11573Srgrimes/*- 21573Srgrimes * Copyright (c) 1992, 1993, 1994 Henry Spencer. 31573Srgrimes * Copyright (c) 1992, 1993, 1994 41573Srgrimes * The Regents of the University of California. All rights reserved. 51573Srgrimes * 61573Srgrimes * This code is derived from software contributed to Berkeley by 71573Srgrimes * Henry Spencer. 81573Srgrimes * 91573Srgrimes * Redistribution and use in source and binary forms, with or without 101573Srgrimes * modification, are permitted provided that the following conditions 111573Srgrimes * are met: 121573Srgrimes * 1. Redistributions of source code must retain the above copyright 131573Srgrimes * notice, this list of conditions and the following disclaimer. 141573Srgrimes * 2. Redistributions in binary form must reproduce the above copyright 151573Srgrimes * notice, this list of conditions and the following disclaimer in the 161573Srgrimes * documentation and/or other materials provided with the distribution. 171573Srgrimes * 4. Neither the name of the University nor the names of its contributors 181573Srgrimes * may be used to endorse or promote products derived from this software 191573Srgrimes * without specific prior written permission. 201573Srgrimes * 211573Srgrimes * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 221573Srgrimes * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 231573Srgrimes * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 241573Srgrimes * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 251573Srgrimes * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 261573Srgrimes * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 271573Srgrimes * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 281573Srgrimes * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 291573Srgrimes * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 301573Srgrimes * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 311573Srgrimes * SUCH DAMAGE. 321573Srgrimes * 331573Srgrimes * @(#)regexec.c 8.3 (Berkeley) 3/20/94 341573Srgrimes */ 351573Srgrimes 361573Srgrimes#if defined(LIBC_SCCS) && !defined(lint) 371573Srgrimesstatic char sccsid[] = "@(#)regexec.c 8.3 (Berkeley) 3/20/94"; 381573Srgrimes#endif /* LIBC_SCCS and not lint */ 3992889Sobrien#include <sys/cdefs.h> 4092889Sobrien__FBSDID("$FreeBSD$"); 411573Srgrimes 421573Srgrimes/* 431573Srgrimes * the outer shell of regexec() 441573Srgrimes * 45132019Stjr * This file includes engine.c three times, after muchos fiddling with the 461573Srgrimes * macros that code uses. This lets the same code operate on two different 47132019Stjr * representations for state sets and characters. 481573Srgrimes */ 491573Srgrimes#include <sys/types.h> 501573Srgrimes#include <stdio.h> 511573Srgrimes#include <stdlib.h> 521573Srgrimes#include <string.h> 531573Srgrimes#include <limits.h> 541573Srgrimes#include <ctype.h> 551573Srgrimes#include <regex.h> 56132019Stjr#include <wchar.h> 57132019Stjr#include <wctype.h> 581573Srgrimes 591573Srgrimes#include "utils.h" 601573Srgrimes#include "regex2.h" 611573Srgrimes 62111010Snectarstatic int nope __unused = 0; /* for use in asserts; shuts lint up */ 631573Srgrimes 64132019Stjrstatic __inline size_t 65170528Sdelphijxmbrtowc(wint_t *wi, const char *s, size_t n, mbstate_t *mbs, wint_t dummy) 66132019Stjr{ 67132019Stjr size_t nr; 68132019Stjr wchar_t wc; 69132019Stjr 70132019Stjr nr = mbrtowc(&wc, s, n, mbs); 71132019Stjr if (wi != NULL) 72132019Stjr *wi = wc; 73132019Stjr if (nr == 0) 74132019Stjr return (1); 75132019Stjr else if (nr == (size_t)-1 || nr == (size_t)-2) { 76132019Stjr memset(mbs, 0, sizeof(*mbs)); 77132019Stjr if (wi != NULL) 78132019Stjr *wi = dummy; 79132019Stjr return (1); 80132019Stjr } else 81132019Stjr return (nr); 82132019Stjr} 83132019Stjr 84132019Stjrstatic __inline size_t 85170528Sdelphijxmbrtowc_dummy(wint_t *wi, 86170528Sdelphij const char *s, 87170528Sdelphij size_t n __unused, 88170528Sdelphij mbstate_t *mbs __unused, 89170528Sdelphij wint_t dummy __unused) 90132019Stjr{ 91132019Stjr 92132019Stjr if (wi != NULL) 93132019Stjr *wi = (unsigned char)*s; 94132019Stjr return (1); 95132019Stjr} 96132019Stjr 971573Srgrimes/* macros for manipulating states, small version */ 981573Srgrimes#define states long 991573Srgrimes#define states1 states /* for later use in regexec() decision */ 1001573Srgrimes#define CLEAR(v) ((v) = 0) 10136043Sjb#define SET0(v, n) ((v) &= ~((unsigned long)1 << (n))) 10236043Sjb#define SET1(v, n) ((v) |= (unsigned long)1 << (n)) 10336043Sjb#define ISSET(v, n) (((v) & ((unsigned long)1 << (n))) != 0) 1041573Srgrimes#define ASSIGN(d, s) ((d) = (s)) 1051573Srgrimes#define EQ(a, b) ((a) == (b)) 10636043Sjb#define STATEVARS long dummy /* dummy version */ 1071573Srgrimes#define STATESETUP(m, n) /* nothing */ 1081573Srgrimes#define STATETEARDOWN(m) /* nothing */ 1091573Srgrimes#define SETUP(v) ((v) = 0) 11036043Sjb#define onestate long 11136043Sjb#define INIT(o, n) ((o) = (unsigned long)1 << (n)) 1121573Srgrimes#define INC(o) ((o) <<= 1) 11336043Sjb#define ISSTATEIN(v, o) (((v) & (o)) != 0) 1141573Srgrimes/* some abbreviations; note that some of these know variable names! */ 1151573Srgrimes/* do "if I'm here, I can also be there" etc without branches */ 11636043Sjb#define FWD(dst, src, n) ((dst) |= ((unsigned long)(src)&(here)) << (n)) 11736043Sjb#define BACK(dst, src, n) ((dst) |= ((unsigned long)(src)&(here)) >> (n)) 11836043Sjb#define ISSETBACK(v, n) (((v) & ((unsigned long)here >> (n))) != 0) 119132019Stjr/* no multibyte support */ 120132019Stjr#define XMBRTOWC xmbrtowc_dummy 121132019Stjr#define ZAPSTATE(mbs) ((void)(mbs)) 1221573Srgrimes/* function names */ 1231573Srgrimes#define SNAMES /* engine.c looks after details */ 1241573Srgrimes 1251573Srgrimes#include "engine.c" 1261573Srgrimes 1271573Srgrimes/* now undo things */ 1281573Srgrimes#undef states 1291573Srgrimes#undef CLEAR 1301573Srgrimes#undef SET0 1311573Srgrimes#undef SET1 1321573Srgrimes#undef ISSET 1331573Srgrimes#undef ASSIGN 1341573Srgrimes#undef EQ 1351573Srgrimes#undef STATEVARS 1361573Srgrimes#undef STATESETUP 1371573Srgrimes#undef STATETEARDOWN 1381573Srgrimes#undef SETUP 1391573Srgrimes#undef onestate 1401573Srgrimes#undef INIT 1411573Srgrimes#undef INC 1421573Srgrimes#undef ISSTATEIN 1431573Srgrimes#undef FWD 1441573Srgrimes#undef BACK 1451573Srgrimes#undef ISSETBACK 1461573Srgrimes#undef SNAMES 147132019Stjr#undef XMBRTOWC 148132019Stjr#undef ZAPSTATE 1491573Srgrimes 1501573Srgrimes/* macros for manipulating states, large version */ 1511573Srgrimes#define states char * 1521573Srgrimes#define CLEAR(v) memset(v, 0, m->g->nstates) 1531573Srgrimes#define SET0(v, n) ((v)[n] = 0) 1541573Srgrimes#define SET1(v, n) ((v)[n] = 1) 1551573Srgrimes#define ISSET(v, n) ((v)[n]) 1561573Srgrimes#define ASSIGN(d, s) memcpy(d, s, m->g->nstates) 1571573Srgrimes#define EQ(a, b) (memcmp(a, b, m->g->nstates) == 0) 15836043Sjb#define STATEVARS long vn; char *space 1591573Srgrimes#define