1f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* $OpenBSD: regexec.c,v 1.11 2005/08/05 13:03:00 espie Exp $ */ 2f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/*- 3f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * Copyright (c) 1992, 1993, 1994 Henry Spencer. 4f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * Copyright (c) 1992, 1993, 1994 5f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * The Regents of the University of California. All rights reserved. 6f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * 7f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * This code is derived from software contributed to Berkeley by 8f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * Henry Spencer. 9f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * 10f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * Redistribution and use in source and binary forms, with or without 11f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * modification, are permitted provided that the following conditions 12f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * are met: 13f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * 1. Redistributions of source code must retain the above copyright 14f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * notice, this list of conditions and the following disclaimer. 15f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * 2. Redistributions in binary form must reproduce the above copyright 16f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * notice, this list of conditions and the following disclaimer in the 17f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * documentation and/or other materials provided with the distribution. 18f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * 3. Neither the name of the University nor the names of its contributors 19f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * may be used to endorse or promote products derived from this software 20f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * without specific prior written permission. 21f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * 22f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 23f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 24f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 25f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 26f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 27f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 28f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 29f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 30f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 31f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 32f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * SUCH DAMAGE. 33f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * 34f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * @(#)regexec.c 8.3 (Berkeley) 3/20/94 35f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine */ 36f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine 37f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* 38f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * the outer shell of regexec() 39f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * 40f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * This file includes engine.c *twice*, after muchos fiddling with the 41f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * macros that code uses. This lets the same code operate on two different 42f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * representations for state sets. 43f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine */ 44f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include <sys/types.h> 45f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include <stdio.h> 46f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include <stdlib.h> 47f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include <string.h> 48f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include <limits.h> 49f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include <ctype.h> 50f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include "./regex.h" 51f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine 52f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include "utils.h" 53f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include "./regex2.h" 54f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine 55f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* macros for manipulating states, small version */ 56f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define states long 57f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define states1 states /* for later use in regexec() decision */ 58f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define CLEAR(v) ((v) = 0) 59f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define SET0(v, n) ((v) &= ~((unsigned long)1 << (n))) 60f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define SET1(v, n) ((v) |= (unsigned long)1 << (n)) 61f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define ISSET(v, n) (((v) & ((unsigned long)1 << (n))) != 0) 62f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define ASSIGN(d, s) ((d) = (s)) 63f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define EQ(a, b) ((a) == (b)) 64f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define STATEVARS long dummy /* dummy version */ 65f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define STATESETUP(m, n) /* nothing */ 66f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define STATETEARDOWN(m) /* nothing */ 67f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define SETUP(v) ((v) = 0) 68f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define onestate long 69f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define INIT(o, n) ((o) = (unsigned long)1 << (n)) 70f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define INC(o) ((o) <<= 1) 71f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define ISSTATEIN(v, o) (((v) & (o)) != 0) 72f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* some abbreviations; note that some of these know variable names! */ 73f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* do "if I'm here, I can also be there" etc without branches */ 74f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define FWD(dst, src, n) ((dst) |= ((unsigned long)(src)&(here)) << (n)) 75f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define BACK(dst, src, n) ((dst) |= ((unsigned long)(src)&(here)) >> (n)) 76f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define ISSETBACK(v, n) (((v) & ((unsigned long)here >> (n))) != 0) 77f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* function names */ 78f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define SNAMES /* engine.c looks after details */ 79f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine 80f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include "engine.c" 81f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine 82f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* now undo things */ 83f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef states 84f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef CLEAR 85f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef SET0 86f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef SET1 87f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef ISSET 88f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef ASSIGN 89f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef EQ 90f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef STATEVARS 91f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef STATESETUP 92f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef STATETEARDOWN 93f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef SETUP 94f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef onestate 95f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef INIT 96f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef INC 97f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef ISSTATEIN 98f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef FWD 99f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef BACK 100f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef ISSETBACK 101f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#undef SNAMES 102f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine 103f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* macros for manipulating states, large version */ 104f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define states char * 105f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define CLEAR(v) memset(v, 0, m->g->nstates) 106f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define SET0(v, n) ((v)[n] = 0) 107f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define SET1(v, n) ((v)[n] = 1) 108f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define ISSET(v, n) ((v)[n]) 109f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define ASSIGN(d, s) memcpy(d, s, m->g->nstates) 110f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define EQ(a, b) (memcmp(a, b, m->g->nstates) == 0) 111f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define STATEVARS long vn; char *space 112f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define STATESETUP(m, nv) { (m)->space = (char*)malloc((nv)*(m)->g->nstates); \ 113f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine if ((m)->space == NULL) return(REG_ESPACE); \ 114f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine (m)->vn = 0; } 115f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define STATETEARDOWN(m) { free((m)->space); } 116f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define SETUP(v) ((v) = &m->space[m->vn++ * m->g->nstates]) 117f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define onestate long 118f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define INIT(o, n) ((o) = (n)) 119f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define INC(o) ((o)++) 120f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define ISSTATEIN(v, o) ((v)[o]) 121f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* some abbreviations; note that some of these know variable names! */ 122f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* do "if I'm here, I can also be there" etc without branches */ 123f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define FWD(dst, src, n) ((dst)[here+(n)] |= (src)[here]) 124f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define BACK(dst, src, n) ((dst)[here-(n)] |= (src)[here]) 125f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define ISSETBACK(v, n) ((v)[here - (n)]) 126f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* function names */ 127f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#define LNAMES /* flag */ 128f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine 129f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#include "engine.c" 130f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine 131f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine/* 132f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine - regexec - interface for matching 133f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * 134f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * We put this here so we can exploit knowledge of the state representation 135f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * when choosing which matcher to call. Also, by this point the matchers 136f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine * have been prototyped. 137f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine */ 138f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkineint /* 0 success, REG_NOMATCH failure */ 139f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkineregexec(const regex_t *preg, const char *string, size_t nmatch, 140f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine regmatch_t pmatch[], int eflags) 141f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine{ 142f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine struct re_guts *g = preg->re_g; 143f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#ifdef REDEBUG 144f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine# define GOODFLAGS(f) (f) 145f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#else 146f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine# define GOODFLAGS(f) ((f)&(REG_NOTBOL|REG_NOTEOL|REG_STARTEND)) 147f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine#endif 148f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine 149f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine if (preg->re_magic != MAGIC1 || g->magic != MAGIC2) 150f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine return(REG_BADPAT); 151f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine assert(!(g->iflags&BAD)); 152f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine if (g->iflags&BAD) /* backstop for no-debug case */ 153f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine return(REG_BADPAT); 154f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine eflags = GOODFLAGS(eflags); 155f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine 156f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine if (g->nstates <= (int)(CHAR_BIT*sizeof(states1)) && !(eflags®_LARGE)) 157f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine return(smatcher(g, (char *)string, nmatch, pmatch, eflags)); 158f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine else 159f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine return(lmatcher(g, (char *)string, nmatch, pmatch, eflags)); 160f9e88edc21c894886747e899dc29c02a28728d65Vladimir Chtchetkine} 161