11573Srgrimes/* 21573Srgrimes * Copyright (c) 1989, 1993, 1994 31573Srgrimes * The Regents of the University of California. All rights reserved. 41573Srgrimes * 51573Srgrimes * This code is derived from software contributed to Berkeley by 61573Srgrimes * Guido van Rossum. 71573Srgrimes * 8227753Stheraven * Copyright (c) 2011 The FreeBSD Foundation 9227753Stheraven * All rights reserved. 10227753Stheraven * Portions of this software were developed by David Chisnall 11227753Stheraven * under sponsorship from the FreeBSD Foundation. 12227753Stheraven * 131573Srgrimes * Redistribution and use in source and binary forms, with or without 141573Srgrimes * modification, are permitted provided that the following conditions 151573Srgrimes * are met: 161573Srgrimes * 1. Redistributions of source code must retain the above copyright 171573Srgrimes * notice, this list of conditions and the following disclaimer. 181573Srgrimes * 2. Redistributions in binary form must reproduce the above copyright 191573Srgrimes * notice, this list of conditions and the following disclaimer in the 201573Srgrimes * documentation and/or other materials provided with the distribution. 211573Srgrimes * 4. Neither the name of the University nor the names of its contributors 221573Srgrimes * may be used to endorse or promote products derived from this software 231573Srgrimes * without specific prior written permission. 241573Srgrimes * 251573Srgrimes * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 261573Srgrimes * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 271573Srgrimes * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 281573Srgrimes * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 291573Srgrimes * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 301573Srgrimes * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 311573Srgrimes * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 321573Srgrimes * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 331573Srgrimes * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 341573Srgrimes * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 351573Srgrimes * SUCH DAMAGE. 361573Srgrimes */ 371573Srgrimes 381573Srgrimes#if defined(LIBC_SCCS) && !defined(lint) 391573Srgrimesstatic char sccsid[] = "@(#)fnmatch.c 8.2 (Berkeley) 4/16/94"; 401573Srgrimes#endif /* LIBC_SCCS and not lint */ 4190045Sobrien#include <sys/cdefs.h> 4290045Sobrien__FBSDID("$FreeBSD: releng/10.3/lib/libc/gen/fnmatch.c 289943 2015-10-25 21:39:23Z jilles $"); 431573Srgrimes 441573Srgrimes/* 451573Srgrimes * Function fnmatch() as specified in POSIX 1003.2-1992, section B.6. 461573Srgrimes * Compares a filename or pathname to a pattern. 471573Srgrimes */ 481573Srgrimes 49132812Stjr/* 50132812Stjr * Some notes on multibyte character support: 51132812Stjr * 1. Patterns with illegal byte sequences match nothing. 52132812Stjr * 2. Illegal byte sequences in the "string" argument are handled by treating 53132812Stjr * them as single-byte characters with a value of the first byte of the 54132812Stjr * sequence cast to wchar_t. 55132812Stjr * 3. Multibyte conversion state objects (mbstate_t) are passed around and 56132812Stjr * used for most, but not all, conversions. Further work will be required 57132812Stjr * to support state-dependent encodings. 