STATESETUP(m, nv) { (m)->space = malloc((nv)*(m)->g->nstates); \ 1601573Srgrimes if ((m)->space == NULL) return(REG_ESPACE); \ 1611573Srgrimes (m)->vn = 0; } 1621573Srgrimes#define STATETEARDOWN(m) { free((m)->space); } 1631573Srgrimes#define SETUP(v) ((v) = &m->space[m->vn++ * m->g->nstates]) 16436043Sjb#define onestate long 1651573Srgrimes#define INIT(o, n) ((o) = (n)) 1661573Srgrimes#define INC(o) ((o)++) 1671573Srgrimes#define ISSTATEIN(v, o) ((v)[o]) 1681573Srgrimes/* some abbreviations; note that some of these know variable names! */ 1691573Srgrimes/* do "if I'm here, I can also be there" etc without branches */ 1701573Srgrimes#define FWD(dst, src, n) ((dst)[here+(n)] |= (src)[here]) 1711573Srgrimes#define BACK(dst, src, n) ((dst)[here-(n)] |= (src)[here]) 1721573Srgrimes#define ISSETBACK(v, n) ((v)[here - (n)]) 173132019Stjr/* no multibyte support */ 174132019Stjr#define XMBRTOWC xmbrtowc_dummy 175132019Stjr#define ZAPSTATE(mbs) ((void)(mbs)) 1761573Srgrimes/* function names */ 1771573Srgrimes#define LNAMES /* flag */ 1781573Srgrimes 1791573Srgrimes#include "engine.c" 1801573Srgrimes 181132019Stjr/* multibyte character & large states version */ 182132019Stjr#undef LNAMES 183132019Stjr#undef XMBRTOWC 184132019Stjr#undef ZAPSTATE 185132019Stjr#define XMBRTOWC xmbrtowc 186132019Stjr#define ZAPSTATE(mbs) memset((mbs), 0, sizeof(*(mbs))) 187132019Stjr#define MNAMES 188132019Stjr 189132019Stjr#include "engine.c" 190132019Stjr 1911573Srgrimes/* 1921573Srgrimes - regexec - interface for matching 1931573Srgrimes = extern int regexec(const regex_t *, const char *, size_t, \ 1941573Srgrimes = regmatch_t [], int); 1951573Srgrimes = #define REG_NOTBOL 00001 1961573Srgrimes = #define REG_NOTEOL 00002 1971573Srgrimes = #define REG_STARTEND 00004 1981573Srgrimes = #define REG_TRACE 00400 // tracing of execution 1991573Srgrimes = #define REG_LARGE 01000 // force large representation 2001573Srgrimes = #define REG_BACKR 02000 // force use of backref code 2011573Srgrimes * 2021573Srgrimes * We put this here so we can exploit knowledge of the state representation 2031573Srgrimes * when choosing which matcher to call. Also, by this point the matchers 2041573Srgrimes * have been prototyped. 2051573Srgrimes */ 2061573Srgrimesint /* 0 success, REG_NOMATCH failure */ 207170528Sdelphijregexec(const regex_t * __restrict preg, 208170528Sdelphij const char * __restrict string, 209170528Sdelphij size_t nmatch, 210170528Sdelphij regmatch_t pmatch[__restrict], 211170528Sdelphij int eflags) 2121573Srgrimes{ 21392889Sobrien struct re_guts *g = preg->re_g; 2141573Srgrimes#ifdef REDEBUG 2151573Srgrimes# define GOODFLAGS(f) (f) 2161573Srgrimes#else 2171573Srgrimes# define GOODFLAGS(f) ((f)&(REG_NOTBOL|REG_NOTEOL|REG_STARTEND)) 2181573Srgrimes#endif 2191573Srgrimes 2201573Srgrimes if (preg->re_magic != MAGIC1 || g->magic != MAGIC2) 2211573Srgrimes return(REG_BADPAT); 2221573Srgrimes assert(!(g->iflags&BAD)); 2231573Srgrimes if (g->iflags&BAD) /* backstop for no-debug case */ 2241573Srgrimes return(REG_BADPAT); 2251573Srgrimes eflags = GOODFLAGS(eflags); 2261573Srgrimes 227132019Stjr if (MB_CUR_MAX > 1) 228132019Stjr return(mmatcher(g, (char *)string, nmatch, pmatch, eflags)); 229132019Stjr else if (g->nstates <= CHAR_BIT*sizeof(states1) && !(eflags®_LARGE)) 2301573Srgrimes return(smatcher(g, (char *)string, nmatch, pmatch, eflags)); 2311573Srgrimes else 2321573Srgrimes return(lmatcher(g, (char *)string, nmatch, pmatch, eflags)); 2331573Srgrimes} 234