posix/tst-regex.c

   1 /* Copyright (C) 2001-2021 Free Software Foundation, Inc.
   2    This file is part of the GNU C Library.
   3
   4    The GNU C Library is free software; you can redistribute it and/or
   5    modify it under the terms of the GNU Lesser General Public
   6    License as published by the Free Software Foundation; either
   7    version 2.1 of the License, or (at your option) any later version.
   8
   9    The GNU C Library is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12    Lesser General Public License for more details.
  13
  14    You should have received a copy of the GNU Lesser General Public
  15    License along with the GNU C Library; if not, see
  16    <https://www.gnu.org/licenses/>.  */
  17
  18 #include <assert.h>
  19 #include <errno.h>
  20 #include <error.h>
  21 #include <fcntl.h>
  22 #include <getopt.h>
  23 #include <iconv.h>
  24 #include <locale.h>
  25 #include <mcheck.h>
  26 #include <stdint.h>
  27 #include <stdio.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30 #include <time.h>
  31 #include <unistd.h>
  32 #include <sys/stat.h>
  33 #include <sys/types.h>
  34 #include <regex.h>
  35
  36
  37 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
  38 static clockid_t cl;
  39 static int use_clock;
  40 #endif
  41 static iconv_t cd;
  42 static char *mem;
  43 static char *umem;
  44 static size_t memlen;
  45 static size_t umemlen;
  46 static int timing;
  47
  48 static int test_expr (const char *expr, int expected, int expectedicase);
  49 static int run_test (const char *expr, const char *mem, size_t memlen,
  50                      int icase, int expected);
  51 static int run_test_backwards (const char *expr, const char *mem,
  52                                size_t memlen, int icase, int expected);
  53
  54
  55 static int
  56 do_test (void)
  57 {
  58   const char *file;
  59   int fd;
  60   struct stat st;
  61   int result;
  62   char *inmem;
  63   char *outmem;
  64   size_t inlen;
  65   size_t outlen;
  66
  67   mtrace ();
  68
  69   /* Make the content of the file available in memory.  */
  70   file = "./tst-regex.input";
  71   fd = open (file, O_RDONLY);
  72   if (fd == -1)
  73     error (EXIT_FAILURE, errno, "cannot open %s", basename (file));
  74
  75   if (fstat (fd, &st) != 0)
  76     error (EXIT_FAILURE, errno, "cannot stat %s", basename (file));
  77   memlen = st.st_size;
  78
  79   mem = (char *) malloc (memlen + 1);
  80   if (mem == NULL)
  81     error (EXIT_FAILURE, errno, "while allocating buffer");
  82
  83   if ((size_t) read (fd, mem, memlen) != memlen)
  84     error (EXIT_FAILURE, 0, "cannot read entire file");
  85   mem[memlen] = '\0';
  86
  87   close (fd);
  88
  89   /* We have to convert a few things from UTF-8 to Latin-1.  */
  90   cd = iconv_open ("ISO-8859-1", "UTF-8");
  91   if (cd == (iconv_t) -1)
  92     error (EXIT_FAILURE, errno, "cannot get conversion descriptor");
  93
  94   /* For the second test we have to convert the file content to Latin-1.
