posix/tst-regex.c

   1 /* Copyright (C) 2001-2015 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    <http://www.gnu.org/licenses/>.  */
  17
  18 #include <spawn.h>
  19 #include "spawn_int.h"
  20
  21 #include <assert.h>
  22 #include <errno.h>
  23 #include <error.h>
  24 #include <fcntl.h>
  25 #include <getopt.h>
  26 #include <iconv.h>
  27 #include <locale.h>
  28 #include <mcheck.h>
  29 #include <stdio.h>
  30 #include <stdlib.h>
  31 #include <string.h>
  32 #include <time.h>
  33 #include <unistd.h>
  34 #include <sys/stat.h>
  35 #include <sys/types.h>
  36 #include <regex.h>
  37
  38
  39 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
  40 static clockid_t cl;
  41 static int use_clock;
  42 #endif
  43 static iconv_t cd;
  44 static char *mem;
  45 static char *umem;
  46 static size_t memlen;
  47 static size_t umemlen;
  48 static int timing;
  49
  50 static int test_expr (const char *expr, int expected, int expectedicase);
  51 static int run_test (const char *expr, const char *mem, size_t memlen,
  52                      int icase, int expected);
  53 static int run_test_backwards (const char *expr, const char *mem,
  54                                size_t memlen, int icase, int expected);
  55
  56
  57 static int
  58 do_test (void)
  59 {
  60   const char *file;
  61   int fd;
  62   struct stat st;
  63   int result;
  64   char *inmem;
  65   char *outmem;
  66   size_t inlen;
  67   size_t outlen;
  68
  69   mtrace ();
  70
  71   /* Make the content of the file available in memory.  */
  72   file = "../ChangeLog.8";
  73   fd = open (file, O_RDONLY);
  74   if (fd == -1)
  75     error (EXIT_FAILURE, errno, "cannot open %s", basename (file));
  76
  77   if (fstat (fd, &st) != 0)
  78     error (EXIT_FAILURE, errno, "cannot stat %s", basename (file));
  79   memlen = st.st_size;
  80
  81   mem = (char *) malloc (memlen + 1);
  82   if (mem == NULL)
  83     error (EXIT_FAILURE, errno, "while allocating buffer");
  84
  85   if ((size_t) read (fd, mem, memlen) != memlen)
  86     error (EXIT_FAILURE, 0, "cannot read entire file");
  87   mem[memlen] = '\0';
  88
  89   close (fd);
  90
  91   /* We have to convert a few things from Latin-1 to UTF-8.  */
  92   cd = iconv_open ("UTF-8", "ISO-8859-1");
  93   if (cd == (iconv_t) -1)
  94     error (EXIT_FAILURE, errno, "cannot get conversion descriptor");
  95
  96   /* For the second test we have to convert the file content to UTF-8.
  97      Since the text is mostly ASCII it should be enough to allocate
  98      twice as much memory for the UTF-8 text than for the Latin-1
  99      text.  */
 100   umem = (char *) calloc (2, memlen);
 101   if (umem == NULL)
 102     error (EXIT_FAILURE, errno, "while allocating buffer");
 103
 104   inmem = mem;
 105   inlen = memlen;
 106   outmem = umem;
 107   outlen = 2 * memlen - 1;
 108   iconv (cd, &inmem, &inlen, &outmem, &outlen);
 109   umemlen = outmem - umem;
 110   if (inlen != 0)
 111     error (EXIT_FAILURE, errno, "cannot convert buffer");
 112
 113 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 114 # if _POSIX_CPUTIME == 0
 115   if (sysconf (_SC_CPUTIME) < 0)
 116     use_clock = 0;
 117   else
 118 # endif
 119     /* See whether we can use the CPU clock.  */
 120     use_clock = clock_getcpuclockid (0, &cl) == 0;
 121 #endif
 122
 123 #ifdef DEBUG
 124   re_set_syntax (RE_DEBUG);
 125 #endif
 126
 127   /* Run the actual tests.  All tests are run in a single-byte and a
 128      multi-byte locale.  */
 129   result = test_expr ("[äáàâéèêíìîñöóòôüúùû]", 2, 2);
 130   result |= test_expr ("G.ran", 2, 3);
 131   result |= test_expr ("G.\\{1\\}ran", 2, 3);
 132   result |= test_expr ("G.*ran", 3, 44);
 133   result |= test_expr ("[äáàâ]", 0, 0);
 134   result |= test_expr ("Uddeborg", 2, 2);
 135   result |= test_expr (".Uddeborg", 2, 2);
 136
 137   /* Free the resources.  */
 138   free (umem);
 139   iconv_close (cd);
 140   free (mem);
 141
 142   return result;
 143 }
 144
 145
 146 static int
 147 test_expr (const char *expr, int expected, int expectedicase)
 148 {
 149   int result;
 150   char *inmem;
 151   char *outmem;
 152   size_t inlen;
 153   size_t outlen;
 154   char *uexpr;
 155
 156   /* First test: search with an ISO-8859-1 locale.  */
 157   if (setlocale (LC_ALL, "de_DE.ISO-8859-1") == NULL)
 158     error (EXIT_FAILURE, 0, "cannot set locale de_DE.ISO-8859-1");
 159
 160   printf ("\nTest \"%s\" with 8-bit locale\n", expr);
 161   result = run_test (expr, mem, memlen, 0, expected);
 162   printf ("\nTest \"%s\" with 8-bit locale, case insensitive\n", expr);
 163   result |= run_test (expr, mem, memlen, 1, expectedicase);
 164   printf ("\nTest \"%s\" backwards with 8-bit locale\n", expr);
 165   result |= run_test_backwards (expr, mem, memlen, 0, expected);
