, end1; +{ + struct group groups[2]; + + groups[0].file = &files[0]; + groups[0].from = beg0; + groups[0].upto = end0; + groups[1].file = &files[1]; + groups[1].from = beg1; + groups[1].upto = end1; + format_group (outfile, format, '\0', groups); +} + +/* Print to file OUT a set of lines according to FORMAT. + The format ends at the first free instance of ENDCHAR. + Yield the address of the terminating character. + GROUPS specifies which lines to print. + If OUT is zero, do not actually print anything; just scan the format. */ + +static char * +format_group (out, format, endchar, groups) + register FILE *out; + char *format; + int endchar; + struct group const *groups; +{ + register char c; + register char *f = format; + + while ((c = *f) != endchar && c != 0) + { + f++; + if (c == '%') + { + char *spec = f; + switch ((c = *f++)) + { + case '%': + break; + + case '(': + /* Print if-then-else format e.g. `%(n=1?thenpart:elsepart)'. */ + { + int i, value[2]; + FILE *thenout, *elseout; + + for (i = 0; i < 2; i++) + { + unsigned char f0 = f[0]; + if (ISDIGIT (f0)) + { + value[i] = atoi (f); + while (ISDIGIT ((unsigned char) *++f)) + continue; + } + else + { + value[i] = groups_letter_value (groups, f0); + if (value[i] < 0) + goto bad_format; + f++; + } + if (*f++ != "=?"[i]) + goto bad_format; + } + if (value[0] == value[1]) + thenout = out, elseout = 0; + else + thenout = 0, elseout = out; + f = format_group (thenout, f, ':', groups); + if (*f) + { + f = format_group (elseout, f + 1, ')', groups); + if (*f) + f++; + } + } + continue; + + case '<': + /* Print lines deleted from first file. */ + print_ifdef_lines (out, line_format[OLD], &groups[0]); + continue; + + case '=': + /* Print common lines. */ + print_ifdef_lines (out, line_format[UNCHANGED], &groups[0]); + continue; + + case '>': + /* Print lines inserted from second file. */ + print_ifdef_lines (out, line_format[NEW], &groups[1]); + continue; + + default: + { + int value; + char *speclim; + + f = scan_printf_spec (spec); + if (!f) + goto bad_format; + speclim = f; + c = *f++; + switch (c) + { + case '\'': + f = scan_char_literal (f, &value); + if (!f) + goto bad_format; + break; + + default: + value = groups_letter_value (groups, c); + if (value < 0) + goto bad_format; + break; + } + if (out) + { + /* Temporarily replace e.g. "%3dnx" with "%3d\0x". */ + *speclim = 0; + fprintf (out, spec - 1, value); + /* Undo the temporary replacement. */ + *speclim = c; + } + } + continue; + + bad_format: + c = '%'; + f = spec; + break; + } + } + if (out) + putc (c, out); + } + return f; +} + +/* For the line group pair G, return the number corresponding to LETTER. + Return -1 if LETTER is not a group format letter. */ +static int +groups_letter_value (g, letter) + struct group const *g; + int letter; +{ + if (ISUPPER (letter)) + { + g++; + letter = tolower (letter); + } + switch (letter) + { + case 'e': return translate_line_number (g->file, g->from) - 1; + case 'f': return translate_line_number (g->file, g->from); + case 'l': return translate_line_number (g->file, g->upto) - 1; + case 'm': return translate_line_number (g->file, g->upto); + case 'n': return g->upto - g->from; + default: return -1; + } +} + +/* Print to file OUT, using FORMAT to print the line group GROUP. + But do nothing if OUT is zero. */ +static void +print_ifdef_lines (out, format, group) + register FILE *out; + char *format; + struct group const *group; +{ + struct file_data const *file = group->file; + char const * const *linbuf = file->linbuf; + int from = group->from, upto = group->upto; + + if (!out) + return; + + /* If possible, use a single fwrite; it's faster. */ + if (!tab_expand_flag && format[0] == '%') + { + if (format[1] == 'l' && format[2] == '\n' && !format[3]) + { + fwrite (linbuf[from], sizeof (char), + linbuf[upto] + (linbuf[upto][-1] != '\n') - linbuf[from], + out); + return; + } + if (format[1] == 'L' && !format[2]) + { + fwrite (linbuf[from], sizeof (char), + linbuf[upto] - linbuf[from], out); + return; + } + } + + for (; from < upto; from++) + { + register char c; + register char *f = format; + + while ((c = *f++) != 0) + { + if (c == '%') + { + char *spec = f; + switch ((c = *f++)) + { + case '%': + break; + + case 'l': + output_1_line (linbuf[from], + linbuf[from + 1] + - (linbuf[from + 1][-1] == '\n'), 0, 0); + continue; + + case 'L': + output_1_line (linbuf[from], linbuf[from + 1], 0, 0); + continue; + + default: + { + int value; + char *speclim; + + f = scan_printf_spec (spec); + if (!f) + goto bad_format; + speclim = f; + c = *f++; + switch (c) + { + case '\'': + f = scan_char_literal (f, &value); + if (!f) + goto bad_format; + break; + + case 'n': + value = translate_line_number (file, from); + break; + + default: + goto bad_format; + } + /* Temporarily replace e.g. "%3dnx" with "%3d\0x". */ + *speclim = 0; + fprintf (out, spec - 1, value); + /* Undo the temporary replacement. */ + *speclim = c; + } + continue; + + bad_format: + c = '%'; + f = spec; + break; + } + } + putc (c, out); + } + } +} + +/* Scan the character literal represented in the string LIT; LIT points just + after the initial apostrophe. Put the literal's value into *INTPTR. + Yield the address of the first character after the closing apostrophe, + or zero if the literal is ill-formed. */ +static char * +scan_char_literal (lit, intptr) + char *lit; + int *intptr; +{ + register char *p = lit; + int value, digits; + char c = *p++; + + switch (c) + { + case 0: + case '\'': + return 0; + + case '\\': + value = 0; + while ((c = *p++) != '\'') + { + unsigned digit = c - '0'; + if (8 <= digit) + return 0; + value = 8 * value + digit; + } + digits = p - lit - 2; + if (! (1 <= digits && digits <= 3)) + return 0; + break; + + default: + value = c; + if (*p++ != '\'') + return 0; + break; + } + *intptr = value; + return p; +} + +/* Scan optional printf-style SPEC of the form `-*[0-9]*(.[0-9]*)?[cdoxX]'. + Return the address of the character following SPEC, or zero if failure. */ +static char * +scan_printf_spec (spec) + register char *spec; +{ + register unsigned char c; + + while ((c = *spec++) == '-') + continue; + while (ISDIGIT (c)) + c = *spec++; + if (c == '.') + while (ISDIGIT (c = *spec++)) + continue; + switch (c) + { + case 'c': case 'd': case 'o': case 'x': case 'X': + return spec; + + default: + return 0; + } +} diff --git a/sys/src/ape/cmd/diff/install-sh b/sys/src/ape/cmd/diff/install-sh new file mode 100755 index 000000000..ab74c882e --- /dev/null +++ b/sys/src/ape/cmd/diff/install-sh @@ -0,0 +1,238 @@ +#!/bin/sh +# +# install - install a program, script, or datafile +# This comes from X11R5. +# +# Calling this script install-sh is preferred over install.sh, to prevent +# `make' implicit rules from creating a file called install from it +# when there is no Makefile. +# +# This script is compatible with the BSD install script, but was written +# from scratch. +# + + +# set DOITPROG to echo to test this script + +# Don't use :- since 4.3BSD and earlier shells don't like it. +doit="${DOITPROG-}" + + +# put in absolute paths if you don't have them in your path; or use env. vars. + +mvprog="${MVPROG-mv}" +cpprog="${CPPROG-cp}" +chmodprog="${CHMODPROG-chmod}" +chownprog="${CHOWNPROG-chown}" +chgrpprog="${CHGRPPROG-chgrp}" +stripprog="${STRIPPROG-strip}" +rmprog="${RMPROG-rm}" +mkdirprog="${MKDIRPROG-mkdir}" + +tranformbasename="" +transform_arg="" +instcmd="$mvprog" +chmodcmd="$chmodprog 0755" +chowncmd="" +chgrpcmd="" +stripcmd="" +rmcmd="$rmprog -f" +mvcmd="$mvprog" +src="" +dst="" +dir_arg="" + +while [ x"$1" != x ]; do + case $1 in + -c) instcmd="$cpprog" + shift + continue;; + + -d) dir_arg=true + shift + continue;; + + -m) chmodcmd="$chmodprog $2" + shift + shift + continue;; + + -o) chowncmd="$chownprog $2" + shift + shift + continue;; + + -g) chgrpcmd="$chgrpprog $2" + shift + shift + continue;; + + -s) stripcmd="$stripprog" + shift + continue;; + + -t=*) transformarg=`echo $1 | sed 's/-t=//'` + shift + continue;; + + -b=*) transformbasename=`echo $1 | sed 's/-b=//'` + shift + continue;; + + *) if [ x"$src" = x ] + then + src=$1 + else + # this colon is to work around a 386BSD /bin/sh bug + : + dst=$1 + fi + shift + continue;; + esac +done + +if [ x"$src" = x ] +then + echo "install: no input file specified" + exit 1 +else + true +fi + +if [ x"$dir_arg" != x ]; then + dst=$src + src="" + + if [ -d $dst ]; then + instcmd=: + else + instcmd=mkdir + fi +else + +# Waiting for this to be detected by the "$instcmd $src $dsttmp" command +# might cause directories to be created, which would be especially bad +# if $src (and thus $dsttmp) contains '*'. + + if [ -f $src -o -d $src ] + then + true + else + echo "install: $src does not exist" + exit 1 + fi + + if [ x"$dst" = x ] + then + echo "install: no destination specified" + exit 1 + else + true + fi + +# If destination is a directory, append the input filename; if your system +# does not like double slashes in filenames, you may need to add some logic + + if [ -d $dst ] + then + dst="$dst"/`basename $src` + else + true + fi +fi + +## this sed command emulates the dirname command +dstdir=`echo $dst | sed -e 's,[^/]*$,,;s,/$,,;s,^$,.,'` + +# Make sure that the destination directory exists. +# this part is taken from Noah Friedman's mkinstalldirs script + +# Skip lots of stat calls in the usual case. +if [ ! -d "$dstdir" ]; then +defaultIFS=' +' +IFS="${IFS-${defaultIFS}}" + +oIFS="${IFS}" +# Some sh's can't handle IFS=/ for some reason. +IFS='%' +set - `echo ${dstdir} | sed -e 's@/@%@g' -e 's@^%@/@'` +IFS="${oIFS}" + +pathcomp='' + +while [ $# -ne 0 ] ; do + pathcomp="${pathcomp}${1}" + shift + + if [ ! -d "${pathcomp}" ] ; + then + $mkdirprog "${pathcomp}" + else + true + fi + + pathcomp="${pathcomp}/" +done +fi + +if [ x"$dir_arg" != x ] +then + $doit $instcmd $dst && + + if [ x"$chowncmd" != x ]; then $doit $chowncmd $dst; else true ; fi && + if [ x"$chgrpcmd" != x ]; then $doit $chgrpcmd $dst; else true ; fi && + if [ x"$stripcmd" != x ]; then $doit $stripcmd $dst; else true ; fi && + if [ x"$chmodcmd" != x ]; then $doit $chmodcmd $dst; else true ; fi +else + +# If we're going to rename the final executable, determine the name now. + + if [ x"$transformarg" = x ] + then + dstfile=`basename $dst` + else + dstfile=`basename $dst $transformbasename | + sed $transformarg`$transformbasename + fi + +# don't allow the sed command to completely eliminate the filename + + if [ x"$dstfile" = x ] + then + dstfile=`basename $dst` + else + true + fi + +# Make a temp file name in the proper directory. + + dsttmp=$dstdir/#inst.$$# + +# Move or copy the file name to the temp name + + $doit $instcmd $src $dsttmp && + + trap "rm -f ${dsttmp}" 0 && + +# and set any options; do chmod last to preserve setuid bits + +# If any of these fail, we abort the whole thing. If we want to +# ignore errors from any of these, just make sure not to ignore +# errors from the above "$doit $instcmd $src $dsttmp" command. + + if [ x"$chowncmd" != x ]; then $doit $chowncmd $dsttmp; else true;fi && + if [ x"$chgrpcmd" != x ]; then $doit $chgrpcmd $dsttmp; else true;fi && + if [ x"$stripcmd" != x ]; then $doit $stripcmd $dsttmp; else true;fi && + if [ x"$chmodcmd" != x ]; then $doit $chmodcmd $dsttmp; else true;fi && + +# Now rename the file to the real destination. + + $doit $rmcmd -f $dstdir/$dstfile && + $doit $mvcmd $dsttmp $dstdir/$dstfile + +fi && + + +exit 0 diff --git a/sys/src/ape/cmd/diff/io.c b/sys/src/ape/cmd/diff/io.c new file mode 100755 index 000000000..660591551 --- /dev/null +++ b/sys/src/ape/cmd/diff/io.c @@ -0,0 +1,714 @@ +/* File I/O for GNU DIFF. + Copyright (C) 1988, 1989, 1992, 1993, 1994 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#include "diff.h" + +/* Rotate a value n bits to the left. */ +#define UINT_BIT (sizeof (unsigned) * CHAR_BIT) +#define ROL(v, n) ((v) << (n) | (v) >> (UINT_BIT - (n))) + +/* Given a hash value and a new character, return a new hash value. */ +#define HASH(h, c) ((c) + ROL (h, 7)) + +/* Guess remaining number of lines from number N of lines so far, + size S so far, and total size T. */ +#define GUESS_LINES(n,s,t) (((t) - (s)) / ((n) < 10 ? 32 : (s) / ((n)-1)) + 5) + +/* Type used for fast prefix comparison in find_identical_ends. */ +#ifndef word +#define word int +#endif + +/* Lines are put into equivalence classes (of lines that match in line_cmp). + Each equivalence class is represented by one of these structures, + but only while the classes are being computed. + Afterward, each class is represented by a number. */ +struct equivclass +{ + int next; /* Next item in this bucket. */ + unsigned hash; /* Hash of lines in this class. */ + char const *line; /* A line that fits this class. */ + size_t length; /* That line's length, not counting its newline. */ +}; + +/* Hash-table: array of buckets, each being a chain of equivalence classes. + buckets[-1] is reserved for incomplete lines. */ +static int *buckets; + +/* Number of buckets in the hash table array, not counting buckets[-1]. */ +static int nbuckets; + +/* Array in which the equivalence classes are allocated. + The bucket-chains go through the elements in this array. + The number of an equivalence class is its index in this array. */ +static struct equivclass *equivs; + +/* Index of first free element in the array `equivs'. */ +static int equivs_index; + +/* Number of elements allocated in the array `equivs'. */ +static int equivs_alloc; + +static void find_and_hash_each_line PARAMS((struct file_data *)); +static void find_identical_ends PARAMS((struct file_data[])); +static void prepare_text_end PARAMS((struct file_data *)); + +/* Check for binary files and compare them for exact identity. */ + +/* Return 1 if BUF contains a non text character. + SIZE is the number of characters in BUF. */ + +#define binary_file_p(buf, size) (memchr (buf, '\0', size) != 0) + +/* Get ready to read the current file. + Return nonzero if SKIP_TEST is zero, + and if it appears to be a binary file. */ + +int +sip (current, skip_test) + struct file_data *current; + int skip_test; +{ + /* If we have a nonexistent file at this stage, treat it as empty. */ + if (current->desc < 0) + { + /* Leave room for a sentinel. */ + current->bufsize = sizeof (word); + current->buffer = xmalloc (current->bufsize); + } + else + { + current->bufsize = STAT_BLOCKSIZE (current->stat); + current->buffer = xmalloc (current->bufsize); + + if (! skip_test) + { + /* Check first part of file to see if it's a binary file. */ +#if HAVE_SETMODE + int oldmode = setmode (current->desc, O_BINARY); +#endif + size_t n = read (current->desc, current->buffer, current->bufsize); + if (n == -1) + pfatal_with_name (current->name); + current->buffered_chars = n; +#if HAVE_SETMODE + if (oldmode != O_BINARY) + { + if (lseek (current->desc, - (off_t) n, SEEK_CUR) == -1) + pfatal_with_name (current->name); + setmode (current->desc, oldmode); + current->buffered_chars = 0; + } +#endif + return binary_file_p (current->buffer, n); + } + } + + current->buffered_chars = 0; + return 0; +} + +/* Slurp the rest of the current file completely into memory. */ + +void +slurp (current) + struct file_data *current; +{ + size_t cc; + + if (current->desc < 0) + /* The file is nonexistent. */ + ; + else if (S_ISREG (current->stat.st_mode)) + { + /* It's a regular file; slurp in the rest all at once. */ + + /* Get the size out of the stat block. + Allocate enough room for appended newline and sentinel. */ + cc = current->stat.st_size + 1 + sizeof (word); + if (current->bufsize < cc) + { + current->bufsize = cc; + current->buffer = xrealloc (current->buffer, cc); + } + + if (current->buffered_chars < current->stat.st_size) + { + cc = read (current->desc, + current->buffer + current->buffered_chars, + current->stat.st_size - current->buffered_chars); + if (cc == -1) + pfatal_with_name (current->name); + current->buffered_chars += cc; + } + } + /* It's not a regular file; read it, growing the buffer as needed. */ + else if (always_text_flag || current->buffered_chars != 0) + { + for (;;) + { + if (current->buffered_chars == current->bufsize) + { + current->bufsize = current->bufsize * 2; + current->buffer = xrealloc (current->buffer, current->bufsize); + } + cc = read (current->desc, + current->buffer + current->buffered_chars, + current->bufsize - current->buffered_chars); + if (cc == 0) + break; + if (cc == -1) + pfatal_with_name (current->name); + current->buffered_chars += cc; + } + /* Allocate just enough room for appended newline and sentinel. */ + current->bufsize = current->buffered_chars + 1 + sizeof (word); + current->buffer = xrealloc (current->buffer, current->bufsize); + } +} + +/* Split the file into lines, simultaneously computing the equivalence class for + each line. */ + +static void +find_and_hash_each_line (current) + struct file_data *current; +{ + unsigned h; + unsigned char const *p = (unsigned char const *) current->prefix_end; + unsigned char c; + int i, *bucket; + size_t length; + + /* Cache often-used quantities in local variables to help the compiler. */ + char const **linbuf = current->linbuf; + int alloc_lines = current->alloc_lines; + int line = 0; + int linbuf_base = current->linbuf_base; + int *cureqs = (int *) xmalloc (alloc_lines * sizeof (int)); + struct equivclass *eqs = equivs; + int eqs_index = equivs_index; + int eqs_alloc = equivs_alloc; + char const *suffix_begin = current->suffix_begin; + char const *bufend = current->buffer + current->buffered_chars; + int use_line_cmp = ignore_some_line_changes; + + while ((char const *) p < suffix_begin) + { + char const *ip = (char const *) p; + + /* Compute the equivalence class for this line. */ + + h = 0; + + /* Hash this line until we find a newline. */ + if (ignore_case_flag) + { + if (ignore_all_space_flag) + while ((c = *p++) != '\n') + { + if (! ISSPACE (c)) + h = HASH (h, ISUPPER (c) ? tolower (c) : c); + } + else if (ignore_space_change_flag) + while ((c = *p++) != '\n') + { + if (ISSPACE (c)) + { + for (;;) + { + c = *p++; + if (!ISSPACE (c)) + break; + if (c == '\n') + goto hashing_done; + } + h = HASH (h, ' '); + } + /* C is now the first non-space. */ + h = HASH (h, ISUPPER (c) ? tolower (c) : c); + } + else + while ((c = *p++) != '\n') + h = HASH (h, ISUPPER (c) ? tolower (c) : c); + } + else + { + if (ignore_all_space_flag) + while ((c = *p++) != '\n') + { + if (! ISSPACE (c)) + h = HASH (h, c); + } + else if (ignore_space_change_flag) + while ((c = *p++) != '\n') + { + if (ISSPACE (c)) + { + for (;;) + { + c = *p++; + if (!ISSPACE (c)) + break; + if (c == '\n') + goto hashing_done; + } + h = HASH (h, ' '); + } + /* C is now the first non-space. */ + h = HASH (h, c); + } + else + while ((c = *p++) != '\n') + h = HASH (h, c); + } + hashing_done:; + + bucket = &buckets[h % nbuckets]; + length = (char const *) p - ip - 1; + + if ((char const *) p == bufend + && current->missing_newline + && ROBUST_OUTPUT_STYLE (output_style)) + { + /* This line is incomplete. If this is significant, + put the line into bucket[-1]. */ + if (! (ignore_space_change_flag | ignore_all_space_flag)) + bucket = &buckets[-1]; + + /* Omit the inserted newline when computing linbuf later. */ + p--; + bufend = suffix_begin = (char const *) p; + } + + for (i = *bucket; ; i = eqs[i].next) + if (!i) + { + /* Create a new equivalence class in this bucket. */ + i = eqs_index++; + if (i == eqs_alloc) + eqs = (struct equivclass *) + xrealloc (eqs, (eqs_alloc*=2) * sizeof(*eqs)); + eqs[i].next = *bucket; + eqs[i].hash = h; + eqs[i].line = ip; + eqs[i].length = length; + *bucket = i; + break; + } + else if (eqs[i].hash == h) + { + char const *eqline = eqs[i].line; + + /* Reuse existing equivalence class if the lines are identical. + This detects the common case of exact identity + faster than complete comparison would. */ + if (eqs[i].length == length && memcmp (eqline, ip, length) == 0) + break; + + /* Reuse existing class if line_cmp reports the lines equal. */ + if (use_line_cmp && line_cmp (eqline, ip) == 0) + break; + } + + /* Maybe increase the size of the line table. */ + if (line == alloc_lines) + { + /* Double (alloc_lines - linbuf_base) by adding to alloc_lines. */ + alloc_lines = 2 * alloc_lines - linbuf_base; + cureqs = (int *) xrealloc (cureqs, alloc_lines * sizeof (*cureqs)); + linbuf = (char const **) xrealloc (linbuf + linbuf_base, + (alloc_lines - linbuf_base) + * sizeof (*linbuf)) + - linbuf_base; + } + linbuf[line] = ip; + cureqs[line] = i; + ++line; + } + + current->buffered_lines = line; + + for (i = 0; ; i++) + { + /* Record the line start for lines in the suffix that we care about. + Record one more line start than lines, + so that we can compute the length of any buffered line. */ + if (line == alloc_lines) + { + /* Double (alloc_lines - linbuf_base) by adding to alloc_lines. */ + alloc_lines = 2 * alloc_lines - linbuf_base; + linbuf = (char const **) xrealloc (linbuf + linbuf_base, + (alloc_lines - linbuf_base) + * sizeof (*linbuf)) + - linbuf_base; + } + linbuf[line] = (char const *) p; + + if ((char const *) p == bufend) + break; + + if (context <= i && no_diff_means_no_output) + break; + + line++; + + while (*p++ != '\n') + ; + } + + /* Done with cache in local variables. */ + current->linbuf = linbuf; + current->valid_lines = line; + current->alloc_lines = alloc_lines; + current->equivs = cureqs; + equivs = eqs; + equivs_alloc = eqs_alloc; + equivs_index = eqs_index; +} + +/* Prepare the end of the text. Make sure it's initialized. + Make sure text ends in a newline, + but remember that we had to add one. */ + +static void +prepare_text_end (current) + struct file_data *current; +{ + size_t buffered_chars = current->buffered_chars; + char *p = current->buffer; + + if (buffered_chars == 0 || p[buffered_chars - 1] == '\n') + current->missing_newline = 0; + else + { + p[buffered_chars++] = '\n'; + current->buffered_chars = buffered_chars; + current->missing_newline = 1; + } + + /* Don't use uninitialized storage when planting or using sentinels. */ + if (p) + bzero (p + buffered_chars, sizeof (word)); +} + +/* Given a vector of two file_data objects, find the identical + prefixes and suffixes of each object. */ + +static void +find_identical_ends (filevec) + struct file_data filevec[]; +{ + word *w0, *w1; + char *p0, *p1, *buffer0, *buffer1; + char const *end0, *beg0; + char const **linbuf0, **linbuf1; + int i, lines; + size_t n0, n1, tem; + int alloc_lines0, alloc_lines1; + int buffered_prefix, prefix_count, prefix_mask; + + slurp (&filevec[0]); + if (filevec[0].desc != filevec[1].desc) + slurp (&filevec[1]); + else + { + filevec[1].buffer = filevec[0].buffer; + filevec[1].bufsize = filevec[0].bufsize; + filevec[1].buffered_chars = filevec[0].buffered_chars; + } + for (i = 0; i < 2; i++) + prepare_text_end (&filevec[i]); + + /* Find identical prefix. */ + + p0 = buffer0 = filevec[0].buffer; + p1 = buffer1 = filevec[1].buffer; + + n0 = filevec[0].buffered_chars; + n1 = filevec[1].buffered_chars; + + if (p0 == p1) + /* The buffers are the same; sentinels won't work. */ + p0 = p1 += n1; + else + { + /* Insert end sentinels, in this case characters that are guaranteed + to make the equality test false, and thus terminate the loop. */ + + if (n0 < n1) + p0[n0] = ~p1[n0]; + else + p1[n1] = ~p0[n1]; + + /* Loop until first mismatch, or to the sentinel characters. */ + + /* Compare a word at a time for speed. */ + w0 = (word *) p0; + w1 = (word *) p1; + while (*w0++ == *w1++) + ; + --w0, --w1; + + /* Do the last few bytes of comparison a byte at a time. */ + p0 = (char *) w0; + p1 = (char *) w1; + while (*p0++ == *p1++) + ; + --p0, --p1; + + /* Don't mistakenly count missing newline as part of prefix. */ + if (ROBUST_OUTPUT_STYLE (output_style) + && (buffer0 + n0 - filevec[0].missing_newline < p0) + != + (buffer1 + n1 - filevec[1].missing_newline < p1)) + --p0, --p1; + } + + /* Now P0 and P1 point at the first nonmatching characters. */ + + /* Skip back to last line-beginning in the prefix, + and then discard up to HORIZON_LINES lines from the prefix. */ + i = horizon_lines; + while (p0 != buffer0 && (p0[-1] != '\n' || i--)) + --p0, --p1; + + /* Record the prefix. */ + filevec[0].prefix_end = p0; + filevec[1].prefix_end = p1; + + /* Find identical suffix. */ + + /* P0 and P1 point beyond the last chars not yet compared. */ + p0 = buffer0 + n0; + p1 = buffer1 + n1; + + if (! ROBUST_OUTPUT_STYLE (output_style) + || filevec[0].missing_newline == filevec[1].missing_newline) + { + end0 = p0; /* Addr of last char in file 0. */ + + /* Get value of P0 at which we should stop scanning backward: + this is when either P0 or P1 points just past the last char + of the identical prefix. */ + beg0 = filevec[0].prefix_end + (n0 < n1 ? 0 : n0 - n1); + + /* Scan back until chars don't match or we reach that point. */ + while (p0 != beg0) + if (*--p0 != *--p1) + { + /* Point at the first char of the matching suffix. */ + ++p0, ++p1; + beg0 = p0; + break; + } + + /* Are we at a line-beginning in both files? If not, add the rest of + this line to the main body. Discard up to HORIZON_LINES lines from + the identical suffix. Also, discard one extra line, + because shift_boundaries may need it. */ + i = horizon_lines + !((buffer0 == p0 || p0[-1] == '\n') + && + (buffer1 == p1 || p1[-1] == '\n')); + while (i-- && p0 != end0) + while (*p0++ != '\n') + ; + + p1 += p0 - beg0; + } + + /* Record the suffix. */ + filevec[0].suffix_begin = p0; + filevec[1].suffix_begin = p1; + + /* Calculate number of lines of prefix to save. + + prefix_count == 0 means save the whole prefix; + we need this with for options like -D that output the whole file. + We also need it for options like -F that output some preceding line; + at least we will need to find the last few lines, + but since we don't know how many, it's easiest to find them all. + + Otherwise, prefix_count != 0. Save just prefix_count lines at start + of the line buffer; they'll be moved to the proper location later. + Handle 1 more line than the context says (because we count 1 too many), + rounded up to the next power of 2 to speed index computation. */ + + if (no_diff_means_no_output && ! function_regexp_list) + { + for (prefix_count = 1; prefix_count < context + 1; prefix_count *= 2) + ; + prefix_mask = prefix_count - 1; + alloc_lines0 + = prefix_count + + GUESS_LINES (0, 0, p0 - filevec[0].prefix_end) + + context; + } + else + { + prefix_count = 0; + prefix_mask = ~0; + alloc_lines0 = GUESS_LINES (0, 0, n0); + } + + lines = 0; + linbuf0 = (char const **) xmalloc (alloc_lines0 * sizeof (*linbuf0)); + + /* If the prefix is needed, find the prefix lines. */ + if (! (no_diff_means_no_output + && filevec[0].prefix_end == p0 + && filevec[1].prefix_end == p1)) + { + p0 = buffer0; + end0 = filevec[0].prefix_end; + while (p0 != end0) + { + int l = lines++ & prefix_mask; + if (l == alloc_lines0) + linbuf0 = (char const **) xrealloc (linbuf0, (alloc_lines0 *= 2) + * sizeof(*linbuf0)); + linbuf0[l] = p0; + while (*p0++ != '\n') + ; + } + } + buffered_prefix = prefix_count && context < lines ? context : lines; + + /* Allocate line buffer 1. */ + tem = prefix_count ? filevec[1].suffix_begin - buffer1 : n1; + + alloc_lines1 + = (buffered_prefix + + GUESS_LINES (lines, filevec[1].prefix_end - buffer1, tem) + + context); + linbuf1 = (char const **) xmalloc (alloc_lines1 * sizeof (*linbuf1)); + + if (buffered_prefix != lines) + { + /* Rotate prefix lines to proper location. */ + for (i = 0; i < buffered_prefix; i++) + linbuf1[i] = linbuf0[(lines - context + i) & prefix_mask]; + for (i = 0; i < buffered_prefix; i++) + linbuf0[i] = linbuf1[i]; + } + + /* Initialize line buffer 1 from line buffer 0. */ + for (i = 0; i < buffered_prefix; i++) + linbuf1[i] = linbuf0[i] - buffer0 + buffer1; + + /* Record the line buffer, adjusted so that + linbuf*[0] points at the first differing line. */ + filevec[0].linbuf = linbuf0 + buffered_prefix; + filevec[1].linbuf = linbuf1 + buffered_prefix; + filevec[0].linbuf_base = filevec[1].linbuf_base = - buffered_prefix; + filevec[0].alloc_lines = alloc_lines0 - buffered_prefix; + filevec[1].alloc_lines = alloc_lines1 - buffered_prefix; + filevec[0].prefix_lines = filevec[1].prefix_lines = lines; +} + +/* Largest primes less than some power of two, for nbuckets. Values range + from useful to preposterous. If one of these numbers isn't prime + after all, don't blame it on me, blame it on primes (6) . . . */ +static int const primes[] = +{ + 509, + 1021, + 2039, + 4093, + 8191, + 16381, + 32749, +#if 32767 < INT_MAX + 65521, + 131071, + 262139, + 524287, + 1048573, + 2097143, + 4194301, + 8388593, + 16777213, + 33554393, + 67108859, /* Preposterously large . . . */ + 134217689, + 268435399, + 536870909, + 1073741789, + 2147483647, +#endif + 0 +}; + +/* Given a vector of two file_data objects, read the file associated + with each one, and build the table of equivalence classes. + Return 1 if either file appears to be a binary file. + If PRETEND_BINARY is nonzero, pretend they are binary regardless. */ + +int +read_files (filevec, pretend_binary) + struct file_data filevec[]; + int pretend_binary; +{ + int i; + int skip_test = always_text_flag | pretend_binary; + int appears_binary = pretend_binary | sip (&filevec[0], skip_test); + + if (filevec[0].desc != filevec[1].desc) + appears_binary |= sip (&filevec[1], skip_test | appears_binary); + else + { + filevec[1].buffer = filevec[0].buffer; + filevec[1].bufsize = filevec[0].bufsize; + filevec[1].buffered_chars = filevec[0].buffered_chars; + } + if (appears_binary) + { +#if HAVE_SETMODE + setmode (filevec[0].desc, O_BINARY); + setmode (filevec[1].desc, O_BINARY); +#endif + return 1; + } + + find_identical_ends (filevec); + + equivs_alloc = filevec[0].alloc_lines + filevec[1].alloc_lines + 1; + equivs = (struct equivclass *) xmalloc (equivs_alloc * sizeof (struct equivclass)); + /* Equivalence class 0 is permanently safe for lines that were not + hashed. Real equivalence classes start at 1. */ + equivs_index = 1; + + for (i = 0; primes[i] < equivs_alloc / 3; i++) + if (! primes[i]) + abort (); + nbuckets = primes[i]; + + buckets = (int *) xmalloc ((nbuckets + 1) * sizeof (*buckets)); + bzero (buckets++, (nbuckets + 1) * sizeof (*buckets)); + + for (i = 0; i < 2; i++) + find_and_hash_each_line (&filevec[i]); + + filevec[0].equiv_max = filevec[1].equiv_max = equivs_index; + + free (equivs); + free (buckets - 1); + + return 0; +} diff --git a/sys/src/ape/cmd/diff/mkfile b/sys/src/ape/cmd/diff/mkfile new file mode 100755 index 000000000..045f93c76 --- /dev/null +++ b/sys/src/ape/cmd/diff/mkfile @@ -0,0 +1,47 @@ +</$objtype/mkfile + +OFILES= +HFILES=\ + system.h\ + config.h\ + +TARG=diff diff3 + +DIFFO=\ + analyze.$O\ + cmpbuf.$O\ + dir.$O\ + io.$O\ + util.$O\ + context.$O\ + ed.$O\ + ifdef.$O\ + normal.$O\ + side.$O\ + fnmatch.$O\ + getopt.$O\ + getopt1.$O\ + regex.$O\ + version.$O\ + prepend_args.$O\ + +DIFF3O=\ + getopt.$O\ + getopt1.$O\ + version.$O\ + +SDIFFO=\ + getopt.$O\ + getopt1.$O\ + version.$O\ + +BIN=/$objtype/bin/ape + +</sys/src/cmd/mkmany + +CC=pcc -c +CFLAGS=-B -p -D_POSIX_SOURCE -DREGEX_MALLOC -I. \ + -DHAVE_CONFIG_H -DDIFF_PROGRAM="/bin/ape/diff" \ + +$O.diff: $DIFFO +$O.diff3: $DIFF3O diff --git a/sys/src/ape/cmd/diff/normal.c b/sys/src/ape/cmd/diff/normal.c new file mode 100755 index 000000000..4d9e23cb7 --- /dev/null +++ b/sys/src/ape/cmd/diff/normal.c @@ -0,0 +1,71 @@ +/* Normal-format output routines for GNU DIFF. + Copyright (C) 1988, 1989, 1993 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + + +#include "diff.h" + +static void print_normal_hunk PARAMS((struct change *)); + +/* Print the edit-script SCRIPT as a normal diff. + INF points to an array of descriptions of the two files. */ + +void +print_normal_script (script) + struct change *script; +{ + print_script (script, find_change, print_normal_hunk); +} + +/* Print a hunk of a normal diff. + This is a contiguous portion of a complete edit script, + describing changes in consecutive lines. */ + +static void +print_normal_hunk (hunk) + struct change *hunk; +{ + int first0, last0, first1, last1, deletes, inserts; + register int i; + + /* Determine range of line numbers involved in each file. */ + analyze_hunk (hunk, &first0, &last0, &first1, &last1, &deletes, &inserts); + if (!deletes && !inserts) + return; + + begin_output (); + + /* Print out the line number header for this hunk */ + print_number_range (',', &files[0], first0, last0); + fprintf (outfile, "%c", change_letter (inserts, deletes)); + print_number_range (',', &files[1], first1, last1); + fprintf (outfile, "\n"); + + /* Print the lines that the first file has. */ + if (deletes) + for (i = first0; i <= last0; i++) + print_1_line ("<", &files[0].linbuf[i]); + + if (inserts && deletes) + fprintf (outfile, "---\n"); + + /* Print the lines that the second file has. */ + if (inserts) + for (i = first1; i <= last1; i++) + print_1_line (">", &files[1].linbuf[i]); +} diff --git a/sys/src/ape/cmd/diff/prepend_args.c b/sys/src/ape/cmd/diff/prepend_args.c new file mode 100755 index 000000000..f3fad167a --- /dev/null +++ b/sys/src/ape/cmd/diff/prepend_args.c @@ -0,0 +1,87 @@ +/* prepend_args.c - utilility programs for manpiulating argv[] + Copyright (C) 1999 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA + 02111-1307, USA. */ + +/* $FreeBSD: src/contrib/diff/prepend_args.c,v 1.1 1999/11/26 02:51:44 obrien Exp $ */ + + +#ifdef HAVE_CONFIG_H +# include "config.h" +#endif +#include "system.h" +#include "prepend_args.h" +#include "diff.h" + + +/* Find the white-space-separated options specified by OPTIONS, and + using BUF to store copies of these options, set ARGV[0], ARGV[1], + etc. to the option copies. Return the number N of options found. + Do not set ARGV[N] to NULL. If ARGV is NULL, do not store ARGV[0] + etc. Backslash can be used to escape whitespace (and backslashes). */ +static int +prepend_args (options, buf, argv) + char const *options; + char *buf; + char **argv; +{ + char const *o = options; + char *b = buf; + int n = 0; + + for (;;) + { + while (ISSPACE ((unsigned char) *o)) + o++; + if (!