  95      This cannot grow the data.  */
  96   umem = (char *) malloc (memlen + 1);
  97   if (umem == NULL)
  98     error (EXIT_FAILURE, errno, "while allocating buffer");
  99
 100   inmem = mem;
 101   inlen = memlen;
 102   outmem = umem;
 103   outlen = memlen;
 104   iconv (cd, &inmem, &inlen, &outmem, &outlen);
 105   umemlen = outmem - umem;
 106   if (inlen != 0)
 107     error (EXIT_FAILURE, errno, "cannot convert buffer");
 108   umem[umemlen] = '\0';
 109
 110 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 111 # if _POSIX_CPUTIME == 0
 112   if (sysconf (_SC_CPUTIME) < 0)
 113     use_clock = 0;
 114   else
 115 # endif
 116     /* See whether we can use the CPU clock.  */
 117     use_clock = clock_getcpuclockid (0, &cl) == 0;
 118 #endif
 119
 120 #ifdef DEBUG
 121   re_set_syntax (RE_DEBUG);
 122 #endif
 123
 124   /* Run the actual tests.  All tests are run in a single-byte and a
 125      multi-byte locale.  */
 126   result = test_expr ("[äáàâéèêíìîñöóòôüúùû]", 4, 4);
 127   result |= test_expr ("G.ran", 2, 3);
 128   result |= test_expr ("G.\\{1\\}ran", 2, 3);
 129   result |= test_expr ("G.*ran", 3, 44);
 130   result |= test_expr ("[äáàâ]", 0, 0);
 131   result |= test_expr ("Uddeborg", 2, 2);
 132   result |= test_expr (".Uddeborg", 2, 2);
 133
 134   /* Free the resources.  */
 135   free (umem);
 136   iconv_close (cd);
 137   free (mem);
 138
 139   return result;
 140 }
 141
 142
 143 static int
 144 test_expr (const char *expr, int expected, int expectedicase)
 145 {
 146   int result;
 147   char *inmem;
 148   char *outmem;
 149   size_t inlen;
 150   size_t outlen;
 151   char *uexpr;
 152
 153   /* First test: search with an UTF-8 locale.  */
 154   if (setlocale (LC_ALL, "de_DE.UTF-8") == NULL)
 155     error (EXIT_FAILURE, 0, "cannot set locale de_DE.UTF-8");
 156
 157   printf ("\nTest \"%s\" with multi-byte locale\n", expr);
 158   result = run_test (expr, mem, memlen, 0, expected);
 159   printf ("\nTest \"%s\" with multi-byte locale, case insensitive\n", expr);
 160   result |= run_test (expr, mem, memlen, 1, expectedicase);
 161   printf ("\nTest \"%s\" backwards with multi-byte locale\n", expr);
 162   result |= run_test_backwards (expr, mem, memlen, 0, expected);
 163   printf ("\nTest \"%s\" backwards with multi-byte locale, case insensitive\n",
 164           expr);
 165   result |= run_test_backwards (expr, mem, memlen, 1, expectedicase);
 166
 167   /* Second test: search with an ISO-8859-1 locale.  */
 168   if (setlocale (LC_ALL, "de_DE.ISO-8859-1") == NULL)
 169     error (EXIT_FAILURE, 0, "cannot set locale de_DE.ISO-8859-1");
 170
 171   inmem = (char *) expr;
 172   inlen = strlen (expr);
 173   outlen = inlen;
 174   outmem = uexpr = alloca (outlen + 1);
 175   memset (outmem, '\0', outlen + 1);
 176   iconv (cd, &inmem, &inlen, &outmem, &outlen);
 177   if (inlen != 0)
 178     error (EXIT_FAILURE, errno, "cannot convert expression");
 179
 180   /* Run the tests.  */
 181   printf ("\nTest \"%s\" with 8-bit locale\n", expr);
 182   result |= run_test (uexpr, umem, umemlen, 0, expected);
 183   printf ("\nTest \"%s\" with 8-bit locale, case insensitive\n", expr);
 184   result |= run_test (uexpr, umem, umemlen, 1, expectedicase);
 185   printf ("\nTest \"%s\" backwards with 8-bit locale\n", expr);
 186   result |= run_test_backwards (uexpr, umem, umemlen, 0, expected);
 187   printf ("\nTest \"%s\" backwards with 8-bit locale, case insensitive\n",
 188           expr);
 189   result |= run_test_backwards (uexpr, umem, umemlen, 1, expectedicase);
 190
 191   return result;
 192 }
 193
 194
 195 static int
 196 run_test (const char *expr, const char *mem, size_t memlen, int icase,
 197           int expected)
 198 {
 199 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 200   struct timespec start;
 201   struct timespec finish;
 202 #endif
 203   regex_t re;
 204   int err;
 205   size_t offset;
 206   int cnt;
 207
 208 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 209   if (use_clock && !timing)
 210     use_clock = clock_gettime (cl, &start) == 0;
 211 #endif
 212
 213   err = regcomp (&re, expr, REG_NEWLINE | (icase ? REG_ICASE : 0));
 214   if (err != REG_NOERROR)
 215     {
 216       char buf[200];
 217       regerror (err, &re, buf, sizeof buf);
 218       error (EXIT_FAILURE, 0, "cannot compile expression: %s", buf);
 219     }
 220
 221   cnt = 0;