58132812Stjr */ 59132812Stjr 601573Srgrimes#include <fnmatch.h> 61132812Stjr#include <limits.h> 621573Srgrimes#include <string.h> 63132812Stjr#include <wchar.h> 64132812Stjr#include <wctype.h> 651573Srgrimes 6619276Sache#include "collate.h" 6719276Sache 681573Srgrimes#define EOS '\0' 691573Srgrimes 7026484Sache#define RANGE_MATCH 1 7126484Sache#define RANGE_NOMATCH 0 7226484Sache#define RANGE_ERROR (-1) 731573Srgrimes 74132812Stjrstatic int rangematch(const char *, wchar_t, int, char **, mbstate_t *); 75206711Sjillesstatic int fnmatch1(const char *, const char *, const char *, int, mbstate_t, 76206711Sjilles mbstate_t); 7726484Sache 781573Srgrimesint 791573Srgrimesfnmatch(pattern, string, flags) 801573Srgrimes const char *pattern, *string; 811573Srgrimes int flags; 821573Srgrimes{ 83132812Stjr static const mbstate_t initial; 84132812Stjr 85206711Sjilles return (fnmatch1(pattern, string, string, flags, initial, initial)); 86132812Stjr} 87132812Stjr 88132812Stjrstatic int 89206711Sjillesfnmatch1(pattern, string, stringstart, flags, patmbs, strmbs) 90206711Sjilles const char *pattern, *string, *stringstart; 91132812Stjr int flags; 92132812Stjr mbstate_t patmbs, strmbs; 93132812Stjr{ 94289943Sjilles const char *bt_pattern, *bt_string; 95289943Sjilles mbstate_t bt_patmbs, bt_strmbs; 9626484Sache char *newp; 97132812Stjr char c; 98132812Stjr wchar_t pc, sc; 99132812Stjr size_t pclen, sclen; 1001573Srgrimes 101289943Sjilles bt_pattern = bt_string = NULL; 102206711Sjilles for (;;) { 103132812Stjr pclen = mbrtowc(&pc, pattern, MB_LEN_MAX, &patmbs); 104132812Stjr if (pclen == (size_t)-1 || pclen == (size_t)-2) 105132812Stjr return (FNM_NOMATCH); 106132812Stjr pattern += pclen; 107132812Stjr sclen = mbrtowc(&sc, string, MB_LEN_MAX, &strmbs); 108132812Stjr if (sclen == (size_t)-1 || sclen == (size_t)-2) { 109132812Stjr sc = (unsigned char)*string; 110132812Stjr sclen = 1; 111132812Stjr memset(&strmbs, 0, sizeof(strmbs)); 112132812Stjr } 113132812Stjr switch (pc) { 1141573Srgrimes case EOS: 115132812Stjr if ((flags & FNM_LEADING_DIR) && sc == '/') 11619132Sache return (0); 117289943Sjilles if (sc == EOS) 118289943Sjilles return (0); 119289943Sjilles goto backtrack; 1201573Srgrimes case '?': 121132812Stjr if (sc == EOS) 1221573Srgrimes return (FNM_NOMATCH); 123132812Stjr if (sc == '/' && (flags & FNM_PATHNAME)) 124289943Sjilles goto backtrack; 125132812Stjr if (sc == '.' && (flags & FNM_PERIOD) && 1261573Srgrimes (string == stringstart || 1271573Srgrimes ((flags & FNM_PATHNAME) && *(string - 1) == '/'))) 128289943Sjilles goto backtrack; 129132812Stjr string += sclen; 1301573Srgrimes break; 1311573Srgrimes case '*': 1321573Srgrimes c = *pattern; 1331573Srgrimes /* Collapse multiple stars. */ 1341573Srgrimes while (c == '*') 1351573Srgrimes c = *++pattern; 1361573Srgrimes 137132812Stjr if (sc == '.' && (flags & FNM_PERIOD) && 1381573Srgrimes (string == stringstart || 1391573Srgrimes ((flags & FNM_PATHNAME) && *(string - 1) == '/'))) 140289943Sjilles goto backtrack; 1411573Srgrimes 1421573Srgrimes /* Optimize for pattern with * at end or before /. */ 1431573Srgrimes if (c == EOS) 1441573Srgrimes if (flags & FNM_PATHNAME) 14525269Sjdp return ((flags & FNM_LEADING_DIR) || 14625269Sjdp strchr(string, '/') == NULL ? 