1ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/*-
2ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * This code is derived from OpenBSD's libc/regex, original license follows:
3ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *
4ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * Copyright (c) 1992, 1993, 1994 Henry Spencer.
5ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * Copyright (c) 1992, 1993, 1994
6ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *	The Regents of the University of California.  All rights reserved.
7ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *
8ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * This code is derived from software contributed to Berkeley by
9ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * Henry Spencer.
10ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *
11ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * Redistribution and use in source and binary forms, with or without
12ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * modification, are permitted provided that the following conditions
13ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * are met:
14ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * 1. Redistributions of source code must retain the above copyright
15ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *    notice, this list of conditions and the following disclaimer.
16ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * 2. Redistributions in binary form must reproduce the above copyright
17ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *    notice, this list of conditions and the following disclaimer in the
18ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *    documentation and/or other materials provided with the distribution.
19ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * 3. Neither the name of the University nor the names of its contributors
20ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *    may be used to endorse or promote products derived from this software
21ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *    without specific prior written permission.
22ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *
23ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * SUCH DAMAGE.
34ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *
35ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *	@(#)regexec.c	8.3 (Berkeley) 3/20/94
36ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin */
37ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin
38ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/*
39ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * the outer shell of llvm_regexec()
40ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *
41ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * This file includes engine.inc *twice*, after muchos fiddling with the
42ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * macros that code uses.  This lets the same code operate on two different
43ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * representations for state sets.
44ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin */
45ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include <sys/types.h>
46ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include <stdio.h>
47ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include <stdlib.h>
48ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include <string.h>
49ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include <limits.h>
50ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include <ctype.h>
51ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include "regex_impl.h"
52ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin
53ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include "regutils.h"
54ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include "regex2.h"
55ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin
56ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/* macros for manipulating states, small version */
57ad3ea3d85f164dc1b2746950e378334197d8688dNAKAMURA Takumi/* FIXME: 'states' is assumed as 'long' on small version. */
58ad3ea3d85f164dc1b2746950e378334197d8688dNAKAMURA Takumi#define	states1	long		/* for later use in llvm_regexec() decision */
59ad3ea3d85f164dc1b2746950e378334197d8688dNAKAMURA Takumi#define	states	states1
60ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	CLEAR(v)	((v) = 0)
61ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	SET0(v, n)	((v) &= ~((unsigned long)1 << (n)))
62ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	SET1(v, n)	((v) |= (unsigned long)1 << (n))
63ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	ISSET(v, n)	(((v) & ((unsigned long)1 << (n))) != 0)
64ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	ASSIGN(d, s)	((d) = (s))
65ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	EQ(a, b)	((a) == (b))
66ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	STATEVARS	long dummy	/* dummy version */
67ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	STATESETUP(m, n)	/* nothing */
68ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	STATETEARDOWN(m)	/* nothing */
69ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	SETUP(v)	((v) = 0)
70ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	onestate	long
71ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	INIT(o, n)	((o) = (unsigned long)1 << (n))
721144af3c9b4da48cd581156e05b24261c8de366aRichard Smith#define	INC(o)		((o) = (unsigned long)(o) << 1)
73ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	ISSTATEIN(v, o)	(((v) & (o)) != 0)
74ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/* some abbreviations; note that some of these know variable names! */
75ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/* do "if I'm here, I can also be there" etc without branches */
76ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	FWD(dst, src, n)	((dst) |= ((unsigned long)(src)&(here)) << (n))
77ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	BACK(dst, src, n)	((dst) |= ((unsigned long)(src)&(here)) >> (n))
78ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	ISSETBACK(v, n)		(((v) & ((unsigned long)here >> (n))) != 0)
79ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/* function names */
80ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define SNAMES			/* engine.inc looks after details */
81ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin
82ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include "regengine.inc"
83ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin
84ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/* now undo things */
85ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	states
86ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	CLEAR
87ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	SET0
88ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	SET1
89ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	ISSET
90ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	ASSIGN
91ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	EQ
92ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	STATEVARS
93ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	STATESETUP
94ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	STATETEARDOWN
95ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	SETUP
96ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	onestate
97ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	INIT
98ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	INC
99ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	ISSTATEIN
100ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	FWD
101ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	BACK
102ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	ISSETBACK
103ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#undef	SNAMES
104ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin
105ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/* macros for manipulating states, large version */
106ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	states	char *
107ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	CLEAR(v)	memset(v, 0, m->g->nstates)
108ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	SET0(v, n)	((v)[n] = 0)
109ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	SET1(v, n)	((v)[n] = 1)
110ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	ISSET(v, n)	((v)[n])
111ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	ASSIGN(d, s)	memmove(d, s, m->g->nstates)
112ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	EQ(a, b)	(memcmp(a, b, m->g->nstates) == 0)
113ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	STATEVARS	long vn; char *space
114ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	STATESETUP(m, nv)	{ (m)->space = malloc((nv)*(m)->g->nstates); \
115ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin				if ((m)->space == NULL) return(REG_ESPACE); \
116ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin				(m)->vn = 0; }
117ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	STATETEARDOWN(m)	{ free((m)->space); }
118ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	SETUP(v)	((v) = &m->space[m->vn++ * m->g->nstates])
119ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	onestate	long
120ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	INIT(o, n)	((o) = (n))
121ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	INC(o)	((o)++)
122ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	ISSTATEIN(v, o)	((v)[o])
123ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/* some abbreviations; note that some of these know variable names! */
124ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/* do "if I'm here, I can also be there" etc without branches */
125ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	FWD(dst, src, n)	((dst)[here+(n)] |= (src)[here])
126ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	BACK(dst, src, n)	((dst)[here-(n)] |= (src)[here])
127ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	ISSETBACK(v, n)	((v)[here - (n)])
128ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/* function names */
129ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#define	LNAMES			/* flag */
130ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin
131ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#include "regengine.inc"
132ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin
133ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin/*
134ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin - llvm_regexec - interface for matching
135ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin *
136ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * We put this here so we can exploit knowledge of the state representation
137ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * when choosing which matcher to call.  Also, by this point the matchers
138ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin * have been prototyped.
139ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin */
140ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwinint				/* 0 success, REG_NOMATCH failure */
141ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwinllvm_regexec(const llvm_regex_t *preg, const char *string, size_t nmatch,
142ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin             llvm_regmatch_t pmatch[], int eflags)
143ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin{
144ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin	struct re_guts *g = preg->re_g;
145ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#ifdef REDEBUG
146ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#	define	GOODFLAGS(f)	(f)
147ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#else
148ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#	define	GOODFLAGS(f)	((f)&(REG_NOTBOL|REG_NOTEOL|REG_STARTEND))
149ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin#endif
150ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin
151ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin	if (preg->re_magic != MAGIC1 || g->magic != MAGIC2)
152ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin		return(REG_BADPAT);
153ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin	assert(!(g->iflags&REGEX_BAD));
154ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin	if (g->iflags&REGEX_BAD)		/* backstop for no-debug case */
155ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin		return(REG_BADPAT);
156ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin	eflags = GOODFLAGS(eflags);
157ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin
158ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin	if (g->nstates <= (long)(CHAR_BIT*sizeof(states1)) && !(eflags&REG_LARGE))
159d2520dcc701da54862e3e73b16d58c25474417beOwen Anderson		return(smatcher(g, string, nmatch, pmatch, eflags));
160ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin	else
161d2520dcc701da54862e3e73b16d58c25474417beOwen Anderson		return(lmatcher(g, string, nmatch, pmatch, eflags));
162ce0c81e7dd321e9f94f628daa5528f56cab0ab88Torok Edwin}
163