lib/glob/smatch.c

   1 /* strmatch.c -- ksh-like extended pattern matching for the shell and filename
   2                 globbing. */
   3
   4 /* Copyright (C) 1991-2005 Free Software Foundation, Inc.
   5
   6    This file is part of GNU Bash, the Bourne Again SHell.
   7
   8    Bash is free software; you can redistribute it and/or modify it under
   9    the terms of the GNU General Public License as published by the Free
  10    Software Foundation; either version 2, or (at your option) any later
  11    version.
  12
  13    Bash is distributed in the hope that it will be useful, but WITHOUT ANY
  14    WARRANTY; without even the implied warranty of MERCHANTABILITY or
  15    FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  16    for more details.
  17
  18    You should have received a copy of the GNU General Public License along
  19    with Bash; see the file COPYING.  If not, write to the Free Software
  20    Foundation, 59 Temple Place, Suite 330, Boston, MA 02111 USA. */
  21
  22 #include <config.h>
  23
  24 #include <stdio.h>      /* for debugging */
  25
  26 #include "strmatch.h"
  27 #include <chartypes.h>
  28
  29 #include "bashansi.h"
  30 #include "shmbutil.h"
  31 #include "xmalloc.h"
  32
  33 /* First, compile `sm_loop.c' for single-byte characters. */
  34 #define CHAR    unsigned char
  35 #define U_CHAR  unsigned char
  36 #define XCHAR   char
  37 #define INT     int
  38 #define L(CS)   CS
  39 #define INVALID -1
  40
  41 #undef STREQ
  42 #undef STREQN
  43 #define STREQ(a, b) ((a)[0] == (b)[0] && strcmp(a, b) == 0)
  44 #define STREQN(a, b, n) ((a)[0] == (b)[0] && strncmp(a, b, n) == 0)
  45
  46 /* We use strcoll(3) for range comparisons in bracket expressions,
  47    even though it can have unwanted side effects in locales
  48    other than POSIX or US.  For instance, in the de locale, [A-Z] matches
  49    all characters. */
  50
  51 #if defined (HAVE_STRCOLL)
  52 /* Helper function for collating symbol equivalence. */
  53 static int rangecmp (c1, c2)
  54      int c1, c2;
  55 {
  56   static char s1[2] = { ' ', '\0' };
  57   static char s2[2] = { ' ', '\0' };
  58   int ret;
  59
  60   /* Eight bits only.  Period. */
  61   c1 &= 0xFF;
  62   c2 &= 0xFF;
  63
  64   if (c1 == c2)
  65     return (0);
  66
  67   s1[0] = c1;
  68   s2[0] = c2;
  69
  70   if ((ret = strcoll (s1, s2)) != 0)
  71     return ret;
  72   return (c1 - c2);
  73 }
  74 #else /* !HAVE_STRCOLL */
  75 #  define rangecmp(c1, c2)      ((int)(c1) - (int)(c2))
  76 #endif /* !HAVE_STRCOLL */
  77
  78 #if defined (HAVE_STRCOLL)
  79 static int
  80 collequiv (c1, c2)
  81      int c1, c2;
  82 {
  83   return (rangecmp (c1, c2) == 0);
  84 }
  85 #else
  86 #  define collequiv(c1, c2)     ((c1) == (c2))
  87 #endif
  88
  89 #define _COLLSYM        _collsym
  90 #define __COLLSYM       __collsym
  91 #define POSIXCOLL       posix_collsyms
  92 #include "collsyms.h"
  93
  94 static int
  95 collsym (s, len)
  96      CHAR *s;
  97      int len;
  98 {
  99   register struct _collsym *csp;
 100   char *x;
 101
 102   x = (char *)s;
 103   for (csp = posix_collsyms; csp->name; csp++)
 104     {
 105       if (STREQN(csp->name, x, len) && csp->name[len] == '\0')
 106         return (csp->code);
 107     }
 108   if (len == 1)
 109     return s[0];
 110   return INVALID;
 111 }
 112
 113 /* unibyte character classification */
 114 #if !defined (isascii) && !defined (HAVE_ISASCII)
 115 #  define isascii(c)    ((unsigned int)(c) <= 0177)
 116 #endif
 117
 118 enum char_class
 119   {
 120     CC_NO_CLASS = 0,
 121     CC_ASCII, CC_ALNUM, CC_ALPHA, CC_BLANK, CC_CNTRL, CC_DIGIT, CC_GRAPH,
 122     CC_LOWER, CC_PRINT, CC_PUNCT, CC_SPACE, CC_UPPER, CC_WORD, CC_XDIGIT
 123   };
 124
 125 static char const *const cclass_name[] =
 126   {
 127     "",
 128     "ascii", "alnum", "alpha", "blank", "cntrl", "digit", "graph",