1471573Srgrimes 0 : FNM_NOMATCH); 1481573Srgrimes else 1491573Srgrimes return (0); 1501573Srgrimes else if (c == '/' && flags & FNM_PATHNAME) { 1511573Srgrimes if ((string = strchr(string, '/')) == NULL) 1521573Srgrimes return (FNM_NOMATCH); 1531573Srgrimes break; 1541573Srgrimes } 1551573Srgrimes 156289943Sjilles /* 157289943Sjilles * First try the shortest match for the '*' that 158289943Sjilles * could work. We can forget any earlier '*' since 159289943Sjilles * there is no way having it match more characters 160289943Sjilles * can help us, given that we are already here. 161289943Sjilles */ 162289943Sjilles bt_pattern = pattern, bt_patmbs = patmbs; 163289943Sjilles bt_string = string, bt_strmbs = strmbs; 164289943Sjilles break; 1651573Srgrimes case '[': 166132812Stjr if (sc == EOS) 1671573Srgrimes return (FNM_NOMATCH); 168132812Stjr if (sc == '/' && (flags & FNM_PATHNAME)) 169289943Sjilles goto backtrack; 170132812Stjr if (sc == '.' && (flags & FNM_PERIOD) && 17126486Sache (string == stringstart || 17226486Sache ((flags & FNM_PATHNAME) && *(string - 1) == '/'))) 173289943Sjilles goto backtrack; 17426486Sache 175132812Stjr switch (rangematch(pattern, sc, flags, &newp, 176132812Stjr &patmbs)) { 17726484Sache case RANGE_ERROR: 17826484Sache goto norm; 17926484Sache case RANGE_MATCH: 18026484Sache pattern = newp; 18126484Sache break; 18226484Sache case RANGE_NOMATCH: 183289943Sjilles goto backtrack; 18426484Sache } 185132812Stjr string += sclen; 1861573Srgrimes break; 1871573Srgrimes case '\\': 1881573Srgrimes if (!(flags & FNM_NOESCAPE)) { 189132812Stjr pclen = mbrtowc(&pc, pattern, MB_LEN_MAX, 190132812Stjr &patmbs); 191132812Stjr if (pclen == (size_t)-1 || pclen == (size_t)-2) 192132812Stjr return (FNM_NOMATCH); 193132812Stjr pattern += pclen; 1941573Srgrimes } 1951573Srgrimes /* FALLTHROUGH */ 1961573Srgrimes default: 19726484Sache norm: 198289943Sjilles string += sclen; 199132812Stjr if (pc == sc) 20019059Swosch ; 20119132Sache else if ((flags & FNM_CASEFOLD) && 202132812Stjr (towlower(pc) == towlower(sc))) 20319059Swosch ; 204289943Sjilles else { 205289943Sjilles backtrack: 206289943Sjilles /* 207289943Sjilles * If we have a mismatch (other than hitting 208289943Sjilles * the end of the string), go back to the last 209289943Sjilles * '*' seen and have it match one additional 210289943Sjilles * character. 211289943Sjilles */ 212289943Sjilles if (bt_pattern == NULL) 213289943Sjilles return (FNM_NOMATCH); 214289943Sjilles sclen = mbrtowc(&sc, bt_string, MB_LEN_MAX, 215289943Sjilles &bt_strmbs); 216289943Sjilles if (sclen == (size_t)-1 || 217289943Sjilles sclen == (size_t)-2) { 218289943Sjilles sc = (unsigned char)*bt_string; 219289943Sjilles sclen = 1; 220289943Sjilles memset(&bt_strmbs, 0, 221289943Sjilles sizeof(bt_strmbs)); 222289943Sjilles } 223289943Sjilles if (sc == EOS) 224289943Sjilles return (FNM_NOMATCH); 225289943Sjilles if (sc == '/' && flags & FNM_PATHNAME) 226289943Sjilles return (FNM_NOMATCH); 227289943Sjilles bt_string += sclen; 228289943Sjilles pattern = bt_pattern, patmbs = bt_patmbs; 229289943Sjilles string = bt_string, strmbs = bt_strmbs; 230289943Sjilles } 2311573Srgrimes break; 2321573Srgrimes } 