*o) + return n; + if (argv) + argv[n] = b; + n++; + + do + if ((*b++ = *o++) == '\\' && *o) + b[-1] = *o++; + while (*o && ! ISSPACE ((unsigned char) *o)); + + *b++ = '\0'; + } +} + +/* Prepend the whitespace-separated options in OPTIONS to the argument + vector of a main program with argument count *PARGC and argument + vector *PARGV. */ +void +prepend_default_options (options, pargc, pargv) + char const *options; + int *pargc; + char ***pargv; +{ + if (options) + { + char *buf = xmalloc (strlen (options) + 1); + int prepended = prepend_args (options, buf, (char **) NULL); + int argc = *pargc; + char * const *argv = *pargv; + char **pp = (char **) xmalloc ((prepended + argc + 1) * sizeof *pp); + *pargc = prepended + argc; + *pargv = pp; + *pp++ = *argv++; + pp += prepend_args (options, buf, pp); + while ((*pp++ = *argv++)) + continue; + } +} diff --git a/sys/src/ape/cmd/diff/prepend_args.h b/sys/src/ape/cmd/diff/prepend_args.h new file mode 100755 index 000000000..a46de62ff --- /dev/null +++ b/sys/src/ape/cmd/diff/prepend_args.h @@ -0,0 +1,21 @@ +/* prepend_args.h - utilility programs for manpiulating argv[] + Copyright (C) 1999 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA + 02111-1307, USA. */ + +/* $FreeBSD: src/contrib/diff/prepend_args.h,v 1.1 1999/11/26 02:51:44 obrien Exp $ */ + +void prepend_default_options PARAMS ((char const *, int *, char ***)); diff --git a/sys/src/ape/cmd/diff/regex.c b/sys/src/ape/cmd/diff/regex.c new file mode 100755 index 000000000..0379be0bd --- /dev/null +++ b/sys/src/ape/cmd/diff/regex.c @@ -0,0 +1,6374 @@ +/* Extended regular expression matching and search library, version + 0.12. (Implements POSIX draft P10003.2/D11.2, except for + internationalization features.) + + Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, + USA. */ + +/* AIX requires this to be the first thing in the file. */ +#if defined (_AIX) && !defined (REGEX_MALLOC) + #pragma alloca +#endif + +#undef _GNU_SOURCE +#define _GNU_SOURCE + +#ifdef emacs +/* Converts the pointer to the char to BEG-based offset from the start. */ +#define PTR_TO_OFFSET(d) \ + POS_AS_IN_BUFFER (MATCHING_IN_FIRST_STRING \ + ? (d) - string1 : (d) - (string2 - size1)) +#define POS_AS_IN_BUFFER(p) ((p) + (NILP (re_match_object) || BUFFERP (re_match_object))) +#else +#define PTR_TO_OFFSET(d) 0 +#endif + +#include "config.h" + +/* We need this for `regex.h', and perhaps for the Emacs include files. */ +#include <sys/types.h> + +/* This is for other GNU distributions with internationalized messages. */ +#if HAVE_LIBINTL_H || defined (_LIBC) +# include <libintl.h> +#else +# define gettext(msgid) (msgid) +#endif + +#ifndef gettext_noop +/* This define is so xgettext can find the internationalizable + strings. */ +#define gettext_noop(String) String +#endif + +/* The `emacs' switch turns on certain matching commands + that make sense only in Emacs. */ +#ifdef emacs + +#include "lisp.h" +#include "buffer.h" + +/* Make syntax table lookup grant data in gl_state. */ +#define SYNTAX_ENTRY_VIA_PROPERTY + +#include "syntax.h" +#include "charset.h" +#include "category.h" + +#define malloc xmalloc +#define realloc xrealloc +#define free xfree + +#else /* not emacs */ + +/* If we are not linking with Emacs proper, + we can't use the relocating allocator + even if config.h says that we can. */ +#undef REL_ALLOC + +#if defined (STDC_HEADERS) || defined (_LIBC) +#include <stdlib.h> +#else +char *malloc (); +char *realloc (); +#endif + +/* When used in Emacs's lib-src, we need to get bzero and bcopy somehow. + If nothing else has been done, use the method below. */ +#ifdef INHIBIT_STRING_HEADER +#if !(defined (HAVE_BZERO) && defined (HAVE_BCOPY)) +#if !defined (bzero) && !defined (bcopy) +#undef INHIBIT_STRING_HEADER +#endif +#endif +#endif + +/* This is the normal way of making sure we have a bcopy and a bzero. + This is used in most programs--a few other programs avoid this + by defining INHIBIT_STRING_HEADER. */ +#ifndef INHIBIT_STRING_HEADER +#if defined (HAVE_STRING_H) || defined (STDC_HEADERS) || defined (_LIBC) +#include <string.h> +#ifndef bcmp +#define bcmp(s1, s2, n) memcmp ((s1), (s2), (n)) +#endif +#ifndef bcopy +#define bcopy(s, d, n) memcpy ((d), (s), (n)) +#endif +#ifndef bzero +#define bzero(s, n) memset ((s), 0, (n)) +#endif +#else +#include <strings.h> +#endif +#endif + +/* Define the syntax stuff for \<, \>, etc. */ + +/* This must be nonzero for the wordchar and notwordchar pattern + commands in re_match_2. */ +#ifndef Sword +#define Sword 1 +#endif + +#ifdef SWITCH_ENUM_BUG +#define SWITCH_ENUM_CAST(x) ((int)(x)) +#else +#define SWITCH_ENUM_CAST(x) (x) +#endif + +#ifdef SYNTAX_TABLE + +extern char *re_syntax_table; + +#else /* not SYNTAX_TABLE */ + +/* How many characters in the character set. */ +#define CHAR_SET_SIZE 256 + +static char re_syntax_table[CHAR_SET_SIZE]; + +static void +init_syntax_once () +{ + register int c; + static int done = 0; + + if (done) + return; + + bzero (re_syntax_table, sizeof re_syntax_table); + + for (c = 'a'; c <= 'z'; c++) + re_syntax_table[c] = Sword; + + for (c = 'A'; c <= 'Z'; c++) + re_syntax_table[c] = Sword; + + for (c = '0'; c <= '9'; c++) + re_syntax_table[c] = Sword; + + re_syntax_table['_'] = Sword; + + done = 1; +} + +#endif /* not SYNTAX_TABLE */ + +#define SYNTAX(c) re_syntax_table[c] + +/* Dummy macros for non-Emacs environments. */ +#define BASE_LEADING_CODE_P(c) (0) +#define WORD_BOUNDARY_P(c1, c2) (0) +#define CHAR_HEAD_P(p) (1) +#define SINGLE_BYTE_CHAR_P(c) (1) +#define SAME_CHARSET_P(c1, c2) (1) +#define MULTIBYTE_FORM_LENGTH(p, s) (1) +#define STRING_CHAR(p, s) (*(p)) +#define STRING_CHAR_AND_LENGTH(p, s, actual_len) ((actual_len) = 1, *(p)) +#define GET_CHAR_AFTER_2(c, p, str1, end1, str2, end2) \ + (c = ((p) == (end1) ? *(str2) : *(p))) +#define GET_CHAR_BEFORE_2(c, p, str1, end1, str2, end2) \ + (c = ((p) == (str2) ? *((end1) - 1) : *((p) - 1))) +#endif /* not emacs */ + +/* Get the interface, including the syntax bits. */ +#include "regex.h" + +/* isalpha etc. are used for the character classes. */ +#include <ctype.h> + +/* Jim Meyering writes: + + "... Some ctype macros are valid only for character codes that + isascii says are ASCII (SGI's IRIX-4.0.5 is one such system --when + using /bin/cc or gcc but without giving an ansi option). So, all + ctype uses should be through macros like ISPRINT... If + STDC_HEADERS is defined, then autoconf has verified that the ctype + macros don't need to be guarded with references to isascii. ... + Defining isascii to 1 should let any compiler worth its salt + eliminate the && through constant folding." */ + +#if defined (STDC_HEADERS) || (!defined (isascii) && !defined (HAVE_ISASCII)) +#define ISASCII(c) 1 +#else +#define ISASCII(c) isascii(c) +#endif + +#ifdef isblank +#define ISBLANK(c) (ISASCII (c) && isblank (c)) +#else +#define ISBLANK(c) ((c) == ' ' || (c) == '\t') +#endif +#ifdef isgraph +#define ISGRAPH(c) (ISASCII (c) && isgraph (c)) +#else +#define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c)) +#endif + +#define ISPRINT(c) (ISASCII (c) && isprint (c)) +#define ISDIGIT(c) (ISASCII (c) && isdigit (c)) +#define ISALNUM(c) (ISASCII (c) && isalnum (c)) +#define ISALPHA(c) (ISASCII (c) && isalpha (c)) +#define ISCNTRL(c) (ISASCII (c) && iscntrl (c)) +#define ISLOWER(c) (ISASCII (c) && islower (c)) +#define ISPUNCT(c) (ISASCII (c) && ispunct (c)) +#define ISSPACE(c) (ISASCII (c) && isspace (c)) +#define ISUPPER(c) (ISASCII (c) && isupper (c)) +#define ISXDIGIT(c) (ISASCII (c) && isxdigit (c)) + +#ifndef NULL +#define NULL (void *)0 +#endif + +/* We remove any previous definition of `SIGN_EXTEND_CHAR', + since ours (we hope) works properly with all combinations of + machines, compilers, `char' and `unsigned char' argument types. + (Per Bothner suggested the basic approach.) */ +#undef SIGN_EXTEND_CHAR +#if __STDC__ +#define SIGN_EXTEND_CHAR(c) ((signed char) (c)) +#else /* not __STDC__ */ +/* As in Harbison and Steele. */ +#define SIGN_EXTEND_CHAR(c) ((((unsigned char) (c)) ^ 128) - 128) +#endif + +/* Should we use malloc or alloca? If REGEX_MALLOC is not defined, we + use `alloca' instead of `malloc'. This is because using malloc in + re_search* or re_match* could cause memory leaks when C-g is used in + Emacs; also, malloc is slower and causes storage fragmentation. On + the other hand, malloc is more portable, and easier to debug. + + Because we sometimes use alloca, some routines have to be macros, + not functions -- `alloca'-allocated space disappears at the end of the + function it is called in. */ + +#ifdef REGEX_MALLOC + +#define REGEX_ALLOCATE malloc +#define REGEX_REALLOCATE(source, osize, nsize) realloc (source, nsize) +#define REGEX_FREE free + +#else /* not REGEX_MALLOC */ + +/* Emacs already defines alloca, sometimes. */ +#ifndef alloca + +/* Make alloca work the best possible way. */ +#ifdef __GNUC__ +#define alloca __builtin_alloca +#else /* not __GNUC__ */ +#if HAVE_ALLOCA_H +#include <alloca.h> +#else /* not __GNUC__ or HAVE_ALLOCA_H */ +#if 0 /* It is a bad idea to declare alloca. We always cast the result. */ +#ifndef _AIX /* Already did AIX, up at the top. */ +char *alloca (); +#endif /* not _AIX */ +#endif +#endif /* not HAVE_ALLOCA_H */ +#endif /* not __GNUC__ */ + +#endif /* not alloca */ + +#define REGEX_ALLOCATE alloca + +/* Assumes a `char *destination' variable. */ +#define REGEX_REALLOCATE(source, osize, nsize) \ + (destination = (char *) alloca (nsize), \ + bcopy (source, destination, osize), \ + destination) + +/* No need to do anything to free, after alloca. */ +#define REGEX_FREE(arg) ((void)0) /* Do nothing! But inhibit gcc warning. */ + +#endif /* not REGEX_MALLOC */ + +/* Define how to allocate the failure stack. */ + +#if defined (REL_ALLOC) && defined (REGEX_MALLOC) + +#define REGEX_ALLOCATE_STACK(size) \ + r_alloc (&failure_stack_ptr, (size)) +#define REGEX_REALLOCATE_STACK(source, osize, nsize) \ + r_re_alloc (&failure_stack_ptr, (nsize)) +#define REGEX_FREE_STACK(ptr) \ + r_alloc_free (&failure_stack_ptr) + +#else /* not using relocating allocator */ + +#ifdef REGEX_MALLOC + +#define REGEX_ALLOCATE_STACK malloc +#define REGEX_REALLOCATE_STACK(source, osize, nsize) realloc (source, nsize) +#define REGEX_FREE_STACK free + +#else /* not REGEX_MALLOC */ + +#define REGEX_ALLOCATE_STACK alloca + +#define REGEX_REALLOCATE_STACK(source, osize, nsize) \ + REGEX_REALLOCATE (source, osize, nsize) +/* No need to explicitly free anything. */ +#define REGEX_FREE_STACK(arg) + +#endif /* not REGEX_MALLOC */ +#endif /* not using relocating allocator */ + + +/* True if `size1' is non-NULL and PTR is pointing anywhere inside + `string1' or just past its end. This works if PTR is NULL, which is + a good thing. */ +#define FIRST_STRING_P(ptr) \ + (size1 && string1 <= (ptr) && (ptr) <= string1 + size1) + +/* (Re)Allocate N items of type T using malloc, or fail. */ +#define TALLOC(n, t) ((t *) malloc ((n) * sizeof (t))) +#define RETALLOC(addr, n, t) ((addr) = (t *) realloc (addr, (n) * sizeof (t))) +#define RETALLOC_IF(addr, n, t) \ + if (addr) RETALLOC((addr), (n), t); else (addr) = TALLOC ((n), t) +#define REGEX_TALLOC(n, t) ((t *) REGEX_ALLOCATE ((n) * sizeof (t))) + +#define BYTEWIDTH 8 /* In bits. */ + +#define STREQ(s1, s2) ((strcmp (s1, s2) == 0)) + +#undef MAX +#undef MIN +#define MAX(a, b) ((a) > (b) ? (a) : (b)) +#define MIN(a, b) ((a) < (b) ? (a) : (b)) + +typedef char boolean; +#define false 0 +#define true 1 + +static int re_match_2_internal (); + +/* These are the command codes that appear in compiled regular + expressions. Some opcodes are followed by argument bytes. A + command code can specify any interpretation whatsoever for its + arguments. Zero bytes may appear in the compiled regular expression. */ + +typedef enum +{ + no_op = 0, + + /* Succeed right away--no more backtracking. */ + succeed, + + /* Followed by one byte giving n, then by n literal bytes. */ + exactn, + + /* Matches any (more or less) character. */ + anychar, + + /* Matches any one char belonging to specified set. First + following byte is number of bitmap bytes. Then come bytes + for a bitmap saying which chars are in. Bits in each byte + are ordered low-bit-first. A character is in the set if its + bit is 1. A character too large to have a bit in the map is + automatically not in the set. */ + charset, + + /* Same parameters as charset, but match any character that is + not one of those specified. */ + charset_not, + + /* Start remembering the text that is matched, for storing in a + register. Followed by one byte with the register number, in + the range 0 to one less than the pattern buffer's re_nsub + field. Then followed by one byte with the number of groups + inner to this one. (This last has to be part of the + start_memory only because we need it in the on_failure_jump + of re_match_2.) */ + start_memory, + + /* Stop remembering the text that is matched and store it in a + memory register. Followed by one byte with the register + number, in the range 0 to one less than `re_nsub' in the + pattern buffer, and one byte with the number of inner groups, + just like `start_memory'. (We need the number of inner + groups here because we don't have any easy way of finding the + corresponding start_memory when we're at a stop_memory.) */ + stop_memory, + + /* Match a duplicate of something remembered. Followed by one + byte containing the register number. */ + duplicate, + + /* Fail unless at beginning of line. */ + begline, + + /* Fail unless at end of line. */ + endline, + + /* Succeeds if at beginning of buffer (if emacs) or at beginning + of string to be matched (if not). */ + begbuf, + + /* Analogously, for end of buffer/string. */ + endbuf, + + /* Followed by two byte relative address to which to jump. */ + jump, + + /* Same as jump, but marks the end of an alternative. */ + jump_past_alt, + + /* Followed by two-byte relative address of place to resume at + in case of failure. */ + on_failure_jump, + + /* Like on_failure_jump, but pushes a placeholder instead of the + current string position when executed. */ + on_failure_keep_string_jump, + + /* Throw away latest failure point and then jump to following + two-byte relative address. */ + pop_failure_jump, + + /* Change to pop_failure_jump if know won't have to backtrack to + match; otherwise change to jump. This is used to jump + back to the beginning of a repeat. If what follows this jump + clearly won't match what the repeat does, such that we can be + sure that there is no use backtracking out of repetitions + already matched, then we change it to a pop_failure_jump. + Followed by two-byte address. */ + maybe_pop_jump, + + /* Jump to following two-byte address, and push a dummy failure + point. This failure point will be thrown away if an attempt + is made to use it for a failure. A `+' construct makes this + before the first repeat. Also used as an intermediary kind + of jump when compiling an alternative. */ + dummy_failure_jump, + + /* Push a dummy failure point and continue. Used at the end of + alternatives. */ + push_dummy_failure, + + /* Followed by two-byte relative address and two-byte number n. + After matching N times, jump to the address upon failure. */ + succeed_n, + + /* Followed by two-byte relative address, and two-byte number n. + Jump to the address N times, then fail. */ + jump_n, + + /* Set the following two-byte relative address to the + subsequent two-byte number. The address *includes* the two + bytes of number. */ + set_number_at, + + wordchar, /* Matches any word-constituent character. */ + notwordchar, /* Matches any char that is not a word-constituent. */ + + wordbeg, /* Succeeds if at word beginning. */ + wordend, /* Succeeds if at word end. */ + + wordbound, /* Succeeds if at a word boundary. */ + notwordbound /* Succeeds if not at a word boundary. */ + +#ifdef emacs + ,before_dot, /* Succeeds if before point. */ + at_dot, /* Succeeds if at point. */ + after_dot, /* Succeeds if after point. */ + + /* Matches any character whose syntax is specified. Followed by + a byte which contains a syntax code, e.g., Sword. */ + syntaxspec, + + /* Matches any character whose syntax is not that specified. */ + notsyntaxspec, + + /* Matches any character whose category-set contains the specified + category. The operator is followed by a byte which contains a + category code (mnemonic ASCII character). */ + categoryspec, + + /* Matches any character whose category-set does not contain the + specified category. The operator is followed by a byte which + contains the category code (mnemonic ASCII character). */ + notcategoryspec +#endif /* emacs */ +} re_opcode_t; + +/* Common operations on the compiled pattern. */ + +/* Store NUMBER in two contiguous bytes starting at DESTINATION. */ + +#define STORE_NUMBER(destination, number) \ + do { \ + (destination)[0] = (number) & 0377; \ + (destination)[1] = (number) >> 8; \ + } while (0) + +/* Same as STORE_NUMBER, except increment DESTINATION to + the byte after where the number is stored. Therefore, DESTINATION + must be an lvalue. */ + +#define STORE_NUMBER_AND_INCR(destination, number) \ + do { \ + STORE_NUMBER (destination, number); \ + (destination) += 2; \ + } while (0) + +/* Put into DESTINATION a number stored in two contiguous bytes starting + at SOURCE. */ + +#define EXTRACT_NUMBER(destination, source) \ + do { \ + (destination) = *(source) & 0377; \ + (destination) += SIGN_EXTEND_CHAR (*((source) + 1)) << 8; \ + } while (0) + +#ifdef DEBUG +static void +extract_number (dest, source) + int *dest; + unsigned char *source; +{ + int temp = SIGN_EXTEND_CHAR (*(source + 1)); + *dest = *source & 0377; + *dest += temp << 8; +} + +#ifndef EXTRACT_MACROS /* To debug the macros. */ +#undef EXTRACT_NUMBER +#define EXTRACT_NUMBER(dest, src) extract_number (&dest, src) +#endif /* not EXTRACT_MACROS */ + +#endif /* DEBUG */ + +/* Same as EXTRACT_NUMBER, except increment SOURCE to after the number. + SOURCE must be an lvalue. */ + +#define EXTRACT_NUMBER_AND_INCR(destination, source) \ + do { \ + EXTRACT_NUMBER (destination, source); \ + (source) += 2; \ + } while (0) + +#ifdef DEBUG +static void +extract_number_and_incr (destination, source) + int *destination; + unsigned char **source; +{ + extract_number (destination, *source); + *source += 2; +} + +#ifndef EXTRACT_MACROS +#undef EXTRACT_NUMBER_AND_INCR +#define EXTRACT_NUMBER_AND_INCR(dest, src) \ + extract_number_and_incr (&dest, &src) +#endif /* not EXTRACT_MACROS */ + +#endif /* DEBUG */ + +/* Store a multibyte character in three contiguous bytes starting + DESTINATION, and increment DESTINATION to the byte after where the + character is stored. Therefore, DESTINATION must be an lvalue. */ + +#define STORE_CHARACTER_AND_INCR(destination, character) \ + do { \ + (destination)[0] = (character) & 0377; \ + (destination)[1] = ((character) >> 8) & 0377; \ + (destination)[2] = (character) >> 16; \ + (destination) += 3; \ + } while (0) + +/* Put into DESTINATION a character stored in three contiguous bytes + starting at SOURCE. */ + +#define EXTRACT_CHARACTER(destination, source) \ + do { \ + (destination) = ((source)[0] \ + | ((source)[1] << 8) \ + | ((source)[2] << 16)); \ + } while (0) + + +/* Macros for charset. */ + +/* Size of bitmap of charset P in bytes. P is a start of charset, + i.e. *P is (re_opcode_t) charset or (re_opcode_t) charset_not. */ +#define CHARSET_BITMAP_SIZE(p) ((p)[1] & 0x7F) + +/* Nonzero if charset P has range table. */ +#define CHARSET_RANGE_TABLE_EXISTS_P(p) ((p)[1] & 0x80) + +/* Return the address of range table of charset P. But not the start + of table itself, but the before where the number of ranges is + stored. `2 +' means to skip re_opcode_t and size of bitmap. */ +#define CHARSET_RANGE_TABLE(p) (&(p)[2 + CHARSET_BITMAP_SIZE (p)]) + +/* Test if C is listed in the bitmap of charset P. */ +#define CHARSET_LOOKUP_BITMAP(p, c) \ + ((c) < CHARSET_BITMAP_SIZE (p) * BYTEWIDTH \ + && (p)[2 + (c) / BYTEWIDTH] & (1 << ((c) % BYTEWIDTH))) + +/* Return the address of end of RANGE_TABLE. COUNT is number of + ranges (which is a pair of (start, end)) in the RANGE_TABLE. `* 2' + is start of range and end of range. `* 3' is size of each start + and end. */ +#define CHARSET_RANGE_TABLE_END(range_table, count) \ + ((range_table) + (count) * 2 * 3) + +/* Test if C is in RANGE_TABLE. A flag NOT is negated if C is in. + COUNT is number of ranges in RANGE_TABLE. */ +#define CHARSET_LOOKUP_RANGE_TABLE_RAW(not, c, range_table, count) \ + do \ + { \ + int range_start, range_end; \ + unsigned char *p; \ + unsigned char *range_table_end \ + = CHARSET_RANGE_TABLE_END ((range_table), (count)); \ + \ + for (p = (range_table); p < range_table_end; p += 2 * 3) \ + { \ + EXTRACT_CHARACTER (range_start, p); \ + EXTRACT_CHARACTER (range_end, p + 3); \ + \ + if (range_start <= (c) && (c) <= range_end) \ + { \ + (not) = !(not); \ + break; \ + } \ + } \ + } \ + while (0) + +/* Test if C is in range table of CHARSET. The flag NOT is negated if + C is listed in it. */ +#define CHARSET_LOOKUP_RANGE_TABLE(not, c, charset) \ + do \ + { \ + /* Number of ranges in range table. */ \ + int count; \ + unsigned char *range_table = CHARSET_RANGE_TABLE (charset); \ + \ + EXTRACT_NUMBER_AND_INCR (count, range_table); \ + CHARSET_LOOKUP_RANGE_TABLE_RAW ((not), (c), range_table, count); \ + } \ + while (0) + +/* If DEBUG is defined, Regex prints many voluminous messages about what + it is doing (if the variable `debug' is nonzero). If linked with the + main program in `iregex.c', you can enter patterns and strings + interactively. And if linked with the main program in `main.c' and + the other test files, you can run the already-written tests. */ + +#ifdef DEBUG + +/* We use standard I/O for debugging. */ +#include <stdio.h> + +/* It is useful to test things that ``must'' be true when debugging. */ +#include <assert.h> + +static int debug = 0; + +#define DEBUG_STATEMENT(e) e +#define DEBUG_PRINT1(x) if (debug) printf (x) +#define DEBUG_PRINT2(x1, x2) if (debug) printf (x1, x2) +#define DEBUG_PRINT3(x1, x2, x3) if (debug) printf (x1, x2, x3) +#define DEBUG_PRINT4(x1, x2, x3, x4) if (debug) printf (x1, x2, x3, x4) +#define DEBUG_PRINT_COMPILED_PATTERN(p, s, e) \ + if (debug) print_partial_compiled_pattern (s, e) +#define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2) \ + if (debug) print_double_string (w, s1, sz1, s2, sz2) + + +/* Print the fastmap in human-readable form. */ + +void +print_fastmap (fastmap) + char *fastmap; +{ + unsigned was_a_range = 0; + unsigned i = 0; + + while (i < (1 << BYTEWIDTH)) + { + if (fastmap[i++]) + { + was_a_range = 0; + putchar (i - 1); + while (i < (1 << BYTEWIDTH) && fastmap[i]) + { + was_a_range = 1; + i++; + } + if (was_a_range) + { + printf ("-"); + putchar (i - 1); + } + } + } + putchar ('\n'); +} + + +/* Print a compiled pattern string in human-readable form, starting at + the START pointer into it and ending just before the pointer END. */ + +void +print_partial_compiled_pattern (start, end) + unsigned char *start; + unsigned char *end; +{ + int mcnt, mcnt2; + unsigned char *p = start; + unsigned char *pend = end; + + if (start == NULL) + { + printf ("(null)\n"); + return; + } + + /* Loop over pattern commands. */ + while (p < pend) + { + printf ("%d:\t", p - start); + + switch ((re_opcode_t) *p++) + { + case no_op: + printf ("/no_op"); + break; + + case exactn: + mcnt = *p++; + printf ("/exactn/%d", mcnt); + do + { + putchar ('/'); + putchar (*p++); + } + while (--mcnt); + break; + + case start_memory: + mcnt = *p++; + printf ("/start_memory/%d/%d", mcnt, *p++); + break; + + case stop_memory: + mcnt = *p++; + printf ("/stop_memory/%d/%d", mcnt, *p++); + break; + + case duplicate: + printf ("/duplicate/%d", *p++); + break; + + case anychar: + printf ("/anychar"); + break; + + case charset: + case charset_not: + { + register int c, last = -100; + register int in_range = 0; + + printf ("/charset [%s", + (re_opcode_t) *(p - 1) == charset_not ? "^" : ""); + + assert (p + *p < pend); + + for (c = 0; c < 256; c++) + if (c / 8 < *p + && (p[1 + (c/8)] & (1 << (c % 8)))) + { + /* Are we starting a range? */ + if (last + 1 == c && ! in_range) + { + putchar ('-'); + in_range = 1; + } + /* Have we broken a range? */ + else if (last + 1 != c && in_range) + { + putchar (last); + in_range = 0; + } + + if (! in_range) + putchar (c); + + last = c; + } + + if (in_range) + putchar (last); + + putchar (']'); + + p += 1 + *p; + } + break; + + case begline: + printf ("/begline"); + break; + + case endline: + printf ("/endline"); + break; + + case on_failure_jump: + extract_number_and_incr (&mcnt, &p); + printf ("/on_failure_jump to %d", p + mcnt - start); + break; + + case on_failure_keep_string_jump: + extract_number_and_incr (&mcnt, &p); + printf ("/on_failure_keep_string_jump to %d", p + mcnt - start); + break; + + case dummy_failure_jump: + extract_number_and_incr (&mcnt, &p); + printf ("/dummy_failure_jump to %d", p + mcnt - start); + break; + + case push_dummy_failure: + printf ("/push_dummy_failure"); + break; + + case maybe_pop_jump: + extract_number_and_incr (&mcnt, &p); + printf ("/maybe_pop_jump to %d", p + mcnt - start); + break; + + case pop_failure_jump: + extract_number_and_incr (&mcnt, &p); + printf ("/pop_failure_jump to %d", p + mcnt - start); + break; + + case jump_past_alt: + extract_number_and_incr (&mcnt, &p); + printf ("/jump_past_alt to %d", p + mcnt - start); + break; + + case jump: + extract_number_and_incr (&mcnt, &p); + printf ("/jump to %d", p + mcnt - start); + break; + + case succeed_n: + extract_number_and_incr (&mcnt, &p); + extract_number_and_incr (&mcnt2, &p); + printf ("/succeed_n to %d, %d times", p + mcnt - start, mcnt2); + break; + + case jump_n: + extract_number_and_incr (&mcnt, &p); + extract_number_and_incr (&mcnt2, &p); + printf ("/jump_n to %d, %d times", p + mcnt - start, mcnt2); + break; + + case set_number_at: + extract_number_and_incr (&mcnt, &p); + extract_number_and_incr (&mcnt2, &p); + printf ("/set_number_at location %d to %d", p + mcnt - start, mcnt2); + break; + + case wordbound: + printf ("/wordbound"); + break; + + case notwordbound: + printf ("/notwordbound"); + break; + + case wordbeg: + printf ("/wordbeg"); + break; + + case wordend: + printf ("/wordend"); + +#ifdef emacs + case before_dot: + printf ("/before_dot"); + break; + + case at_dot: + printf ("/at_dot"); + break; + + case after_dot: + printf ("/after_dot"); + break; + + case syntaxspec: + printf ("/syntaxspec"); + mcnt = *p++; + printf ("/%d", mcnt); + break; + + case notsyntaxspec: + printf ("/notsyntaxspec"); + mcnt = *p++; + printf ("/%d", mcnt); + break; +#endif /* emacs */ + + case wordchar: + printf ("/wordchar"); + break; + + case notwordchar: + printf ("/notwordchar"); + break; + + case begbuf: + printf ("/begbuf"); + break; + + case endbuf: + printf ("/endbuf"); + break; + + default: + printf ("?%d", *(p-1)); + } + + putchar ('\n'); + } + + printf ("%d:\tend of pattern.\n", p - start); +} + + +void +print_compiled_pattern (bufp) + struct re_pattern_buffer *bufp; +{ + unsigned char *buffer = bufp->buffer; + + print_partial_compiled_pattern (buffer, buffer + bufp->used); + printf ("%d bytes used/%d bytes allocated.\n", bufp->used, bufp->allocated); + + if (bufp->fastmap_accurate && bufp->fastmap) + { + printf ("fastmap: "); + print_fastmap (bufp->fastmap); + } + + printf ("re_nsub: %d\t", bufp->re_nsub); + printf ("regs_alloc: %d\t", bufp->regs_allocated); + printf ("can_be_null: %d\t", bufp->can_be_null); + printf ("newline_anchor: %d\n", bufp->newline_anchor); + printf ("no_sub: %d\t", bufp->no_sub); + printf ("not_bol: %d\t", bufp->not_bol); + printf ("not_eol: %d\t", bufp->not_eol); + printf ("syntax: %d\n", bufp->syntax); + /* Perhaps we should print the translate table? */ +} + + +void +print_double_string (where, string1, size1, string2, size2) + const char *where; + const char *string1; + const char *string2; + int size1; + int size2; +{ + unsigned this_char; + + if (where == NULL) + printf ("(null)"); + else + { + if (FIRST_STRING_P (where)) + { + for (this_char = where - string1; this_char < size1; this_char++) + putchar (string1[this_char]); + + where = string2; + } + + for (this_char = where - string2; this_char < size2; this_char++) + putchar (string2[this_char]); + } +} + +#else /* not DEBUG */ + +#undef assert +#define assert(e) + +#define DEBUG_STATEMENT(e) +#define DEBUG_PRINT1(x) +#define DEBUG_PRINT2(x1, x2) +#define DEBUG_PRINT3(x1, x2, x3) +#define DEBUG_PRINT4(x1, x2, x3, x4) +#define DEBUG_PRINT_COMPILED_PATTERN(p, s, e) +#define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2) + +#endif /* not DEBUG */ + +/* Set by `re_set_syntax' to the current regexp syntax to recognize. Can + also be assigned to arbitrarily: each pattern buffer stores its own + syntax, so it can be changed between regex compilations. */ +/* This has no initializer because initialized variables in Emacs + become read-only after dumping. */ +reg_syntax_t re_syntax_options; + + +/* Specify the precise syntax of regexps for compilation. This provides + for compatibility for various utilities which historically have + different, incompatible syntaxes. + + The argument SYNTAX is a bit mask comprised of the various bits + defined in regex.h. We return the old syntax. */ + +reg_syntax_t +re_set_syntax (syntax) + reg_syntax_t syntax; +{ + reg_syntax_t ret = re_syntax_options; + + re_syntax_options = syntax; + return ret; +} + +/* This table gives an error message for each of the error codes listed + in regex.h. Obviously the order here has to be same as there. + POSIX doesn't require that we do anything for REG_NOERROR, + but why not be nice? */ + +static const char *re_error_msgid[] = + { + gettext_noop ("Success"), /* REG_NOERROR */ + gettext_noop ("No match"), /* REG_NOMATCH */ + gettext_noop ("Invalid regular expression"), /* REG_BADPAT */ + gettext_noop ("Invalid collation character"), /* REG_ECOLLATE */ + gettext_noop ("Invalid character class name"), /* REG_ECTYPE */ + gettext_noop ("Trailing backslash"), /* REG_EESCAPE */ + gettext_noop ("Invalid back reference"), /* REG_ESUBREG */ + gettext_noop ("Unmatched [ or [^"), /* REG_EBRACK */ + gettext_noop ("Unmatched ( or \\("), /* REG_EPAREN */ + gettext_noop ("Unmatched \\{"), /* REG_EBRACE */ + gettext_noop ("Invalid content of \\{\\}"), /* REG_BADBR */ + gettext_noop ("Invalid range end"), /* REG_ERANGE */ + gettext_noop ("Memory exhausted"), /* REG_ESPACE */ + gettext_noop ("Invalid preceding regular expression"), /* REG_BADRPT */ + gettext_noop ("Premature end of regular expression"), /* REG_EEND */ + gettext_noop ("Regular expression too big"), /* REG_ESIZE */ + gettext_noop ("Unmatched ) or \\)"), /* REG_ERPAREN */ + }; + +/* Avoiding alloca during matching, to placate r_alloc. */ + +/* Define MATCH_MAY_ALLOCATE unless we need to make sure that the + searching and matching functions should not call alloca. On some + systems, alloca is implemented in terms of malloc, and if we're + using the relocating allocator routines, then malloc could cause a + relocation, which might (if the strings being searched are in the + ralloc heap) shift the data out from underneath the regexp + routines. + + Here's another reason to avoid allocation: Emacs + processes input from X in a signal handler; processing X input may + call malloc; if input arrives while a matching routine is calling + malloc, then we're scrod. But Emacs can't just block input while + calling matching routines; then we don't notice interrupts when + they come in. So, Emacs blocks input around all regexp calls + except the matching calls, which it leaves unprotected, in the + faith that they will not malloc. */ + +/* Normally, this is fine. */ +#define MATCH_MAY_ALLOCATE + +/* When using GNU C, we are not REALLY using the C alloca, no matter + what config.h may say. So don't take precautions for it. */ +#ifdef __GNUC__ +#undef C_ALLOCA +#endif + +/* The match routines may not allocate if (1) they would do it with malloc + and (2) it's not safe for them to use malloc. + Note that if REL_ALLOC is defined, matching would not use malloc for the + failure stack, but we would still use it for the register vectors; + so REL_ALLOC should not affect this. */ +#if (defined (C_ALLOCA) || defined (REGEX_MALLOC)) && defined (emacs) +#undef MATCH_MAY_ALLOCATE +#endif + + +/* Failure stack declarations and macros; both re_compile_fastmap and + re_match_2 use a failure stack. These have to be macros because of + REGEX_ALLOCATE_STACK. */ + + +/* Approximate number of failure points for which to initially allocate space + when matching. If this number is exceeded, we allocate more + space, so it is not a hard limit. */ +#ifndef INIT_FAILURE_ALLOC +#define INIT_FAILURE_ALLOC 20 +#endif + +/* Roughly the maximum number of failure points on the stack. Would be + exactly that if always used TYPICAL_FAILURE_SIZE items each time we failed. + This is a variable only so users of regex can assign to it; we never + change it ourselves. */ +#if defined (MATCH_MAY_ALLOCATE) +/* Note that 4400 is enough to cause a crash on Alpha OSF/1, + whose default stack limit is 2mb. In order for a larger + value to work reliably, you have to try to make it accord + with the process stack limit. */ +int re_max_failures = 40000; +#else +int re_max_failures = 4000; +#endif + +union fail_stack_elt +{ + unsigned char *pointer; + int integer; +}; + +typedef union fail_stack_elt fail_stack_elt_t; + +typedef struct +{ + fail_stack_elt_t *stack; + unsigned size; + unsigned avail; /* Offset of next open position. */ +} fail_stack_type; + +#define FAIL_STACK_EMPTY() (fail_stack.avail == 0) +#define FAIL_STACK_PTR_EMPTY() (fail_stack_ptr->avail == 0) +#define FAIL_STACK_FULL() (fail_stack.avail == fail_stack.size) + + +/* Define macros to initialize and free the failure stack. + Do `return -2' if the alloc fails. */ + +#ifdef MATCH_MAY_ALLOCATE +#define INIT_FAIL_STACK() \ + do { \ + fail_stack.stack = (fail_stack_elt_t *) \ + REGEX_ALLOCATE_STACK (INIT_FAILURE_ALLOC * TYPICAL_FAILURE_SIZE \ + * sizeof (fail_stack_elt_t)); \ + \ + if (fail_stack.stack == NULL) \ + return -2; \ + \ + fail_stack.size = INIT_FAILURE_ALLOC; \ + fail_stack.avail = 0; \ + } while (0) + +#define RESET_FAIL_STACK() REGEX_FREE_STACK (fail_stack.stack) +#else +#define INIT_FAIL_STACK() \ + do { \ + fail_stack.avail = 0; \ + } while (0) + +#define RESET_FAIL_STACK() +#endif + + +/* Double the size of FAIL_STACK, up to a limit + which allows approximately `re_max_failures' items. + + Return 1 if succeeds, and 0 if either ran out of memory + allocating space for it or it was already too large. + + REGEX_REALLOCATE_STACK requires `destination' be declared. */ + +/* Factor to increase the failure stack size by + when we increase it. + This used to be 2, but 2 was too wasteful + because the old discarded stacks added up to as much space + were as ultimate, maximum-size stack. */ +#define FAIL_STACK_GROWTH_FACTOR 4 + +#define GROW_FAIL_STACK(fail_stack) \ + (((fail_stack).size * sizeof (fail_stack_elt_t) \ + >= re_max_failures * TYPICAL_FAILURE_SIZE) \ + ? 0 \ + : ((fail_stack).stack \ + = (fail_stack_elt_t *) \ + REGEX_REALLOCATE_STACK ((fail_stack).stack, \ + (fail_stack).size * sizeof (fail_stack_elt_t), \ + MIN (re_max_failures * TYPICAL_FAILURE_SIZE, \ + ((fail_stack).size * sizeof (fail_stack_elt_t) \ + * FAIL_STACK_GROWTH_FACTOR))), \ + \ + (fail_stack).stack == NULL \ + ? 0 \ + : ((fail_stack).size \ + = (MIN (re_max_failures * TYPICAL_FAILURE_SIZE, \ + ((fail_stack).size * sizeof (fail_stack_elt_t) \ + * FAIL_STACK_GROWTH_FACTOR)) \ + / sizeof (fail_stack_elt_t)), \ + 1))) + + +/* Push pointer POINTER on FAIL_STACK. + Return 1 if was able to do so and 0 if ran out of memory allocating + space to do so. */ +#define PUSH_PATTERN_OP(POINTER, FAIL_STACK) \ + ((FAIL_STACK_FULL () \ + && !GROW_FAIL_STACK (FAIL_STACK)) \ + ? 