 222   offset = 0;
 223   assert (mem[memlen] == '\0');
 224   while (offset < memlen)
 225     {
 226       regmatch_t ma[1];
 227       const char *sp;
 228       const char *ep;
 229
 230       err = regexec (&re, mem + offset, 1, ma, 0);
 231       if (err == REG_NOMATCH)
 232         break;
 233
 234       if (err != REG_NOERROR)
 235         {
 236           char buf[200];
 237           regerror (err, &re, buf, sizeof buf);
 238           error (EXIT_FAILURE, 0, "cannot use expression: %s", buf);
 239         }
 240
 241       assert (ma[0].rm_so >= 0);
 242       sp = mem + offset + ma[0].rm_so;
 243       while (sp > mem && sp[-1] != '\n')
 244         --sp;
 245
 246       ep = mem + offset + ma[0].rm_so;
 247       while (*ep != '\0' && *ep != '\n')
 248         ++ep;
 249
 250       printf ("match %d: \"%.*s\"\n", ++cnt, (int) (ep - sp), sp);
 251
 252       offset = ep + 1 - mem;
 253     }
 254
 255   regfree (&re);
 256
 257 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 258   if (use_clock && !timing)
 259     {
 260       use_clock = clock_gettime (cl, &finish) == 0;
 261       if (use_clock)
 262         {
 263           if (finish.tv_nsec < start.tv_nsec)
 264             {
 265               finish.tv_nsec -= start.tv_nsec - 1000000000;
 266               finish.tv_sec -= 1 + start.tv_sec;
 267             }
 268           else
 269             {
 270               finish.tv_nsec -= start.tv_nsec;
 271               finish.tv_sec -= start.tv_sec;
 272             }
 273
 274           printf ("elapsed time: %jd.%09jd sec\n",
 275                   (intmax_t) finish.tv_sec, (intmax_t) finish.tv_nsec);
 276         }
 277     }
 278
 279   if (use_clock && timing)
 280     {
 281       struct timespec mintime = { .tv_sec = 24 * 60 * 60 };
 282
 283       for (int i = 0; i < 10; ++i)
 284         {
 285           offset = 0;
 286           use_clock = clock_gettime (cl, &start) == 0;
 287
 288           if (!use_clock)
 289             continue;
 290
 291           err = regcomp (&re, expr, REG_NEWLINE | (icase ? REG_ICASE : 0));
 292           if (err != REG_NOERROR)
 293             continue;
 294
 295           while (offset < memlen)
 296             {
 297               regmatch_t ma[1];
 298
 299               err = regexec (&re, mem + offset, 1, ma, 0);
 300               if (err != REG_NOERROR)
 301                 break;
 302
 303               offset += ma[0].rm_eo;
 304             }
 305
 306           regfree (&re);
 307
 308           use_clock = clock_gettime (cl, &finish) == 0;
 309           if (use_clock)
 310             {
 311               if (finish.tv_nsec < start.tv_nsec)
 312                 {
 313                   finish.tv_nsec -= start.tv_nsec - 1000000000;
 314                   finish.tv_sec -= 1 + start.tv_sec;
 315                 }
 316               else
 317                 {
 318                   finish.tv_nsec -= start.tv_nsec;
 319                   finish.tv_sec -= start.tv_sec;
 320                 }
 321               if (finish.tv_sec < mintime.tv_sec
 322                   || (finish.tv_sec == mintime.tv_sec
 323                       && finish.tv_nsec < mintime.tv_nsec))
 324                 mintime = finish;
 325             }
 326         }
 327       printf ("elapsed time: %jd.%09jd sec\n",
 328               (intmax_t) mintime.tv_sec, (intmax_t) mintime.tv_nsec);
 329     }
 330 #endif
 331
 332   /* Return an error if the number of matches found is not match we
 333      expect.  */
 334   return cnt != expected;
 335 }
 336
 337
 338 static int
 339 run_test_backwards (const char *expr, const char *mem, size_t memlen,
 340                     int icase, int expected)
 341 {
 342 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 343   struct timespec start;
 344   struct timespec finish;
 345 #endif
 346   struct re_pattern_buffer re;
 347   const char *err;
 348   size_t offset;
 349   int cnt;
 350
 351 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 352   if (use_clock && !timing)
 353     use_clock = clock_gettime (cl, &start) == 0;
 354 #endif
 355
 356   re_set_syntax ((RE_SYNTAX_POSIX_BASIC & ~RE_DOT_NEWLINE)
 357                  | RE_HAT_LISTS_NOT_NEWLINE
 358                  | (icase ? RE_ICASE : 0));
 359
 360   memset (&re, 0, sizeof (re));
 361   re.fastmap = malloc (256);
 362   if (re.fastmap == NULL)