233132812Stjr } 2341573Srgrimes /* NOTREACHED */ 2351573Srgrimes} 2361573Srgrimes 23726484Sachestatic int 238132812Stjrrangematch(pattern, test, flags, newp, patmbs) 2391573Srgrimes const char *pattern; 240132812Stjr wchar_t test; 24119132Sache int flags; 24226484Sache char **newp; 243132812Stjr mbstate_t *patmbs; 2441573Srgrimes{ 24526492Sache int negate, ok; 246132812Stjr wchar_t c, c2; 247132812Stjr size_t pclen; 248132812Stjr const char *origpat; 249227753Stheraven struct xlocale_collate *table = 250227753Stheraven (struct xlocale_collate*)__get_locale()->components[XLC_COLLATE]; 2511573Srgrimes 2521573Srgrimes /* 2531573Srgrimes * A bracket expression starting with an unquoted circumflex 2541573Srgrimes * character produces unspecified results (IEEE 1003.2-1992, 2551573Srgrimes * 3.13.2). This implementation treats it like '!', for 2561573Srgrimes * consistency with the regular expression syntax. 2571573Srgrimes * J.T. Conklin (conklin@ngai.kaleida.com) 2581573Srgrimes */ 25926486Sache if ( (negate = (*pattern == '!' || *pattern == '^')) ) 2601573Srgrimes ++pattern; 2618870Srgrimes 26219132Sache if (flags & FNM_CASEFOLD) 263132812Stjr test = towlower(test); 26419059Swosch 26526484Sache /* 26626484Sache * A right bracket shall lose its special meaning and represent 26726484Sache * itself in a bracket expression if it occurs first in the list. 26826484Sache * -- POSIX.2 2.8.3.2 26926484Sache */ 27026492Sache ok = 0; 271132812Stjr origpat = pattern; 272132812Stjr for (;;) { 273132812Stjr if (*pattern == ']' && pattern > origpat) { 274132812Stjr pattern++; 275132812Stjr break; 276132812Stjr } else if (*pattern == '\0') { 27726484Sache return (RANGE_ERROR); 278132812Stjr } else if (*pattern == '/' && (flags & FNM_PATHNAME)) { 27926486Sache return (RANGE_NOMATCH); 280132812Stjr } else if (*pattern == '\\' && !(flags & FNM_NOESCAPE)) 281132812Stjr pattern++; 282132812Stjr pclen = mbrtowc(&c, pattern, MB_LEN_MAX, patmbs); 283132812Stjr if (pclen == (size_t)-1 || pclen == (size_t)-2) 284132812Stjr return (RANGE_NOMATCH); 285132812Stjr pattern += pclen; 28626486Sache 28719132Sache if (flags & FNM_CASEFOLD) 288132812Stjr c = towlower(c); 28919059Swosch 290132812Stjr if (*pattern == '-' && *(pattern + 1) != EOS && 291132812Stjr *(pattern + 1) != ']') { 292132812Stjr if (*++pattern == '\\' && !(flags & FNM_NOESCAPE)) 293132812Stjr if (*pattern != EOS) 294132812Stjr pattern++; 295132812Stjr pclen = mbrtowc(&c2, pattern, MB_LEN_MAX, patmbs); 296132812Stjr if (pclen == (size_t)-1 || pclen == (size_t)-2) 297132812Stjr return (RANGE_NOMATCH); 298132812Stjr pattern += pclen; 2991573Srgrimes if (c2 == EOS) 30026484Sache return (RANGE_ERROR); 30119059Swosch 30219132Sache if (flags & FNM_CASEFOLD) 303132812Stjr c2 = towlower(c2); 30419059Swosch 305227753Stheraven if (table->__collate_load_error ? 30624632Sache c <= test && test <= c2 : 307227753Stheraven __collate_range_cmp(table, c, test) <= 0 308227753Stheraven && __collate_range_cmp(table, test, c2) <= 0 30917533Sache ) 3101573Srgrimes ok = 1; 3111573Srgrimes } else if (c == test) 3121573Srgrimes ok = 1; 313132812Stjr } 31426492Sache 31526484Sache *newp = (char *)pattern; 31626484Sache return (ok == negate ? RANGE_NOMATCH : RANGE_MATCH); 3171573Srgrimes} 318