0 \ + : ((FAIL_STACK).stack[(FAIL_STACK).avail++].pointer = POINTER, \ + 1)) + +/* Push a pointer value onto the failure stack. + Assumes the variable `fail_stack'. Probably should only + be called from within `PUSH_FAILURE_POINT'. */ +#define PUSH_FAILURE_POINTER(item) \ + fail_stack.stack[fail_stack.avail++].pointer = (unsigned char *) (item) + +/* This pushes an integer-valued item onto the failure stack. + Assumes the variable `fail_stack'. Probably should only + be called from within `PUSH_FAILURE_POINT'. */ +#define PUSH_FAILURE_INT(item) \ + fail_stack.stack[fail_stack.avail++].integer = (item) + +/* Push a fail_stack_elt_t value onto the failure stack. + Assumes the variable `fail_stack'. Probably should only + be called from within `PUSH_FAILURE_POINT'. */ +#define PUSH_FAILURE_ELT(item) \ + fail_stack.stack[fail_stack.avail++] = (item) + +/* These three POP... operations complement the three PUSH... operations. + All assume that `fail_stack' is nonempty. */ +#define POP_FAILURE_POINTER() fail_stack.stack[--fail_stack.avail].pointer +#define POP_FAILURE_INT() fail_stack.stack[--fail_stack.avail].integer +#define POP_FAILURE_ELT() fail_stack.stack[--fail_stack.avail] + +/* Used to omit pushing failure point id's when we're not debugging. */ +#ifdef DEBUG +#define DEBUG_PUSH PUSH_FAILURE_INT +#define DEBUG_POP(item_addr) *(item_addr) = POP_FAILURE_INT () +#else +#define DEBUG_PUSH(item) +#define DEBUG_POP(item_addr) +#endif + + +/* Push the information about the state we will need + if we ever fail back to it. + + Requires variables fail_stack, regstart, regend, reg_info, and + num_regs be declared. GROW_FAIL_STACK requires `destination' be + declared. + + Does `return FAILURE_CODE' if runs out of memory. */ + +#define PUSH_FAILURE_POINT(pattern_place, string_place, failure_code) \ + do { \ + char *destination; \ + /* Must be int, so when we don't save any registers, the arithmetic \ + of 0 + -1 isn't done as unsigned. */ \ + int this_reg; \ + \ + DEBUG_STATEMENT (failure_id++); \ + DEBUG_STATEMENT (nfailure_points_pushed++); \ + DEBUG_PRINT2 ("\nPUSH_FAILURE_POINT #%u:\n", failure_id); \ + DEBUG_PRINT2 (" Before push, next avail: %d\n", (fail_stack).avail);\ + DEBUG_PRINT2 (" size: %d\n", (fail_stack).size);\ + \ + DEBUG_PRINT2 (" slots needed: %d\n", NUM_FAILURE_ITEMS); \ + DEBUG_PRINT2 (" available: %d\n", REMAINING_AVAIL_SLOTS); \ + \ + /* Ensure we have enough space allocated for what we will push. */ \ + while (REMAINING_AVAIL_SLOTS < NUM_FAILURE_ITEMS) \ + { \ + if (!GROW_FAIL_STACK (fail_stack)) \ + return failure_code; \ + \ + DEBUG_PRINT2 ("\n Doubled stack; size now: %d\n", \ + (fail_stack).size); \ + DEBUG_PRINT2 (" slots available: %d\n", REMAINING_AVAIL_SLOTS);\ + } \ + \ + /* Push the info, starting with the registers. */ \ + DEBUG_PRINT1 ("\n"); \ + \ + if (1) \ + for (this_reg = lowest_active_reg; this_reg <= highest_active_reg; \ + this_reg++) \ + { \ + DEBUG_PRINT2 (" Pushing reg: %d\n", this_reg); \ + DEBUG_STATEMENT (num_regs_pushed++); \ + \ + DEBUG_PRINT2 (" start: 0x%x\n", regstart[this_reg]); \ + PUSH_FAILURE_POINTER (regstart[this_reg]); \ + \ + DEBUG_PRINT2 (" end: 0x%x\n", regend[this_reg]); \ + PUSH_FAILURE_POINTER (regend[this_reg]); \ + \ + DEBUG_PRINT2 (" info: 0x%x\n ", reg_info[this_reg]); \ + DEBUG_PRINT2 (" match_null=%d", \ + REG_MATCH_NULL_STRING_P (reg_info[this_reg])); \ + DEBUG_PRINT2 (" active=%d", IS_ACTIVE (reg_info[this_reg])); \ + DEBUG_PRINT2 (" matched_something=%d", \ + MATCHED_SOMETHING (reg_info[this_reg])); \ + DEBUG_PRINT2 (" ever_matched=%d", \ + EVER_MATCHED_SOMETHING (reg_info[this_reg])); \ + DEBUG_PRINT1 ("\n"); \ + PUSH_FAILURE_ELT (reg_info[this_reg].word); \ + } \ + \ + DEBUG_PRINT2 (" Pushing low active reg: %d\n", lowest_active_reg);\ + PUSH_FAILURE_INT (lowest_active_reg); \ + \ + DEBUG_PRINT2 (" Pushing high active reg: %d\n", highest_active_reg);\ + PUSH_FAILURE_INT (highest_active_reg); \ + \ + DEBUG_PRINT2 (" Pushing pattern 0x%x: ", pattern_place); \ + DEBUG_PRINT_COMPILED_PATTERN (bufp, pattern_place, pend); \ + PUSH_FAILURE_POINTER (pattern_place); \ + \ + DEBUG_PRINT2 (" Pushing string 0x%x: `", string_place); \ + DEBUG_PRINT_DOUBLE_STRING (string_place, string1, size1, string2, \ + size2); \ + DEBUG_PRINT1 ("'\n"); \ + PUSH_FAILURE_POINTER (string_place); \ + \ + DEBUG_PRINT2 (" Pushing failure id: %u\n", failure_id); \ + DEBUG_PUSH (failure_id); \ + } while (0) + +/* This is the number of items that are pushed and popped on the stack + for each register. */ +#define NUM_REG_ITEMS 3 + +/* Individual items aside from the registers. */ +#ifdef DEBUG +#define NUM_NONREG_ITEMS 5 /* Includes failure point id. */ +#else +#define NUM_NONREG_ITEMS 4 +#endif + +/* Estimate the size of data pushed by a typical failure stack entry. + An estimate is all we need, because all we use this for + is to choose a limit for how big to make the failure stack. */ + +#define TYPICAL_FAILURE_SIZE 20 + +/* This is how many items we actually use for a failure point. + It depends on the regexp. */ +#define NUM_FAILURE_ITEMS \ + (((0 \ + ? 0 : highest_active_reg - lowest_active_reg + 1) \ + * NUM_REG_ITEMS) \ + + NUM_NONREG_ITEMS) + +/* How many items can still be added to the stack without overflowing it. */ +#define REMAINING_AVAIL_SLOTS ((fail_stack).size - (fail_stack).avail) + + +/* Pops what PUSH_FAIL_STACK pushes. + + We restore into the parameters, all of which should be lvalues: + STR -- the saved data position. + PAT -- the saved pattern position. + LOW_REG, HIGH_REG -- the highest and lowest active registers. + REGSTART, REGEND -- arrays of string positions. + REG_INFO -- array of information about each subexpression. + + Also assumes the variables `fail_stack' and (if debugging), `bufp', + `pend', `string1', `size1', `string2', and `size2'. */ + +#define POP_FAILURE_POINT(str, pat, low_reg, high_reg, regstart, regend, reg_info)\ +{ \ + DEBUG_STATEMENT (fail_stack_elt_t failure_id;) \ + int this_reg; \ + const unsigned char *string_temp; \ + \ + assert (!FAIL_STACK_EMPTY ()); \ + \ + /* Remove failure points and point to how many regs pushed. */ \ + DEBUG_PRINT1 ("POP_FAILURE_POINT:\n"); \ + DEBUG_PRINT2 (" Before pop, next avail: %d\n", fail_stack.avail); \ + DEBUG_PRINT2 (" size: %d\n", fail_stack.size); \ + \ + assert (fail_stack.avail >= NUM_NONREG_ITEMS); \ + \ + DEBUG_POP (&failure_id); \ + DEBUG_PRINT2 (" Popping failure id: %u\n", failure_id); \ + \ + /* If the saved string location is NULL, it came from an \ + on_failure_keep_string_jump opcode, and we want to throw away the \ + saved NULL, thus retaining our current position in the string. */ \ + string_temp = POP_FAILURE_POINTER (); \ + if (string_temp != NULL) \ + str = (const char *) string_temp; \ + \ + DEBUG_PRINT2 (" Popping string 0x%x: `", str); \ + DEBUG_PRINT_DOUBLE_STRING (str, string1, size1, string2, size2); \ + DEBUG_PRINT1 ("'\n"); \ + \ + pat = (unsigned char *) POP_FAILURE_POINTER (); \ + DEBUG_PRINT2 (" Popping pattern 0x%x: ", pat); \ + DEBUG_PRINT_COMPILED_PATTERN (bufp, pat, pend); \ + \ + /* Restore register info. */ \ + high_reg = (unsigned) POP_FAILURE_INT (); \ + DEBUG_PRINT2 (" Popping high active reg: %d\n", high_reg); \ + \ + low_reg = (unsigned) POP_FAILURE_INT (); \ + DEBUG_PRINT2 (" Popping low active reg: %d\n", low_reg); \ + \ + if (1) \ + for (this_reg = high_reg; this_reg >= low_reg; this_reg--) \ + { \ + DEBUG_PRINT2 (" Popping reg: %d\n", this_reg); \ + \ + reg_info[this_reg].word = POP_FAILURE_ELT (); \ + DEBUG_PRINT2 (" info: 0x%x\n", reg_info[this_reg]); \ + \ + regend[this_reg] = (const char *) POP_FAILURE_POINTER (); \ + DEBUG_PRINT2 (" end: 0x%x\n", regend[this_reg]); \ + \ + regstart[this_reg] = (const char *) POP_FAILURE_POINTER (); \ + DEBUG_PRINT2 (" start: 0x%x\n", regstart[this_reg]); \ + } \ + else \ + { \ + for (this_reg = highest_active_reg; this_reg > high_reg; this_reg--) \ + { \ + reg_info[this_reg].word.integer = 0; \ + regend[this_reg] = 0; \ + regstart[this_reg] = 0; \ + } \ + highest_active_reg = high_reg; \ + } \ + \ + set_regs_matched_done = 0; \ + DEBUG_STATEMENT (nfailure_points_popped++); \ +} /* POP_FAILURE_POINT */ + + + +/* Structure for per-register (a.k.a. per-group) information. + Other register information, such as the + starting and ending positions (which are addresses), and the list of + inner groups (which is a bits list) are maintained in separate + variables. + + We are making a (strictly speaking) nonportable assumption here: that + the compiler will pack our bit fields into something that fits into + the type of `word', i.e., is something that fits into one item on the + failure stack. */ + +typedef union +{ + fail_stack_elt_t word; + struct + { + /* This field is one if this group can match the empty string, + zero if not. If not yet determined, `MATCH_NULL_UNSET_VALUE'. */ +#define MATCH_NULL_UNSET_VALUE 3 + unsigned match_null_string_p : 2; + unsigned is_active : 1; + unsigned matched_something : 1; + unsigned ever_matched_something : 1; + } bits; +} register_info_type; + +#define REG_MATCH_NULL_STRING_P(R) ((R).bits.match_null_string_p) +#define IS_ACTIVE(R) ((R).bits.is_active) +#define MATCHED_SOMETHING(R) ((R).bits.matched_something) +#define EVER_MATCHED_SOMETHING(R) ((R).bits.ever_matched_something) + + +/* Call this when have matched a real character; it sets `matched' flags + for the subexpressions which we are currently inside. Also records + that those subexprs have matched. */ +#define SET_REGS_MATCHED() \ + do \ + { \ + if (!set_regs_matched_done) \ + { \ + unsigned r; \ + set_regs_matched_done = 1; \ + for (r = lowest_active_reg; r <= highest_active_reg; r++) \ + { \ + MATCHED_SOMETHING (reg_info[r]) \ + = EVER_MATCHED_SOMETHING (reg_info[r]) \ + = 1; \ + } \ + } \ + } \ + while (0) + +/* Registers are set to a sentinel when they haven't yet matched. */ +static char reg_unset_dummy; +#define REG_UNSET_VALUE (®_unset_dummy) +#define REG_UNSET(e) ((e) == REG_UNSET_VALUE) + +/* Subroutine declarations and macros for regex_compile. */ + +static void store_op1 (), store_op2 (); +static void insert_op1 (), insert_op2 (); +static boolean at_begline_loc_p (), at_endline_loc_p (); +static boolean group_in_compile_stack (); +static reg_errcode_t compile_range (); + +/* Fetch the next character in the uncompiled pattern---translating it + if necessary. Also cast from a signed character in the constant + string passed to us by the user to an unsigned char that we can use + as an array index (in, e.g., `translate'). */ +#ifndef PATFETCH +#define PATFETCH(c) \ + do {if (p == pend) return REG_EEND; \ + c = (unsigned char) *p++; \ + if (RE_TRANSLATE_P (translate)) c = RE_TRANSLATE (translate, c); \ + } while (0) +#endif + +/* Fetch the next character in the uncompiled pattern, with no + translation. */ +#define PATFETCH_RAW(c) \ + do {if (p == pend) return REG_EEND; \ + c = (unsigned char) *p++; \ + } while (0) + +/* Go backwards one character in the pattern. */ +#define PATUNFETCH p-- + + +/* If `translate' is non-null, return translate[D], else just D. We + cast the subscript to translate because some data is declared as + `char *', to avoid warnings when a string constant is passed. But + when we use a character as a subscript we must make it unsigned. */ +#ifndef TRANSLATE +#define TRANSLATE(d) \ + (RE_TRANSLATE_P (translate) \ + ? (unsigned) RE_TRANSLATE (translate, (unsigned) (d)) : (d)) +#endif + + +/* Macros for outputting the compiled pattern into `buffer'. */ + +/* If the buffer isn't allocated when it comes in, use this. */ +#define INIT_BUF_SIZE 32 + +/* Make sure we have at least N more bytes of space in buffer. */ +#define GET_BUFFER_SPACE(n) \ + while (b - bufp->buffer + (n) > bufp->allocated) \ + EXTEND_BUFFER () + +/* Make sure we have one more byte of buffer space and then add C to it. */ +#define BUF_PUSH(c) \ + do { \ + GET_BUFFER_SPACE (1); \ + *b++ = (unsigned char) (c); \ + } while (0) + + +/* Ensure we have two more bytes of buffer space and then append C1 and C2. */ +#define BUF_PUSH_2(c1, c2) \ + do { \ + GET_BUFFER_SPACE (2); \ + *b++ = (unsigned char) (c1); \ + *b++ = (unsigned char) (c2); \ + } while (0) + + +/* As with BUF_PUSH_2, except for three bytes. */ +#define BUF_PUSH_3(c1, c2, c3) \ + do { \ + GET_BUFFER_SPACE (3); \ + *b++ = (unsigned char) (c1); \ + *b++ = (unsigned char) (c2); \ + *b++ = (unsigned char) (c3); \ + } while (0) + + +/* Store a jump with opcode OP at LOC to location TO. We store a + relative address offset by the three bytes the jump itself occupies. */ +#define STORE_JUMP(op, loc, to) \ + store_op1 (op, loc, (to) - (loc) - 3) + +/* Likewise, for a two-argument jump. */ +#define STORE_JUMP2(op, loc, to, arg) \ + store_op2 (op, loc, (to) - (loc) - 3, arg) + +/* Like `STORE_JUMP', but for inserting. Assume `b' is the buffer end. */ +#define INSERT_JUMP(op, loc, to) \ + insert_op1 (op, loc, (to) - (loc) - 3, b) + +/* Like `STORE_JUMP2', but for inserting. Assume `b' is the buffer end. */ +#define INSERT_JUMP2(op, loc, to, arg) \ + insert_op2 (op, loc, (to) - (loc) - 3, arg, b) + + +/* This is not an arbitrary limit: the arguments which represent offsets + into the pattern are two bytes long. So if 2^16 bytes turns out to + be too small, many things would have to change. */ +#define MAX_BUF_SIZE (1L << 16) + + +/* Extend the buffer by twice its current size via realloc and + reset the pointers that pointed into the old block to point to the + correct places in the new one. If extending the buffer results in it + being larger than MAX_BUF_SIZE, then flag memory exhausted. */ +#define EXTEND_BUFFER() \ + do { \ + unsigned char *old_buffer = bufp->buffer; \ + if (bufp->allocated == MAX_BUF_SIZE) \ + return REG_ESIZE; \ + bufp->allocated <<= 1; \ + if (bufp->allocated > MAX_BUF_SIZE) \ + bufp->allocated = MAX_BUF_SIZE; \ + bufp->buffer = (unsigned char *) realloc (bufp->buffer, bufp->allocated);\ + if (bufp->buffer == NULL) \ + return REG_ESPACE; \ + /* If the buffer moved, move all the pointers into it. */ \ + if (old_buffer != bufp->buffer) \ + { \ + b = (b - old_buffer) + bufp->buffer; \ + begalt = (begalt - old_buffer) + bufp->buffer; \ + if (fixup_alt_jump) \ + fixup_alt_jump = (fixup_alt_jump - old_buffer) + bufp->buffer;\ + if (laststart) \ + laststart = (laststart - old_buffer) + bufp->buffer; \ + if (pending_exact) \ + pending_exact = (pending_exact - old_buffer) + bufp->buffer; \ + } \ + } while (0) + + +/* Since we have one byte reserved for the register number argument to + {start,stop}_memory, the maximum number of groups we can report + things about is what fits in that byte. */ +#define MAX_REGNUM 255 + +/* But patterns can have more than `MAX_REGNUM' registers. We just + ignore the excess. */ +typedef unsigned regnum_t; + + +/* Macros for the compile stack. */ + +/* Since offsets can go either forwards or backwards, this type needs to + be able to hold values from -(MAX_BUF_SIZE - 1) to MAX_BUF_SIZE - 1. */ +typedef int pattern_offset_t; + +typedef struct +{ + pattern_offset_t begalt_offset; + pattern_offset_t fixup_alt_jump; + pattern_offset_t inner_group_offset; + pattern_offset_t laststart_offset; + regnum_t regnum; +} compile_stack_elt_t; + + +typedef struct +{ + compile_stack_elt_t *stack; + unsigned size; + unsigned avail; /* Offset of next open position. */ +} compile_stack_type; + + +#define INIT_COMPILE_STACK_SIZE 32 + +#define COMPILE_STACK_EMPTY (compile_stack.avail == 0) +#define COMPILE_STACK_FULL (compile_stack.avail == compile_stack.size) + +/* The next available element. */ +#define COMPILE_STACK_TOP (compile_stack.stack[compile_stack.avail]) + + +/* Structure to manage work area for range table. */ +struct range_table_work_area +{ + int *table; /* actual work area. */ + int allocated; /* allocated size for work area in bytes. */ + int used; /* actually used size in words. */ +}; + +/* Make sure that WORK_AREA can hold more N multibyte characters. */ +#define EXTEND_RANGE_TABLE_WORK_AREA(work_area, n) \ + do { \ + if (((work_area).used + (n)) * sizeof (int) > (work_area).allocated) \ + { \ + (work_area).allocated += 16 * sizeof (int); \ + if ((work_area).table) \ + (work_area).table \ + = (int *) realloc ((work_area).table, (work_area).allocated); \ + else \ + (work_area).table \ + = (int *) malloc ((work_area).allocated); \ + if ((work_area).table == 0) \ + FREE_STACK_RETURN (REG_ESPACE); \ + } \ + } while (0) + +/* Set a range (RANGE_START, RANGE_END) to WORK_AREA. */ +#define SET_RANGE_TABLE_WORK_AREA(work_area, range_start, range_end) \ + do { \ + EXTEND_RANGE_TABLE_WORK_AREA ((work_area), 2); \ + (work_area).table[(work_area).used++] = (range_start); \ + (work_area).table[(work_area).used++] = (range_end); \ + } while (0) + +/* Free allocated memory for WORK_AREA. */ +#define FREE_RANGE_TABLE_WORK_AREA(work_area) \ + do { \ + if ((work_area).table) \ + free ((work_area).table); \ + } while (0) + +#define CLEAR_RANGE_TABLE_WORK_USED(work_area) ((work_area).used = 0) +#define RANGE_TABLE_WORK_USED(work_area) ((work_area).used) +#define RANGE_TABLE_WORK_ELT(work_area, i) ((work_area).table[i]) + + +/* Set the bit for character C in a list. */ +#define SET_LIST_BIT(c) \ + (b[((unsigned char) (c)) / BYTEWIDTH] \ + |= 1 << (((unsigned char) c) % BYTEWIDTH)) + + +/* Get the next unsigned number in the uncompiled pattern. */ +#define GET_UNSIGNED_NUMBER(num) \ + { if (p != pend) \ + { \ + PATFETCH (c); \ + while (ISDIGIT (c)) \ + { \ + if (num < 0) \ + num = 0; \ + num = num * 10 + c - '0'; \ + if (p == pend) \ + break; \ + PATFETCH (c); \ + } \ + } \ + } + +#define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */ + +#define IS_CHAR_CLASS(string) \ + (STREQ (string, "alpha") || STREQ (string, "upper") \ + || STREQ (string, "lower") || STREQ (string, "digit") \ + || STREQ (string, "alnum") || STREQ (string, "xdigit") \ + || STREQ (string, "space") || STREQ (string, "print") \ + || STREQ (string, "punct") || STREQ (string, "graph") \ + || STREQ (string, "cntrl") || STREQ (string, "blank")) + +#ifndef MATCH_MAY_ALLOCATE + +/* If we cannot allocate large objects within re_match_2_internal, + we make the fail stack and register vectors global. + The fail stack, we grow to the maximum size when a regexp + is compiled. + The register vectors, we adjust in size each time we + compile a regexp, according to the number of registers it needs. */ + +static fail_stack_type fail_stack; + +/* Size with which the following vectors are currently allocated. + That is so we can make them bigger as needed, + but never make them smaller. */ +static int regs_allocated_size; + +static const char ** regstart, ** regend; +static const char ** old_regstart, ** old_regend; +static const char **best_regstart, **best_regend; +static register_info_type *reg_info; +static const char **reg_dummy; +static register_info_type *reg_info_dummy; + +/* Make the register vectors big enough for NUM_REGS registers, + but don't make them smaller. */ + +static +regex_grow_registers (num_regs) + int num_regs; +{ + if (num_regs > regs_allocated_size) + { + RETALLOC_IF (regstart, num_regs, const char *); + RETALLOC_IF (regend, num_regs, const char *); + RETALLOC_IF (old_regstart, num_regs, const char *); + RETALLOC_IF (old_regend, num_regs, const char *); + RETALLOC_IF (best_regstart, num_regs, const char *); + RETALLOC_IF (best_regend, num_regs, const char *); + RETALLOC_IF (reg_info, num_regs, register_info_type); + RETALLOC_IF (reg_dummy, num_regs, const char *); + RETALLOC_IF (reg_info_dummy, num_regs, register_info_type); + + regs_allocated_size = num_regs; + } +} + +#endif /* not MATCH_MAY_ALLOCATE */ + +/* `regex_compile' compiles PATTERN (of length SIZE) according to SYNTAX. + Returns one of error codes defined in `regex.h', or zero for success. + + Assumes the `allocated' (and perhaps `buffer') and `translate' + fields are set in BUFP on entry. + + If it succeeds, results are put in BUFP (if it returns an error, the + contents of BUFP are undefined): + `buffer' is the compiled pattern; + `syntax' is set to SYNTAX; + `used' is set to the length of the compiled pattern; + `fastmap_accurate' is zero; + `re_nsub' is the number of subexpressions in PATTERN; + `not_bol' and `not_eol' are zero; + + The `fastmap' and `newline_anchor' fields are neither + examined nor set. */ + +/* Return, freeing storage we allocated. */ +#define FREE_STACK_RETURN(value) \ + do { \ + FREE_RANGE_TABLE_WORK_AREA (range_table_work); \ + free (compile_stack.stack); \ + return value; \ + } while (0) + +static reg_errcode_t +regex_compile (pattern, size, syntax, bufp) + const char *pattern; + int size; + reg_syntax_t syntax; + struct re_pattern_buffer *bufp; +{ + /* We fetch characters from PATTERN here. Even though PATTERN is + `char *' (i.e., signed), we declare these variables as unsigned, so + they can be reliably used as array indices. */ + register unsigned int c, c1; + + /* A random temporary spot in PATTERN. */ + const char *p1; + + /* Points to the end of the buffer, where we should append. */ + register unsigned char *b; + + /* Keeps track of unclosed groups. */ + compile_stack_type compile_stack; + + /* Points to the current (ending) position in the pattern. */ +#ifdef AIX + /* `const' makes AIX compiler fail. */ + char *p = pattern; +#else + const char *p = pattern; +#endif + const char *pend = pattern + size; + + /* How to translate the characters in the pattern. */ + RE_TRANSLATE_TYPE translate = bufp->translate; + + /* Address of the count-byte of the most recently inserted `exactn' + command. This makes it possible to tell if a new exact-match + character can be added to that command or if the character requires + a new `exactn' command. */ + unsigned char *pending_exact = 0; + + /* Address of start of the most recently finished expression. + This tells, e.g., postfix * where to find the start of its + operand. Reset at the beginning of groups and alternatives. */ + unsigned char *laststart = 0; + + /* Address of beginning of regexp, or inside of last group. */ + unsigned char *begalt; + + /* Place in the uncompiled pattern (i.e., the {) to + which to go back if the interval is invalid. */ + const char *beg_interval; + + /* Address of the place where a forward jump should go to the end of + the containing expression. Each alternative of an `or' -- except the + last -- ends with a forward jump of this sort. */ + unsigned char *fixup_alt_jump = 0; + + /* Counts open-groups as they are encountered. Remembered for the + matching close-group on the compile stack, so the same register + number is put in the stop_memory as the start_memory. */ + regnum_t regnum = 0; + + /* Work area for range table of charset. */ + struct range_table_work_area range_table_work; + +#ifdef DEBUG + DEBUG_PRINT1 ("\nCompiling pattern: "); + if (debug) + { + unsigned debug_count; + + for (debug_count = 0; debug_count < size; debug_count++) + putchar (pattern[debug_count]); + putchar ('\n'); + } +#endif /* DEBUG */ + + /* Initialize the compile stack. */ + compile_stack.stack = TALLOC (INIT_COMPILE_STACK_SIZE, compile_stack_elt_t); + if (compile_stack.stack == NULL) + return REG_ESPACE; + + compile_stack.size = INIT_COMPILE_STACK_SIZE; + compile_stack.avail = 0; + + range_table_work.table = 0; + range_table_work.allocated = 0; + + /* Initialize the pattern buffer. */ + bufp->syntax = syntax; + bufp->fastmap_accurate = 0; + bufp->not_bol = bufp->not_eol = 0; + + /* Set `used' to zero, so that if we return an error, the pattern + printer (for debugging) will think there's no pattern. We reset it + at the end. */ + bufp->used = 0; + + /* Always count groups, whether or not bufp->no_sub is set. */ + bufp->re_nsub = 0; + +#ifdef emacs + /* bufp->multibyte is set before regex_compile is called, so don't alter + it. */ +#else /* not emacs */ + /* Nothing is recognized as a multibyte character. */ + bufp->multibyte = 0; +#endif + +#if !defined (emacs) && !defined (SYNTAX_TABLE) + /* Initialize the syntax table. */ + init_syntax_once (); +#endif + + if (bufp->allocated == 0) + { + if (bufp->buffer) + { /* If zero allocated, but buffer is non-null, try to realloc + enough space. This loses if buffer's address is bogus, but + that is the user's responsibility. */ + RETALLOC (bufp->buffer, INIT_BUF_SIZE, unsigned char); + } + else + { /* Caller did not allocate a buffer. Do it for them. */ + bufp->buffer = TALLOC (INIT_BUF_SIZE, unsigned char); + } + if (!bufp->buffer) FREE_STACK_RETURN (REG_ESPACE); + + bufp->allocated = INIT_BUF_SIZE; + } + + begalt = b = bufp->buffer; + + /* Loop through the uncompiled pattern until we're at the end. */ + while (p != pend) + { + PATFETCH (c); + + switch (c) + { + case '^': + { + if ( /* If at start of pattern, it's an operator. */ + p == pattern + 1 + /* If context independent, it's an operator. */ + || syntax & RE_CONTEXT_INDEP_ANCHORS + /* Otherwise, depends on what's come before. */ + || at_begline_loc_p (pattern, p, syntax)) + BUF_PUSH (begline); + else + goto normal_char; + } + break; + + + case '$': + { + if ( /* If at end of pattern, it's an operator. */ + p == pend + /* If context independent, it's an operator. */ + || syntax & RE_CONTEXT_INDEP_ANCHORS + /* Otherwise, depends on what's next. */ + || at_endline_loc_p (p, pend, syntax)) + BUF_PUSH (endline); + else + goto normal_char; + } + break; + + + case '+': + case '?': + if ((syntax & RE_BK_PLUS_QM) + || (syntax & RE_LIMITED_OPS)) + goto normal_char; + handle_plus: + case '*': + /* If there is no previous pattern... */ + if (!laststart) + { + if (syntax & RE_CONTEXT_INVALID_OPS) + FREE_STACK_RETURN (REG_BADRPT); + else if (!(syntax & RE_CONTEXT_INDEP_OPS)) + goto normal_char; + } + + { + /* Are we optimizing this jump? */ + boolean keep_string_p = false; + + /* 1 means zero (many) matches is allowed. */ + char zero_times_ok = 0, many_times_ok = 0; + + /* If there is a sequence of repetition chars, collapse it + down to just one (the right one). We can't combine + interval operators with these because of, e.g., `a{2}*', + which should only match an even number of `a's. */ + + for (;;) + { + zero_times_ok |= c != '+'; + many_times_ok |= c != '?'; + + if (p == pend) + break; + + PATFETCH (c); + + if (c == '*' + || (!(syntax & RE_BK_PLUS_QM) && (c == '+' || c == '?'))) + ; + + else if (syntax & RE_BK_PLUS_QM && c == '\\') + { + if (p == pend) FREE_STACK_RETURN (REG_EESCAPE); + + PATFETCH (c1); + if (!(c1 == '+' || c1 == '?')) + { + PATUNFETCH; + PATUNFETCH; + break; + } + + c = c1; + } + else + { + PATUNFETCH; + break; + } + + /* If we get here, we found another repeat character. */ + } + + /* Star, etc. applied to an empty pattern is equivalent + to an empty pattern. */ + if (!laststart) + break; + + /* Now we know whether or not zero matches is allowed + and also whether or not two or more matches is allowed. */ + if (many_times_ok) + { /* More than one repetition is allowed, so put in at the + end a backward relative jump from `b' to before the next + jump we're going to put in below (which jumps from + laststart to after this jump). + + But if we are at the `*' in the exact sequence `.*\n', + insert an unconditional jump backwards to the ., + instead of the beginning of the loop. This way we only + push a failure point once, instead of every time + through the loop. */ + assert (p - 1 > pattern); + + /* Allocate the space for the jump. */ + GET_BUFFER_SPACE (3); + + /* We know we are not at the first character of the pattern, + because laststart was nonzero. And we've already + incremented `p', by the way, to be the character after + the `*'. Do we have to do something analogous here + for null bytes, because of RE_DOT_NOT_NULL? */ + if (TRANSLATE ((unsigned char)*(p - 2)) == TRANSLATE ('.') + && zero_times_ok + && p < pend + && TRANSLATE ((unsigned char)*p) == TRANSLATE ('\n') + && !(syntax & RE_DOT_NEWLINE)) + { /* We have .*\n. */ + STORE_JUMP (jump, b, laststart); + keep_string_p = true; + } + else + /* Anything else. */ + STORE_JUMP (maybe_pop_jump, b, laststart - 3); + + /* We've added more stuff to the buffer. */ + b += 3; + } + + /* On failure, jump from laststart to b + 3, which will be the + end of the buffer after this jump is inserted. */ + GET_BUFFER_SPACE (3); + INSERT_JUMP (keep_string_p ? on_failure_keep_string_jump + : on_failure_jump, + laststart, b + 3); + pending_exact = 0; + b += 3; + + if (!zero_times_ok) + { + /* At least one repetition is required, so insert a + `dummy_failure_jump' before the initial + `on_failure_jump' instruction of the loop. This + effects a skip over that instruction the first time + we hit that loop. */ + GET_BUFFER_SPACE (3); + INSERT_JUMP (dummy_failure_jump, laststart, laststart + 6); + b += 3; + } + } + break; + + + case '.': + laststart = b; + BUF_PUSH (anychar); + break; + + + case '[': + { + CLEAR_RANGE_TABLE_WORK_USED (range_table_work); + + if (p == pend) FREE_STACK_RETURN (REG_EBRACK); + + /* Ensure that we have enough space to push a charset: the + opcode, the length count, and the bitset; 34 bytes in all. */ + GET_BUFFER_SPACE (34); + + laststart = b; + + /* We test `*p == '^' twice, instead of using an if + statement, so we only need one BUF_PUSH. */ + BUF_PUSH (*p == '^' ? charset_not : charset); + if (*p == '^') + p++; + + /* Remember the first position in the bracket expression. */ + p1 = p; + + /* Push the number of bytes in the bitmap. */ + BUF_PUSH ((1 << BYTEWIDTH) / BYTEWIDTH); + + /* Clear the whole map. */ + bzero (b, (1 << BYTEWIDTH) / BYTEWIDTH); + + /* charset_not matches newline according to a syntax bit. */ + if ((re_opcode_t) b[-2] == charset_not + && (syntax & RE_HAT_LISTS_NOT_NEWLINE)) + SET_LIST_BIT ('\n'); + + /* Read in characters and ranges, setting map bits. */ + for (;;) + { + int len; + boolean escaped_char = false; + + if (p == pend) FREE_STACK_RETURN (REG_EBRACK); + + PATFETCH (c); + + /* \ might escape characters inside [...] and [^...]. */ + if ((syntax & RE_BACKSLASH_ESCAPE_IN_LISTS) && c == '\\') + { + if (p == pend) FREE_STACK_RETURN (REG_EESCAPE); + + PATFETCH (c); + escaped_char = true; + } + else + { + /* Could be the end of the bracket expression. If it's + not (i.e., when the bracket expression is `[]' so + far), the ']' character bit gets set way below. */ + if (c == ']' && p != p1 + 1) + break; + } + + /* If C indicates start of multibyte char, get the + actual character code in C, and set the pattern + pointer P to the next character boundary. */ + if (bufp->multibyte && BASE_LEADING_CODE_P (c)) + { + PATUNFETCH; + c = STRING_CHAR_AND_LENGTH (p, pend - p, len); + p += len; + } + /* What should we do for the character which is + greater than 0x7F, but not BASE_LEADING_CODE_P? + XXX */ + + /* See if we're at the beginning of a possible character + class. */ + + else if (!escaped_char && + syntax & RE_CHAR_CLASSES && c == '[' && *p == ':') + { + /* Leave room for the null. */ + char str[CHAR_CLASS_MAX_LENGTH + 1]; + + PATFETCH (c); + c1 = 0; + + /* If pattern is `[[:'. */ + if (p == pend) FREE_STACK_RETURN (REG_EBRACK); + + for (;;) + { + PATFETCH (c); + if (c == ':' || c == ']' || p == pend + || c1 == CHAR_CLASS_MAX_LENGTH) + break; + str[c1++] = c; + } + str[c1] = '\0'; + + /* If isn't a word bracketed by `[:' and `:]': + undo the ending character, the letters, and + leave the leading `:' and `[' (but set bits for + them). */ + if (c == ':' && *p == ']') + { + int ch; + boolean is_alnum = STREQ (str, "alnum"); + boolean is_alpha = STREQ (str, "alpha"); + boolean is_blank = STREQ (str, "blank"); + boolean is_cntrl = STREQ (str, "cntrl"); + boolean is_digit = STREQ (str, "digit"); + boolean is_graph = STREQ (str, "graph"); + boolean is_lower = STREQ (str, "lower"); + boolean is_print = STREQ (str, "print"); + boolean is_punct = STREQ (str, "punct"); + boolean is_space = STREQ (str, "space"); + boolean is_upper = STREQ (str, "upper"); + boolean is_xdigit = STREQ (str, "xdigit"); + + if (!IS_CHAR_CLASS (str)) + FREE_STACK_RETURN (REG_ECTYPE); + + /* Throw away the ] at the end of the character + class. */ + PATFETCH (c); + + if (p == pend) FREE_STACK_RETURN (REG_EBRACK); + + for (ch = 0; ch < 1 << BYTEWIDTH; ch++) + { + int translated = TRANSLATE (ch); + /* This was split into 3 if's to + avoid an arbitrary limit in some compiler. */ + if ( (is_alnum && ISALNUM (ch)) + || (is_alpha && ISALPHA (ch)) + || (is_blank && ISBLANK (ch)) + || (is_cntrl && ISCNTRL (ch))) + SET_LIST_BIT (translated); + if ( (is_digit && ISDIGIT (ch)) + || (is_graph && ISGRAPH (ch)) + || (is_lower && ISLOWER (ch)) + || (is_print && ISPRINT (ch))) + SET_LIST_BIT (translated); + if ( (is_punct && ISPUNCT (ch)) + || (is_space && ISSPACE (ch)) + || (is_upper && ISUPPER (ch)) + || (is_xdigit && ISXDIGIT (ch))) + SET_LIST_BIT (translated); + } + + /* Repeat the loop. */ + continue; + } + else + { + c1++; + while (c1--) + PATUNFETCH; + SET_LIST_BIT ('['); + + /* Because the `:' may starts the range, we + can't simply set bit and repeat the loop. + Instead, just set it to C and handle below. */ + c = ':'; + } + } + + if (p < pend && p[0] == '-' && p[1] != ']') + { + + /* Discard the `-'. */ + PATFETCH (c1); + + /* Fetch the character which ends the range. */ + PATFETCH (c1); + if (bufp->multibyte && BASE_LEADING_CODE_P (c1)) + { + PATUNFETCH; + c1 = STRING_CHAR_AND_LENGTH (p, pend - p, len); + p += len; + } + + if (SINGLE_BYTE_CHAR_P (c) + && ! SINGLE_BYTE_CHAR_P (c1)) + { + /* Handle a range such as \177-\377 in multibyte mode. + Split that into two ranges,, + the low one ending at 0237, and the high one + starting at ...040. */ + int c1_base = (c1 & ~0177) | 040; + SET_RANGE_TABLE_WORK_AREA (range_table_work, c, c1); + c1 = 0237; + } + else if (!SAME_CHARSET_P (c, c1)) + FREE_STACK_RETURN (REG_ERANGE); + } + else + /* Range from C to C. */ + c1 = c; + + /* Set the range ... */ + if (SINGLE_BYTE_CHAR_P (c)) + /* ... into bitmap. */ + { + unsigned this_char; + int range_start = c, range_end = c1; + + /* If the start is after the end, the range is empty. */ + if (range_start > range_end) + { + if (syntax & RE_NO_EMPTY_RANGES) + FREE_STACK_RETURN (REG_ERANGE); + /* Else, repeat the loop. */ + } + else + { + for (this_char = range_start; this_char <= range_end; + this_char++) + SET_LIST_BIT (TRANSLATE (this_char)); + } + } + else + /* ... into range table. */ + SET_RANGE_TABLE_WORK_AREA (range_table_work, c, c1); + } + + /* Discard any (non)matching list bytes that are all 0 at the + end of the map. Decrease the map-length byte too. */ + while ((int) b[-1] > 0 && b[b[-1] - 1] == 0) + b[-1]--; + b += b[-1]; + + /* Build real range table from work area. */ + if (RANGE_TABLE_WORK_USED (range_table_work)) + { + int i; + int used = RANGE_TABLE_WORK_USED (range_table_work); + + /* Allocate space for COUNT + RANGE_TABLE. Needs two + bytes for COUNT and three bytes for each character. */ + GET_BUFFER_SPACE (2 + used * 3); + + /* Indicate the existence of range table. */ + laststart[1] |= 0x80; + + STORE_NUMBER_AND_INCR (b, used / 2); + for (i = 0; i < used; i++) + STORE_CHARACTER_AND_INCR + (b, RANGE_TABLE_WORK_ELT (range_table_work, i)); + } + } + break; + + + case '(': + if (syntax & RE_NO_BK_PARENS) + goto handle_open; + else + goto normal_char; + + + case ')': + if (syntax & RE_NO_BK_PARENS) + goto handle_close; + else + goto normal_char; + + + case '\n': + if (syntax & RE_NEWLINE_ALT) + goto handle_alt; + else + goto normal_char; + + + case '|': + if (syntax & RE_NO_BK_VBAR) + goto handle_alt; + else + goto normal_char; + + + case '{': + if (syntax & RE_INTERVALS && syntax & RE_NO_BK_BRACES) + goto handle_interval; + else + goto normal_char; + + + case '\\': + if (p == pend) FREE_STACK_RETURN (REG_EESCAPE); + + /* Do not translate the character after the \, so that we can + distinguish, e.g., \B from \b, even if we normally would + translate, e.g., B to b. */ + PATFETCH_RAW (c); + + switch (c) + { + case '(': + if (syntax & RE_NO_BK_PARENS) + goto normal_backslash; + + handle_open: + bufp->re_nsub++; + regnum++; + + if (COMPILE_STACK_FULL) + { + RETALLOC (compile_stack.stack, compile_stack.size << 1, + compile_stack_elt_t); + if (compile_stack.stack == NULL) return REG_ESPACE; + + compile_stack.size <<= 1; + } + + /* These are the values to restore when we hit end of this + group. They are all relative offsets, so that if the + whole pattern moves because of realloc, they will still + be valid. */ + COMPILE_STACK_TOP.begalt_offset = begalt - bufp->buffer; + COMPILE_STACK_TOP.fixup_alt_jump + = fixup_alt_jump ? fixup_alt_jump - bufp->buffer + 1 : 0; + COMPILE_STACK_TOP.laststart_offset = b - bufp->buffer; + COMPILE_STACK_TOP.regnum = regnum; + + /* We will eventually replace the 0 with the number of + groups inner to this one. But do not push a + start_memory for groups beyond the last one we can + represent in the compiled pattern. */ + if (regnum <= MAX_REGNUM) + { + COMPILE_STACK_TOP.inner_group_offset = b - bufp->buffer + 2; + BUF_PUSH_3 (start_memory, regnum, 0); + } + + compile_stack.avail++; + + fixup_alt_jump = 0; + laststart = 0; + begalt = b; + /* If we've reached MAX_REGNUM groups, then this open + won't actually generate any code, so we'll have to + clear pending_exact explicitly. */ + pending_exact = 0; + break; + + + case ')': + if (syntax & RE_NO_BK_PARENS) goto normal_backslash; + + if (COMPILE_STACK_EMPTY) + if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD) + goto normal_backslash; + else + FREE_STACK_RETURN (REG_ERPAREN); + + handle_close: + if (fixup_alt_jump) + { /* Push a dummy failure point at the end of the + alternative for a possible future + `pop_failure_jump' to pop. See comments at + `push_dummy_failure' in `re_match_2'. */ + BUF_PUSH (push_dummy_failure); + + /* We allocated space for this jump when we assigned + to `fixup_alt_jump', in the `handle_alt' case below. */ + STORE_JUMP (jump_past_alt, fixup_alt_jump, b - 1); + } + + /* See similar code for backslashed left paren above. */ + if (COMPILE_STACK_EMPTY) + if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD) + goto normal_char; + else + FREE_STACK_RETURN (REG_ERPAREN); + + /* Since we just checked for an empty stack above, this + ``can't happen''. */ + assert (compile_stack.avail != 0); + { + /* We don't just want to restore into `regnum', because + later groups should continue to be numbered higher, + as in `(ab)c(de)' -- the second group is #2. */ + regnum_t this_group_regnum; + + compile_stack.avail--; + begalt = bufp->buffer + COMPILE_STACK_TOP.begalt_offset; + fixup_alt_jump + = COMPILE_STACK_TOP.fixup_alt_jump + ? bufp->buffer + COMPILE_STACK_TOP.fixup_alt_jump - 1 + : 0; + laststart = bufp->buffer + COMPILE_STACK_TOP.laststart_offset; + this_group_regnum = COMPILE_STACK_TOP.regnum; + /* If we've reached MAX_REGNUM groups, then this open + won't actually generate any code, so we'll have to + clear pending_exact explicitly. */ + pending_exact = 0; + + /* We're at the end of the group, so now we know how many + groups were inside this one. */ + if (this_group_regnum <= MAX_REGNUM) + { + unsigned char *inner_group_loc + = bufp->buffer + COMPILE_STACK_TOP.inner_group_offset; + + *inner_group_loc = regnum - this_group_regnum; + BUF_PUSH_3 (stop_memory, this_group_regnum, + regnum - this_group_regnum); + } + } + break; + + + case '|': /* `\|'. */ + if (syntax & RE_LIMITED_OPS || syntax & RE_NO_BK_VBAR) + goto normal_backslash; + handle_alt: + if (syntax & RE_LIMITED_OPS) + goto normal_char; + + /* Insert before the previous alternative a jump which + jumps to this alternative if the former fails. */ + GET_BUFFER_SPACE (3); + INSERT_JUMP (on_failure_jump, begalt, b + 6); + pending_exact = 0; + b += 3; + + /* The alternative before this one has a jump after it + which gets executed if it gets matched. Adjust that + jump so it will jump to this alternative's analogous + jump (put in below, which in turn will jump to the next + (if any) alternative's such jump, etc.). The last such + jump jumps to the correct final destination. A picture: + _____ _____ + | | | | + | v | v + a | b | c + + If we are at `b', then fixup_alt_jump right now points to a + three-byte space after `a'. We'll put in the jump, set + fixup_alt_jump to right after `b', and leave behind three + bytes which we'll fill in when we get to after `c'. */ + + if (fixup_alt_jump) + STORE_JUMP (jump_past_alt, fixup_alt_jump, b); + + /* Mark and leave space for a jump after this alternative, + to be filled in later either by next alternative or + when know we're at the end of a series of alternatives. */ + fixup_alt_jump = b; + GET_BUFFER_SPACE (3); + b += 3; + + laststart = 0; + begalt = b; + break; + + + case '{': + /* If \{ is a literal. */ + if (!