 363     error (EXIT_FAILURE, errno, "cannot allocate fastmap");
 364
 365   err = re_compile_pattern (expr, strlen (expr), &re);
 366   if (err != NULL)
 367     error (EXIT_FAILURE, 0, "cannot compile expression: %s", err);
 368
 369   if (re_compile_fastmap (&re))
 370     error (EXIT_FAILURE, 0, "couldn't compile fastmap");
 371
 372   cnt = 0;
 373   offset = memlen;
 374   assert (mem[memlen] == '\0');
 375   while (offset <= memlen)
 376     {
 377       int start;
 378       const char *sp;
 379       const char *ep;
 380
 381       start = re_search (&re, mem, memlen, offset, -offset, NULL);
 382       if (start == -1)
 383         break;
 384
 385       if (start == -2)
 386         error (EXIT_FAILURE, 0, "internal error in re_search");
 387
 388       sp = mem + start;
 389       while (sp > mem && sp[-1] != '\n')
 390         --sp;
 391
 392       ep = mem + start;
 393       while (*ep != '\0' && *ep != '\n')
 394         ++ep;
 395
 396       printf ("match %d: \"%.*s\"\n", ++cnt, (int) (ep - sp), sp);
 397
 398       offset = sp - 1 - mem;
 399     }
 400
 401   regfree (&re);
 402
 403 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 404   if (use_clock && !timing)
 405     {
 406       use_clock = clock_gettime (cl, &finish) == 0;
 407       if (use_clock)
 408         {
 409           if (finish.tv_nsec < start.tv_nsec)
 410             {
 411               finish.tv_nsec -= start.tv_nsec - 1000000000;
 412               finish.tv_sec -= 1 + start.tv_sec;
 413             }
 414           else
 415             {
 416               finish.tv_nsec -= start.tv_nsec;
 417               finish.tv_sec -= start.tv_sec;
 418             }
 419
 420           printf ("elapsed time: %jd.%09jd sec\n",
 421                   (intmax_t) finish.tv_sec, (intmax_t) finish.tv_nsec);
 422         }
 423     }
 424
 425   if (use_clock && timing)
 426     {
 427       struct timespec mintime = { .tv_sec = 24 * 60 * 60 };
 428
 429       for (int i = 0; i < 10; ++i)
 430         {
 431           offset = memlen;
 432           use_clock = clock_gettime (cl, &start) == 0;
 433
 434           if (!use_clock)
 435             continue;
 436
 437           memset (&re, 0, sizeof (re));
 438           re.fastmap = malloc (256);
 439           if (re.fastmap == NULL)
 440             continue;
 441
 442           err = re_compile_pattern (expr, strlen (expr), &re);
 443           if (err != NULL)
 444             continue;
 445
 446           if (re_compile_fastmap (&re))
 447             {
 448               regfree (&re);
 449               continue;
 450             }
 451
 452           while (offset <= memlen)
 453             {
 454               int start;
 455               const char *sp;
 456
 457               start = re_search (&re, mem, memlen, offset, -offset, NULL);
 458               if (start < -1)
 459                 break;
 460
 461               sp = mem + start;
 462               while (sp > mem && sp[-1] != '\n')
 463                 --sp;
 464
 465               offset = sp - 1 - mem;
 466             }
 467
 468           regfree (&re);
 469
 470           use_clock = clock_gettime (cl, &finish) == 0;
 471           if (use_clock)
 472             {
 473               if (finish.tv_nsec < start.tv_nsec)
 474                 {
 475                   finish.tv_nsec -= start.tv_nsec - 1000000000;
 476                   finish.tv_sec -= 1 + start.tv_sec;
 477                 }
 478               else
 479                 {
 480                   finish.tv_nsec -= start.tv_nsec;
 481                   finish.tv_sec -= start.tv_sec;
 482                 }
 483               if (finish.tv_sec < mintime.tv_sec
 484                   || (finish.tv_sec == mintime.tv_sec
 485                       && finish.tv_nsec < mintime.tv_nsec))
 486                 mintime = finish;
 487             }
 488         }
 489       printf ("elapsed time: %jd.%09jd sec\n",
 490               (intmax_t) mintime.tv_sec, (intmax_t) mintime.tv_nsec);
 491     }
 492 #endif
 493
 494   /* Return an error if the number of matches found is not match we
 495      expect.  */
 496   return cnt != expected;
 497 }
 498
 499 /* If --timing is used we will need a larger timout.  */
 500 #define TIMEOUT 50
 501 #define CMDLINE_OPTIONS \
 502    {"timing", no_argument, &timing, 1 },
 503 #define TEST_FUNCTION do_test ()
 504 #include "../test-skeleton.c"