 129     "lower", "print", "punct", "space", "upper", "word", "xdigit"
 130   };
 131
 132 #define N_CHAR_CLASS (sizeof(cclass_name) / sizeof (cclass_name[0]))
 133
 134 static int
 135 is_cclass (c, name)
 136      int c;
 137      const char *name;
 138 {
 139   enum char_class char_class = CC_NO_CLASS;
 140   int i, result;
 141
 142   for (i = 1; i < N_CHAR_CLASS; i++)
 143     {
 144       if (STREQ (name, cclass_name[i]))
 145         {
 146           char_class = (enum char_class)i;
 147           break;
 148         }
 149     }
 150
 151   if (char_class == 0)
 152     return -1;
 153
 154   switch (char_class)
 155     {
 156       case CC_ASCII:
 157         result = isascii (c);
 158         break;
 159       case CC_ALNUM:
 160         result = ISALNUM (c);
 161         break;
 162       case CC_ALPHA:
 163         result = ISALPHA (c);
 164         break;
 165       case CC_BLANK:
 166         result = ISBLANK (c);
 167         break;
 168       case CC_CNTRL:
 169         result = ISCNTRL (c);
 170         break;
 171       case CC_DIGIT:
 172         result = ISDIGIT (c);
 173         break;
 174       case CC_GRAPH:
 175         result = ISGRAPH (c);
 176         break;
 177       case CC_LOWER:
 178         result = ISLOWER (c);
 179         break;
 180       case CC_PRINT:
 181         result = ISPRINT (c);
 182         break;
 183       case CC_PUNCT:
 184         result = ISPUNCT (c);
 185         break;
 186       case CC_SPACE:
 187         result = ISSPACE (c);
 188         break;
 189       case CC_UPPER:
 190         result = ISUPPER (c);
 191         break;
 192       case CC_WORD:
 193         result = (ISALNUM (c) || c == '_');
 194         break;
 195       case CC_XDIGIT:
 196         result = ISXDIGIT (c);
 197         break;
 198       default:
 199         result = -1;
 200         break;
 201     }
 202
 203   return result;
 204 }
 205
 206 /* Now include `sm_loop.c' for single-byte characters. */
 207 /* The result of FOLD is an `unsigned char' */
 208 # define FOLD(c) ((flags & FNM_CASEFOLD) \
 209         ? TOLOWER ((unsigned char)c) \
 210         : ((unsigned char)c))
 211
 212 #define FCT                     internal_strmatch
 213 #define GMATCH                  gmatch
 214 #define COLLSYM                 collsym
 215 #define PARSE_COLLSYM           parse_collsym
 216 #define BRACKMATCH              brackmatch
 217 #define PATSCAN                 patscan
 218 #define STRCOMPARE              strcompare
 219 #define EXTMATCH                extmatch
 220 #define STRCHR(S, C)            strchr((S), (C))
 221 #define STRCOLL(S1, S2)         strcoll((S1), (S2))
 222 #define STRLEN(S)               strlen(S)
 223 #define STRCMP(S1, S2)          strcmp((S1), (S2))
 224 #define RANGECMP(C1, C2)        rangecmp((C1), (C2))
 225 #define COLLEQUIV(C1, C2)       collequiv((C1), (C2))
 226 #define CTYPE_T                 enum char_class
 227 #define IS_CCLASS(C, S)         is_cclass((C), (S))
 228 #include "sm_loop.c"
 229
 230 #if HANDLE_MULTIBYTE
 231
 232 #  define CHAR          wchar_t
 233 #  define U_CHAR        wint_t
 234 #  define XCHAR         wchar_t
 235 #  define INT           wint_t
 236 #  define L(CS)         L##CS
 237 #  define INVALID       WEOF
 238
 239 #  undef STREQ
 240 #  undef STREQN
 241 #  define STREQ(s1, s2) ((wcscmp (s1, s2) == 0))
 242 #  define STREQN(a, b, n) ((a)[0] == (b)[0] && wcsncmp(a, b, n) == 0)
 243
 244 static int
 245 rangecmp_wc (c1, c2)
 246      wint_t c1, c2;
 247 {
 248   static wchar_t s1[2] = { L' ', L'\0' };
 249   static wchar_t s2[2] = { L' ', L'\0' };
 250
 251   if (c1 == c2)
 252     return 0;
 253
 254   s1[0] = c1;
 255   s2[0] = c2;
 256
 257   return (wcscoll (s1, s2));
 258 }
 259
 260 static int
 261 collequiv_wc (c, equiv)
 262      wint_t c, equiv;
 263 {
 264   return (!(c - equiv));
 265 }
 266