(syntax & RE_INTERVALS) + /* If we're at `\{' and it's not the open-interval + operator. */ + || ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES)) + || (p - 2 == pattern && p == pend)) + goto normal_backslash; + + handle_interval: + { + /* If got here, then the syntax allows intervals. */ + + /* At least (most) this many matches must be made. */ + int lower_bound = -1, upper_bound = -1; + + beg_interval = p - 1; + + if (p == pend) + { + if (syntax & RE_NO_BK_BRACES) + goto unfetch_interval; + else + FREE_STACK_RETURN (REG_EBRACE); + } + + GET_UNSIGNED_NUMBER (lower_bound); + + if (c == ',') + { + GET_UNSIGNED_NUMBER (upper_bound); + if (upper_bound < 0) upper_bound = RE_DUP_MAX; + } + else + /* Interval such as `{1}' => match exactly once. */ + upper_bound = lower_bound; + + if (lower_bound < 0 || upper_bound > RE_DUP_MAX + || lower_bound > upper_bound) + { + if (syntax & RE_NO_BK_BRACES) + goto unfetch_interval; + else + FREE_STACK_RETURN (REG_BADBR); + } + + if (!(syntax & RE_NO_BK_BRACES)) + { + if (c != '\\') FREE_STACK_RETURN (REG_EBRACE); + + PATFETCH (c); + } + + if (c != '}') + { + if (syntax & RE_NO_BK_BRACES) + goto unfetch_interval; + else + FREE_STACK_RETURN (REG_BADBR); + } + + /* We just parsed a valid interval. */ + + /* If it's invalid to have no preceding re. */ + if (!laststart) + { + if (syntax & RE_CONTEXT_INVALID_OPS) + FREE_STACK_RETURN (REG_BADRPT); + else if (syntax & RE_CONTEXT_INDEP_OPS) + laststart = b; + else + goto unfetch_interval; + } + + /* If the upper bound is zero, don't want to succeed at + all; jump from `laststart' to `b + 3', which will be + the end of the buffer after we insert the jump. */ + if (upper_bound == 0) + { + GET_BUFFER_SPACE (3); + INSERT_JUMP (jump, laststart, b + 3); + b += 3; + } + + /* Otherwise, we have a nontrivial interval. When + we're all done, the pattern will look like: + set_number_at <jump count> <upper bound> + set_number_at <succeed_n count> <lower bound> + succeed_n <after jump addr> <succeed_n count> + <body of loop> + jump_n <succeed_n addr> <jump count> + (The upper bound and `jump_n' are omitted if + `upper_bound' is 1, though.) */ + else + { /* If the upper bound is > 1, we need to insert + more at the end of the loop. */ + unsigned nbytes = 10 + (upper_bound > 1) * 10; + + GET_BUFFER_SPACE (nbytes); + + /* Initialize lower bound of the `succeed_n', even + though it will be set during matching by its + attendant `set_number_at' (inserted next), + because `re_compile_fastmap' needs to know. + Jump to the `jump_n' we might insert below. */ + INSERT_JUMP2 (succeed_n, laststart, + b + 5 + (upper_bound > 1) * 5, + lower_bound); + b += 5; + + /* Code to initialize the lower bound. Insert + before the `succeed_n'. The `5' is the last two + bytes of this `set_number_at', plus 3 bytes of + the following `succeed_n'. */ + insert_op2 (set_number_at, laststart, 5, lower_bound, b); + b += 5; + + if (upper_bound > 1) + { /* More than one repetition is allowed, so + append a backward jump to the `succeed_n' + that starts this interval. + + When we've reached this during matching, + we'll have matched the interval once, so + jump back only `upper_bound - 1' times. */ + STORE_JUMP2 (jump_n, b, laststart + 5, + upper_bound - 1); + b += 5; + + /* The location we want to set is the second + parameter of the `jump_n'; that is `b-2' as + an absolute address. `laststart' will be + the `set_number_at' we're about to insert; + `laststart+3' the number to set, the source + for the relative address. But we are + inserting into the middle of the pattern -- + so everything is getting moved up by 5. + Conclusion: (b - 2) - (laststart + 3) + 5, + i.e., b - laststart. + + We insert this at the beginning of the loop + so that if we fail during matching, we'll + reinitialize the bounds. */ + insert_op2 (set_number_at, laststart, b - laststart, + upper_bound - 1, b); + b += 5; + } + } + pending_exact = 0; + beg_interval = NULL; + } + break; + + unfetch_interval: + /* If an invalid interval, match the characters as literals. */ + assert (beg_interval); + p = beg_interval; + beg_interval = NULL; + + /* normal_char and normal_backslash need `c'. */ + PATFETCH (c); + + if (!(syntax & RE_NO_BK_BRACES)) + { + if (p > pattern && p[-1] == '\\') + goto normal_backslash; + } + goto normal_char; + +#ifdef emacs + /* There is no way to specify the before_dot and after_dot + operators. rms says this is ok. --karl */ + case '=': + BUF_PUSH (at_dot); + break; + + case 's': + laststart = b; + PATFETCH (c); + BUF_PUSH_2 (syntaxspec, syntax_spec_code[c]); + break; + + case 'S': + laststart = b; + PATFETCH (c); + BUF_PUSH_2 (notsyntaxspec, syntax_spec_code[c]); + break; + + case 'c': + laststart = b; + PATFETCH_RAW (c); + BUF_PUSH_2 (categoryspec, c); + break; + + case 'C': + laststart = b; + PATFETCH_RAW (c); + BUF_PUSH_2 (notcategoryspec, c); + break; +#endif /* emacs */ + + + case 'w': + laststart = b; + BUF_PUSH (wordchar); + break; + + + case 'W': + laststart = b; + BUF_PUSH (notwordchar); + break; + + + case '<': + BUF_PUSH (wordbeg); + break; + + case '>': + BUF_PUSH (wordend); + break; + + case 'b': + BUF_PUSH (wordbound); + break; + + case 'B': + BUF_PUSH (notwordbound); + break; + + case '`': + BUF_PUSH (begbuf); + break; + + case '\'': + BUF_PUSH (endbuf); + break; + + case '1': case '2': case '3': case '4': case '5': + case '6': case '7': case '8': case '9': + if (syntax & RE_NO_BK_REFS) + goto normal_char; + + c1 = c - '0'; + + if (c1 > regnum) + FREE_STACK_RETURN (REG_ESUBREG); + + /* Can't back reference to a subexpression if inside of it. */ + if (group_in_compile_stack (compile_stack, c1)) + goto normal_char; + + laststart = b; + BUF_PUSH_2 (duplicate, c1); + break; + + + case '+': + case '?': + if (syntax & RE_BK_PLUS_QM) + goto handle_plus; + else + goto normal_backslash; + + default: + normal_backslash: + /* You might think it would be useful for \ to mean + not to translate; but if we don't translate it + it will never match anything. */ + c = TRANSLATE (c); + goto normal_char; + } + break; + + + default: + /* Expects the character in `c'. */ + normal_char: + p1 = p - 1; /* P1 points the head of C. */ +#ifdef emacs + if (bufp->multibyte) + { + c = STRING_CHAR (p1, pend - p1); + c = TRANSLATE (c); + /* Set P to the next character boundary. */ + p += MULTIBYTE_FORM_LENGTH (p1, pend - p1) - 1; + } +#endif + /* If no exactn currently being built. */ + if (!pending_exact + + /* If last exactn not at current position. */ + || pending_exact + *pending_exact + 1 != b + + /* We have only one byte following the exactn for the count. */ + || *pending_exact >= (1 << BYTEWIDTH) - (p - p1) + + /* If followed by a repetition operator. */ + || (p != pend && (*p == '*' || *p == '^')) + || ((syntax & RE_BK_PLUS_QM) + ? p + 1 < pend && *p == '\\' && (p[1] == '+' || p[1] == '?') + : p != pend && (*p == '+' || *p == '?')) + || ((syntax & RE_INTERVALS) + && ((syntax & RE_NO_BK_BRACES) + ? p != pend && *p == '{' + : p + 1 < pend && p[0] == '\\' && p[1] == '{'))) + { + /* Start building a new exactn. */ + + laststart = b; + + BUF_PUSH_2 (exactn, 0); + pending_exact = b - 1; + } + +#ifdef emacs + if (! SINGLE_BYTE_CHAR_P (c)) + { + unsigned char work[4], *str; + int i = CHAR_STRING (c, work, str); + int j; + for (j = 0; j < i; j++) + { + BUF_PUSH (str[j]); + (*pending_exact)++; + } + } + else +#endif + { + BUF_PUSH (c); + (*pending_exact)++; + } + break; + } /* switch (c) */ + } /* while p != pend */ + + + /* Through the pattern now. */ + + if (fixup_alt_jump) + STORE_JUMP (jump_past_alt, fixup_alt_jump, b); + + if (!COMPILE_STACK_EMPTY) + FREE_STACK_RETURN (REG_EPAREN); + + /* If we don't want backtracking, force success + the first time we reach the end of the compiled pattern. */ + if (syntax & RE_NO_POSIX_BACKTRACKING) + BUF_PUSH (succeed); + + free (compile_stack.stack); + + /* We have succeeded; set the length of the buffer. */ + bufp->used = b - bufp->buffer; + +#ifdef DEBUG + if (debug) + { + DEBUG_PRINT1 ("\nCompiled pattern: \n"); + print_compiled_pattern (bufp); + } +#endif /* DEBUG */ + +#ifndef MATCH_MAY_ALLOCATE + /* Initialize the failure stack to the largest possible stack. This + isn't necessary unless we're trying to avoid calling alloca in + the search and match routines. */ + { + int num_regs = bufp->re_nsub + 1; + + if (fail_stack.size < re_max_failures * TYPICAL_FAILURE_SIZE) + { + fail_stack.size = re_max_failures * TYPICAL_FAILURE_SIZE; + +#ifdef emacs + if (! fail_stack.stack) + fail_stack.stack + = (fail_stack_elt_t *) xmalloc (fail_stack.size + * sizeof (fail_stack_elt_t)); + else + fail_stack.stack + = (fail_stack_elt_t *) xrealloc (fail_stack.stack, + (fail_stack.size + * sizeof (fail_stack_elt_t))); +#else /* not emacs */ + if (! fail_stack.stack) + fail_stack.stack + = (fail_stack_elt_t *) malloc (fail_stack.size + * sizeof (fail_stack_elt_t)); + else + fail_stack.stack + = (fail_stack_elt_t *) realloc (fail_stack.stack, + (fail_stack.size + * sizeof (fail_stack_elt_t))); +#endif /* not emacs */ + } + + regex_grow_registers (num_regs); + } +#endif /* not MATCH_MAY_ALLOCATE */ + + return REG_NOERROR; +} /* regex_compile */ + +/* Subroutines for `regex_compile'. */ + +/* Store OP at LOC followed by two-byte integer parameter ARG. */ + +static void +store_op1 (op, loc, arg) + re_opcode_t op; + unsigned char *loc; + int arg; +{ + *loc = (unsigned char) op; + STORE_NUMBER (loc + 1, arg); +} + + +/* Like `store_op1', but for two two-byte parameters ARG1 and ARG2. */ + +static void +store_op2 (op, loc, arg1, arg2) + re_opcode_t op; + unsigned char *loc; + int arg1, arg2; +{ + *loc = (unsigned char) op; + STORE_NUMBER (loc + 1, arg1); + STORE_NUMBER (loc + 3, arg2); +} + + +/* Copy the bytes from LOC to END to open up three bytes of space at LOC + for OP followed by two-byte integer parameter ARG. */ + +static void +insert_op1 (op, loc, arg, end) + re_opcode_t op; + unsigned char *loc; + int arg; + unsigned char *end; +{ + register unsigned char *pfrom = end; + register unsigned char *pto = end + 3; + + while (pfrom != loc) + *--pto = *--pfrom; + + store_op1 (op, loc, arg); +} + + +/* Like `insert_op1', but for two two-byte parameters ARG1 and ARG2. */ + +static void +insert_op2 (op, loc, arg1, arg2, end) + re_opcode_t op; + unsigned char *loc; + int arg1, arg2; + unsigned char *end; +{ + register unsigned char *pfrom = end; + register unsigned char *pto = end + 5; + + while (pfrom != loc) + *--pto = *--pfrom; + + store_op2 (op, loc, arg1, arg2); +} + + +/* P points to just after a ^ in PATTERN. Return true if that ^ comes + after an alternative or a begin-subexpression. We assume there is at + least one character before the ^. */ + +static boolean +at_begline_loc_p (pattern, p, syntax) + const char *pattern, *p; + reg_syntax_t syntax; +{ + const char *prev = p - 2; + boolean prev_prev_backslash = prev > pattern && prev[-1] == '\\'; + + return + /* After a subexpression? */ + (*prev == '(' && (syntax & RE_NO_BK_PARENS || prev_prev_backslash)) + /* After an alternative? */ + || (*prev == '|' && (syntax & RE_NO_BK_VBAR || prev_prev_backslash)); +} + + +/* The dual of at_begline_loc_p. This one is for $. We assume there is + at least one character after the $, i.e., `P < PEND'. */ + +static boolean +at_endline_loc_p (p, pend, syntax) + const char *p, *pend; + int syntax; +{ + const char *next = p; + boolean next_backslash = *next == '\\'; + const char *next_next = p + 1 < pend ? p + 1 : 0; + + return + /* Before a subexpression? */ + (syntax & RE_NO_BK_PARENS ? *next == ')' + : next_backslash && next_next && *next_next == ')') + /* Before an alternative? */ + || (syntax & RE_NO_BK_VBAR ? *next == '|' + : next_backslash && next_next && *next_next == '|'); +} + + +/* Returns true if REGNUM is in one of COMPILE_STACK's elements and + false if it's not. */ + +static boolean +group_in_compile_stack (compile_stack, regnum) + compile_stack_type compile_stack; + regnum_t regnum; +{ + int this_element; + + for (this_element = compile_stack.avail - 1; + this_element >= 0; + this_element--) + if (compile_stack.stack[this_element].regnum == regnum) + return true; + + return false; +} + +/* re_compile_fastmap computes a ``fastmap'' for the compiled pattern in + BUFP. A fastmap records which of the (1 << BYTEWIDTH) possible + characters can start a string that matches the pattern. This fastmap + is used by re_search to skip quickly over impossible starting points. + + The caller must supply the address of a (1 << BYTEWIDTH)-byte data + area as BUFP->fastmap. + + We set the `fastmap', `fastmap_accurate', and `can_be_null' fields in + the pattern buffer. + + Returns 0 if we succeed, -2 if an internal error. */ + +int +re_compile_fastmap (bufp) + struct re_pattern_buffer *bufp; +{ + int i, j, k; +#ifdef MATCH_MAY_ALLOCATE + fail_stack_type fail_stack; +#endif +#ifndef REGEX_MALLOC + char *destination; +#endif + /* We don't push any register information onto the failure stack. */ + unsigned num_regs = 0; + + register char *fastmap = bufp->fastmap; + unsigned char *pattern = bufp->buffer; + unsigned long size = bufp->used; + unsigned char *p = pattern; + register unsigned char *pend = pattern + size; + + /* This holds the pointer to the failure stack, when + it is allocated relocatably. */ + fail_stack_elt_t *failure_stack_ptr; + + /* Assume that each path through the pattern can be null until + proven otherwise. We set this false at the bottom of switch + statement, to which we get only if a particular path doesn't + match the empty string. */ + boolean path_can_be_null = true; + + /* We aren't doing a `succeed_n' to begin with. */ + boolean succeed_n_p = false; + + /* If all elements for base leading-codes in fastmap is set, this + flag is set true. */ + boolean match_any_multibyte_characters = false; + + /* Maximum code of simple (single byte) character. */ + int simple_char_max; + + assert (fastmap != NULL && p != NULL); + + INIT_FAIL_STACK (); + bzero (fastmap, 1 << BYTEWIDTH); /* Assume nothing's valid. */ + bufp->fastmap_accurate = 1; /* It will be when we're done. */ + bufp->can_be_null = 0; + + while (1) + { + if (p == pend || *p == succeed) + { + /* We have reached the (effective) end of pattern. */ + if (!FAIL_STACK_EMPTY ()) + { + bufp->can_be_null |= path_can_be_null; + + /* Reset for next path. */ + path_can_be_null = true; + + p = fail_stack.stack[--fail_stack.avail].pointer; + + continue; + } + else + break; + } + + /* We should never be about to go beyond the end of the pattern. */ + assert (p < pend); + + switch (SWITCH_ENUM_CAST ((re_opcode_t) *p++)) + { + + /* I guess the idea here is to simply not bother with a fastmap + if a backreference is used, since it's too hard to figure out + the fastmap for the corresponding group. Setting + `can_be_null' stops `re_search_2' from using the fastmap, so + that is all we do. */ + case duplicate: + bufp->can_be_null = 1; + goto done; + + + /* Following are the cases which match a character. These end + with `break'. */ + + case exactn: + fastmap[p[1]] = 1; + break; + + +#ifndef emacs + case charset: + for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--) + if (p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH))) + fastmap[j] = 1; + break; + + + case charset_not: + /* Chars beyond end of map must be allowed. */ + for (j = *p * BYTEWIDTH; j < (1 << BYTEWIDTH); j++) + fastmap[j] = 1; + + for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--) + if (!(p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH)))) + fastmap[j] = 1; + break; + + + case wordchar: + for (j = 0; j < (1 << BYTEWIDTH); j++) + if (SYNTAX (j) == Sword) + fastmap[j] = 1; + break; + + + case notwordchar: + for (j = 0; j < (1 << BYTEWIDTH); j++) + if (SYNTAX (j) != Sword) + fastmap[j] = 1; + break; +#else /* emacs */ + case charset: + for (j = CHARSET_BITMAP_SIZE (&p[-1]) * BYTEWIDTH - 1, p++; + j >= 0; j--) + if (p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH))) + fastmap[j] = 1; + + if (CHARSET_RANGE_TABLE_EXISTS_P (&p[-2]) + && match_any_multibyte_characters == false) + { + /* Set fastmap[I] 1 where I is a base leading code of each + multibyte character in the range table. */ + int c, count; + + /* Make P points the range table. */ + p += CHARSET_BITMAP_SIZE (&p[-2]); + + /* Extract the number of ranges in range table into + COUNT. */ + EXTRACT_NUMBER_AND_INCR (count, p); + for (; count > 0; count--, p += 2 * 3) /* XXX */ + { + /* Extract the start of each range. */ + EXTRACT_CHARACTER (c, p); + j = CHAR_CHARSET (c); + fastmap[CHARSET_LEADING_CODE_BASE (j)] = 1; + } + } + break; + + + case charset_not: + /* Chars beyond end of bitmap are possible matches. + All the single-byte codes can occur in multibyte buffers. + So any that are not listed in the charset + are possible matches, even in multibyte buffers. */ + simple_char_max = (1 << BYTEWIDTH); + for (j = CHARSET_BITMAP_SIZE (&p[-1]) * BYTEWIDTH; + j < simple_char_max; j++) + fastmap[j] = 1; + + for (j = CHARSET_BITMAP_SIZE (&p[-1]) * BYTEWIDTH - 1, p++; + j >= 0; j--) + if (!(p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH)))) + fastmap[j] = 1; + + if (bufp->multibyte) + /* Any character set can possibly contain a character + which doesn't match the specified set of characters. */ + { + set_fastmap_for_multibyte_characters: + if (match_any_multibyte_characters == false) + { + for (j = 0x80; j < 0xA0; j++) /* XXX */ + if (BASE_LEADING_CODE_P (j)) + fastmap[j] = 1; + match_any_multibyte_characters = true; + } + } + break; + + + case wordchar: + /* All the single-byte codes can occur in multibyte buffers, + and they may have word syntax. So do consider them. */ + simple_char_max = (1 << BYTEWIDTH); + for (j = 0; j < simple_char_max; j++) + if (SYNTAX (j) == Sword) + fastmap[j] = 1; + + if (bufp->multibyte) + /* Any character set can possibly contain a character + whose syntax is `Sword'. */ + goto set_fastmap_for_multibyte_characters; + break; + + + case notwordchar: + /* All the single-byte codes can occur in multibyte buffers, + and they may not have word syntax. So do consider them. */ + simple_char_max = (1 << BYTEWIDTH); + for (j = 0; j < simple_char_max; j++) + if (SYNTAX (j) != Sword) + fastmap[j] = 1; + + if (bufp->multibyte) + /* Any character set can possibly contain a character + whose syntax is not `Sword'. */ + goto set_fastmap_for_multibyte_characters; + break; +#endif + + case anychar: + { + int fastmap_newline = fastmap['\n']; + + /* `.' matches anything, except perhaps newline. + Even in a multibyte buffer, it should match any + conceivable byte value for the fastmap. */ + if (bufp->multibyte) + match_any_multibyte_characters = true; + + simple_char_max = (1 << BYTEWIDTH); + for (j = 0; j < simple_char_max; j++) + fastmap[j] = 1; + + /* ... except perhaps newline. */ + if (!(bufp->syntax & RE_DOT_NEWLINE)) + fastmap['\n'] = fastmap_newline; + + /* Return if we have already set `can_be_null'; if we have, + then the fastmap is irrelevant. Something's wrong here. */ + else if (bufp->can_be_null) + goto done; + + /* Otherwise, have to check alternative paths. */ + break; + } + +#ifdef emacs + case wordbound: + case notwordbound: + case wordbeg: + case wordend: + case notsyntaxspec: + case syntaxspec: + /* This match depends on text properties. These end with + aborting optimizations. */ + bufp->can_be_null = 1; + goto done; +#if 0 + k = *p++; + simple_char_max = bufp->multibyte ? 0x80 : (1 << BYTEWIDTH); + for (j = 0; j < simple_char_max; j++) + if (SYNTAX (j) == (enum syntaxcode) k) + fastmap[j] = 1; + + if (bufp->multibyte) + /* Any character set can possibly contain a character + whose syntax is K. */ + goto set_fastmap_for_multibyte_characters; + break; + + case notsyntaxspec: + k = *p++; + simple_char_max = bufp->multibyte ? 0x80 : (1 << BYTEWIDTH); + for (j = 0; j < simple_char_max; j++) + if (SYNTAX (j) != (enum syntaxcode) k) + fastmap[j] = 1; + + if (bufp->multibyte) + /* Any character set can possibly contain a character + whose syntax is not K. */ + goto set_fastmap_for_multibyte_characters; + break; +#endif + + + case categoryspec: + k = *p++; + simple_char_max = (1 << BYTEWIDTH); + for (j = 0; j < simple_char_max; j++) + if (CHAR_HAS_CATEGORY (j, k)) + fastmap[j] = 1; + + if (bufp->multibyte) + /* Any character set can possibly contain a character + whose category is K. */ + goto set_fastmap_for_multibyte_characters; + break; + + + case notcategoryspec: + k = *p++; + simple_char_max = (1 << BYTEWIDTH); + for (j = 0; j < simple_char_max; j++) + if (!CHAR_HAS_CATEGORY (j, k)) + fastmap[j] = 1; + + if (bufp->multibyte) + /* Any character set can possibly contain a character + whose category is not K. */ + goto set_fastmap_for_multibyte_characters; + break; + + /* All cases after this match the empty string. These end with + `continue'. */ + + + case before_dot: + case at_dot: + case after_dot: + continue; +#endif /* emacs */ + + + case no_op: + case begline: + case endline: + case begbuf: + case endbuf: +#ifndef emacs + case wordbound: + case notwordbound: + case wordbeg: + case wordend: +#endif + case push_dummy_failure: + continue; + + + case jump_n: + case pop_failure_jump: + case maybe_pop_jump: + case jump: + case jump_past_alt: + case dummy_failure_jump: + EXTRACT_NUMBER_AND_INCR (j, p); + p += j; + if (j > 0) + continue; + + /* Jump backward implies we just went through the body of a + loop and matched nothing. Opcode jumped to should be + `on_failure_jump' or `succeed_n'. Just treat it like an + ordinary jump. For a * loop, it has pushed its failure + point already; if so, discard that as redundant. */ + if ((re_opcode_t) *p != on_failure_jump + && (re_opcode_t) *p != succeed_n) + continue; + + p++; + EXTRACT_NUMBER_AND_INCR (j, p); + p += j; + + /* If what's on the stack is where we are now, pop it. */ + if (!FAIL_STACK_EMPTY () + && fail_stack.stack[fail_stack.avail - 1].pointer == p) + fail_stack.avail--; + + continue; + + + case on_failure_jump: + case on_failure_keep_string_jump: + handle_on_failure_jump: + EXTRACT_NUMBER_AND_INCR (j, p); + + /* For some patterns, e.g., `(a?)?', `p+j' here points to the + end of the pattern. We don't want to push such a point, + since when we restore it above, entering the switch will + increment `p' past the end of the pattern. We don't need + to push such a point since we obviously won't find any more + fastmap entries beyond `pend'. Such a pattern can match + the null string, though. */ + if (p + j < pend) + { + if (!PUSH_PATTERN_OP (p + j, fail_stack)) + { + RESET_FAIL_STACK (); + return -2; + } + } + else + bufp->can_be_null = 1; + + if (succeed_n_p) + { + EXTRACT_NUMBER_AND_INCR (k, p); /* Skip the n. */ + succeed_n_p = false; + } + + continue; + + + case succeed_n: + /* Get to the number of times to succeed. */ + p += 2; + + /* Increment p past the n for when k != 0. */ + EXTRACT_NUMBER_AND_INCR (k, p); + if (k == 0) + { + p -= 4; + succeed_n_p = true; /* Spaghetti code alert. */ + goto handle_on_failure_jump; + } + continue; + + + case set_number_at: + p += 4; + continue; + + + case start_memory: + case stop_memory: + p += 2; + continue; + + + default: + abort (); /* We have listed all the cases. */ + } /* switch *p++ */ + + /* Getting here means we have found the possible starting + characters for one path of the pattern -- and that the empty + string does not match. We need not follow this path further. + Instead, look at the next alternative (remembered on the + stack), or quit if no more. The test at the top of the loop + does these things. */ + path_can_be_null = false; + p = pend; + } /* while p */ + + /* Set `can_be_null' for the last path (also the first path, if the + pattern is empty). */ + bufp->can_be_null |= path_can_be_null; + + done: + RESET_FAIL_STACK (); + return 0; +} /* re_compile_fastmap */ + +/* Set REGS to hold NUM_REGS registers, storing them in STARTS and + ENDS. Subsequent matches using PATTERN_BUFFER and REGS will use + this memory for recording register information. STARTS and ENDS + must be allocated using the malloc library routine, and must each + be at least NUM_REGS * sizeof (regoff_t) bytes long. + + If NUM_REGS == 0, then subsequent matches should allocate their own + register data. + + Unless this function is called, the first search or match using + PATTERN_BUFFER will allocate its own register data, without + freeing the old data. */ + +void +re_set_registers (bufp, regs, num_regs, starts, ends) + struct re_pattern_buffer *bufp; + struct re_registers *regs; + unsigned num_regs; + regoff_t *starts, *ends; +{ + if (num_regs) + { + bufp->regs_allocated = REGS_REALLOCATE; + regs->num_regs = num_regs; + regs->start = starts; + regs->end = ends; + } + else + { + bufp->regs_allocated = REGS_UNALLOCATED; + regs->num_regs = 0; + regs->start = regs->end = (regoff_t *) 0; + } +} + +/* Searching routines. */ + +/* Like re_search_2, below, but only one string is specified, and + doesn't let you say where to stop matching. */ + +int +re_search (bufp, string, size, startpos, range, regs) + struct re_pattern_buffer *bufp; + const char *string; + int size, startpos, range; + struct re_registers *regs; +{ + return re_search_2 (bufp, NULL, 0, string, size, startpos, range, + regs, size); +} + +/* End address of virtual concatenation of string. */ +#define STOP_ADDR_VSTRING(P) \ + (((P) >= size1 ? string2 + size2 : string1 + size1)) + +/* Address of POS in the concatenation of virtual string. */ +#define POS_ADDR_VSTRING(POS) \ + (((POS) >= size1 ? string2 - size1 : string1) + (POS)) + +/* Using the compiled pattern in BUFP->buffer, first tries to match the + virtual concatenation of STRING1 and STRING2, starting first at index + STARTPOS, then at STARTPOS + 1, and so on. + + STRING1 and STRING2 have length SIZE1 and SIZE2, respectively. + + RANGE is how far to scan while trying to match. RANGE = 0 means try + only at STARTPOS; in general, the last start tried is STARTPOS + + RANGE. + + In REGS, return the indices of the virtual concatenation of STRING1 + and STRING2 that matched the entire BUFP->buffer and its contained + subexpressions. + + Do not consider matching one past the index STOP in the virtual + concatenation of STRING1 and STRING2. + + We return either the position in the strings at which the match was + found, -1 if no match, or -2 if error (such as failure + stack overflow). */ + +int +re_search_2 (bufp, string1, size1, string2, size2, startpos, range, regs, stop) + struct re_pattern_buffer *bufp; + const char *string1, *string2; + int size1, size2; + int startpos; + int range; + struct re_registers *regs; + int stop; +{ + int val; + register char *fastmap = bufp->fastmap; + register RE_TRANSLATE_TYPE translate = bufp->translate; + int total_size = size1 + size2; + int endpos = startpos + range; + int anchored_start = 0; + + /* Nonzero if we have to concern multibyte character. */ + int multibyte = bufp->multibyte; + + /* Check for out-of-range STARTPOS. */ + if (startpos < 0 || startpos > total_size) + return -1; + + /* Fix up RANGE if it might eventually take us outside + the virtual concatenation of STRING1 and STRING2. + Make sure we won't move STARTPOS below 0 or above TOTAL_SIZE. */ + if (endpos < 0) + range = 0 - startpos; + else if (endpos > total_size) + range = total_size - startpos; + + /* If the search isn't to be a backwards one, don't waste time in a + search for a pattern anchored at beginning of buffer. */ + if (bufp->used > 0 && (re_opcode_t) bufp->buffer[0] == begbuf && range > 0) + { + if (startpos > 0) + return -1; + else + range = 0; + } + +#ifdef emacs + /* In a forward search for something that starts with \=. + don't keep searching past point. */ + if (bufp->used > 0 && (re_opcode_t) bufp->buffer[0] == at_dot && range > 0) + { + range = PT_BYTE - BEGV_BYTE - startpos; + if (range < 0) + return -1; + } +#endif /* emacs */ + + /* Update the fastmap now if not correct already. */ + if (fastmap && !bufp->fastmap_accurate) + if (re_compile_fastmap (bufp) == -2) + return -2; + + /* See whether the pattern is anchored. */ + if (bufp->buffer[0] == begline) + anchored_start = 1; + +#ifdef emacs + gl_state.object = re_match_object; + { + int adjpos = NILP (re_match_object) || BUFFERP (re_match_object); + int charpos = SYNTAX_TABLE_BYTE_TO_CHAR (startpos + adjpos); + + SETUP_SYNTAX_TABLE_FOR_OBJECT (re_match_object, charpos, 1); + } +#endif + + /* Loop through the string, looking for a place to start matching. */ + for (;;) + { + /* If the pattern is anchored, + skip quickly past places we cannot match. + We don't bother to treat startpos == 0 specially + because that case doesn't repeat. */ + if (anchored_start && startpos > 0) + { + if (! (bufp->newline_anchor + && ((startpos <= size1 ? string1[startpos - 1] + : string2[startpos - size1 - 1]) + == '\n'))) + goto advance; + } + + /* If a fastmap is supplied, skip quickly over characters that + cannot be the start of a match. If the pattern can match the + null string, however, we don't need to skip characters; we want + the first null string. */ + if (fastmap && startpos < total_size && !bufp->can_be_null) + { + register const char *d; + register unsigned int buf_ch; + + d = POS_ADDR_VSTRING (startpos); + + if (range > 0) /* Searching forwards. */ + { + register int lim = 0; + int irange = range; + + if (startpos < size1 && startpos + range >= size1) + lim = range - (size1 - startpos); + + /* Written out as an if-else to avoid testing `translate' + inside the loop. */ + if (RE_TRANSLATE_P (translate)) + { + if (multibyte) + while (range > lim) + { + int buf_charlen; + + buf_ch = STRING_CHAR_AND_LENGTH (d, range - lim, + buf_charlen); + + buf_ch = RE_TRANSLATE (translate, buf_ch); + if (buf_ch >= 0400 + || fastmap[buf_ch]) + break; + + range -= buf_charlen; + d += buf_charlen; + } + else + while (range > lim + && !fastmap[(unsigned char) + RE_TRANSLATE (translate, (unsigned char) *d)]) + { + d++; + range--; + } + } + else + while (range > lim && !fastmap[(unsigned char) *d]) + { + d++; + range--; + } + + startpos += irange - range; + } + else /* Searching backwards. */ + { + int room = (size1 == 0 || startpos >= size1 + ? size2 + size1 - startpos + : size1 - startpos); + + buf_ch = STRING_CHAR (d, room); + if (RE_TRANSLATE_P (translate)) + buf_ch = RE_TRANSLATE (translate, buf_ch); + + if (! (buf_ch >= 0400 + || fastmap[buf_ch])) + goto advance; + } + } + + /* If can't match the null string, and that's all we have left, fail. */ + if (range >= 0 && startpos == total_size && fastmap + && !bufp->can_be_null) + return -1; + + val = re_match_2_internal (bufp, string1, size1, string2, size2, + startpos, regs, stop); +#ifndef REGEX_MALLOC +#ifdef C_ALLOCA + alloca (0); +#endif +#endif + + if (val >= 0) + return startpos; + + if (val == -2) + return -2; + + advance: + if (!range) + break; + else if (range > 0) + { + /* Update STARTPOS to the next character boundary. */ + if (multibyte) + { + const unsigned char *p + = (const unsigned char *) POS_ADDR_VSTRING (startpos); + const unsigned char *pend + = (const unsigned char *) STOP_ADDR_VSTRING (startpos); + int len = MULTIBYTE_FORM_LENGTH (p, pend - p); + + range -= len; + if (range < 0) + break; + startpos += len; + } + else + { + range--; + startpos++; + } + } + else + { + range++; + startpos--; + + /* Update STARTPOS to the previous character boundary. */ + if (multibyte) + { + const unsigned char *p + = (const unsigned char *) POS_ADDR_VSTRING (startpos); + int len = 0; + + /* Find the head of multibyte form. */ + while (!CHAR_HEAD_P (*p)) + p--, len++; + + /* Adjust it. */ +#if 0 /* XXX */ + if (MULTIBYTE_FORM_LENGTH (p, len + 1) != (len + 1)) + ; + else +#endif + { + range += len; + if (range > 0) + break; + + startpos -= len; + } + } + } + } + return -1; +} /* re_search_2 */ + +/* Declarations and macros for re_match_2. */ + +static int bcmp_translate (); +static boolean alt_match_null_string_p (), + common_op_match_null_string_p (), + group_match_null_string_p (); + +/* This converts PTR, a pointer into one of the search strings `string1' + and `string2' into an offset from the beginning of that string. */ +#define POINTER_TO_OFFSET(ptr) \ + (FIRST_STRING_P (ptr) \ + ? ((regoff_t) ((ptr) - string1)) \ + : ((regoff_t) ((ptr) - string2 + size1))) + +/* Macros for dealing with the split strings in re_match_2. */ + +#define MATCHING_IN_FIRST_STRING (dend == end_match_1) + +/* Call before fetching a character with *d. This switches over to + string2 if necessary. */ +#define PREFETCH() \ + while (d == dend) \ + { \ + /* End of string2 => fail. */ \ + if (dend == end_match_2) \ + goto fail; \ + /* End of string1 => advance to string2. */ \ + d = string2; \ + dend = end_match_2; \ + } + + +/* Test if at very beginning or at very end of the virtual concatenation + of `string1' and `string2'. If only one string, it's `string2'. */ +#define AT_STRINGS_BEG(d) ((d) == (size1 ? string1 : string2) || !size2) +#define AT_STRINGS_END(d) ((d) == end2) + + +/* Test if D points to a character which is word-constituent. We have + two special cases to check for: if past the end of string1, look at + the first character in string2; and if before the beginning of + string2, look at the last character in string1. */ +#define WORDCHAR_P(d) \ + (SYNTAX ((d) == end1 ? *string2 \ + : (d) == string2 - 1 ? *(end1 - 1) : *(d)) \ + == Sword) + +/* Disabled due to a compiler bug -- see comment at case wordbound */ + +/* The comment at case wordbound is following one, but we don't use + AT_WORD_BOUNDARY anymore to support multibyte form. + + The DEC Alpha C compiler 3.x generates incorrect code for the + test WORDCHAR_P (d - 1) != WORDCHAR_P (d) in the expansion of + AT_WORD_BOUNDARY, so this code is disabled. Expanding the + macro and introducing temporary variables works around the bug. */ + +#if 0 +/* Test if the character before D and the one at D differ with respect + to being word-constituent. */ +#define AT_WORD_BOUNDARY(d) \ + (AT_STRINGS_BEG (d) || AT_STRINGS_END (d) \ + || WORDCHAR_P (d - 1) != WORDCHAR_P (d)) +#endif + +/* Free everything we malloc. */ +#ifdef MATCH_MAY_ALLOCATE +#define FREE_VAR(var) if (var) { REGEX_FREE (var); var = NULL; } else +#define FREE_VARIABLES() \ + do { \ + REGEX_FREE_STACK (fail_stack.stack); \ + FREE_VAR (regstart); \ + FREE_VAR (regend); \ + FREE_VAR (old_regstart); \ + FREE_VAR (old_regend); \ + FREE_VAR (best_regstart); \ + FREE_VAR (best_regend); \ + FREE_VAR (reg_info); \ + FREE_VAR (reg_dummy); \ + FREE_VAR (reg_info_dummy); \ + } while (0) +#else +#define FREE_VARIABLES() ((void)0) /* Do nothing! But inhibit gcc warning. */ +#endif /* not MATCH_MAY_ALLOCATE */ + +/* These values must meet several constraints. They must not be valid + register values; since we have a limit of 255 registers (because + we use only one byte in the pattern for the register number), we can + use numbers larger than 255. They must differ by 1, because of + NUM_FAILURE_ITEMS above. And the value for the lowest register must + be larger than the value for the highest register, so we do not try + to actually save any registers when none are active. */ +#define NO_HIGHEST_ACTIVE_REG (1 << BYTEWIDTH) +#define NO_LOWEST_ACTIVE_REG (NO_HIGHEST_ACTIVE_REG + 1) + +/* Matching routines. */ + +#ifndef emacs /* Emacs never uses this. */ +/* re_match is like re_match_2 except it takes only a single string. */ + +int +re_match (bufp, string, size, pos, regs) + struct re_pattern_buffer *bufp; + const char *string; + int size, pos; + struct re_registers *regs; +{ + int result = re_match_2_internal (bufp, NULL, 0, string, size, + pos, regs, size); +#ifndef REGEX_MALLOC /* CVS */ +#ifdef C_ALLOCA /* CVS */ + alloca (0); +#endif /* CVS */ +#endif /* CVS */ + return result; +} +#endif /* not emacs */ + +#ifdef emacs +/* In Emacs, this is the string or buffer in which we + are matching. It is used for looking up syntax properties. */ +Lisp_Object re_match_object; +#endif + +/* re_match_2 matches the compiled pattern in BUFP against the + the (virtual) concatenation of STRING1 and STRING2 (of length SIZE1 + and SIZE2, respectively). We start matching at POS, and stop + matching at STOP. + + If REGS is non-null and the `no_sub' field of BUFP is nonzero, we + store offsets for the substring each group matched in REGS. See the + documentation for exactly how many groups we fill. + + We return -1 if no match, -2 if an internal error (such as the + failure stack overflowing). Otherwise, we return the length of the + matched substring. */ + +int +re_match_2 (bufp, string1, size1, string2, size2, pos, regs, stop) + struct re_pattern_buffer *bufp; + const char *string1, *string2; + int size1, size2; + int pos; + struct re_registers *regs; + int stop; +{ + int result; + +#ifdef emacs + int charpos; + int adjpos = NILP (re_match_object) || BUFFERP (re_match_object); + gl_state.object = re_match_object; + charpos = SYNTAX_TABLE_BYTE_TO_CHAR (pos + adjpos); + SETUP_SYNTAX_TABLE_FOR_OBJECT (re_match_object, charpos, 1); +#endif + + result = re_match_2_internal (bufp, string1, size1, string2, size2, + pos, regs, stop); +#ifndef REGEX_MALLOC /* CVS */ +#ifdef C_ALLOCA /* CVS */ + alloca (0); +#endif /* CVS */ +#endif /* CVS */ + return result; +} + +/* This is a separate function so that we can force an alloca cleanup + afterwards. */ +static int +re_match_2_internal (bufp, string1, size1, string2, size2, pos, regs, stop) + struct re_pattern_buffer *bufp; + const char *string1, *string2; + int size1, size2; + int pos; + struct re_registers *regs; + int stop; +{ + /* General temporaries. */ + int mcnt; + unsigned char *p1; + + /* Just past the end of the corresponding string. */ + const char *end1, *end2; + + /* Pointers into string1 and string2, just past the last characters in + each to consider matching. */ + const char *end_match_1, *end_match_2; + + /* Where we are in the data, and the end of the current string. */ + const char *d, *dend; + + /* Where we are in the pattern, and the end of the pattern. */ + unsigned char *p = bufp->buffer; + register unsigned char *pend = p + bufp->used; + + /* Mark the opcode just after a start_memory, so we can test for an + empty subpattern when we get to the stop_memory. */ + unsigned char *just_past_start_mem = 0; + + /* We use this to map every character in the string. */ + RE_TRANSLATE_TYPE translate = bufp->translate; + + /* Nonzero if we have to concern multibyte character. */ + int multibyte = bufp->multibyte; + + /* Failure point stack. Each place that can handle a failure further + down the line pushes a failure point on this stack. It consists of + restart, regend, and reg_info for all registers corresponding to + the subexpressions we're currently inside, plus the number of such + registers, and, finally, two char *'s. The first char * is where + to resume scanning the pattern; the second one is where to resume + scanning the strings. If the latter is zero, the failure point is + a ``dummy''; if a failure happens and the failure point is a dummy, + it gets discarded and the next next one is tried. */ +#ifdef MATCH_MAY_ALLOCATE /* otherwise, this is global. */ + fail_stack_type fail_stack; +#endif +#ifdef DEBUG + static unsigned failure_id = 0; + unsigned nfailure_points_pushed = 0, nfailure_points_popped = 0; +#endif + + /* This holds the pointer to the failure stack, when + it is allocated relocatably. */ + fail_stack_elt_t *failure_stack_ptr; + + /* We fill all the registers internally, independent of what we + return, for use in backreferences. The number here includes + an element for register zero. */ + unsigned num_regs = bufp->re_nsub + 1; + + /* The currently active registers. */ + unsigned lowest_active_reg = NO_LOWEST_ACTIVE_REG; + unsigned highest_active_reg = NO_HIGHEST_ACTIVE_REG; + + /* Information on the contents of registers. These are pointers into + the input strings; they record just what was matched (on this + attempt) by a subexpression part of the pattern, that is, the + regnum-th regstart pointer points to where in the pattern we began + matching and the regnum-th regend points to right after where we + stopped matching the regnum-th subexpression. (The zeroth register + keeps track of what the whole pattern matches.) */ +#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */ + const char **regstart, **regend; +#endif + + /* If a group that's operated upon by a repetition operator fails to + match anything, then the register for its start will need to be + restored because it will have been set to wherever in the string we + are when we last see its open-group operator. Similarly for a + register's end. */ +#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */ + const char **old_regstart, **old_regend; +#endif + + /* The is_active field of reg_info helps us keep track of which (possibly + nested) subexpressions we are currently in. The matched_something + field of reg_info[reg_num] helps us tell whether or not we have + matched any of the pattern so far this time through the reg_num-th + subexpression. These two fields get reset each time through any + loop their register is in. */ +#ifdef MATCH_MAY_ALLOCATE /* otherwise, this is global. */ + register_info_type *reg_info; +#endif + + /* The following record the register info as found in the above + variables when we find a match better than any we've seen before. + This happens as we backtrack through the failure points, which in + turn happens only if we have not yet matched the entire string. */ + unsigned best_regs_set = false; +#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */ + const char **best_regstart, **best_regend; +#endif + + /* Logically, this is `best_regend[0]'. But we don't want to have to + allocate space for that if we're not allocating space for anything + else (see below). Also, we never need info about register 0 for + any of the other register vectors, and it seems rather a kludge to + treat `best_regend' differently than the rest. So we keep track of + the end of the best match so far in a separate variable. We + initialize this to NULL so that when we backtrack the first time + and need to test it, it's not garbage. */ + const char *match_end = NULL; + + /* This helps SET_REGS_MATCHED avoid doing redundant work. */ + int set_regs_matched_done = 0; + + /* Used when we pop values we don't care about. */ +#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */ + const char **reg_dummy; + register_info_type *reg_info_dummy; +#endif + +#ifdef DEBUG + /* Counts the total number of registers pushed. */ + unsigned num_regs_pushed = 0; +#endif + + DEBUG_PRINT1 ("\n\nEntering re_match_2.\n"); + + INIT_FAIL_STACK (); + +#ifdef MATCH_MAY_ALLOCATE + /* Do not bother to initialize all the register variables if there are + no groups in the pattern, as it takes a fair amount of time. If + there are groups, we include space for register 0 (the whole + pattern), even though we never use it, since it simplifies the + array indexing. We should fix this. */ + if (bufp->re_nsub) + { + regstart = REGEX_TALLOC (num_regs, const char *); + regend = REGEX_TALLOC (num_regs, const char *); + old_regstart = REGEX_TALLOC (num_regs, const char *); + old_regend = REGEX_TALLOC (num_regs, const char *); + best_regstart = REGEX_TALLOC (num_regs, const char *); + best_regend = REGEX_TALLOC (num_regs, const char *); + reg_info = REGEX_TALLOC (num_regs, register_info_type); + reg_dummy = REGEX_TALLOC (num_regs, const char *); + reg_info_dummy = REGEX_TALLOC (num_regs, register_info_type); + + if (!(regstart && regend && old_regstart && old_regend && reg_info + && best_regstart && best_regend && reg_dummy && reg_info_dummy)) + { + FREE_VARIABLES (); + return -2; + } + } + else + { + /* We must initialize all our variables to NULL, so that + `FREE_VARIABLES' doesn't try to free them. */ + regstart = regend = old_regstart = old_regend = best_regstart + = best_regend = reg_dummy = NULL; + reg_info = reg_info_dummy = (register_info_type *) NULL; + } +#endif /* MATCH_MAY_ALLOCATE */ + + /* The starting position is bogus. */ + if (pos < 0 || pos > size1 + size2) + { + FREE_VARIABLES (); + return -1; + } + + /* Initialize subexpression text positions to -1 to mark ones that no + start_memory/stop_memory has been seen for. Also initialize the + register information struct. */ + for (mcnt = 1; mcnt < num_regs; mcnt++) + { + regstart[mcnt] = regend[mcnt] + = old_regstart[mcnt] = old_regend[mcnt] = REG_UNSET_VALUE; + + REG_MATCH_NULL_STRING_P (reg_info[mcnt]) = MATCH_NULL_UNSET_VALUE; + IS_ACTIVE (reg_info[mcnt]) = 0; + MATCHED_SOMETHING (reg_info[mcnt]) = 0; + EVER_MATCHED_SOMETHING (reg_info[mcnt]) = 0; + } + + /* We move `string1' into `string2' if the latter's empty -- but not if + `string1' is null. */ + if (size2 == 0 && string1 != NULL) + { + string2 = string1; + size2 = size1; + string1 = 0; + size1 = 0; + } + end1 = string1 + size1; + end2 = string2 + size2; + + /* Compute where to stop matching, within the two strings. */ + if (stop <= size1) + { + end_match_1 = string1 + stop; + end_match_2 = string2; + } + else + { + end_match_1 = end1; + end_match_2 = string2 + stop - size1; + } + + /* `p' scans through the pattern as `d' scans through the data. + `dend' is the end of the input string that `d' points within. `d' + is advanced into the following input string whenever necessary, but + this happens before fetching; therefore, at the beginning of the + loop, `d' can be pointing at the end of a string, but it cannot + equal `string2'. */ + if (size1 > 0 && pos <= size1) + { + d = string1 + pos; + dend = end_match_1; + } + else + { + d = string2 + pos - size1; + dend = end_match_2; + } + + DEBUG_PRINT1 ("The compiled pattern is: "); + DEBUG_PRINT_COMPILED_PATTERN (bufp, p, pend); + DEBUG_PRINT1 ("The string to match is: `"); + DEBUG_PRINT_DOUBLE_STRING (d, string1, size1, string2, size2); + DEBUG_PRINT1 ("'\n"); + + /* This loops over pattern commands. It exits by returning from the + function if the match is complete, or it drops through if the match + fails at this starting point in the input data. */ + for (;;) + { + DEBUG_PRINT2 ("\n0x%x: ", p); + + if (p == pend) + { /* End of pattern means we might have succeeded. */ + DEBUG_PRINT1 ("end of pattern ... "); + + /* If we haven't matched the entire string, and we want the + longest match, try backtracking. */ + if (d != end_match_2) + { + /* 1 if this match ends in the same string (string1 or string2) + as the best previous match. */ + boolean same_str_p = (FIRST_STRING_P (match_end) + == MATCHING_IN_FIRST_STRING); + /* 1 if this match is the best seen so far. */ + boolean best_match_p; + + /* AIX compiler got confused when this was combined + with the previous declaration. */ + if (same_str_p) + best_match_p = d > match_end; + else + best_match_p = !MATCHING_IN_FIRST_STRING; + + DEBUG_PRINT1 ("backtracking.\n"); + + if (!FAIL_STACK_EMPTY ()) + { /* More failure points to try. */ + + /* If exceeds best match so far, save it. */ + if (!best_regs_set || best_match_p) + { + best_regs_set = true; + match_end = d; + + DEBUG_PRINT1 ("\nSAVING match as best so far.\n"); + + for (mcnt = 1; mcnt < num_regs; mcnt++) + { + best_regstart[mcnt] = regstart[mcnt]; + best_regend[mcnt] = regend[mcnt]; + } + } + goto fail; + } + + /* If no failure points, don't restore garbage. And if + last match is real best match, don't restore second + best one. */ + else if (best_regs_set && !best_match_p) + { + restore_best_regs: + /* Restore best match. It may happen that `dend == + end_match_1' while the restored d is in string2. + For example, the pattern `x.*y.*z' against the + strings `x-' and `y-z-', if the two strings are + not consecutive in memory. */ + DEBUG_PRINT1 ("Restoring best registers.\n"); + + d = match_end; + dend = ((d >= string1 && d <= end1) + ? end_match_1 : end_match_2); + + for (mcnt = 1; mcnt < num_regs; mcnt++) + { + regstart[mcnt] = best_regstart[mcnt]; + regend[mcnt] = best_regend[mcnt]; + } + } + } /* d != end_match_2 */ + + succeed_label: + DEBUG_PRINT1 ("Accepting match.\n"); + + /* If caller wants register contents data back, do it. */ + if (regs && !bufp->no_sub) + { + /* Have the register data arrays been allocated? */ + if (bufp->regs_allocated == REGS_UNALLOCATED) + { /* No. So allocate them with malloc. We need one + extra element beyond `num_regs' for the `-1' marker + GNU code uses. */ + regs->num_regs = MAX (RE_NREGS, num_regs + 1); + regs->start = TALLOC (regs->num_regs, regoff_t); + regs->end = TALLOC (regs->num_regs, regoff_t); + if (regs->start == NULL || regs->end == NULL) + { + FREE_VARIABLES (); + return -2; + } + bufp->regs_allocated = REGS_REALLOCATE; + } + else if (bufp->regs_allocated == REGS_REALLOCATE) + { /* Yes. If we need more elements than were already + allocated, reallocate them. If we need fewer, just + leave it alone. */ + if (regs->num_regs < num_regs + 1) + { + regs->num_regs = num_regs + 1; + RETALLOC (regs->start, regs->num_regs, regoff_t); + RETALLOC (regs->end, regs->num_regs, regoff_t); + if (regs->start == NULL || regs->end == NULL) + { + FREE_VARIABLES (); + return -2; + } + } + } + else + { + /* These braces fend off a "empty body in an else-statement" + warning under GCC when assert expands to nothing. */ + assert (bufp->regs_allocated == REGS_FIXED); + } + + /* Convert the pointer data in `regstart' and `regend' to + indices. Register zero has to be set differently, + since we haven't kept track of any info for it. */ + if (regs->num_regs > 0) + { + regs->start[0] = pos; + regs->end[0] = (MATCHING_IN_FIRST_STRING + ? ((regoff_t) (d - string1)) + : ((regoff_t) (d - string2 + size1))); + } + + /* Go through the first `min (num_regs, regs->num_regs)' + registers, since that is all we initialized. */ + for (mcnt = 1; mcnt < MIN (num_regs, regs->num_regs); mcnt++) + { + if (REG_UNSET (regstart[mcnt]) || REG_UNSET (regend[mcnt])) + regs->start[mcnt] = regs->end[mcnt] = -1; + else + { + regs->start[mcnt] + = (regoff_t) POINTER_TO_OFFSET (regstart[mcnt]); + regs->end[mcnt] + = (regoff_t) POINTER_TO_OFFSET (regend[mcnt]); + } + } + + /* If the regs structure we return has more elements than + were in the pattern, set the extra elements to -1. If + we (re)allocated the registers, this is the case, + because we always allocate enough to have at least one + -1 at the end. */ + for (mcnt = num_regs; mcnt < regs->num_regs; mcnt++) + regs->start[mcnt] = regs->end[mcnt] = -1; + } /* regs && !bufp->no_sub */ + + DEBUG_PRINT4 ("%u failure points pushed, %u popped (%u remain).\n", + nfailure_points_pushed, nfailure_points_popped, + nfailure_points_pushed - nfailure_points_popped); + DEBUG_PRINT2 ("%u registers pushed.\n", num_regs_pushed); + + mcnt = d - pos - (MATCHING_IN_FIRST_STRING + ? string1 + : string2 - size1); + + DEBUG_PRINT2 ("Returning %d from re_match_2.\n", mcnt); + + FREE_VARIABLES (); + return mcnt; + } + + /* Otherwise match next pattern command. */ + switch (SWITCH_ENUM_CAST ((re_opcode_t) *p++)) + { + /* Ignore these. Used to ignore the n of succeed_n's which + currently have n == 0. */ + case no_op: + DEBUG_PRINT1 ("EXECUTING no_op.\n"); + break; + + case succeed: + DEBUG_PRINT1 ("EXECUTING succeed.\n"); + goto succeed_label; + + /* Match the next n pattern characters exactly. The following + byte in the pattern defines n, and the n bytes after that + are the characters to match. */ + case exactn: + mcnt = *p++; + DEBUG_PRINT2 ("EXECUTING exactn %d.\n", mcnt); + + /* This is written out as an if-else so we don't waste time + testing `translate' inside the loop. */ + if (RE_TRANSLATE_P (translate)) + { +#ifdef emacs + if (multibyte) + do + { + int pat_charlen, buf_charlen; + unsigned int pat_ch, buf_ch; + + PREFETCH (); + pat_ch = STRING_CHAR_AND_LENGTH (p, pend - p, pat_charlen); + buf_ch = STRING_CHAR_AND_LENGTH (d, dend - d, buf_charlen); + + if (RE_TRANSLATE (translate, buf_ch) + != pat_ch) + goto fail; + + p += pat_charlen; + d += buf_charlen; + mcnt -= pat_charlen; + } + while (mcnt > 0); + else +#endif /* not emacs */ + do + { + PREFETCH (); + if ((unsigned char) RE_TRANSLATE (translate, (unsigned char) *d) + != (unsigned char) *p++) + goto fail; + d++; + } + while (--mcnt); + } + else + { + do + { + PREFETCH (); + if (*d++ != (char) *p++) goto fail; + } + while (--mcnt); + } + SET_REGS_MATCHED (); + break; + + + /* Match any character except possibly a newline or a null. */ + case anychar: + { + int buf_charlen; + unsigned int buf_ch; + + DEBUG_PRINT1 ("EXECUTING anychar.\n"); + + PREFETCH (); + +#ifdef emacs + if (multibyte) + buf_ch = STRING_CHAR_AND_LENGTH (d, dend - d, buf_charlen); + else +#endif /* not emacs */ + { + buf_ch = (unsigned char) *d; + buf_charlen = 1; + } + + buf_ch = TRANSLATE (buf_ch); + + if ((!(bufp->syntax & RE_DOT_NEWLINE) + && buf_ch == '\n') + || ((bufp->syntax & RE_DOT_NOT_NULL) + && buf_ch == '\000')) + goto fail; + + SET_REGS_MATCHED (); + DEBUG_PRINT2 (" Matched `%d'.\n", *d); + d += buf_charlen; + } + break; + + + case charset: + case charset_not: + { + register unsigned int c; + boolean not = (re_opcode_t) *(p - 1) == charset_not; + int len; + + /* Start of actual range_table, or end of bitmap if there is no + range table. */ + unsigned char *range_table; + + /* Nonzero if there is range table. */ + int range_table_exists; + + /* Number of ranges of range table. Not in bytes. */ + int count; + + DEBUG_PRINT2 ("EXECUTING charset%s.\n", not ? "_not" : ""); + + PREFETCH (); + c = (unsigned char) *d; + + range_table = CHARSET_RANGE_TABLE (&p[-1]); /* Past the bitmap. */ + range_table_exists = CHARSET_RANGE_TABLE_EXISTS_P (&p[-1]); + if (range_table_exists) + EXTRACT_NUMBER_AND_INCR (count, range_table); + else + count = 0; + + if (multibyte && BASE_LEADING_CODE_P (c)) + c = STRING_CHAR_AND_LENGTH (d, dend - d, len); + + if (SINGLE_BYTE_CHAR_P (c)) + { /* Lookup bitmap. */ + c = TRANSLATE (c); /* The character to match. */ + len = 1; + + /* Cast to `unsigned' instead of `unsigned char' in + case the bit list is a full 32 bytes long. */ + if (c < (unsigned) (CHARSET_BITMAP_SIZE (&p[-1]) * BYTEWIDTH) + && p[1 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH))) + not = !not; + } + else if (range_table_exists) + CHARSET_LOOKUP_RANGE_TABLE_RAW (not, c, range_table, count); + + p = CHARSET_RANGE_TABLE_END (range_table, count); + + if (!not) goto fail; + + SET_REGS_MATCHED (); + d += len; + break; + } + + + /* The beginning of a group is represented by start_memory. + The arguments are the register number in the next byte, and the + number of groups inner to this one in the next. The text + matched within the group is recorded (in the internal + registers data structure) under the register number. */ + case start_memory: + DEBUG_PRINT3 ("EXECUTING start_memory %d (%d):\n", *p, p[1]); + + /* Find out if this group can match the empty string. */ + p1 = p; /* To send to group_match_null_string_p. */ + + if (REG_MATCH_NULL_STRING_P (reg_info[*p]) == MATCH_NULL_UNSET_VALUE) + REG_MATCH_NULL_STRING_P (reg_info[*p]) + = group_match_null_string_p (&p1, pend, reg_info); + + /* Save the position in the string where we were the last time + we were at this open-group operator in case the group is + operated upon by a repetition operator, e.g., with `(a*)*b' + against `ab'; then we want to ignore where we are now in + the string in case this attempt to match fails. */ + old_regstart[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p]) + ? REG_UNSET (regstart[*p]) ? d : regstart[*p] + : regstart[*p]; + DEBUG_PRINT2 (" old_regstart: %d\n", + POINTER_TO_OFFSET (old_regstart[*p])); + + regstart[*p] = d; + DEBUG_PRINT2 (" regstart: %d\n", POINTER_TO_OFFSET (regstart[*p])); + + IS_ACTIVE (reg_info[*p]) = 1; + MATCHED_SOMETHING (reg_info[*p]) = 0; + + /* Clear this whenever we change the register activity status. */ + set_regs_matched_done = 0; + + /* This is the new highest active register. */ + highest_active_reg = *p; + + /* If nothing was active before, this is the new lowest active + register. */ + if (lowest_active_reg == NO_LOWEST_ACTIVE_REG) + lowest_active_reg = *p; + + /* Move past the register number and inner group count. */ + p += 2; + just_past_start_mem = p; + + break; + + + /* The stop_memory opcode represents the end of a group. Its + arguments are the same as start_memory's: the register + number, and the number of inner groups. */ + case stop_memory: + DEBUG_PRINT3 ("EXECUTING stop_memory %d (%d):\n", *p, p[1]); + + /* We need to save the string position the last time we were at + this close-group operator in case the group is operated + upon by a repetition operator, e.g., with `((a*)*(b*)*)*' + against `aba'; then we want to ignore where we are now in + the string in case this attempt to match fails. */ + old_regend[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p]) + ? REG_UNSET (regend[*p]) ? d : regend[*p] + : regend[*p]; + DEBUG_PRINT2 (" old_regend: %d\n", + POINTER_TO_OFFSET (old_regend[*p])); + + regend[*p] = d; + DEBUG_PRINT2 (" regend: %d\n", POINTER_TO_OFFSET (regend[*p])); + + /* This register isn't active anymore. */ + IS_ACTIVE (reg_info[*p]) = 0; + + /* Clear this whenever we change the register activity status. */ + set_regs_matched_done = 0; + + /* If this was the only register active, nothing is active + anymore. */ + if (lowest_active_reg == highest_active_reg) + { + lowest_active_reg = NO_LOWEST_ACTIVE_REG; + highest_active_reg = NO_HIGHEST_ACTIVE_REG; + } + else + { /* We must scan for the new highest active register, since + it isn't necessarily one less than now: consider + (a(b)c(d(e)f)g). When group 3 ends, after the f), the + new highest active register is 1. */ + unsigned char r = *p - 1; + while (r > 0 && !IS_ACTIVE (reg_info[r])) + r--; + + /* If we end up at register zero, that means that we saved + the registers as the result of an `on_failure_jump', not + a `start_memory', and we jumped to past the innermost + `stop_memory'. For example, in ((.)*) we save + registers 1 and 2 as a result of the *, but when we pop + back to the second ), we are at the stop_memory 1. + Thus, nothing is active. */ + if (r == 0) + { + lowest_active_reg = NO_LOWEST_ACTIVE_REG; + highest_active_reg = NO_HIGHEST_ACTIVE_REG; + } + else + highest_active_reg = r; + } + + /* If just failed to match something this time around with a + group that's operated on by a repetition operator, try to + force exit from the ``loop'', and restore the register + information for this group that we had before trying this + last match. */ + if ((!MATCHED_SOMETHING (reg_info[*p]) + || just_past_start_mem == p - 1) + && (p + 2) < pend) + { + boolean is_a_jump_n = false; + + p1 = p + 2; + mcnt = 0; + switch ((re_opcode_t) *p1++) + { + case jump_n: + is_a_jump_n = true; + case pop_failure_jump: + case maybe_pop_jump: + case jump: + case dummy_failure_jump: + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + if (is_a_jump_n) + p1 += 2; + break; + + default: + /* do nothing */ ; + } + p1 += mcnt; + + /* If the next operation is a jump backwards in the pattern + to an on_failure_jump right before the start_memory + corresponding to this stop_memory, exit from the loop + by forcing a failure after pushing on the stack the + on_failure_jump's jump in the pattern, and d. */ + if (mcnt < 0 && (re_opcode_t) *p1 == on_failure_jump + && (re_opcode_t) p1[3] == start_memory && p1[4] == *p) + { + /* If this group ever matched anything, then restore + what its registers were before trying this last + failed match, e.g., with `(a*)*b' against `ab' for + regstart[1], and, e.g., with `((a*)*(b*)*)*' + against `aba' for regend[3]. + + Also restore the registers for inner groups for, + e.g., `((a*)(b*))*' against `aba' (register 3 would + otherwise get trashed). */ + + if (EVER_MATCHED_SOMETHING (reg_info[*p])) + { + unsigned r; + + EVER_MATCHED_SOMETHING (reg_info[*p]) = 0; + + /* Restore this and inner groups' (if any) registers. */ + for (r = *p; r < *p + *(p + 1); r++) + { + regstart[r] = old_regstart[r]; + + /* xx why this test? */ + if (old_regend[r] >= regstart[r]) + regend[r] = old_regend[r]; + } + } + p1++; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + PUSH_FAILURE_POINT (p1 + mcnt, d, -2); + + goto fail; + } + } + + /* Move past the register number and the inner group count. */ + p += 2; + break; + + + /* \<digit> has been turned into a `duplicate' command which is + followed by the numeric value of <digit> as the register number. */ + case duplicate: + { + register const char *d2, *dend2; + int regno = *p++; /* Get which register to match against. */ + DEBUG_PRINT2 ("EXECUTING duplicate %d.\n", regno); + + /* Can't back reference a group which we've never matched. */ + if (REG_UNSET (regstart[regno]) || REG_UNSET (regend[regno])) + goto fail; + + /* Where in input to try to start matching. */ + d2 = regstart[regno]; + + /* Where to stop matching; if both the place to start and + the place to stop matching are in the same string, then + set to the place to stop, otherwise, for now have to use + the end of the first string. */ + + dend2 = ((FIRST_STRING_P (regstart[regno]) + == FIRST_STRING_P (regend[regno])) + ? regend[regno] : end_match_1); + for (;;) + { + /* If necessary, advance to next segment in register + contents. */ + while (d2 == dend2) + { + if (dend2 == end_match_2) break; + if (dend2 == regend[regno]) break; + + /* End of string1 => advance to string2. */ + d2 = string2; + dend2 = regend[regno]; + } + /* At end of register contents => success */ + if (d2 == dend2) break; + + /* If necessary, advance to next segment in data. */ + PREFETCH (); + + /* How many characters left in this segment to match. */ + mcnt = dend - d; + + /* Want how many consecutive characters we can match in + one shot, so, if necessary, adjust the count. */ + if (mcnt > dend2 - d2) + mcnt = dend2 - d2; + + /* Compare that many; failure if mismatch, else move + past them. */ + if (RE_TRANSLATE_P (translate) + ? bcmp_translate (d, d2, mcnt, translate) + : bcmp (d, d2, mcnt)) + goto fail; + d += mcnt, d2 += mcnt; + + /* Do this because we've match some characters. */ + SET_REGS_MATCHED (); + } + } + break; + + + /* begline matches the empty string at the beginning of the string + (unless `not_bol' is set in `bufp'), and, if + `newline_anchor' is set, after newlines. */ + case begline: + DEBUG_PRINT1 ("EXECUTING begline.\n"); + + if (AT_STRINGS_BEG (d)) + { + if (!bufp->not_bol) break; + } + else if (d[-1] == '\n' && bufp->newline_anchor) + { + break; + } + /* In all other cases, we fail. */ + goto fail; + + + /* endline is the dual of begline. */ + case endline: + DEBUG_PRINT1 ("EXECUTING endline.\n"); + + if (AT_STRINGS_END (d)) + { + if (!bufp->not_eol) break; + } + + /* We have to ``prefetch'' the next character. */ + else if ((d == end1 ? *string2 : *d) == '\n' + && bufp->newline_anchor) + { + break; + } + goto fail; + + + /* Match at the very beginning of the data. */ + case begbuf: + DEBUG_PRINT1 ("EXECUTING begbuf.\n"); + if (AT_STRINGS_BEG (d)) + break; + goto fail; + + + /* Match at the very end of the data. */ + case endbuf: + DEBUG_PRINT1 ("EXECUTING endbuf.\n"); + if (AT_STRINGS_END (d)) + break; + goto fail; + + + /* on_failure_keep_string_jump is used to optimize `.*\n'. It + pushes NULL as the value for the string on the stack. Then + `pop_failure_point' will keep the current value for the + string, instead of restoring it. To see why, consider + matching `foo\nbar' against `.*\n'. The .* matches the foo; + then the . fails against the \n. But the next thing we want + to do is match the \n against the \n; if we restored the + string value, we would be back at the foo. + + Because this is used only in specific cases, we don't need to + check all the things that `on_failure_jump' does, to make + sure the right things get saved on the stack. Hence we don't + share its code. The only reason to push anything on the + stack at all is that otherwise we would have to change + `anychar's code to do something besides goto fail in this + case; that seems worse than this. */ + case on_failure_keep_string_jump: + DEBUG_PRINT1 ("EXECUTING on_failure_keep_string_jump"); + + EXTRACT_NUMBER_AND_INCR (mcnt, p); + DEBUG_PRINT3 (" %d (to 0x%x):\n", mcnt, p + mcnt); + + PUSH_FAILURE_POINT (p + mcnt, NULL, -2); + break; + + + /* Uses of on_failure_jump: + + Each alternative starts with an on_failure_jump that points + to the beginning of the next alternative. Each alternative + except the last ends with a jump that in effect jumps past + the rest of the alternatives. (They really jump to the + ending jump of the following alternative, because tensioning + these jumps is a hassle.) + + Repeats start with an on_failure_jump that points past both + the repetition text and either the following jump or + pop_failure_jump back to this on_failure_jump. */ + case on_failure_jump: + on_failure: + DEBUG_PRINT1 ("EXECUTING on_failure_jump"); + +#if defined (WINDOWSNT) && defined (emacs) + QUIT; +#endif + + EXTRACT_NUMBER_AND_INCR (mcnt, p); + DEBUG_PRINT3 (" %d (to 0x%x)", mcnt, p + mcnt); + + /* If this on_failure_jump comes right before a group (i.e., + the original * applied to a group), save the information + for that group and all inner ones, so that if we fail back + to this point, the group's information will be correct. + For example, in \(a*\)*\1, we need the preceding group, + and in \(zz\(a*\)b*\)\2, we need the inner group. */ + + /* We can't use `p' to check ahead because we push + a failure point to `p + mcnt' after we do this. */ + p1 = p; + + /* We need to skip no_op's before we look for the + start_memory in case this on_failure_jump is happening as + the result of a completed succeed_n, as in \(a\)\{1,3\}b\1 + against aba. */ + while (p1 < pend && (re_opcode_t) *p1 == no_op) + p1++; + + if (p1 < pend && (re_opcode_t) *p1 == start_memory) + { + /* We have a new highest active register now. This will + get reset at the start_memory we are about to get to, + but we will have saved all the registers relevant to + this repetition op, as described above. */ + highest_active_reg = *(p1 + 1) + *(p1 + 2); + if (lowest_active_reg == NO_LOWEST_ACTIVE_REG) + lowest_active_reg = *(p1 + 1); + } + + DEBUG_PRINT1 (":\n"); + PUSH_FAILURE_POINT (p + mcnt, d, -2); + break; + + + /* A smart repeat ends with `maybe_pop_jump'. + We change it to either `pop_failure_jump' or `jump'. */ + case maybe_pop_jump: +#if defined (WINDOWSNT) && defined (emacs) + QUIT; +#endif + EXTRACT_NUMBER_AND_INCR (mcnt, p); + DEBUG_PRINT2 ("EXECUTING maybe_pop_jump %d.