remove old libregexp files; add headers for upas/bayes

2 files changed, 129 insertions, 0 deletions

diff --git a/sys/src/cmd/upas/bayes/regcomp.h b/sys/src/cmd/upas/bayes/regcomp.h
new file mode 100644
index 000000000..402fe7d5c
--- /dev/null
+++ b/sys/src/cmd/upas/bayes/regcomp.h
@@ -0,0 +1,63 @@
+/*
+ *  substitution list
+ */
+#define NSUBEXP 32
+typedef struct Resublist	Resublist;
+struct	Resublist
+{
+	Resub	m[NSUBEXP];
+};
+
+/*
+ * Actions and Tokens (Reinst types)
+ *
+ *	02xx are operators, value == precedence
+ *	03xx are tokens, i.e. operands for operators
+ */
+#define RUNE		0177
+#define	OPERATOR	0200	/* Bitmask of all operators */
+#define	START		0200	/* Start, used for marker on stack */
+#define	RBRA		0201	/* Right bracket, ) */
+#define	LBRA		0202	/* Left bracket, ( */
+#define	OR		0203	/* Alternation, | */
+#define	CAT		0204	/* Concatentation, implicit operator */
+#define	STAR		0205	/* Closure, * */
+#define	PLUS		0206	/* a+ == aa* */
+#define	QUEST		0207	/* a? == a|nothing, i.e. 0 or 1 a's */
+#define	ANY		0300	/* Any character except newline, . */
+#define	ANYNL		0301	/* Any character including newline, . */
+#define	NOP		0302	/* No operation, internal use only */
+#define	BOL		0303	/* Beginning of line, ^ */
+#define	EOL		0304	/* End of line, $ */
+#define	CCLASS		0305	/* Character class, [] */
+#define	NCCLASS		0306	/* Negated character class, [] */
+#define	END		0377	/* Terminate: match found */
+
+/*
+ *  regexec execution lists
+ */
+#define LISTSIZE	10
+#define BIGLISTSIZE	(25*LISTSIZE)
+typedef struct Relist	Relist;
+struct Relist
+{
+	Reinst*		inst;		/* Reinstruction of the thread */
+	Resublist	se;		/* matched subexpressions in this thread */
+};
+typedef struct Reljunk	Reljunk;
+struct	Reljunk
+{
+	Relist*	relist[2];
+	Relist*	reliste[2];
+	int	starttype;
+	Rune	startchar;
+	char*	starts;
+	char*	eol;
+	Rune*	rstarts;
+	Rune*	reol;
+};
+
+extern Relist*	_renewthread(Relist*, Reinst*, int, Resublist*);
+extern void	_renewmatch(Resub*, int, Resublist*);
+extern Relist*	_renewemptythread(Relist*, Reinst*, int, char*);
+extern Relist*	_rrenewemptythread(Relist*, Reinst*, int, Rune*);
diff --git a/sys/src/cmd/upas/bayes/regexp.h b/sys/src/cmd/upas/bayes/regexp.h
new file mode 100644
index 000000000..780dc8001
--- /dev/null
+++ b/sys/src/cmd/upas/bayes/regexp.h
@@ -0,0 +1,66 @@
+#pragma	src	"/sys/src/oldlibregexp"
+#pragma	lib	"oldlibregexp.a"
+
+typedef struct Resub		Resub;
+typedef struct Reclass		Reclass;
+typedef struct Reinst		Reinst;
+typedef struct Reprog		Reprog;
+
+/*
+ *	Sub expression matches
+ */
+struct Resub{
+	union
+	{
+		char *sp;
+		Rune *rsp;
+	};
+	union
+	{
+		char *ep;
+		Rune *rep;
+	};
+};
+
+/*
+ *	character class, each pair of rune's defines a range
+ */
+struct Reclass{
+	Rune	*end;
+	Rune	spans[64];
+};
+
+/*
+ *	Machine instructions
+ */
+struct Reinst{
+	int	type;
+	union	{
+		Reclass	*cp;		/* class pointer */
+		Rune	r;		/* character */
+		int	subid;		/* sub-expression id for RBRA and LBRA */
+		Reinst	*right;		/* right child of OR */
+	};
+	union {	/* regexp relies on these two being in the same union */
+		Reinst *left;		/* left child of OR */
+		Reinst *next;		/* next instruction for CAT & LBRA */
+	};
+};
+
+/*
+ *	Reprogram definition
+ */
+struct Reprog{
+	Reinst	*startinst;	/* start pc */
+	Reclass	class[16];	/* .data */
+	Reinst	firstinst[5];	/* .text */
+};
+
+extern Reprog	*regcomp(char*);
+extern Reprog	*regcomplit(char*);
+extern Reprog	*regcompnl(char*);
+extern void	regerror(char*);
+extern int	regexec(Reprog*, char*, Resub*, int);
+extern void	regsub(char*, char*, int, Resub*, int);
+extern int	rregexec(Reprog*, Rune*, Resub*, int);
+extern void	rregsub(Rune*, Rune*, int, Resub*, int);