 267 /* Helper function for collating symbol. */
 268 #  define _COLLSYM      _collwcsym
 269 #  define __COLLSYM     __collwcsym
 270 #  define POSIXCOLL     posix_collwcsyms
 271 #  include "collsyms.h"
 272
 273 static wint_t
 274 collwcsym (s, len)
 275      wchar_t *s;
 276      int len;
 277 {
 278   register struct _collwcsym *csp;
 279
 280   for (csp = posix_collwcsyms; csp->name; csp++)
 281     {
 282       if (STREQN(csp->name, s, len) && csp->name[len] == L'\0')
 283         return (csp->code);
 284     }
 285   if (len == 1)
 286     return s[0];
 287   return INVALID;
 288 }
 289
 290 static int
 291 is_wcclass (wc, name)
 292      wint_t wc;
 293      wchar_t *name;
 294 {
 295   char *mbs;
 296   mbstate_t state;
 297   size_t mbslength;
 298   wctype_t desc;
 299   int want_word;
 300
 301   if ((wctype ("ascii") == (wctype_t)0) && (wcscmp (name, L"ascii") == 0))
 302     {
 303       int c;
 304
 305       if ((c = wctob (wc)) == EOF)
 306         return 0;
 307       else
 308         return (c <= 0x7F);
 309     }
 310
 311   want_word = (wcscmp (name, L"word") == 0);
 312   if (want_word)
 313     name = L"alnum";
 314
 315   memset (&state, '\0', sizeof (mbstate_t));
 316   mbs = (char *) malloc (wcslen(name) * MB_CUR_MAX + 1);
 317   mbslength = wcsrtombs(mbs, (const wchar_t **)&name, (wcslen(name) * MB_CUR_MAX + 1), &state);
 318
 319   if (mbslength == (size_t)-1 || mbslength == (size_t)-2)
 320     {
 321       free (mbs);
 322       return -1;
 323     }
 324   desc = wctype (mbs);
 325   free (mbs);
 326
 327   if (desc == (wctype_t)0)
 328     return -1;
 329
 330   if (want_word)
 331     return (iswctype (wc, desc) || wc == L'_');
 332   else
 333     return (iswctype (wc, desc));
 334 }
 335
 336 /* Now include `sm_loop.c' for multibyte characters. */
 337 #define FOLD(c) ((flags & FNM_CASEFOLD) && iswupper (c) ? towlower (c) : (c))
 338 #define FCT                     internal_wstrmatch
 339 #define GMATCH                  gmatch_wc
 340 #define COLLSYM                 collwcsym
 341 #define PARSE_COLLSYM           parse_collwcsym
 342 #define BRACKMATCH              brackmatch_wc
 343 #define PATSCAN                 patscan_wc
 344 #define STRCOMPARE              wscompare
 345 #define EXTMATCH                extmatch_wc
 346 #define STRCHR(S, C)            wcschr((S), (C))
 347 #define STRCOLL(S1, S2)         wcscoll((S1), (S2))
 348 #define STRLEN(S)               wcslen(S)
 349 #define STRCMP(S1, S2)          wcscmp((S1), (S2))
 350 #define RANGECMP(C1, C2)        rangecmp_wc((C1), (C2))
 351 #define COLLEQUIV(C1, C2)       collequiv_wc((C1), (C2))
 352 #define CTYPE_T                 enum char_class
 353 #define IS_CCLASS(C, S)         is_wcclass((C), (S))
 354 #include "sm_loop.c"
 355
 356 #endif /* HAVE_MULTIBYTE */
 357
 358 int
 359 xstrmatch (pattern, string, flags)
 360      char *pattern;
 361      char *string;
 362      int flags;
 363 {
 364 #if HANDLE_MULTIBYTE
 365   int ret;
 366   size_t n;
 367   wchar_t *wpattern, *wstring;
 368
 369   if (MB_CUR_MAX == 1)
 370     return (internal_strmatch ((unsigned char *)pattern, (unsigned char *)string, flags));
 371
 372   n = xdupmbstowcs (&wpattern, NULL, pattern);
 373   if (n == (size_t)-1 || n == (size_t)-2)
 374     return (internal_strmatch ((unsigned char *)pattern, (unsigned char *)string, flags));
 375
 376   n = xdupmbstowcs (&wstring, NULL, string);
 377   if (n == (size_t)-1 || n == (size_t)-2)
 378     {
 379       free (wpattern);
 380       return (internal_strmatch ((unsigned char *)pattern, (unsigned char *)string, flags));
 381     }
 382
 383   ret = internal_wstrmatch (wpattern, wstring, flags);
 384
 385   free (wpattern);
 386   free (wstring);
 387
 388   return ret;
 389 #else
 390   return (internal_strmatch ((unsigned char *)pattern, (unsigned char *)string, flags));
 391 #endif /* !HANDLE_MULTIBYTE */
 392 }