\n", mcnt); + { + register unsigned char *p2 = p; + + /* Compare the beginning of the repeat with what in the + pattern follows its end. If we can establish that there + is nothing that they would both match, i.e., that we + would have to backtrack because of (as in, e.g., `a*a') + then we can change to pop_failure_jump, because we'll + never have to backtrack. + + This is not true in the case of alternatives: in + `(a|ab)*' we do need to backtrack to the `ab' alternative + (e.g., if the string was `ab'). But instead of trying to + detect that here, the alternative has put on a dummy + failure point which is what we will end up popping. */ + + /* Skip over open/close-group commands. + If what follows this loop is a ...+ construct, + look at what begins its body, since we will have to + match at least one of that. */ + while (1) + { + if (p2 + 2 < pend + && ((re_opcode_t) *p2 == stop_memory + || (re_opcode_t) *p2 == start_memory)) + p2 += 3; + else if (p2 + 6 < pend + && (re_opcode_t) *p2 == dummy_failure_jump) + p2 += 6; + else + break; + } + + p1 = p + mcnt; + /* p1[0] ... p1[2] are the `on_failure_jump' corresponding + to the `maybe_finalize_jump' of this case. Examine what + follows. */ + + /* If we're at the end of the pattern, we can change. */ + if (p2 == pend) + { + /* Consider what happens when matching ":\(.*\)" + against ":/". I don't really understand this code + yet. */ + p[-3] = (unsigned char) pop_failure_jump; + DEBUG_PRINT1 + (" End of pattern: change to `pop_failure_jump'.\n"); + } + + else if ((re_opcode_t) *p2 == exactn + || (bufp->newline_anchor && (re_opcode_t) *p2 == endline)) + { + register unsigned int c + = *p2 == (unsigned char) endline ? '\n' : p2[2]; + + if ((re_opcode_t) p1[3] == exactn) + { + if (!(multibyte /* && (c != '\n') */ + && BASE_LEADING_CODE_P (c)) + ? c != p1[5] + : (STRING_CHAR (&p2[2], pend - &p2[2]) + != STRING_CHAR (&p1[5], pend - &p1[5]))) + { + p[-3] = (unsigned char) pop_failure_jump; + DEBUG_PRINT3 (" %c != %c => pop_failure_jump.\n", + c, p1[5]); + } + } + + else if ((re_opcode_t) p1[3] == charset + || (re_opcode_t) p1[3] == charset_not) + { + int not = (re_opcode_t) p1[3] == charset_not; + + if (multibyte /* && (c != '\n') */ + && BASE_LEADING_CODE_P (c)) + c = STRING_CHAR (&p2[2], pend - &p2[2]); + + /* Test if C is listed in charset (or charset_not) + at `&p1[3]'. */ + if (SINGLE_BYTE_CHAR_P (c)) + { + if (c < CHARSET_BITMAP_SIZE (&p1[3]) * BYTEWIDTH + && p1[5 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH))) + not = !not; + } + else if (CHARSET_RANGE_TABLE_EXISTS_P (&p1[3])) + CHARSET_LOOKUP_RANGE_TABLE (not, c, &p1[3]); + + /* `not' is equal to 1 if c would match, which means + that we can't change to pop_failure_jump. */ + if (!not) + { + p[-3] = (unsigned char) pop_failure_jump; + DEBUG_PRINT1 (" No match => pop_failure_jump.\n"); + } + } + } + else if ((re_opcode_t) *p2 == charset) + { + if ((re_opcode_t) p1[3] == exactn) + { + register unsigned int c = p1[5]; + int not = 0; + + if (multibyte && BASE_LEADING_CODE_P (c)) + c = STRING_CHAR (&p1[5], pend - &p1[5]); + + /* Test if C is listed in charset at `p2'. */ + if (SINGLE_BYTE_CHAR_P (c)) + { + if (c < CHARSET_BITMAP_SIZE (p2) * BYTEWIDTH + && (p2[2 + c / BYTEWIDTH] + & (1 << (c % BYTEWIDTH)))) + not = !not; + } + else if (CHARSET_RANGE_TABLE_EXISTS_P (p2)) + CHARSET_LOOKUP_RANGE_TABLE (not, c, p2); + + if (!not) + { + p[-3] = (unsigned char) pop_failure_jump; + DEBUG_PRINT1 (" No match => pop_failure_jump.\n"); + } + } + + /* It is hard to list up all the character in charset + P2 if it includes multibyte character. Give up in + such case. */ + else if (!multibyte || !CHARSET_RANGE_TABLE_EXISTS_P (p2)) + { + /* Now, we are sure that P2 has no range table. + So, for the size of bitmap in P2, `p2[1]' is + enough. But P1 may have range table, so the + size of bitmap table of P1 is extracted by + using macro `CHARSET_BITMAP_SIZE'. + + Since we know that all the character listed in + P2 is ASCII, it is enough to test only bitmap + table of P1. */ + + if ((re_opcode_t) p1[3] == charset_not) + { + int idx; + /* We win if the charset_not inside the loop lists + every character listed in the charset after. */ + for (idx = 0; idx < (int) p2[1]; idx++) + if (! (p2[2 + idx] == 0 + || (idx < CHARSET_BITMAP_SIZE (&p1[3]) + && ((p2[2 + idx] & ~ p1[5 + idx]) == 0)))) + break; + + if (idx == p2[1]) + { + p[-3] = (unsigned char) pop_failure_jump; + DEBUG_PRINT1 (" No match => pop_failure_jump.\n"); + } + } + else if ((re_opcode_t) p1[3] == charset) + { + int idx; + /* We win if the charset inside the loop + has no overlap with the one after the loop. */ + for (idx = 0; + (idx < (int) p2[1] + && idx < CHARSET_BITMAP_SIZE (&p1[3])); + idx++) + if ((p2[2 + idx] & p1[5 + idx]) != 0) + break; + + if (idx == p2[1] + || idx == CHARSET_BITMAP_SIZE (&p1[3])) + { + p[-3] = (unsigned char) pop_failure_jump; + DEBUG_PRINT1 (" No match => pop_failure_jump.\n"); + } + } + } + } + } + p -= 2; /* Point at relative address again. */ + if ((re_opcode_t) p[-1] != pop_failure_jump) + { + p[-1] = (unsigned char) jump; + DEBUG_PRINT1 (" Match => jump.\n"); + goto unconditional_jump; + } + /* Note fall through. */ + + + /* The end of a simple repeat has a pop_failure_jump back to + its matching on_failure_jump, where the latter will push a + failure point. The pop_failure_jump takes off failure + points put on by this pop_failure_jump's matching + on_failure_jump; we got through the pattern to here from the + matching on_failure_jump, so didn't fail. */ + case pop_failure_jump: + { + /* We need to pass separate storage for the lowest and + highest registers, even though we don't care about the + actual values. Otherwise, we will restore only one + register from the stack, since lowest will == highest in + `pop_failure_point'. */ + unsigned dummy_low_reg, dummy_high_reg; + unsigned char *pdummy; + const char *sdummy; + + DEBUG_PRINT1 ("EXECUTING pop_failure_jump.\n"); + POP_FAILURE_POINT (sdummy, pdummy, + dummy_low_reg, dummy_high_reg, + reg_dummy, reg_dummy, reg_info_dummy); + } + /* Note fall through. */ + + + /* Unconditionally jump (without popping any failure points). */ + case jump: + unconditional_jump: +#if defined (WINDOWSNT) && defined (emacs) + QUIT; +#endif + EXTRACT_NUMBER_AND_INCR (mcnt, p); /* Get the amount to jump. */ + DEBUG_PRINT2 ("EXECUTING jump %d ", mcnt); + p += mcnt; /* Do the jump. */ + DEBUG_PRINT2 ("(to 0x%x).\n", p); + break; + + + /* We need this opcode so we can detect where alternatives end + in `group_match_null_string_p' et al. */ + case jump_past_alt: + DEBUG_PRINT1 ("EXECUTING jump_past_alt.\n"); + goto unconditional_jump; + + + /* Normally, the on_failure_jump pushes a failure point, which + then gets popped at pop_failure_jump. We will end up at + pop_failure_jump, also, and with a pattern of, say, `a+', we + are skipping over the on_failure_jump, so we have to push + something meaningless for pop_failure_jump to pop. */ + case dummy_failure_jump: + DEBUG_PRINT1 ("EXECUTING dummy_failure_jump.\n"); + /* It doesn't matter what we push for the string here. What + the code at `fail' tests is the value for the pattern. */ + PUSH_FAILURE_POINT (0, 0, -2); + goto unconditional_jump; + + + /* At the end of an alternative, we need to push a dummy failure + point in case we are followed by a `pop_failure_jump', because + we don't want the failure point for the alternative to be + popped. For example, matching `(a|ab)*' against `aab' + requires that we match the `ab' alternative. */ + case push_dummy_failure: + DEBUG_PRINT1 ("EXECUTING push_dummy_failure.\n"); + /* See comments just above at `dummy_failure_jump' about the + two zeroes. */ + PUSH_FAILURE_POINT (0, 0, -2); + break; + + /* Have to succeed matching what follows at least n times. + After that, handle like `on_failure_jump'. */ + case succeed_n: + EXTRACT_NUMBER (mcnt, p + 2); + DEBUG_PRINT2 ("EXECUTING succeed_n %d.\n", mcnt); + + assert (mcnt >= 0); + /* Originally, this is how many times we HAVE to succeed. */ + if (mcnt > 0) + { + mcnt--; + p += 2; + STORE_NUMBER_AND_INCR (p, mcnt); + DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p, mcnt); + } + else if (mcnt == 0) + { + DEBUG_PRINT2 (" Setting two bytes from 0x%x to no_op.\n", p+2); + p[2] = (unsigned char) no_op; + p[3] = (unsigned char) no_op; + goto on_failure; + } + break; + + case jump_n: + EXTRACT_NUMBER (mcnt, p + 2); + DEBUG_PRINT2 ("EXECUTING jump_n %d.\n", mcnt); + + /* Originally, this is how many times we CAN jump. */ + if (mcnt) + { + mcnt--; + STORE_NUMBER (p + 2, mcnt); + goto unconditional_jump; + } + /* If don't have to jump any more, skip over the rest of command. */ + else + p += 4; + break; + + case set_number_at: + { + DEBUG_PRINT1 ("EXECUTING set_number_at.\n"); + + EXTRACT_NUMBER_AND_INCR (mcnt, p); + p1 = p + mcnt; + EXTRACT_NUMBER_AND_INCR (mcnt, p); + DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p1, mcnt); + STORE_NUMBER (p1, mcnt); + break; + } + + case wordbound: + DEBUG_PRINT1 ("EXECUTING wordbound.\n"); + + /* We SUCCEED in one of the following cases: */ + + /* Case 1: D is at the beginning or the end of string. */ + if (AT_STRINGS_BEG (d) || AT_STRINGS_END (d)) + break; + else + { + /* C1 is the character before D, S1 is the syntax of C1, C2 + is the character at D, and S2 is the syntax of C2. */ + int c1, c2, s1, s2; + int pos1 = PTR_TO_OFFSET (d - 1); + int charpos; + + GET_CHAR_BEFORE_2 (c1, d, string1, end1, string2, end2); + GET_CHAR_AFTER_2 (c2, d, string1, end1, string2, end2); +#ifdef emacs + charpos = SYNTAX_TABLE_BYTE_TO_CHAR (pos1); + UPDATE_SYNTAX_TABLE (charpos); +#endif + s1 = SYNTAX (c1); +#ifdef emacs + UPDATE_SYNTAX_TABLE_FORWARD (charpos + 1); +#endif + s2 = SYNTAX (c2); + + if (/* Case 2: Only one of S1 and S2 is Sword. */ + ((s1 == Sword) != (s2 == Sword)) + /* Case 3: Both of S1 and S2 are Sword, and macro + WORD_BOUNDARY_P (C1, C2) returns nonzero. */ + || ((s1 == Sword) && WORD_BOUNDARY_P (c1, c2))) + break; + } + goto fail; + + case notwordbound: + DEBUG_PRINT1 ("EXECUTING notwordbound.\n"); + + /* We FAIL in one of the following cases: */ + + /* Case 1: D is at the beginning or the end of string. */ + if (AT_STRINGS_BEG (d) || AT_STRINGS_END (d)) + goto fail; + else + { + /* C1 is the character before D, S1 is the syntax of C1, C2 + is the character at D, and S2 is the syntax of C2. */ + int c1, c2, s1, s2; + int pos1 = PTR_TO_OFFSET (d - 1); + int charpos; + + GET_CHAR_BEFORE_2 (c1, d, string1, end1, string2, end2); + GET_CHAR_AFTER_2 (c2, d, string1, end1, string2, end2); +#ifdef emacs + charpos = SYNTAX_TABLE_BYTE_TO_CHAR (pos1); + UPDATE_SYNTAX_TABLE (charpos); +#endif + s1 = SYNTAX (c1); +#ifdef emacs + UPDATE_SYNTAX_TABLE_FORWARD (charpos + 1); +#endif + s2 = SYNTAX (c2); + + if (/* Case 2: Only one of S1 and S2 is Sword. */ + ((s1 == Sword) != (s2 == Sword)) + /* Case 3: Both of S1 and S2 are Sword, and macro + WORD_BOUNDARY_P (C1, C2) returns nonzero. */ + || ((s1 == Sword) && WORD_BOUNDARY_P (c1, c2))) + goto fail; + } + break; + + case wordbeg: + DEBUG_PRINT1 ("EXECUTING wordbeg.\n"); + + /* We FAIL in one of the following cases: */ + + /* Case 1: D is at the end of string. */ + if (AT_STRINGS_END (d)) + goto fail; + else + { + /* C1 is the character before D, S1 is the syntax of C1, C2 + is the character at D, and S2 is the syntax of C2. */ + int c1, c2, s1, s2; + int pos1 = PTR_TO_OFFSET (d); + int charpos; + + GET_CHAR_AFTER_2 (c2, d, string1, end1, string2, end2); +#ifdef emacs + charpos = SYNTAX_TABLE_BYTE_TO_CHAR (pos1); + UPDATE_SYNTAX_TABLE (charpos); +#endif + s2 = SYNTAX (c2); + + /* Case 2: S2 is not Sword. */ + if (s2 != Sword) + goto fail; + + /* Case 3: D is not at the beginning of string ... */ + if (!AT_STRINGS_BEG (d)) + { + GET_CHAR_BEFORE_2 (c1, d, string1, end1, string2, end2); +#ifdef emacs + UPDATE_SYNTAX_TABLE_BACKWARD (charpos - 1); +#endif + s1 = SYNTAX (c1); + + /* ... and S1 is Sword, and WORD_BOUNDARY_P (C1, C2) + returns 0. */ + if ((s1 == Sword) && !WORD_BOUNDARY_P (c1, c2)) + goto fail; + } + } + break; + + case wordend: + DEBUG_PRINT1 ("EXECUTING wordend.\n"); + + /* We FAIL in one of the following cases: */ + + /* Case 1: D is at the beginning of string. */ + if (AT_STRINGS_BEG (d)) + goto fail; + else + { + /* C1 is the character before D, S1 is the syntax of C1, C2 + is the character at D, and S2 is the syntax of C2. */ + int c1, c2, s1, s2; + int pos1 = PTR_TO_OFFSET (d); + int charpos; + + GET_CHAR_BEFORE_2 (c1, d, string1, end1, string2, end2); +#ifdef emacs + charpos = SYNTAX_TABLE_BYTE_TO_CHAR (pos1 - 1); + UPDATE_SYNTAX_TABLE (charpos); +#endif + s1 = SYNTAX (c1); + + /* Case 2: S1 is not Sword. */ + if (s1 != Sword) + goto fail; + + /* Case 3: D is not at the end of string ... */ + if (!AT_STRINGS_END (d)) + { + GET_CHAR_AFTER_2 (c2, d, string1, end1, string2, end2); +#ifdef emacs + UPDATE_SYNTAX_TABLE_FORWARD (charpos); +#endif + s2 = SYNTAX (c2); + + /* ... and S2 is Sword, and WORD_BOUNDARY_P (C1, C2) + returns 0. */ + if ((s2 == Sword) && !WORD_BOUNDARY_P (c1, c2)) + goto fail; + } + } + break; + +#ifdef emacs + case before_dot: + DEBUG_PRINT1 ("EXECUTING before_dot.\n"); + if (PTR_BYTE_POS ((unsigned char *) d) >= PT_BYTE) + goto fail; + break; + + case at_dot: + DEBUG_PRINT1 ("EXECUTING at_dot.\n"); + if (PTR_BYTE_POS ((unsigned char *) d) != PT_BYTE) + goto fail; + break; + + case after_dot: + DEBUG_PRINT1 ("EXECUTING after_dot.\n"); + if (PTR_BYTE_POS ((unsigned char *) d) <= PT_BYTE) + goto fail; + break; + + case syntaxspec: + DEBUG_PRINT2 ("EXECUTING syntaxspec %d.\n", mcnt); + mcnt = *p++; + goto matchsyntax; + + case wordchar: + DEBUG_PRINT1 ("EXECUTING Emacs wordchar.\n"); + mcnt = (int) Sword; + matchsyntax: + PREFETCH (); +#ifdef emacs + { + int pos1 = SYNTAX_TABLE_BYTE_TO_CHAR (PTR_TO_OFFSET (d)); + UPDATE_SYNTAX_TABLE (pos1); + } +#endif + { + int c, len; + + if (multibyte) + /* we must concern about multibyte form, ... */ + c = STRING_CHAR_AND_LENGTH (d, dend - d, len); + else + /* everything should be handled as ASCII, even though it + looks like multibyte form. */ + c = *d, len = 1; + + if (SYNTAX (c) != (enum syntaxcode) mcnt) + goto fail; + d += len; + } + SET_REGS_MATCHED (); + break; + + case notsyntaxspec: + DEBUG_PRINT2 ("EXECUTING notsyntaxspec %d.\n", mcnt); + mcnt = *p++; + goto matchnotsyntax; + + case notwordchar: + DEBUG_PRINT1 ("EXECUTING Emacs notwordchar.\n"); + mcnt = (int) Sword; + matchnotsyntax: + PREFETCH (); +#ifdef emacs + { + int pos1 = SYNTAX_TABLE_BYTE_TO_CHAR (PTR_TO_OFFSET (d)); + UPDATE_SYNTAX_TABLE (pos1); + } +#endif + { + int c, len; + + if (multibyte) + c = STRING_CHAR_AND_LENGTH (d, dend - d, len); + else + c = *d, len = 1; + + if (SYNTAX (c) == (enum syntaxcode) mcnt) + goto fail; + d += len; + } + SET_REGS_MATCHED (); + break; + + case categoryspec: + DEBUG_PRINT2 ("EXECUTING categoryspec %d.\n", *p); + mcnt = *p++; + PREFETCH (); + { + int c, len; + + if (multibyte) + c = STRING_CHAR_AND_LENGTH (d, dend - d, len); + else + c = *d, len = 1; + + if (!CHAR_HAS_CATEGORY (c, mcnt)) + goto fail; + d += len; + } + SET_REGS_MATCHED (); + break; + + case notcategoryspec: + DEBUG_PRINT2 ("EXECUTING notcategoryspec %d.\n", *p); + mcnt = *p++; + PREFETCH (); + { + int c, len; + + if (multibyte) + c = STRING_CHAR_AND_LENGTH (d, dend - d, len); + else + c = *d, len = 1; + + if (CHAR_HAS_CATEGORY (c, mcnt)) + goto fail; + d += len; + } + SET_REGS_MATCHED (); + break; + +#else /* not emacs */ + case wordchar: + DEBUG_PRINT1 ("EXECUTING non-Emacs wordchar.\n"); + PREFETCH (); + if (!WORDCHAR_P (d)) + goto fail; + SET_REGS_MATCHED (); + d++; + break; + + case notwordchar: + DEBUG_PRINT1 ("EXECUTING non-Emacs notwordchar.\n"); + PREFETCH (); + if (WORDCHAR_P (d)) + goto fail; + SET_REGS_MATCHED (); + d++; + break; +#endif /* not emacs */ + + default: + abort (); + } + continue; /* Successfully executed one pattern command; keep going. */ + + + /* We goto here if a matching operation fails. */ + fail: +#if defined (WINDOWSNT) && defined (emacs) + QUIT; +#endif + if (!FAIL_STACK_EMPTY ()) + { /* A restart point is known. Restore to that state. */ + DEBUG_PRINT1 ("\nFAIL:\n"); + POP_FAILURE_POINT (d, p, + lowest_active_reg, highest_active_reg, + regstart, regend, reg_info); + + /* If this failure point is a dummy, try the next one. */ + if (!p) + goto fail; + + /* If we failed to the end of the pattern, don't examine *p. */ + assert (p <= pend); + if (p < pend) + { + boolean is_a_jump_n = false; + + /* If failed to a backwards jump that's part of a repetition + loop, need to pop this failure point and use the next one. */ + switch ((re_opcode_t) *p) + { + case jump_n: + is_a_jump_n = true; + case maybe_pop_jump: + case pop_failure_jump: + case jump: + p1 = p + 1; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + p1 += mcnt; + + if ((is_a_jump_n && (re_opcode_t) *p1 == succeed_n) + || (!is_a_jump_n + && (re_opcode_t) *p1 == on_failure_jump)) + goto fail; + break; + default: + /* do nothing */ ; + } + } + + if (d >= string1 && d <= end1) + dend = end_match_1; + } + else + break; /* Matching at this starting point really fails. */ + } /* for (;;) */ + + if (best_regs_set) + goto restore_best_regs; + + FREE_VARIABLES (); + + return -1; /* Failure to match. */ +} /* re_match_2 */ + +/* Subroutine definitions for re_match_2. */ + + +/* We are passed P pointing to a register number after a start_memory. + + Return true if the pattern up to the corresponding stop_memory can + match the empty string, and false otherwise. + + If we find the matching stop_memory, sets P to point to one past its number. + Otherwise, sets P to an undefined byte less than or equal to END. + + We don't handle duplicates properly (yet). */ + +static boolean +group_match_null_string_p (p, end, reg_info) + unsigned char **p, *end; + register_info_type *reg_info; +{ + int mcnt; + /* Point to after the args to the start_memory. */ + unsigned char *p1 = *p + 2; + + while (p1 < end) + { + /* Skip over opcodes that can match nothing, and return true or + false, as appropriate, when we get to one that can't, or to the + matching stop_memory. */ + + switch ((re_opcode_t) *p1) + { + /* Could be either a loop or a series of alternatives. */ + case on_failure_jump: + p1++; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + + /* If the next operation is not a jump backwards in the + pattern. */ + + if (mcnt >= 0) + { + /* Go through the on_failure_jumps of the alternatives, + seeing if any of the alternatives cannot match nothing. + The last alternative starts with only a jump, + whereas the rest start with on_failure_jump and end + with a jump, e.g., here is the pattern for `a|b|c': + + /on_failure_jump/0/6/exactn/1/a/jump_past_alt/0/6 + /on_failure_jump/0/6/exactn/1/b/jump_past_alt/0/3 + /exactn/1/c + + So, we have to first go through the first (n-1) + alternatives and then deal with the last one separately. */ + + + /* Deal with the first (n-1) alternatives, which start + with an on_failure_jump (see above) that jumps to right + past a jump_past_alt. */ + + while ((re_opcode_t) p1[mcnt-3] == jump_past_alt) + { + /* `mcnt' holds how many bytes long the alternative + is, including the ending `jump_past_alt' and + its number. */ + + if (!alt_match_null_string_p (p1, p1 + mcnt - 3, + reg_info)) + return false; + + /* Move to right after this alternative, including the + jump_past_alt. */ + p1 += mcnt; + + /* Break if it's the beginning of an n-th alternative + that doesn't begin with an on_failure_jump. */ + if ((re_opcode_t) *p1 != on_failure_jump) + break; + + /* Still have to check that it's not an n-th + alternative that starts with an on_failure_jump. */ + p1++; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + if ((re_opcode_t) p1[mcnt-3] != jump_past_alt) + { + /* Get to the beginning of the n-th alternative. */ + p1 -= 3; + break; + } + } + + /* Deal with the last alternative: go back and get number + of the `jump_past_alt' just before it. `mcnt' contains + the length of the alternative. */ + EXTRACT_NUMBER (mcnt, p1 - 2); + + if (!alt_match_null_string_p (p1, p1 + mcnt, reg_info)) + return false; + + p1 += mcnt; /* Get past the n-th alternative. */ + } /* if mcnt > 0 */ + break; + + + case stop_memory: + assert (p1[1] == **p); + *p = p1 + 2; + return true; + + + default: + if (!common_op_match_null_string_p (&p1, end, reg_info)) + return false; + } + } /* while p1 < end */ + + return false; +} /* group_match_null_string_p */ + + +/* Similar to group_match_null_string_p, but doesn't deal with alternatives: + It expects P to be the first byte of a single alternative and END one + byte past the last. The alternative can contain groups. */ + +static boolean +alt_match_null_string_p (p, end, reg_info) + unsigned char *p, *end; + register_info_type *reg_info; +{ + int mcnt; + unsigned char *p1 = p; + + while (p1 < end) + { + /* Skip over opcodes that can match nothing, and break when we get + to one that can't. */ + + switch ((re_opcode_t) *p1) + { + /* It's a loop. */ + case on_failure_jump: + p1++; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + p1 += mcnt; + break; + + default: + if (!common_op_match_null_string_p (&p1, end, reg_info)) + return false; + } + } /* while p1 < end */ + + return true; +} /* alt_match_null_string_p */ + + +/* Deals with the ops common to group_match_null_string_p and + alt_match_null_string_p. + + Sets P to one after the op and its arguments, if any. */ + +static boolean +common_op_match_null_string_p (p, end, reg_info) + unsigned char **p, *end; + register_info_type *reg_info; +{ + int mcnt; + boolean ret; + int reg_no; + unsigned char *p1 = *p; + + switch ((re_opcode_t) *p1++) + { + case no_op: + case begline: + case endline: + case begbuf: + case endbuf: + case wordbeg: + case wordend: + case wordbound: + case notwordbound: +#ifdef emacs + case before_dot: + case at_dot: + case after_dot: +#endif + break; + + case start_memory: + reg_no = *p1; + assert (reg_no > 0 && reg_no <= MAX_REGNUM); + ret = group_match_null_string_p (&p1, end, reg_info); + + /* Have to set this here in case we're checking a group which + contains a group and a back reference to it. */ + + if (REG_MATCH_NULL_STRING_P (reg_info[reg_no]) == MATCH_NULL_UNSET_VALUE) + REG_MATCH_NULL_STRING_P (reg_info[reg_no]) = ret; + + if (!ret) + return false; + break; + + /* If this is an optimized succeed_n for zero times, make the jump. */ + case jump: + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + if (mcnt >= 0) + p1 += mcnt; + else + return false; + break; + + case succeed_n: + /* Get to the number of times to succeed. */ + p1 += 2; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + + if (mcnt == 0) + { + p1 -= 4; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + p1 += mcnt; + } + else + return false; + break; + + case duplicate: + if (!REG_MATCH_NULL_STRING_P (reg_info[*p1])) + return false; + break; + + case set_number_at: + p1 += 4; + + default: + /* All other opcodes mean we cannot match the empty string. */ + return false; + } + + *p = p1; + return true; +} /* common_op_match_null_string_p */ + + +/* Return zero if TRANSLATE[S1] and TRANSLATE[S2] are identical for LEN + bytes; nonzero otherwise. */ + +static int +bcmp_translate (s1, s2, len, translate) + unsigned char *s1, *s2; + register int len; + RE_TRANSLATE_TYPE translate; +{ + register unsigned char *p1 = s1, *p2 = s2; + unsigned char *p1_end = s1 + len; + unsigned char *p2_end = s2 + len; + + while (p1 != p1_end && p2 != p2_end) + { + int p1_charlen, p2_charlen; + int p1_ch, p2_ch; + + p1_ch = STRING_CHAR_AND_LENGTH (p1, p1_end - p1, p1_charlen); + p2_ch = STRING_CHAR_AND_LENGTH (p2, p2_end - p2, p2_charlen); + + if (RE_TRANSLATE (translate, p1_ch) + != RE_TRANSLATE (translate, p2_ch)) + return 1; + + p1 += p1_charlen, p2 += p2_charlen; + } + + if (p1 != p1_end || p2 != p2_end) + return 1; + + return 0; +} + +/* Entry points for GNU code. */ + +/* re_compile_pattern is the GNU regular expression compiler: it + compiles PATTERN (of length SIZE) and puts the result in BUFP. + Returns 0 if the pattern was valid, otherwise an error string. + + Assumes the `allocated' (and perhaps `buffer') and `translate' fields + are set in BUFP on entry. + + We call regex_compile to do the actual compilation. */ + +const char * +re_compile_pattern (pattern, length, bufp) + const char *pattern; + int length; + struct re_pattern_buffer *bufp; +{ + reg_errcode_t ret; + + /* GNU code is written to assume at least RE_NREGS registers will be set + (and at least one extra will be -1). */ + bufp->regs_allocated = REGS_UNALLOCATED; + + /* And GNU code determines whether or not to get register information + by passing null for the REGS argument to re_match, etc., not by + setting no_sub. */ + bufp->no_sub = 0; + + /* Match anchors at newline. */ + bufp->newline_anchor = 1; + + ret = regex_compile (pattern, length, re_syntax_options, bufp); + + if (!ret) + return NULL; + return gettext (re_error_msgid[(int) ret]); +} + +/* Entry points compatible with 4.2 BSD regex library. We don't define + them unless specifically requested. */ + +#if defined (_REGEX_RE_COMP) || defined (_LIBC) + +/* BSD has one and only one pattern buffer. */ +static struct re_pattern_buffer re_comp_buf; + +char * +#ifdef _LIBC +/* Make these definitions weak in libc, so POSIX programs can redefine + these names if they don't use our functions, and still use + regcomp/regexec below without link errors. */ +weak_function +#endif +re_comp (s) + const char *s; +{ + reg_errcode_t ret; + + if (!s) + { + if (!re_comp_buf.buffer) + return gettext ("No previous regular expression"); + return 0; + } + + if (!re_comp_buf.buffer) + { + re_comp_buf.buffer = (unsigned char *) malloc (200); + if (re_comp_buf.buffer == NULL) + /* CVS: Yes, we're discarding `const' here if !HAVE_LIBINTL. */ + return (char *) gettext (re_error_msgid[(int) REG_ESPACE]); + re_comp_buf.allocated = 200; + + re_comp_buf.fastmap = (char *) malloc (1 << BYTEWIDTH); + if (re_comp_buf.fastmap == NULL) + /* CVS: Yes, we're discarding `const' here if !HAVE_LIBINTL. */ + return (char *) gettext (re_error_msgid[(int) REG_ESPACE]); + } + + /* Since `re_exec' always passes NULL for the `regs' argument, we + don't need to initialize the pattern buffer fields which affect it. */ + + /* Match anchors at newlines. */ + re_comp_buf.newline_anchor = 1; + + ret = regex_compile (s, strlen (s), re_syntax_options, &re_comp_buf); + + if (!ret) + return NULL; + + /* Yes, we're discarding `const' here if !HAVE_LIBINTL. */ + return (char *) gettext (re_error_msgid[(int) ret]); +} + + +int +#ifdef _LIBC +weak_function +#endif +re_exec (s) + const char *s; +{ + const int len = strlen (s); + return + 0 <= re_search (&re_comp_buf, s, len, 0, len, (struct re_registers *) 0); +} +#endif /* _REGEX_RE_COMP */ + +/* POSIX.2 functions. Don't define these for Emacs. */ + +#ifndef emacs + +/* regcomp takes a regular expression as a string and compiles it. + + PREG is a regex_t *. We do not expect any fields to be initialized, + since POSIX says we shouldn't. Thus, we set + + `buffer' to the compiled pattern; + `used' to the length of the compiled pattern; + `syntax' to RE_SYNTAX_POSIX_EXTENDED if the + REG_EXTENDED bit in CFLAGS is set; otherwise, to + RE_SYNTAX_POSIX_BASIC; + `newline_anchor' to REG_NEWLINE being set in CFLAGS; + `fastmap' and `fastmap_accurate' to zero; + `re_nsub' to the number of subexpressions in PATTERN. + + PATTERN is the address of the pattern string. + + CFLAGS is a series of bits which affect compilation. + + If REG_EXTENDED is set, we use POSIX extended syntax; otherwise, we + use POSIX basic syntax. + + If REG_NEWLINE is set, then . and [^...] don't match newline. + Also, regexec will try a match beginning after every newline. + + If REG_ICASE is set, then we considers upper- and lowercase + versions of letters to be equivalent when matching. + + If REG_NOSUB is set, then when PREG is passed to regexec, that + routine will report only success or failure, and nothing about the + registers. + + It returns 0 if it succeeds, nonzero if it doesn't. (See regex.h for + the return codes and their meanings.) */ + +int +regcomp (preg, pattern, cflags) + regex_t *preg; + const char *pattern; + int cflags; +{ + reg_errcode_t ret; + unsigned syntax + = (cflags & REG_EXTENDED) ? + RE_SYNTAX_POSIX_EXTENDED : RE_SYNTAX_POSIX_BASIC; + + /* regex_compile will allocate the space for the compiled pattern. */ + preg->buffer = 0; + preg->allocated = 0; + preg->used = 0; + + /* Don't bother to use a fastmap when searching. This simplifies the + REG_NEWLINE case: if we used a fastmap, we'd have to put all the + characters after newlines into the fastmap. This way, we just try + every character. */ + preg->fastmap = 0; + + if (cflags & REG_ICASE) + { + unsigned i; + + preg->translate + = (RE_TRANSLATE_TYPE) malloc (CHAR_SET_SIZE + * sizeof (*(RE_TRANSLATE_TYPE)0)); + if (preg->translate == NULL) + return (int) REG_ESPACE; + + /* Map uppercase characters to corresponding lowercase ones. */ + for (i = 0; i < CHAR_SET_SIZE; i++) + preg->translate[i] = ISUPPER (i) ? tolower (i) : i; + } + else + preg->translate = NULL; + + /* If REG_NEWLINE is set, newlines are treated differently. */ + if (cflags & REG_NEWLINE) + { /* REG_NEWLINE implies neither . nor [^...] match newline. */ + syntax &= ~RE_DOT_NEWLINE; + syntax |= RE_HAT_LISTS_NOT_NEWLINE; + /* It also changes the matching behavior. */ + preg->newline_anchor = 1; + } + else + preg->newline_anchor = 0; + + preg->no_sub = !!(cflags & REG_NOSUB); + + /* POSIX says a null character in the pattern terminates it, so we + can use strlen here in compiling the pattern. */ + ret = regex_compile (pattern, strlen (pattern), syntax, preg); + + /* POSIX doesn't distinguish between an unmatched open-group and an + unmatched close-group: both are REG_EPAREN. */ + if (ret == REG_ERPAREN) ret = REG_EPAREN; + + return (int) ret; +} + + +/* regexec searches for a given pattern, specified by PREG, in the + string STRING. + + If NMATCH is zero or REG_NOSUB was set in the cflags argument to + `regcomp', we ignore PMATCH. Otherwise, we assume PMATCH has at + least NMATCH elements, and we set them to the offsets of the + corresponding matched substrings. + + EFLAGS specifies `execution flags' which affect matching: if + REG_NOTBOL is set, then ^ does not match at the beginning of the + string; if REG_NOTEOL is set, then $ does not match at the end. + + We return 0 if we find a match and REG_NOMATCH if not. */ + +int +regexec (preg, string, nmatch, pmatch, eflags) + const regex_t *preg; + const char *string; + size_t nmatch; + regmatch_t pmatch[]; + int eflags; +{ + int ret; + struct re_registers regs; + regex_t private_preg; + int len = strlen (string); + boolean want_reg_info = !preg->no_sub && nmatch > 0; + + private_preg = *preg; + + private_preg.not_bol = !!(eflags & REG_NOTBOL); + private_preg.not_eol = !!(eflags & REG_NOTEOL); + + /* The user has told us exactly how many registers to return + information about, via `nmatch'. We have to pass that on to the + matching routines. */ + private_preg.regs_allocated = REGS_FIXED; + + if (want_reg_info) + { + regs.num_regs = nmatch; + regs.start = TALLOC (nmatch, regoff_t); + regs.end = TALLOC (nmatch, regoff_t); + if (regs.start == NULL || regs.end == NULL) + return (int) REG_NOMATCH; + } + + /* Perform the searching operation. */ + ret = re_search (&private_preg, string, len, + /* start: */ 0, /* range: */ len, + want_reg_info ? ®s : (struct re_registers *) 0); + + /* Copy the register information to the POSIX structure. */ + if (want_reg_info) + { + if (ret >= 0) + { + unsigned r; + + for (r = 0; r < nmatch; r++) + { + pmatch[r].rm_so = regs.start[r]; + pmatch[r].rm_eo = regs.end[r]; + } + } + + /* If we needed the temporary register info, free the space now. */ + free (regs.start); + free (regs.end); + } + + /* We want zero return to mean success, unlike `re_search'. */ + return ret >= 0 ? (int) REG_NOERROR : (int) REG_NOMATCH; +} + + +/* Returns a message corresponding to an error code, ERRCODE, returned + from either regcomp or regexec. We don't use PREG here. */ + +size_t +regerror (errcode, preg, errbuf, errbuf_size) + int errcode; + const regex_t *preg; + char *errbuf; + size_t errbuf_size; +{ + const char *msg; + size_t msg_size; + + if (errcode < 0 + || errcode >= (sizeof (re_error_msgid) / sizeof (re_error_msgid[0]))) + /* Only error codes returned by the rest of the code should be passed + to this routine. If we are given anything else, or if other regex + code generates an invalid error code, then the program has a bug. + Dump core so we can fix it. */ + abort (); + + msg = gettext (re_error_msgid[errcode]); + + msg_size = strlen (msg) + 1; /* Includes the null. */ + + if (errbuf_size != 0) + { + if (msg_size > errbuf_size) + { + strncpy (errbuf, msg, errbuf_size - 1); + errbuf[errbuf_size - 1] = 0; + } + else + strcpy (errbuf, msg); + } + + return msg_size; +} + + +/* Free dynamically allocated space used by PREG. */ + +void +regfree (preg) + regex_t *preg; +{ + if (preg->buffer != NULL) + free (preg->buffer); + preg->buffer = NULL; + + preg->allocated = 0; + preg->used = 0; + + if (preg->fastmap != NULL) + free (preg->fastmap); + preg->fastmap = NULL; + preg->fastmap_accurate = 0; + + if (preg->translate != NULL) + free (preg->translate); + preg->translate = NULL; +} + +#endif /* not emacs */ diff --git a/sys/src/ape/cmd/diff/regex.h b/sys/src/ape/cmd/diff/regex.h new file mode 100755 index 000000000..411e7d26e --- /dev/null +++ b/sys/src/ape/cmd/diff/regex.h @@ -0,0 +1,510 @@ +/* Definitions for data structures and routines for the regular + expression library, version 0.12. + + Copyright (C) 1985, 89, 90, 91, 92, 93, 95 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, + USA. */ + +#ifndef __REGEXP_LIBRARY_H__ +#define __REGEXP_LIBRARY_H__ + +/* POSIX says that <sys/types.h> must be included (by the caller) before + <regex.h>. */ + +#if !defined (_POSIX_C_SOURCE) && !defined (_POSIX_SOURCE) && defined (VMS) +/* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it + should be there. */ +#include <stddef.h> +#endif + + +/* The following bits are used to determine the regexp syntax we + recognize. The set/not-set meanings are chosen so that Emacs syntax + remains the value 0. The bits are given in alphabetical order, and + the definitions shifted by one from the previous bit; thus, when we + add or remove a bit, only one other definition need change. */ +typedef unsigned reg_syntax_t; + +/* If this bit is not set, then \ inside a bracket expression is literal. + If set, then such a \ quotes the following character. */ +#define RE_BACKSLASH_ESCAPE_IN_LISTS (1) + +/* If this bit is not set, then + and ? are operators, and \+ and \? are + literals. + If set, then \+ and \? are operators and + and ? are literals. */ +#define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) + +/* If this bit is set, then character classes are supported. They are: + [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], + [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. + If not set, then character classes are not supported. */ +#define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) + +/* If this bit is set, then ^ and $ are always anchors (outside bracket + expressions, of course). + If this bit is not set, then it depends: + ^ is an anchor if it is at the beginning of a regular + expression or after an open-group or an alternation operator; + $ is an anchor if it is at the end of a regular expression, or + before a close-group or an alternation operator. + + This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because + POSIX draft 11.2 says that * etc. in leading positions is undefined. + We already implemented a previous draft which made those constructs + invalid, though, so we haven't changed the code back. */ +#define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) + +/* If this bit is set, then special characters are always special + regardless of where they are in the pattern. + If this bit is not set, then special characters are special only in + some contexts; otherwise they are ordinary. Specifically, + * + ? and intervals are only special when not after the beginning, + open-group, or alternation operator. */ +#define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) + +/* If this bit is set, then *, +, ?, and { cannot be first in an re or + immediately after an alternation or begin-group operator. */ +#define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) + +/* If this bit is set, then . matches newline. + If not set, then it doesn't. */ +#define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) + +/* If this bit is set, then . doesn't match NUL. + If not set, then it does. */ +#define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) + +/* If this bit is set, nonmatching lists [^...] do not match newline. + If not set, they do. */ +#define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) + +/* If this bit is set, either \{...\} or {...} defines an + interval, depending on RE_NO_BK_BRACES. + If not set, \{, \}, {, and } are literals. */ +#define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) + +/* If this bit is set, +, ? and | aren't recognized as operators. + If not set, they are. */ +#define RE_LIMITED_OPS (RE_INTERVALS << 1) + +/* If this bit is set, newline is an alternation operator. + If not set, newline is literal. */ +#define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) + +/* If this bit is set, then `{...}' defines an interval, and \{ and \} + are literals. + If not set, then `\{...\}' defines an interval. */ +#define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) + +/* If this bit is set, (...) defines a group, and \( and \) are literals. + If not set, \(...\) defines a group, and ( and ) are literals. */ +#define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) + +/* If this bit is set, then \<digit> matches <digit>. + If not set, then \<digit> is a back-reference. */ +#define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) + +/* If this bit is set, then | is an alternation operator, and \| is literal. + If not set, then \| is an alternation operator, and | is literal. */ +#define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) + +/* If this bit is set, then an ending range point collating higher + than the starting range point, as in [z-a], is invalid. + If not set, then when ending range point collates higher than the + starting range point, the range is ignored. */ +#define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) + +/* If this bit is set, then an unmatched ) is ordinary. + If not set, then an unmatched ) is invalid. */ +#define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) + +/* If this bit is set, succeed as soon as we match the whole pattern, + without further backtracking. */ +#define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) + +/* This global variable defines the particular regexp syntax to use (for + some interfaces). When a regexp is compiled, the syntax used is + stored in the pattern buffer, so changing this does not affect + already-compiled regexps. */ +extern reg_syntax_t re_syntax_options; + +#ifdef emacs +/* In Emacs, this is the string or buffer in which we + are matching. It is used for looking up syntax properties. */ +extern Lisp_Object re_match_object; +#endif + + +/* Define combinations of the above bits for the standard possibilities. + (The [[[ comments delimit what gets put into the Texinfo file, so + don't delete them!) */ +/* [[[begin syntaxes]]] */ +#define RE_SYNTAX_EMACS 0 + +#define RE_SYNTAX_AWK \ + (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ + | RE_NO_BK_PARENS | RE_NO_BK_REFS \ + | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ + | RE_UNMATCHED_RIGHT_PAREN_ORD) + +#define RE_SYNTAX_POSIX_AWK \ + (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS) + +#define RE_SYNTAX_GREP \ + (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ + | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ + | RE_NEWLINE_ALT) + +#define RE_SYNTAX_EGREP \ + (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ + | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ + | RE_NO_BK_VBAR) + +#define RE_SYNTAX_POSIX_EGREP \ + (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES) + +/* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ +#define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC + +#define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC + +/* Syntax bits common to both basic and extended POSIX regex syntax. */ +#define _RE_SYNTAX_POSIX_COMMON \ + (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ + | RE_INTERVALS | RE_NO_EMPTY_RANGES) + +#define RE_SYNTAX_POSIX_BASIC \ + (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM) + +/* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes + RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this + isn't minimal, since other operators, such as \`, aren't disabled. */ +#define RE_SYNTAX_POSIX_MINIMAL_BASIC \ + (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) + +#define RE_SYNTAX_POSIX_EXTENDED \ + (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ + | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ + | RE_UNMATCHED_RIGHT_PAREN_ORD) + +/* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS + replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */ +#define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ + (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ + | RE_NO_BK_PARENS | RE_NO_BK_REFS \ + | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) +/* [[[end syntaxes]]] */ + +/* Maximum number of duplicates an interval can allow. Some systems + (erroneously) define this in other header files, but we want our + value, so remove any previous define. */ +#ifdef RE_DUP_MAX +#undef RE_DUP_MAX +#endif +#define RE_DUP_MAX ((1 << 15) - 1) + + +/* POSIX `cflags' bits (i.e., information for `regcomp'). */ + +/* If this bit is set, then use extended regular expression syntax. + If not set, then use basic regular expression syntax. */ +#define REG_EXTENDED 1 + +/* If this bit is set, then ignore case when matching. + If not set, then case is significant. */ +#define REG_ICASE (REG_EXTENDED << 1) + +/* If this bit is set, then anchors do not match at newline + characters in the string. + If not set, then anchors do match at newlines. */ +#define REG_NEWLINE (REG_ICASE << 1) + +/* If this bit is set, then report only success or fail in regexec. + If not set, then returns differ between not matching and errors. */ +#define REG_NOSUB (REG_NEWLINE << 1) + + +/* POSIX `eflags' bits (i.e., information for regexec). */ + +/* If this bit is set, then the beginning-of-line operator doesn't match + the beginning of the string (presumably because it's not the + beginning of a line). + If not set, then the beginning-of-line operator does match the + beginning of the string. */ +#define REG_NOTBOL 1 + +/* Like REG_NOTBOL, except for the end-of-line. */ +#define REG_NOTEOL (1 << 1) + + +/* If any error codes are removed, changed, or added, update the + `re_error_msg' table in regex.c. */ +typedef enum +{ + REG_NOERROR = 0, /* Success. */ + REG_NOMATCH, /* Didn't find a match (for regexec). */ + + /* POSIX regcomp return error codes. (In the order listed in the + standard.) */ + REG_BADPAT, /* Invalid pattern. */ + REG_ECOLLATE, /* Not implemented. */ + REG_ECTYPE, /* Invalid character class name. */ + REG_EESCAPE, /* Trailing backslash. */ + REG_ESUBREG, /* Invalid back reference. */ + REG_EBRACK, /* Unmatched left bracket. */ + REG_EPAREN, /* Parenthesis imbalance. */ + REG_EBRACE, /* Unmatched \{. */ + REG_BADBR, /* Invalid contents of \{\}. */ + REG_ERANGE, /* Invalid range end. */ + REG_ESPACE, /* Ran out of memory. */ + REG_BADRPT, /* No preceding re for repetition op. */ + + /* Error codes we've added. */ + REG_EEND, /* Premature end. */ + REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ + REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ +} reg_errcode_t; + +/* This data structure represents a compiled pattern. Before calling + the pattern compiler, the fields `buffer', `allocated', `fastmap', + `translate', and `no_sub' can be set. After the pattern has been + compiled, the `re_nsub' field is available. All other fields are + private to the regex routines. */ + +#ifndef RE_TRANSLATE_TYPE +#define RE_TRANSLATE_TYPE char * +#define RE_TRANSLATE(TBL, C) ((TBL)[C]) +#define RE_TRANSLATE_P(TBL) (TBL) +#endif + +struct re_pattern_buffer +{ +/* [[[begin pattern_buffer]]] */ + /* Space that holds the compiled pattern. It is declared as + `unsigned char *' because its elements are + sometimes used as array indexes. */ + unsigned char *buffer; + + /* Number of bytes to which `buffer' points. */ + unsigned long allocated; + + /* Number of bytes actually used in `buffer'. */ + unsigned long used; + + /* Syntax setting with which the pattern was compiled. */ + reg_syntax_t syntax; + + /* Pointer to a fastmap, if any, otherwise zero. re_search uses + the fastmap, if there is one, to skip over impossible + starting points for matches. */ + char *fastmap; + + /* Either a translate table to apply to all characters before + comparing them, or zero for no translation. The translation + is applied to a pattern when it is compiled and to a string + when it is matched. */ + RE_TRANSLATE_TYPE translate; + + /* Number of subexpressions found by the compiler. */ + size_t re_nsub; + + /* Zero if this pattern cannot match the empty string, one else. + Well, in truth it's used only in `re_search_2', to see + whether or not we should use the fastmap, so we don't set + this absolutely perfectly; see `re_compile_fastmap' (the + `duplicate' case). */ + unsigned can_be_null : 1; + + /* If REGS_UNALLOCATED, allocate space in the `regs' structure + for `max (RE_NREGS, re_nsub + 1)' groups. + If REGS_REALLOCATE, reallocate space if necessary. + If REGS_FIXED, use what's there. */ +#define REGS_UNALLOCATED 0 +#define REGS_REALLOCATE 1 +#define REGS_FIXED 2 + unsigned regs_allocated : 2; + + /* Set to zero when `regex_compile' compiles a pattern; set to one + by `re_compile_fastmap' if it updates the fastmap. */ + unsigned fastmap_accurate : 1; + + /* If set, `re_match_2' does not return information about + subexpressions. */ + unsigned no_sub : 1; + + /* If set, a beginning-of-line anchor doesn't match at the + beginning of the string. */ + unsigned not_bol : 1; + + /* Similarly for an end-of-line anchor. */ + unsigned not_eol : 1; + + /* If true, an anchor at a newline matches. */ + unsigned newline_anchor : 1; + + /* If true, multi-byte form in the `buffer' should be recognized as a + multibyte character. */ + unsigned multibyte : 1; + +/* [[[end pattern_buffer]]] */ +}; + +typedef struct re_pattern_buffer regex_t; + +/* Type for byte offsets within the string. POSIX mandates this. */ +typedef int regoff_t; + + +/* This is the structure we store register match data in. See + regex.texinfo for a full description of what registers match. */ +struct re_registers +{ + unsigned num_regs; + regoff_t *start; + regoff_t *end; +}; + + +/* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, + `re_match_2' returns information about at least this many registers + the first time a `regs' structure is passed. */ +#ifndef RE_NREGS +#define RE_NREGS 30 +#endif + + +/* POSIX specification for registers. Aside from the different names than + `re_registers', POSIX uses an array of structures, instead of a + structure of arrays. */ +typedef struct +{ + regoff_t rm_so; /* Byte offset from string's start to substring's start. */ + regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ +} regmatch_t; + +/* Declarations for routines. */ + +/* To avoid duplicating every routine declaration -- once with a + prototype (if we are ANSI), and once without (if we aren't) -- we + use the following macro to declare argument types. This + unfortunately clutters up the declarations a bit, but I think it's + worth it. */ + +#if __STDC__ + +#define _RE_ARGS(args) args + +#else /* not __STDC__ */ + +#define _RE_ARGS(args) () + +#endif /* not __STDC__ */ + +/* Sets the current default syntax to SYNTAX, and return the old syntax. + You can also simply assign to the `re_syntax_options' variable. */ +extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax)); + +/* Compile the regular expression PATTERN, with length LENGTH + and syntax given by the global `re_syntax_options', into the buffer + BUFFER. Return NULL if successful, and an error string if not. */ +extern const char *re_compile_pattern + _RE_ARGS ((const char *pattern, int length, + struct re_pattern_buffer *buffer)); + + +/* Compile a fastmap for the compiled pattern in BUFFER; used to + accelerate searches. Return 0 if successful and -2 if was an + internal error. */ +extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer)); + + +/* Search in the string STRING (with length LENGTH) for the pattern + compiled into BUFFER. Start searching at position START, for RANGE + characters. Return the starting position of the match, -1 for no + match, or -2 for an internal error. Also return register + information in REGS (if REGS and BUFFER->no_sub are nonzero). */ +extern int re_search + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, + int length, int start, int range, struct re_registers *regs)); + + +/* Like `re_search', but search in the concatenation of STRING1 and + STRING2. Also, stop searching at index START + STOP. */ +extern int re_search_2 + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, + int length1, const char *string2, int length2, + int start, int range, struct re_registers *regs, int stop)); + + +/* Like `re_search', but return how many characters in STRING the regexp + in BUFFER matched, starting at position START. */ +extern int re_match + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, + int length, int start, struct re_registers *regs)); + + +/* Relates to `re_match' as `re_search_2' relates to `re_search'. */ +extern int re_match_2 + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, + int length1, const char *string2, int length2, + int start, struct re_registers *regs, int stop)); + + +/* Set REGS to hold NUM_REGS registers, storing them in STARTS and + ENDS. Subsequent matches using BUFFER and REGS will use this memory + for recording register information. STARTS and ENDS must be + allocated with malloc, and must each be at least `NUM_REGS * sizeof + (regoff_t)' bytes long. + + If NUM_REGS == 0, then subsequent matches should allocate their own + register data. + + Unless this function is called, the first search or match using + PATTERN_BUFFER will allocate its own register data, without + freeing the old data. */ +extern void re_set_registers + _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs, + unsigned num_regs, regoff_t *starts, regoff_t *ends)); + +#ifdef _REGEX_RE_COMP +/* 4.2 bsd compatibility. */ +/* CVS: don't use prototypes: they may conflict with system headers. */ +extern char *re_comp _RE_ARGS (()); +extern int re_exec _RE_ARGS (()); +#endif + +/* POSIX compatibility. */ +extern int regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags)); +extern int regexec + _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch, + regmatch_t pmatch[], int eflags)); +extern size_t regerror + _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf, + size_t errbuf_size)); +extern void regfree _RE_ARGS ((regex_t *preg)); + +#endif /* not __REGEXP_LIBRARY_H__ */ + +/* +Local variables: +make-backup-files: t +version-control: t +trim-versions-without-asking: nil +End: +*/ diff --git a/sys/src/ape/cmd/diff/sdiff.c b/sys/src/ape/cmd/diff/sdiff.c new file mode 100755 index 000000000..cd49aeac5 --- /dev/null +++ b/sys/src/ape/cmd/diff/sdiff.c @@ -0,0 +1,1109 @@ +/* SDIFF -- interactive merge front end to diff + Copyright (C) 1992, 1993, 1994 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* GNU SDIFF was written by Thomas Lord. */ + +#include <sys/cdefs.h> +__FBSDID("$FreeBSD: src/contrib/diff/sdiff.c,v 1.1.1.1.12.1 2002/01/28 01:26:35 nectar Exp $"); + +#include "system.h" +#include <stdio.h> +#include <signal.h> +#include "getopt.h" + +/* Size of chunks read from files which must be parsed into lines. */ +#define SDIFF_BUFSIZE ((size_t) 65536) + +/* Default name of the diff program */ +#ifndef DIFF_PROGRAM +#define DIFF_PROGRAM "/usr/bin/diff" +#endif + +/* Users' editor of nonchoice */ +#ifndef DEFAULT_EDITOR_PROGRAM +#define DEFAULT_EDITOR_PROGRAM "ed" +#endif + +extern char version_string[]; +static char const *program_name; +static char const *diffbin = DIFF_PROGRAM; +static char const *edbin = DEFAULT_EDITOR_PROGRAM; +static char const **diffargv; + +static char *tmpname; +static int volatile tmpmade; + +#if HAVE_FORK +static pid_t volatile diffpid; +#endif + +struct line_filter; + +static FILE *ck_fopen PARAMS((char const *, char const *)); +static RETSIGTYPE catchsig PARAMS((int)); +static VOID *xmalloc PARAMS((size_t)); +static char const *expand_name PARAMS((char *, int, char const *)); +static int edit PARAMS((struct line_filter *, int, struct line_filter *, int, FILE*)); +static int interact PARAMS((struct line_filter *, struct line_filter *, struct line_filter *, FILE*)); +static int lf_snarf PARAMS((struct line_filter *, char *, size_t)); +static int skip_white PARAMS((void)); +static size_t ck_fread PARAMS((char *, size_t, FILE *)); +static size_t lf_refill PARAMS((struct line_filter *)); +static void checksigs PARAMS((void)); +static void ck_fclose PARAMS((FILE *)); +static void ck_fflush PARAMS((FILE *)); +static void ck_fwrite PARAMS((char const *, size_t, FILE *)); +static void cleanup PARAMS((void)); +static void diffarg PARAMS((char const *)); +static void execdiff PARAMS((void)); +static void exiterr PARAMS((void)); +static void fatal PARAMS((char const *)); +static void flush_line PARAMS((void)); +static void give_help PARAMS((void)); +static void lf_copy PARAMS((struct line_filter *, int, FILE *)); +static void lf_init PARAMS((struct line_filter *, FILE *)); +static void lf_skip PARAMS((struct line_filter *, int)); +static void perror_fatal PARAMS((char const *)); +static void trapsigs PARAMS((void)); +static void try_help PARAMS((char const *)); +static void untrapsig PARAMS((int)); +static void usage PARAMS((void)); + +static int diraccess PARAMS((char const *)); + +/* Options: */ + +/* name of output file if -o spec'd */ +static char *out_file; + +/* do not print common lines if true, set by -s option */ +static int suppress_common_flag; + +static struct option const longopts[] = +{ + {"ignore-blank-lines", 0, 0, 'B'}, + {"speed-large-files", 0, 0, 'H'}, + {"ignore-matching-lines", 1, 0, 'I'}, + {"ignore-all-space", 0, 0, 'W'}, /* swap W and w for historical reasons */ + {"text", 0, 0, 'a'}, + {"ignore-space-change", 0, 0, 'b'}, + {"minimal", 0, 0, 'd'}, + {"ignore-case", 0, 0, 'i'}, + {"left-column", 0, 0, 'l'}, + {"output", 1, 0, 'o'}, + {"suppress-common-lines", 0, 0, 's'}, + {"expand-tabs", 0, 0, 't'}, + {"width", 1, 0, 'w'}, + {"version", 0, 0, 'v'}, + {"help", 0, 0, 129}, + {0, 0, 0, 0} +}; + +static void +try_help (reason) + char const *reason; +{ + if (reason) + fprintf (stderr, "%s: %s\n", program_name, reason); + fprintf (stderr, "%s: Try `%s --help' for more information.\n", + program_name, program_name); + exit (2); +} + +static void +usage () +{ + printf ("Usage: %s [OPTIONS]... FILE1 FILE2\n\n", program_name); + printf ("%s", "\ + -o FILE --output=FILE Operate interactively, sending output to FILE.\n\n"); + printf ("%s", "\ + -i --ignore-case Consider upper- and lower-case to be the same.\n\ + -W --ignore-all-space Ignore all white space.\n\ + -b --ignore-space-change Ignore changes in the amount of white space.\n\ + -B --ignore-blank-lines Ignore changes whose lines are all blank.\n\ + -I RE --ignore-matching-lines=RE Ignore changes whose lines all match RE.\n\ + -a --text Treat all files as text.\n\n"); + printf ("%s", "\ + -w NUM --width=NUM Output at most NUM (default 130) characters per line.\n\ + -l --left-column Output only the left column of common lines.\n\ + -s --suppress-common-lines Do not output common lines.\n\n"); + printf ("\ + -t --expand-tabs Expand tabs to spaces in output.\n\n"); + printf ("%s", "\ + -d --minimal Try hard to find a smaller set of changes.\n\ + -H --speed-large-files Assume large files and many scattered small changes.\n\n"); + printf ("%s", "\ + -v --version Output version info.\n\ + --help Output this help.\n\n\ +If FILE1 or FILE2 is `-', read standard input.\n"); +} + +static void +cleanup () +{ +#if HAVE_FORK + if (0 < diffpid) + kill (diffpid, SIGPIPE); +#endif + if (tmpmade) + unlink (tmpname); +} + +static void +exiterr () +{ + cleanup (); + untrapsig (0); + checksigs (); + exit (2); +} + +static void +fatal (msg) + char const *msg; +{ + fprintf (stderr, "%s: %s\n", program_name, msg); + exiterr (); +} + +static void +perror_fatal (msg) + char const *msg; +{ + int e = errno; + checksigs (); + fprintf (stderr, "%s: ", program_name); + errno = e; + perror (msg); + exiterr (); +} + + +/* malloc freely or DIE! */ +static VOID * +xmalloc (size) + size_t size; +{ + VOID *r = (VOID *) malloc (size); + if (!r) + fatal ("memory exhausted"); + return r; +} + +static FILE * +ck_fopen (fname, type) + char const *fname, *type; +{ + FILE *r = fopen (fname, type); + if (!r) + perror_fatal (fname); + return r; +} + +static void +ck_fclose (f) + FILE *f; +{ + if (fclose (f)) + perror_fatal ("input/output error"); +} + +static size_t +ck_fread (buf, size, f) + char *buf; + size_t size; + FILE *f; +{ + size_t r = fread (buf, sizeof (char), size, f); + if (r == 0 && ferror (f)) + perror_fatal ("input error"); + return r; +} + +static void +ck_fwrite (buf, size, f) + char const *buf; + size_t size; + FILE *f; +{ + if (fwrite (buf, sizeof (char), size, f) != size) + perror_fatal ("output error"); +} + +static void +ck_fflush (f) + FILE *f; +{ + if (fflush (f) != 0) + perror_fatal ("output error"); +} + +static char const * +expand_name (name, is_dir, other_name) + char *name; + int is_dir; + char const *other_name; +{ + if (strcmp (name, "-") == 0) + fatal ("cannot interactively merge standard input"); + if (!is_dir) + return name; + else + { + /* Yield NAME/BASE, where BASE is OTHER_NAME's basename. */ + char const *p = filename_lastdirchar (other_name); + char const *base = p ? p+1 : other_name; + size_t namelen = strlen (name), baselen = strlen (base); + char *r = xmalloc (namelen + baselen + 2); + memcpy (r, name, namelen); + r[namelen] = '/'; + memcpy (r + namelen + 1, base, baselen + 1); + return r; + } +} + + + +struct line_filter { + FILE *infile; + char *bufpos; + char *buffer; + char *buflim; +}; + +static void +lf_init (lf, infile) + struct line_filter *lf; + FILE *infile; +{ + lf->infile = infile; + lf->bufpos = lf->buffer = lf->buflim = xmalloc (SDIFF_BUFSIZE + 1); + lf->buflim[0] = '\n'; +} + +/* Fill an exhausted line_filter buffer from its INFILE */ +static size_t +lf_refill (lf) + struct line_filter *lf; +{ + size_t s = ck_fread (lf->buffer, SDIFF_BUFSIZE, lf->infile); + lf->bufpos = lf->buffer; + lf->buflim = lf->buffer + s; + lf->buflim[0] = '\n'; + checksigs (); + return s; +} + +/* Advance LINES on LF's infile, copying lines to OUTFILE */ +static void +lf_copy (lf, lines, outfile) + struct line_filter *lf; + int lines; + FILE *outfile; +{ + char *start = lf->bufpos; + + while (lines) + { + lf->bufpos = (char *) memchr (lf->bufpos, '\n', lf->buflim - lf->bufpos); + if (! lf->bufpos) + { + ck_fwrite (start, lf->buflim - start, outfile); + if (! lf_refill (lf)) + return; + start = lf->bufpos; + } + else + { + --lines; + ++lf->bufpos; + } + } + + ck_fwrite (start, lf->bufpos - start, outfile); +} + +/* Advance LINES on LF's infile without doing output */ +static void +lf_skip (lf, lines) + struct line_filter *lf; + int lines; +{ + while (lines) + { + lf->bufpos = (char *) memchr (lf->bufpos, '\n', lf->buflim - lf->bufpos); + if (! lf->bufpos) + { + if (! lf_refill (lf)) + break; + } + else + { + --lines; + ++lf->bufpos; + } + } +} + +/* Snarf a line into a buffer. Return EOF if EOF, 0 if error, 1 if OK. */ +static int +lf_snarf (lf, buffer, bufsize) + struct line_filter *lf; + char *buffer; + size_t bufsize; +{ + char *start = lf->bufpos; + + for (;;) + { + char *next = (char *) memchr (start, '\n', lf->buflim + 1 - start); + size_t s = next - start; + if (bufsize <= s) + return 0; + memcpy (buffer, start, s); + if (next < lf->buflim) + { + buffer[s] = 0; + lf->bufpos = next + 1; + return 1; + } + if (! lf_refill (lf)) + return s ? 0 : EOF; + buffer += s; + bufsize -= s; + start = next; + } +} + + + +int +main (argc, argv) + int argc; + char *argv[]; +{ + int opt; + char *editor; + char *differ; + + initialize_main (&argc, &argv); + program_name = argv[0]; + + editor = getenv ("EDITOR"); + if (editor) + edbin = editor; + differ = getenv ("DIFF"); + if (differ) + diffbin = differ; + + diffarg ("diff"); + + /* parse command line args */ + while ((opt = getopt_long (argc, argv, "abBdHiI:lo:stvw:W", longopts, 0)) + != EOF) + { + switch (opt) + { + case 'a': + diffarg ("-a"); + break; + + case 'b': + diffarg ("-b"); + break; + + case 'B': + diffarg ("-B"); + break; + + case 'd': + diffarg ("-d"); + break; + + case 'H': + diffarg ("-H"); + break; + + case 'i': + diffarg ("-i"); + break; + + case 'I': + diffarg ("-I"); + diffarg (optarg); + break; + + case 'l': + diffarg ("--left-column"); + break; + + case 'o': + out_file = optarg; + break; + + case 's': + suppress_common_flag = 1; + break; + + case 't': + diffarg ("-t"); + break; + + case 'v': + printf ("sdiff - GNU diffutils version %s\n", version_string); + exit (0); + + case 'w': + diffarg ("-W"); + diffarg (optarg); + break; + + case 'W': + diffarg ("-w"); + break; + + case 129: + usage (); + if (ferror (stdout) || fclose (stdout) != 0) + fatal ("write error"); + exit (0); + + default: + try_help (0); + } + } + + if (argc - optind != 2) + try_help (argc - optind < 2 ? "missing operand" : "extra operand"); + + if (! out_file) + { + /* easy case: diff does everything for us */ + if (suppress_common_flag) + diffarg ("--suppress-common-lines"); + diffarg ("-y"); + diffarg ("--"); + diffarg (argv[optind]); + diffarg (argv[optind + 1]); + diffarg (0); + execdiff (); + } + else + { + FILE *left, *right, *out, *diffout; + int interact_ok; + struct line_filter lfilt; + struct line_filter rfilt; + struct line_filter diff_filt; + int leftdir = diraccess (argv[optind]); + int rightdir = diraccess (argv[optind + 1]); + + if (leftdir && rightdir) + fatal ("both files to be compared are directories"); + + left = ck_fopen (expand_name (argv[optind], leftdir, argv[optind + 1]), "r"); + ; + right = ck_fopen (expand_name (argv[optind + 1], rightdir, argv[optind]), "r"); + out = ck_fopen (out_file, "w"); + + diffarg ("--sdiff-merge-assist"); + diffarg ("--"); + diffarg (argv[optind]); + diffarg (argv[optind + 1]); + diffarg (0); + + trapsigs (); + +#if ! HAVE_FORK + { + size_t cmdsize = 1; + char *p, *command; + int i; + + for (i = 0; diffargv[i]; i++) + cmdsize += 4 * strlen (diffargv[i]) + 3; + command = p = xmalloc (cmdsize); + for (i = 0; diffargv[i]; i++) + { + char const *a = diffargv[i]; + SYSTEM_QUOTE_ARG (p, a); + *p++ = ' '; + } + p[-1] = '\0'; + diffout = popen (command, "r"); + if (!diffout) + perror_fatal (command); + free (command); + } +#else /* HAVE_FORK */ + { + int diff_fds[2]; + + if (pipe (diff_fds) != 0) + perror_fatal ("pipe"); + + diffpid = fork (); + if (diffpid < 0) + perror_fatal ("fork failed"); + if (!diffpid) + { + signal (SIGINT, SIG_IGN); /* in case user interrupts editor */ + signal (SIGPIPE, SIG_DFL); + + close (diff_fds[0]); + if (diff_fds[1] != STDOUT_FILENO) + { + dup2 (diff_fds[1], STDOUT_FILENO); + close (diff_fds[1]); + } + + execdiff (); + } + + close (diff_fds[1]); + diffout = fdopen (diff_fds[0], "r"); + if (!diffout) + perror_fatal ("fdopen"); + } +#endif /* HAVE_FORK */ + + lf_init (&diff_filt, diffout); + lf_init (&lfilt, left); + lf_init (&rfilt, right); + + interact_ok = interact (&diff_filt, &lfilt, &rfilt, out); + + ck_fclose (left); + ck_fclose (right); + ck_fclose (out); + + { + int wstatus; + +#if ! HAVE_FORK + wstatus = pclose (diffout); +#else + ck_fclose (diffout); + while (waitpid (diffpid, &wstatus, 0) < 0) + if (errno == EINTR) + checksigs (); + else + perror_fatal ("wait failed"); + diffpid = 0; +#endif + + if (tmpmade) + { + unlink (tmpname); + tmpmade = 0; + } + + if (! interact_ok) + exiterr (); + + if (! (WIFEXITED (wstatus) && WEXITSTATUS (wstatus) < 2)) + fatal ("Subsidiary diff failed"); + + untrapsig (0); + checksigs (); + exit (WEXITSTATUS (wstatus)); + } + } + return 0; /* Fool -Wall . . . */ +} + +static void +diffarg (a) + char const *a; +{ + static unsigned diffargs, diffargsmax; + + if (diffargs == diffargsmax) + { + if (! diffargsmax) + { + diffargv = (char const **) xmalloc (sizeof (char)); + diffargsmax = 8; + } + diffargsmax *= 2; + diffargv = (char const **) realloc (diffargv, + diffargsmax * sizeof (char const *)); + if (! diffargv) + fatal ("out of memory"); + } + diffargv[diffargs++] = a; +} + +static void +execdiff () +{ + execvp (diffbin, (char **) diffargv); + write (STDERR_FILENO, diffbin, strlen (diffbin)); + write (STDERR_FILENO, ": not found\n", 12); + _exit (2); +} + + + + +/* Signal handling */ + +#define NUM_SIGS (sizeof (sigs) / sizeof (*sigs)) +static int const sigs[] = { +#ifdef SIGHUP + SIGHUP, +#endif +#ifdef SIGQUIT + SIGQUIT, +#endif +#ifdef SIGTERM + SIGTERM, +#endif +#ifdef SIGXCPU + SIGXCPU, +#endif +#ifdef SIGXFSZ + SIGXFSZ, +#endif + SIGINT, + SIGPIPE +}; + +/* Prefer `sigaction' if it is available, since `signal' can lose signals. */ +#if HAVE_SIGACTION +static struct sigaction initial_action[NUM_SIGS]; +#define initial_handler(i) (initial_action[i].sa_handler) +#else +static RETSIGTYPE (*initial_action[NUM_SIGS]) (); +#define initial_handler(i) (initial_action[i]) +#endif + +static int volatile ignore_SIGINT; +static int volatile signal_received; +static int sigs_trapped; + +static RETSIGTYPE +catchsig (s) + int s; +{ +#if ! HAVE_SIGACTION + signal (s, SIG_IGN); +#endif + if (! (s == SIGINT && ignore_SIGINT)) + signal_received = s; +} + +static void +trapsigs () +{ + int i; + +#if HAVE_SIGACTION + struct sigaction catchaction; + bzero (&catchaction, sizeof (catchaction)); + catchaction.sa_handler = catchsig; +#ifdef SA_INTERRUPT + /* Non-Posix BSD-style systems like SunOS 4.1.x need this + so that `read' calls are interrupted properly. */ + catchaction.sa_flags = SA_INTERRUPT; +#endif + sigemptyset (&catchaction.sa_mask); + for (i = 0; i < NUM_SIGS; i++) + sigaddset (&catchaction.sa_mask, sigs[i]); + for (i = 0; i < NUM_SIGS; i++) + { + sigaction (sigs[i], 0, &initial_action[i]); + if (initial_handler (i) != SIG_IGN + && sigaction (sigs[i], &catchaction, 0) != 0) + fatal ("signal error"); + } +#else /* ! HAVE_SIGACTION */ + for (i = 0; i < NUM_SIGS; i++) + { + initial_action[i] = signal (sigs[i], SIG_IGN); + if (initial_handler (i) != SIG_IGN + && signal (sigs[i], catchsig) != SIG_IGN) + fatal ("signal error"); + } +#endif /* ! HAVE_SIGACTION */ + +#if !defined(SIGCHLD) && defined(SIGCLD) +#define SIGCHLD SIGCLD +#endif +#ifdef SIGCHLD + /* System V fork+wait does not work if SIGCHLD is ignored. */ + signal (SIGCHLD, SIG_DFL); +#endif + + sigs_trapped = 1; +} + +/* Untrap signal S, or all trapped signals if S is zero. */ +static void +untrapsig (s) + int s; +{ + int i; + + if (sigs_trapped) + for (i = 0; i < NUM_SIGS; i++) + if ((!s || sigs[i] == s) && initial_handler (i) != SIG_IGN) +#if HAVE_SIGACTION + sigaction (sigs[i], &initial_action[i], 0); +#else + signal (sigs[i], initial_action[i]); +#endif +} + +/* Exit if a signal has been received. */ +static void +checksigs () +{ + int s = signal_received; + if (s) + { + cleanup (); + + /* Yield an exit status indicating that a signal was received. */ + untrapsig (s); + kill (getpid (), s); + + /* That didn't work, so exit with error status. */ + exit (2); + } +} + + + +static void +give_help () +{ + fprintf (stderr,"l:\tuse the left version\n"); + fprintf (stderr,"r:\tuse the right version\n"); + fprintf (stderr,"e l:\tedit then use the left version\n"); + fprintf (stderr,"e r:\tedit then use the right version\n"); + fprintf (stderr,"e b:\tedit then use the left and right versions concatenated\n"); + fprintf (stderr,"e:\tedit a new version\n"); + fprintf (stderr,"s:\tsilently include common lines\n"); + fprintf (stderr,"v:\tverbosely include common lines\n"); + fprintf (stderr,"q:\tquit\n"); +} + +static int +skip_white () +{ + int c; + for (;;) + { + c = getchar (); + if (!ISSPACE (c) || c == '\n') + break; + checksigs (); + } + if (ferror (stdin)) + perror_fatal ("input error"); + return c; +} + +static void +flush_line () +{ + int c; + while ((c = getchar ()) != '\n' && c != EOF) + ; + if (ferror (stdin)) + perror_fatal ("input error"); +} + + +/* interpret an edit command */ +static int +edit (left, lenl, right, lenr, outfile) + struct line_filter *left; + int lenl; + struct line_filter *right; + int lenr; + FILE *outfile; +{ + for (;;) + { + int cmd0, cmd1; + int gotcmd = 0; + + cmd1 = 0; /* Pacify `gcc -W'. */ + + while (!gotcmd) + { + if (putchar ('%') != '%') + perror_fatal ("output error"); + ck_fflush (stdout); + + cmd0 = skip_white (); + switch (cmd0) + { + case 'l': case 'r': case 's': case 'v': case 'q': + if (skip_white () != '\n') + { + give_help (); + flush_line (); + continue; + } + gotcmd = 1; + break; + + case 'e': + cmd1 = skip_white (); + switch (cmd1) + { + case 'l': case 'r': case 'b': + if (skip_white () != '\n') + { + give_help (); + flush_line (); + continue; + } + gotcmd = 1; + break; + case '\n': + gotcmd = 1; + break; + default: + give_help (); + flush_line (); + continue; + } + break; + case EOF: + if (feof (stdin)) + { + gotcmd = 1; + cmd0 = 'q'; + break; + } + /* falls through */ + default: + flush_line (); + /* falls through */ + case '\n': + give_help (); + continue; + } + } + + switch (cmd0) + { + case 'l': + lf_copy (left, lenl, outfile); + lf_skip (right, lenr); + return 1; + case 'r': + lf_copy (right, lenr, outfile); + lf_skip (left, lenl); + return 1; + case 's': + suppress_common_flag = 1; + break; + case 'v': + suppress_common_flag = 0; + break; + case 'q': + return 0; + case 'e': + { + int tfd; + FILE *tmp; + + if (tmpmade) + { + unlink (tmpname); + tmpmade = 0; + free (tmpname); + } + + asprintf (&tmpname, "%s/sdiff.XXXXXX", + getenv("TMPDIR") ?: P_tmpdir); + if (tmpname == NULL) + perror_fatal ("temporary file name"); + tfd = mkstemp(tmpname); + if (tfd == -1) + perror_fatal ("temporary file name"); + tmp = fdopen (tfd, "w+"); + if (tmp == NULL) + perror_fatal ("temporary file name"); + + tmpmade = 1; + + if (cmd1 == 'l' || cmd1 == 'b') + lf_copy (left, lenl, tmp); + else + lf_skip (left, lenl); + + if (cmd1 == 'r' || cmd1 == 'b') + lf_copy (right, lenr, tmp); + else + lf_skip (right, lenr); + + ck_fflush (tmp); + + { + int wstatus; +#if ! HAVE_FORK + char *command = xmalloc (strlen (edbin) + strlen (tmpname) + 2); + sprintf (command, "%s %s", edbin, tmpname); + wstatus = system (command); + free (command); +#else /* HAVE_FORK */ + pid_t pid; + + ignore_SIGINT = 1; + checksigs (); + + pid = fork (); + if (pid == 0) + { + char const *argv[3]; + int i = 0; + + argv[i++] = edbin; + argv[i++] = tmpname; + argv[i++] = 0; + + execvp (edbin, (char **) argv); + write (STDERR_FILENO, edbin, strlen (edbin)); + write (STDERR_FILENO, ": not found\n", 12); + _exit (1); + } + + if (pid < 0) + perror_fatal ("fork failed"); + + while (waitpid (pid, &wstatus, 0) < 0) + if (errno == EINTR) + checksigs (); + else + perror_fatal ("wait failed"); + + ignore_SIGINT = 0; +#endif /* HAVE_FORK */ + + if (wstatus != 0) + fatal ("Subsidiary editor failed"); + } + + if (fseek (tmp, 0L, SEEK_SET) != 0) + perror_fatal ("fseek"); + { + /* SDIFF_BUFSIZE is too big for a local var + in some compilers, so we allocate it dynamically. */ + char *buf = xmalloc (SDIFF_BUFSIZE); + size_t size; + + while ((size = ck_fread (buf, SDIFF_BUFSIZE, tmp)) != 0) + { + checksigs (); + ck_fwrite (buf, size, outfile); + } + ck_fclose (tmp); + + free (buf); + } + return 1; + } + default: + give_help (); + break; + } + } +} + + + +/* Alternately reveal bursts of diff output and handle user commands. */ +static int +interact (diff, left, right, outfile) + struct line_filter *diff; + struct line_filter *left; + struct line_filter *right; + FILE *outfile; +{ + for (;;) + { + char diff_help[256]; + int snarfed = lf_snarf (diff, diff_help, sizeof (diff_help)); + + if (snarfed <= 0) + return snarfed; + + checksigs (); + + switch (diff_help[0]) + { + case ' ': + puts (diff_help + 1); + break; + case 'i': + { + int lenl = atoi (diff_help + 1), lenr, lenmax; + char *p = strchr (diff_help, ','); + + if (!p) + fatal (diff_help); + lenr = atoi (p + 1); + lenmax = max (lenl, lenr); + + if (suppress_common_flag) + lf_skip (diff, lenmax); + else + lf_copy (diff, lenmax, stdout); + + lf_copy (left, lenl, outfile); + lf_skip (right, lenr); + break; + } + case 'c': + { + int lenl = atoi (diff_help + 1), lenr; + char *p = strchr (diff_help, ','); + + if (!p) + fatal (diff_help); + lenr = atoi (p + 1); + lf_copy (diff, max (lenl, lenr), stdout); + if (! edit (left, lenl, right, lenr, outfile)) + return 0; + break; + } + default: + fatal (diff_help); + break; + } + } +} + + + +/* temporary lossage: this is torn from gnu libc */ +/* Return nonzero if DIR is an existing directory. */ +static int +diraccess (dir) + char const *dir; +{ + struct stat buf; + return stat (dir, &buf) == 0 && S_ISDIR (buf.st_mode); +} diff --git a/sys/src/ape/cmd/diff/side.c b/sys/src/ape/cmd/diff/side.c new file mode 100755 index 000000000..a150b5e70 --- /dev/null +++ b/sys/src/ape/cmd/diff/side.c @@ -0,0 +1,284 @@ +/* sdiff-format output routines for GNU DIFF. + Copyright (C) 1991, 1992, 1993 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY. No author or distributor +accepts responsibility to anyone for the consequences of using it +or for whether it serves any particular purpose or works at all, +unless he says so in writing. Refer to the GNU DIFF General Public +License for full details. + +Everyone is granted permission to copy, modify and redistribute +GNU DIFF, but only under the conditions described in the +GNU DIFF General Public License. A copy of this license is +supposed to have been given to you along with GNU DIFF so you +can know your rights and responsibilities. It should be in a +file named COPYING. Among other things, the copyright notice +and this notice must be preserved on all copies. */ + + +#include "diff.h" + +static unsigned print_half_line PARAMS((char const * const *, unsigned, unsigned)); +static unsigned tab_from_to PARAMS((unsigned, unsigned)); +static void print_1sdiff_line PARAMS((char const * const *, int, char const * const *)); +static void print_sdiff_common_lines PARAMS((int, int)); +static void print_sdiff_hunk PARAMS((struct change *)); + +/* Next line number to be printed in the two input files. */ +static int next0, next1; + +/* Print the edit-script SCRIPT as a sdiff style output. */ + +void +print_sdiff_script (script) + struct change *script; +{ + begin_output (); + + next0 = next1 = - files[0].prefix_lines; + print_script (script, find_change, print_sdiff_hunk); + + print_sdiff_common_lines (files[0].valid_lines, files[1].valid_lines); +} + +/* Tab from column FROM to column TO, where FROM <= TO. Yield TO. */ + +static unsigned +tab_from_to (from, to) + unsigned from, to; +{ + FILE *out = outfile; + unsigned tab; + + if (! tab_expand_flag) + for (tab = from + TAB_WIDTH - from % TAB_WIDTH; tab <= to; tab += TAB_WIDTH) + { + putc ('\t', out); + from = tab; + } + while (from++ < to) + putc (' ', out); + return to; +} + +/* + * Print the text for half an sdiff line. This means truncate to width + * observing tabs, and trim a trailing newline. Returns the last column + * written (not the number of chars). + */ +static unsigned +print_half_line (line, indent, out_bound) + char const * const *line; + unsigned indent, out_bound; +{ + FILE *out = outfile; + register unsigned in_position = 0, out_position = 0; + register char const + *text_pointer = line[0], + *text_limit = line[1]; + + while (text_pointer < text_limit) + { + register unsigned char c = *text_pointer++; + + switch (c) + { + case '\t': + { + unsigned spaces = TAB_WIDTH - in_position % TAB_WIDTH; + if (in_position == out_position) + { + unsigned tabstop = out_position + spaces; + if (tab_expand_flag) + { + if (out_bound < tabstop) + tabstop = out_bound; + for (; out_position < tabstop; out_position++) + putc (' ', out); + } + else + if (tabstop < out_bound) + { + out_position = tabstop; + putc (c, out); + } + } + in_position += spaces; + } + break; + + case '\r': + { + putc (c, out); + tab_from_to (0, indent); + in_position = out_position = 0; + } + break; + + case '\b': + if (in_position != 0 && --in_position < out_bound) + if (out_position <= in_position) + /* Add spaces to make up for suppressed tab past out_bound. */ + for (; out_position < in_position; out_position++) + putc (' ', out); + else + { + out_position = in_position; + putc (c, out); + } + break; + + case '\f': + case '\v': + control_char: + if (in_position < out_bound) + putc (c, out); + break; + + default: + if (! ISPRINT (c)) + goto control_char; + /* falls through */ + case ' ': + if (in_position++ < out_bound) + { + out_position = in_position; + putc (c, out); + } + break; + + case '\n': + return out_position; + } + } + + return out_position; +} + +/* + * Print side by side lines with a separator in the middle. + * 0 parameters are taken to indicate white space text. + * Blank lines that can easily be caught are reduced to a single newline. + */ + +static void +print_1sdiff_line (left, sep, right) + char const * const *left; + int sep; + char const * const *right; +{ + FILE *out = outfile; + unsigned hw = sdiff_half_width, c2o = sdiff_column2_offset; + unsigned col = 0; + int put_newline = 0; + + if (left) + { + if (left[1][-1] == '\n') + put_newline = 1; + col = print_half_line (left, 0, hw); + } + + if (sep != ' ') + { + col = tab_from_to (col, (hw + c2o - 1) / 2) + 1; + if (sep == '|' && put_newline != (right[1][-1] == '\n')) + sep = put_newline ? '/' : '\\'; + putc (sep, out); + } + + if (right) + { + if (right[1][-1] == '\n') + put_newline = 1; + if (**right != '\n') + { + col = tab_from_to (col, c2o); + print_half_line (right, col, hw); + } + } + + if (put_newline) + putc ('\n', out); +} + +/* Print lines common to both files in side-by-side format. */ +static void +print_sdiff_common_lines (limit0, limit1) + int limit0, limit1; +{ + int i0 = next0, i1 = next1; + + if (! sdiff_skip_common_lines && (i0 != limit0 || i1 != limit1)) + { + if (sdiff_help_sdiff) + fprintf (outfile, "i%d,%d\n", limit0 - i0, limit1 - i1); + + if (! sdiff_left_only) + { + while (i0 != limit0 && i1 != limit1) + print_1sdiff_line (&files[0].linbuf[i0++], ' ', &files[1].linbuf[i1++]); + while (i1 != limit1) + print_1sdiff_line (0, ')', &files[1].linbuf[i1++]); + } + while (i0 != limit0) + print_1sdiff_line (&files[0].linbuf[i0++], '(', 0); + } + + next0 = limit0; + next1 = limit1; +} + +/* Print a hunk of an sdiff diff. + This is a contiguous portion of a complete edit script, + describing changes in consecutive lines. */ + +static void +print_sdiff_hunk (hunk) + struct change *hunk; +{ + int first0, last0, first1, last1, deletes, inserts; + register int i, j; + + /* Determine range of line numbers involved in each file. */ + analyze_hunk (hunk, &first0, &last0, &first1, &last1, &deletes, &inserts); + if (!deletes && !inserts) + return; + + /* Print out lines up to this change. */ + print_sdiff_common_lines (first0, first1); + + if (sdiff_help_sdiff) + fprintf (outfile, "c%d,%d\n", last0 - first0 + 1, last1 - first1 + 1); + + /* Print ``xxx | xxx '' lines */ + if (inserts && deletes) + { + for (i = first0, j = first1; i <= last0 && j <= last1; ++i, ++j) + print_1sdiff_line (&files[0].linbuf[i], '|', &files[1].linbuf[j]); + deletes = i <= last0; + inserts = j <= last1; + next0 = first0 = i; + next1 = first1 = j; + } + + + /* Print `` > xxx '' lines */ + if (inserts) + { + for (j = first1; j <= last1; ++j) + print_1sdiff_line (0, '>', &files[1].linbuf[j]); + next1 = j; + } + + /* Print ``xxx < '' lines */ + if (deletes) + { + for (i = first0; i <= last0; ++i) + print_1sdiff_line (&files[0].linbuf[i], '<', 0); + next0 = i; + } +} diff --git a/sys/src/ape/cmd/diff/system.h b/sys/src/ape/cmd/diff/system.h new file mode 100755 index 000000000..f59c883f8 --- /dev/null +++ b/sys/src/ape/cmd/diff/system.h @@ -0,0 +1,273 @@ +/* System dependent declarations. + Copyright (C) 1988, 1989, 1992, 1993, 1994 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* We must define `volatile' and `const' first (the latter inside config.h), + so that they're used consistently in all system includes. */ +#if !__STDC__ +#ifndef volatile +#define volatile +#endif +#endif +#include <config.h> + +#include <sys/types.h> +#include <sys/stat.h> + +#if __STDC__ +#define PARAMS(args) args +#define VOID void +#else +#define PARAMS(args) () +#define VOID char +#endif + +#if STAT_MACROS_BROKEN +#undef S_ISBLK +#undef S_ISCHR +#undef S_ISDIR +#undef S_ISFIFO +#undef S_ISREG +#undef S_ISSOCK +#endif +#ifndef S_ISDIR +#define S_ISDIR(mode) (((mode) & S_IFMT) == S_IFDIR) +#endif +#ifndef S_ISREG +#define S_ISREG(mode) (((mode) & S_IFMT) == S_IFREG) +#endif +#if !defined(S_ISBLK) && defined(S_IFBLK) +#define S_ISBLK(mode) (((mode) & S_IFMT) == S_IFBLK) +#endif +#if !defined(S_ISCHR) && defined(S_IFCHR) +#define S_ISCHR(mode) (((mode) & S_IFMT) == S_IFCHR) +#endif +#if !defined(S_ISFIFO) && defined(S_IFFIFO) +#define S_ISFIFO(mode) (((mode) & S_IFMT) == S_IFFIFO) +#endif +#if !defined(S_ISSOCK) && defined(S_IFSOCK) +#define S_ISSOCK(mode) (((mode) & S_IFMT) == S_IFSOCK) +#endif + +#if HAVE_UNISTD_H +#include <unistd.h> +#endif + +#ifndef SEEK_SET +#define SEEK_SET 0 +#endif +#ifndef SEEK_CUR +#define SEEK_CUR 1 +#endif + +#ifndef STDIN_FILENO +#define STDIN_FILENO 0 +#endif +#ifndef STDOUT_FILENO +#define STDOUT_FILENO 1 +#endif +#ifndef STDERR_FILENO +#define STDERR_FILENO 2 +#endif + +#if HAVE_TIME_H +#include <time.h> +#else +#include <sys/time.h> +#endif + +#if HAVE_FCNTL_H +#include <fcntl.h> +#else +#if HAVE_SYS_FILE_H +#include <sys/file.h> +#endif +#endif + +#if !HAVE_DUP2 +#define dup2(f,t) (close (t), fcntl (f,F_DUPFD,t)) +#endif + +#ifndef O_RDONLY +#define O_RDONLY 0 +#endif + +#if HAVE_SYS_WAIT_H +#include <sys/wait.h> +#endif +#ifndef WEXITSTATUS +#define WEXITSTATUS(stat_val) ((unsigned) (stat_val) >> 8) +#endif +#ifndef WIFEXITED +#define WIFEXITED(stat_val) (((stat_val) & 255) == 0) +#endif + +#ifndef STAT_BLOCKSIZE +#if HAVE_ST_BLKSIZE +#define STAT_BLOCKSIZE(s) (s).st_blksize +#else +#define STAT_BLOCKSIZE(s) (8 * 1024) +#endif +#endif + +#if HAVE_DIRENT_H +# include <dirent.h> +# define NAMLEN(dirent) strlen((dirent)->d_name) +#else +# define dirent direct +# define NAMLEN(dirent) ((dirent)->d_namlen) +# if HAVE_SYS_NDIR_H +# include <sys/ndir.h> +# endif +# if HAVE_SYS_DIR_H +# include <sys/dir.h> +# endif +# if HAVE_NDIR_H +# include <ndir.h> +# endif +#endif + +#if HAVE_VFORK_H +#include <vfork.h> +#endif + +#if HAVE_STDLIB_H +#include <stdlib.h> +#else +VOID *malloc (); +VOID *realloc (); +#endif +#ifndef getenv +char *getenv (); +#endif + +#if HAVE_LIMITS_H +#include <limits.h> +#endif +#ifndef INT_MAX +#define INT_MAX 2147483647 +#endif +#ifndef CHAR_BIT +#define CHAR_BIT 8 +#endif + +#if STDC_HEADERS || HAVE_STRING_H +# include <string.h> +# ifndef bzero +# define bzero(s, n) memset (s, 0, n) +# endif +#else +# if !HAVE_STRCHR +# define strchr index +# define strrchr rindex +# endif +char *strchr (), *strrchr (); +# if !HAVE_MEMCHR +# define memcmp(s1, s2, n) bcmp (s1, s2, n) +# define memcpy(d, s, n) bcopy (s, d, n) +void *memchr (); +# endif +#endif + +#include <ctype.h> +/* CTYPE_DOMAIN (C) is nonzero if the unsigned char C can safely be given + as an argument to <ctype.h> macros like `isspace'. */ +#if STDC_HEADERS +#define CTYPE_DOMAIN(c) 1 +#else +#define CTYPE_DOMAIN(c) ((unsigned) (c) <= 0177) +#endif +#ifndef ISPRINT +#define ISPRINT(c) (CTYPE_DOMAIN (c) && isprint (c)) +#endif +#ifndef ISSPACE +#define ISSPACE(c) (CTYPE_DOMAIN (c) && isspace (c)) +#endif +#ifndef ISUPPER +#define ISUPPER(c) (CTYPE_DOMAIN (c) && isupper (c)) +#endif + +#ifndef ISDIGIT +#define ISDIGIT(c) ((unsigned) (c) - '0' <= 9) +#endif + +#include <errno.h> +#if !STDC_HEADERS +extern int errno; +#endif + +#ifdef min +#undef min +#endif +#ifdef max +#undef max +#endif +#define min(a,b) ((a) <= (b) ? (a) : (b)) +#define max(a,b) ((a) >= (b) ? (a) : (b)) + +/* This section contains Posix-compliant defaults for macros + that are meant to be overridden by hand in config.h as needed. */ + +#ifndef filename_cmp +#define filename_cmp(a, b) strcmp (a, b) +#endif + +#ifndef filename_lastdirchar +#define filename_lastdirchar(filename) strrchr (filename, '/') +#endif + +#ifndef HAVE_FORK +#define HAVE_FORK 1 +#endif + +#ifndef HAVE_SETMODE +#define HAVE_SETMODE 0 +#endif + +#ifndef initialize_main +#define initialize_main(argcp, argvp) +#endif + +/* Do struct stat *S, *T describe the same file? Answer -1 if unknown. */ +#ifndef same_file +/* #define same_file(s,t) ((s)->st_ino==(t)->st_ino && (s)->st_dev==(t)->st_dev) */ +#define same_file(s,t) 0 +#endif + +/* Place into Q a quoted version of A suitable for `popen' or `system', + incrementing Q and junking A. + Do not increment Q by more than 4 * strlen (A) + 2. */ +#ifndef SYSTEM_QUOTE_ARG +#define SYSTEM_QUOTE_ARG(q, a) \ + { \ + *(q)++ = '\''; \ + for (; *(a); *(q)++ = *(a)++) \ + if (*(a) == '\'') \ + { \ + *(q)++ = '\''; \ + *(q)++ = '\\'; \ + *(q)++ = '\''; \ + } \ + *(q)++ = '\''; \ + } +#endif +#ifndef FOLD_FN_CHAR +#define FOLD_FN_CHAR(c) (c) +#define fnfold(filename) (filename) +#define fncmp strcmp +#endif diff --git a/sys/src/ape/cmd/diff/util.c b/sys/src/ape/cmd/diff/util.c new file mode 100755 index 000000000..ba6ea79d9 --- /dev/null +++ b/sys/src/ape/cmd/diff/util.c @@ -0,0 +1,759 @@ +/* Support routines for GNU DIFF. + Copyright (C) 1988, 1989, 1992, 1993, 1994 Free Software Foundation, Inc. + +This file is part of GNU DIFF. + +GNU DIFF is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU DIFF is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU DIFF; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* $FreeBSD: src/contrib/diff/util.c,v 1.2.6.2 2000/09/20 02:24:32 jkh Exp $ */ + +#include "diff.h" + +#ifndef PR_PROGRAM +#define PR_PROGRAM "/bin/pr" +#endif + +/* Queue up one-line messages to be printed at the end, + when -l is specified. Each message is recorded with a `struct msg'. */ + +struct msg +{ + struct msg *next; + char const *format; + char const *arg1; + char const *arg2; + char const *arg3; + char const *arg4; +}; + +/* Head of the chain of queues messages. */ + +static struct msg *msg_chain; + +/* Tail of the chain of queues messages. */ + +static struct msg **msg_chain_end = &msg_chain; + +/* Use when a system call returns non-zero status. + TEXT should normally be the file name. */ + +void +perror_with_name (text) + char const *text; +{ + int e = errno; + fprintf (stderr, "%s: ", program_name); + errno = e; + perror (text); +} + +/* Use when a system call returns non-zero status and that is fatal. */ + +void +pfatal_with_name (text) + char const *text; +{ + int e = errno; + print_message_queue (); + fprintf (stderr, "%s: ", program_name); + errno = e; + perror (text); + exit (2); +} + +/* Print an error message from the format-string FORMAT + with args ARG1 and ARG2. */ + +void +error (format, arg, arg1) + char const *format, *arg, *arg1; +{ + fprintf (stderr, "%s: ", program_name); + fprintf (stderr, format, arg, arg1); + fprintf (stderr, "\n"); +} + +/* Print an error message containing the string TEXT, then exit. */ + +void +fatal (m) + char const *m; +{ + print_message_queue (); + error ("%s", m, 0); + exit (2); +} + +/* Like printf, except if -l in effect then save the message and print later. + This is used for things like "binary files differ" and "Only in ...". */ + +void +message (format, arg1, arg2) + char const *format, *arg1, *arg2; +{ + message5 (format, arg1, arg2, 0, 0); +} + +void +message5 (format, arg1, arg2, arg3, arg4) + char const *format, *arg1, *arg2, *arg3, *arg4; +{ + if (paginate_flag) + { + struct msg *new = (struct msg *) xmalloc (sizeof (struct msg)); + new->format = format; + new->arg1 = concat (arg1, "", ""); + new->arg2 = concat (arg2, "", ""); + new->arg3 = arg3 ? concat (arg3, "", "") : 0; + new->arg4 = arg4 ? concat (arg4, "", "") : 0; + new->next = 0; + *msg_chain_end = new; + msg_chain_end = &new->next; + } + else + { + if (sdiff_help_sdiff) + putchar (' '); + printf (format, arg1, arg2, arg3, arg4); + } +} + +/* Output all the messages that were saved up by calls to `message'. */ + +void +print_message_queue () +{ + struct msg *m; + + for (m = msg_chain; m; m = m->next) + printf (m->format, m->arg1, m->arg2, m->arg3, m->arg4); +} + +/* Call before outputting the results of comparing files NAME0 and NAME1 + to set up OUTFILE, the stdio stream for the output to go to. + + Usually, OUTFILE is just stdout. But when -l was specified + we fork off a `pr' and make OUTFILE a pipe to it. + `pr' then outputs to our stdout. */ + +static char const *current_name0; +static char const *current_name1; +static int current_depth; + +void +setup_output (name0, name1, depth) + char const *name0, *name1; + int depth; +{ + current_name0 = name0; + current_name1 = name1; + current_depth = depth; + outfile = 0; +} + +#if HAVE_FORK +static pid_t pr_pid; +#endif + +void +begin_output () +{ + char *name; + + if (outfile != 0) + return; + + /* Construct the header of this piece of diff. */ + name = xmalloc (strlen (current_name0) + strlen (current_name1) + + strlen (switch_string) + 7); + /* Posix.2 section 4.17.6.1.1 specifies this format. But there is a + bug in the first printing (IEEE Std 1003.2-1992 p 251 l 3304): + it says that we must print only the last component of the pathnames. + This requirement is silly and does not match historical practice. */ + sprintf (name, "diff%s %s %s", switch_string, current_name0, current_name1); + + if (paginate_flag) + { + /* Make OUTFILE a pipe to a subsidiary `pr'. */ + +#if HAVE_FORK + int pipes[2]; + + if (pipe (pipes) != 0) + pfatal_with_name ("pipe"); + + fflush (stdout); + + pr_pid = fork (); + if (pr_pid < 0) + pfatal_with_name ("vfork"); + + if (pr_pid == 0) + { + close (pipes[1]); + if (pipes[0] != STDIN_FILENO) + { + if (dup2 (pipes[0], STDIN_FILENO) < 0) + pfatal_with_name ("dup2"); + close (pipes[0]); + } +#ifdef __FreeBSD__ + execl (PR_PROGRAM, PR_PROGRAM, "-F", "-h", name, 0); +#else + execl (PR_PROGRAM, PR_PROGRAM, "-f", "-h", name, 0); +#endif + pfatal_with_name (PR_PROGRAM); + } + else + { + close (pipes[0]); + outfile = fdopen (pipes[1], "w"); + if (!outfile) + pfatal_with_name ("fdopen"); + } +#else /* ! HAVE_FORK */ + char *command = xmalloc (4 * strlen (name) + strlen (PR_PROGRAM) + 10); + char *p; + char const *a = name; + sprintf (command, "%s -f -h ", PR_PROGRAM); + p = command + strlen (command); + SYSTEM_QUOTE_ARG (p, a); + *p = 0; + outfile = popen (command, "w"); + if (!outfile) + pfatal_with_name (command); + free (command); +#endif /* ! HAVE_FORK */ + } + else + { + + /* If -l was not specified, output the diff straight to `stdout'. */ + + outfile = stdout; + + /* If handling multiple files (because scanning a directory), + print which files the following output is about. */ + if (current_depth > 0) + printf ("%s\n", name); + } + + free (name); + + /* A special header is needed at the beginning of context output. */ + switch (output_style) + { + case OUTPUT_CONTEXT: + print_context_header (files, 0); + break; + + case OUTPUT_UNIFIED: + print_context_header (files, 1); + break; + + default: + break; + } +} + +/* Call after the end of output of diffs for one file. + Close OUTFILE and get rid of the `pr' subfork. */ + +void +finish_output () +{ + if (outfile != 0 && outfile != stdout) + { + int wstatus; + if (ferror (outfile)) + fatal ("write error"); +#if ! HAVE_FORK + wstatus = pclose (outfile); +#else /* HAVE_FORK */ + if (fclose (outfile) != 0) + pfatal_with_name ("write error"); + if (waitpid (pr_pid, &wstatus, 0) < 0) + pfatal_with_name ("waitpid"); +#endif /* HAVE_FORK */ + if (wstatus != 0) + fatal ("subsidiary pr failed"); + } + + outfile = 0; +} + +/* Compare two lines (typically one from each input file) + according to the command line options. + For efficiency, this is invoked only when the lines do not match exactly + but an option like -i might cause us to ignore the difference. + Return nonzero if the lines differ. */ + +int +line_cmp (s1, s2) + char const *s1, *s2; +{ + register unsigned char const *t1 = (unsigned char const *) s1; + register unsigned char const *t2 = (unsigned char const *) s2; + + while (1) + { + register unsigned char c1 = *t1++; + register unsigned char c2 = *t2++; + + /* Test for exact char equality first, since it's a common case. */ + if (c1 != c2) + { + /* Ignore horizontal white space if -b or -w is specified. */ + + if (ignore_all_space_flag) + { + /* For -w, just skip past any white space. */ + while (ISSPACE (c1) && c1 != '\n') c1 = *t1++; + while (ISSPACE (c2) && c2 != '\n') c2 = *t2++; + } + else if (ignore_space_change_flag) + { + /* For -b, advance past any sequence of white space in line 1 + and consider it just one Space, or nothing at all + if it is at the end of the line. */ + if (ISSPACE (c1)) + { + while (c1 != '\n') + { + c1 = *t1++; + if (! ISSPACE (c1)) + { + --t1; + c1 = ' '; + break; + } + } + } + + /* Likewise for line 2. */ + if (ISSPACE (c2)) + { + while (c2 != '\n') + { + c2 = *t2++; + if (! ISSPACE (c2)) + { + --t2; + c2 = ' '; + break; + } + } + } + + if (c1 != c2) + { + /* If we went too far when doing the simple test + for equality, go back to the first non-white-space + character in both sides and try again. */ + if (c2 == ' ' && c1 != '\n' + && (unsigned char const *) s1 + 1 < t1 + && ISSPACE(t1[-2])) + { + --t1; + continue; + } + if (c1 == ' ' && c2 != '\n' + && (unsigned char const *) s2 + 1 < t2 + && ISSPACE(t2[-2])) + { + --t2; + continue; + } + } + } + + /* Lowercase all letters if -i is specified. */ + + if (ignore_case_flag) + { + if (ISUPPER (c1)) + c1 = tolower (c1); + if (ISUPPER (c2)) + c2 = tolower (c2); + } + + if (c1 != c2) + break; + } + if (c1 == '\n') + return 0; + } + + return (1); +} + +/* Find the consecutive changes at the start of the script START. + Return the last link before the first gap. */ + +struct change * +find_change (start) + struct change *start; +{ + return start; +} + +struct change * +find_reverse_change (start) + struct change *start; +{ + return start; +} + +/* Divide SCRIPT into pieces by calling HUNKFUN and + print each piece with PRINTFUN. + Both functions take one arg, an edit script. + + HUNKFUN is called with the tail of the script + and returns the last link that belongs together with the start + of the tail. + + PRINTFUN takes a subscript which belongs together (with a null + link at the end) and prints it. */ + +void +print_script (script, hunkfun, printfun) + struct change *script; + struct change * (*hunkfun) PARAMS((struct change *)); + void (*printfun) PARAMS((struct change *)); +{ + struct change *next = script; + + while (next) + { + struct change *this, *end; + + /* Find a set of changes that belong together. */ + this = next; + end = (*hunkfun) (next); + + /* Disconnect them from the rest of the changes, + making them a hunk, and remember the rest for next iteration. */ + next = end->link; + end->link = 0; +#ifdef DEBUG + debug_script (this); +#endif + + /* Print this hunk. */ + (*printfun) (this); + + /* Reconnect the script so it will all be freed properly. */ + end->link = next; + } +} + +/* Print the text of a single line LINE, + flagging it with the characters in LINE_FLAG (which say whether + the line is inserted, deleted, changed, etc.). */ + +void +print_1_line (line_flag, line) + char const *line_flag; + char const * const *line; +{ + char const *text = line[0], *limit = line[1]; /* Help the compiler. */ + FILE *out = outfile; /* Help the compiler some more. */ + char const *flag_format = 0; + + /* If -T was specified, use a Tab between the line-flag and the text. + Otherwise use a Space (as Unix diff does). + Print neither space nor tab if line-flags are empty. */ + + if (line_flag && *line_flag) + { + flag_format = tab_align_flag ? "%s\t" : "%s "; + fprintf (out, flag_format, line_flag); + } + + output_1_line (text, limit, flag_format, line_flag); + + if ((!line_flag || line_flag[0]) && limit[-1] != '\n') + fputc ('\n', out); +} + +/* Output a line from TEXT up to LIMIT. Without -t, output verbatim. + With -t, expand white space characters to spaces, and if FLAG_FORMAT + is nonzero, output it with argument LINE_FLAG after every + internal carriage return, so that tab stops continue to line up. */ + +void +output_1_line (text, limit, flag_format, line_flag) + char const *text, *limit, *flag_format, *line_flag; +{ + if (!tab_expand_flag) + fwrite (text, sizeof (char), limit - text, outfile); + else + { + register FILE *out = outfile; + register unsigned char c; + register char const *t = text; + register unsigned column = 0; + + while (t < limit) + switch ((c = *t++)) + { + case '\t': + { + unsigned spaces = TAB_WIDTH - column % TAB_WIDTH; + column += spaces; + do + putc (' ', out); + while (--spaces); + } + break; + + case '\r': + putc (c, out); + if (flag_format && t < limit && *t != '\n') + fprintf (out, flag_format, line_flag); + column = 0; + break; + + case '\b': + if (column == 0) + continue; + column--; + putc (c, out); + break; + + default: + if (ISPRINT (c)) + column++; + putc (c, out); + break; + } + } +} + +int +change_letter (inserts, deletes) + int inserts, deletes; +{ + if (!inserts) + return 'd'; + else if (!deletes) + return 'a'; + else + return 'c'; +} + +/* Translate an internal line number (an index into diff's table of lines) + into an actual line number in the input file. + The internal line number is LNUM. FILE points to the data on the file. + + Internal line numbers count from 0 starting after the prefix. + Actual line numbers count from 1 within the entire file. */ + +int +translate_line_number (file, lnum) + struct file_data const *file; + int lnum; +{ + return lnum + file->prefix_lines + 1; +} + +void +translate_range (file, a, b, aptr, bptr) + struct file_data const *file; + int a, b; + int *aptr, *bptr; +{ + *aptr = translate_line_number (file, a - 1) + 1; + *bptr = translate_line_number (file, b + 1) - 1; +} + +/* Print a pair of line numbers with SEPCHAR, translated for file FILE. + If the two numbers are identical, print just one number. + + Args A and B are internal line numbers. + We print the translated (real) line numbers. */ + +void +print_number_range (sepchar, file, a, b) + int sepchar; + struct file_data *file; + int a, b; +{ + int trans_a, trans_b; + translate_range (file, a, b, &trans_a, &trans_b); + + /* Note: we can have B < A in the case of a range of no lines. + In this case, we should print the line number before the range, + which is B. */ + if (trans_b > trans_a) + fprintf (outfile, "%d%c%d", trans_a, sepchar, trans_b); + else + fprintf (outfile, "%d", trans_b); +} + +/* Look at a hunk of edit script and report the range of lines in each file + that it applies to. HUNK is the start of the hunk, which is a chain + of `struct change'. The first and last line numbers of file 0 are stored in + *FIRST0 and *LAST0, and likewise for file 1 in *FIRST1 and *LAST1. + Note that these are internal line numbers that count from 0. + + If no lines from file 0 are deleted, then FIRST0 is LAST0+1. + + Also set *DELETES nonzero if any lines of file 0 are deleted + and set *INSERTS nonzero if any lines of file 1 are inserted. + If only ignorable lines are inserted or deleted, both are + set to 0. */ + +void +analyze_hunk (hunk, first0, last0, first1, last1, deletes, inserts) + struct change *hunk; + int *first0, *last0, *first1, *last1; + int *deletes, *inserts; +{ + int l0, l1, show_from, show_to; + int i; + int trivial = ignore_blank_lines_flag || ignore_regexp_list; + struct change *next; + + show_from = show_to = 0; + + *first0 = hunk->line0; + *first1 = hunk->line1; + + next = hunk; + do + { + l0 = next->line0 + next->deleted - 1; + l1 = next->line1 + next->inserted - 1; + show_from += next->deleted; + show_to += next->inserted; + + for (i = next->line0; i <= l0 && trivial; i++) + if (!ignore_blank_lines_flag || files[0].linbuf[i][0] != '\n') + { + struct regexp_list *r; + char const *line = files[0].linbuf[i]; + int len = files[0].linbuf[i + 1] - line; + + for (r = ignore_regexp_list; r; r = r->next) + if (0 <= re_search (&r->buf, line, len, 0, len, 0)) + break; /* Found a match. Ignore this line. */ + /* If we got all the way through the regexp list without + finding a match, then it's nontrivial. */ + if (!r) + trivial = 0; + } + + for (i = next->line1; i <= l1 && trivial; i++) + if (!ignore_blank_lines_flag || files[1].linbuf[i][0] != '\n') + { + struct regexp_list *r; + char const *line = files[1].linbuf[i]; + int len = files[1].linbuf[i + 1] - line; + + for (r = ignore_regexp_list; r; r = r->next) + if (0 <= re_search (&r->buf, line, len, 0, len, 0)) + break; /* Found a match. Ignore this line. */ + /* If we got all the way through the regexp list without + finding a match, then it's nontrivial. */ + if (!r) + trivial = 0; + } + } + while ((next = next->link) != 0); + + *last0 = l0; + *last1 = l1; + + /* If all inserted or deleted lines are ignorable, + tell the caller to ignore this hunk. */ + + if (trivial) + show_from = show_to = 0; + + *deletes = show_from; + *inserts = show_to; +} + +/* malloc a block of memory, with fatal error message if we can't do it. */ + +VOID * +xmalloc (size) + size_t size; +{ + register VOID *value; + + if (size == 0) + size = 1; + + value = (VOID *) malloc (size); + + if (!value) + fatal ("memory exhausted"); + return value; +} + +/* realloc a block of memory, with fatal error message if we can't do it. */ + +VOID * +xrealloc (old, size) + VOID *old; + size_t size; +{ + register VOID *value; + + if (size == 0) + size = 1; + + value = (VOID *) realloc (old, size); + + if (!value) + fatal ("memory exhausted"); + return value; +} + +/* Concatenate three strings, returning a newly malloc'd string. */ + +char * +concat (s1, s2, s3) + char const *s1, *s2, *s3; +{ + size_t len = strlen (s1) + strlen (s2) + strlen (s3); + char *new = xmalloc (len + 1); + sprintf (new, "%s%s%s", s1, s2, s3); + return new; +} + +/* Yield the newly malloc'd pathname + of the file in DIR whose filename is FILE. */ + +char * +dir_file_pathname (dir, file) + char const *dir, *file; +{ + char const *p = filename_lastdirchar (dir); + return concat (dir, "/" + (p && !p[1]), file); +} + +void +debug_script (sp) + struct change *sp; +{ + fflush (stdout); + for (; sp; sp = sp->link) + fprintf (stderr, "%3d %3d delete %d insert %d\n", + sp->line0, sp->line1, sp->deleted, sp->inserted); + fflush (stderr); +} diff --git a/sys/src/ape/cmd/diff/version.c b/sys/src/ape/cmd/diff/version.c new file mode 100755 index 000000000..c2b6a8ae5 --- /dev/null +++ b/sys/src/ape/cmd/diff/version.c @@ -0,0 +1,5 @@ +/* Version number of GNU diff. */ + +#include "config.h" + +char const version_string[] = "2.7"; diff --git a/sys/src/ape/cmd/diff/xmalloc.c b/sys/src/ape/cmd/diff/xmalloc.c new file mode 100755 index 000000000..dc44ba4cf --- /dev/null +++ b/sys/src/ape/cmd/diff/xmalloc.c @@ -0,0 +1,81 @@ +/* xmalloc.c -- malloc with out of memory checking + Copyright (C) 1990, 1991, 1993 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#if __STDC__ +#define VOID void +#else +#define VOID char +#endif + +#include <sys/types.h> + +#if STDC_HEADERS +#include <stdlib.h> +#else +VOID *malloc (); +VOID *realloc (); +void free (); +#endif + +#if __STDC__ && defined (HAVE_VPRINTF) +void error (int, int, char const *, ...); +#else +void error (); +#endif + +/* Allocate N bytes of memory dynamically, with error checking. */ + +VOID * +xmalloc (n) + size_t n; +{ + VOID *p; + + p = malloc (n); + if (p == 0) + /* Must exit with 2 for `cmp'. */ + error (2, 0, "memory exhausted"); + return p; +} + +/* Change the size of an allocated block of memory P to N bytes, + with error checking. + If P is NULL, run xmalloc. + If N is 0, run free and return NULL. */ + +VOID * +xrealloc (p, n) + VOID *p; + size_t n; +{ + if (p == 0) + return xmalloc (n); + if (n == 0) + { + free (p); + return 0; + } + p = realloc (p, n); + if (p == 0) + /* Must exit with 2 for `cmp'. */ + error (2, 0, "memory exhausted"); + return p; +} |