 166   printf ("\nTest \"%s\" backwards with 8-bit locale, case insensitive\n",
 167           expr);
 168   result |= run_test_backwards (expr, mem, memlen, 1, expectedicase);
 169
 170   /* Second test: search with an UTF-8 locale.  */
 171   if (setlocale (LC_ALL, "de_DE.UTF-8") == NULL)
 172     error (EXIT_FAILURE, 0, "cannot set locale de_DE.UTF-8");
 173
 174   inmem = (char *) expr;
 175   inlen = strlen (expr);
 176   outlen = inlen * MB_CUR_MAX;
 177   outmem = uexpr = alloca (outlen + 1);
 178   memset (outmem, '\0', outlen + 1);
 179   iconv (cd, &inmem, &inlen, &outmem, &outlen);
 180   if (inlen != 0)
 181     error (EXIT_FAILURE, errno, "cannot convert expression");
 182
 183   /* Run the tests.  */
 184   printf ("\nTest \"%s\" with multi-byte locale\n", expr);
 185   result |= run_test (uexpr, umem, umemlen, 0, expected);
 186   printf ("\nTest \"%s\" with multi-byte locale, case insensitive\n", expr);
 187   result |= run_test (uexpr, umem, umemlen, 1, expectedicase);
 188   printf ("\nTest \"%s\" backwards with multi-byte locale\n", expr);
 189   result |= run_test_backwards (uexpr, umem, umemlen, 0, expected);
 190   printf ("\nTest \"%s\" backwards with multi-byte locale, case insensitive\n",
 191           expr);
 192   result |= run_test_backwards (uexpr, umem, umemlen, 1, expectedicase);
 193
 194   return result;
 195 }
 196
 197
 198 static int
 199 run_test (const char *expr, const char *mem, size_t memlen, int icase,
 200           int expected)
 201 {
 202 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 203   struct timespec start;
 204   struct timespec finish;
 205 #endif
 206   regex_t re;
 207   int err;
 208   size_t offset;
 209   int cnt;
 210
 211 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 212   if (use_clock && !timing)
 213     use_clock = clock_gettime (cl, &start) == 0;
 214 #endif
 215
 216   err = regcomp (&re, expr, REG_NEWLINE | (icase ? REG_ICASE : 0));
 217   if (err != REG_NOERROR)
 218     {
 219       char buf[200];
 220       regerror (err, &re, buf, sizeof buf);
 221       error (EXIT_FAILURE, 0, "cannot compile expression: %s", buf);
 222     }
 223
 224   cnt = 0;
 225   offset = 0;
 226   assert (mem[memlen] == '\0');
 227   while (offset < memlen)
 228     {
 229       regmatch_t ma[1];
 230       const char *sp;
 231       const char *ep;
 232
 233       err = regexec (&re, mem + offset, 1, ma, 0);
 234       if (err == REG_NOMATCH)
 235         break;
 236
 237       if (err != REG_NOERROR)
 238         {
 239           char buf[200];
 240           regerror (err, &re, buf, sizeof buf);
 241           error (EXIT_FAILURE, 0, "cannot use expression: %s", buf);
 242         }
 243
 244       assert (ma[0].rm_so >= 0);
 245       sp = mem + offset + ma[0].rm_so;
 246       while (sp > mem && sp[-1] != '\n')
 247         --sp;
 248
 249       ep = mem + offset + ma[0].rm_so;
 250       while (*ep != '\0' && *ep != '\n')
 251         ++ep;
 252
 253       printf ("match %d: \"%.*s\"\n", ++cnt, (int) (ep - sp), sp);
 254
 255       offset = ep + 1 - mem;
 256     }
 257
 258   regfree (&re);
 259
 260 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 261   if (use_clock && !timing)
 262     {
 263       use_clock = clock_gettime (cl, &finish) == 0;
 264       if (use_clock)
 265         {
 266           if (finish.tv_nsec < start.tv_nsec)
 267             {
 268               finish.tv_nsec -= start.tv_nsec - 1000000000;
 269               finish.tv_sec -= 1 + start.tv_sec;
 270             }
 271           else
 272             {
 273               finish.tv_nsec -= start.tv_nsec;
 274               finish.tv_sec -= start.tv_sec;
 275             }
 276
 277           printf ("elapsed time: %jd.%09jd sec\n",
 278                   (intmax_t) finish.tv_sec, (intmax_t) finish.tv_nsec);
 279         }
 280     }
 281
 282   if (use_clock && timing)
 283     {
 284       struct timespec mintime = { .tv_sec = 24 * 60 * 60 };
 285
 286       for (int i = 0; i < 10; ++i)
 287         {
 288           offset = 0;
 289           use_clock = clock_gettime (cl, &start) == 0;
 290
 291           if (!use_clock)
 292             continue;
 293
 294           err = regcomp (&re, expr, REG_NEWLINE | (icase ? REG_ICASE : 0));
 295           if (err != REG_NOERROR)
 296             continue;
 297
 298           while (offset < memlen)
 299             {
 300               regmatch_t ma[1];
 301
 302               err = regexec (&re, mem + offset, 1, ma, 0);
 303               if (err != REG_NOERROR)
 304                 break;
 305
 306               offset += ma[0].rm_eo;
 307             }
 308
 309           regfree (&re);
 310
 311           use_clock = clock_gettime (cl, &finish) == 0;
 312           if (use_clock)
 313             {
 314               if (finish.tv_nsec < start.tv_nsec)
 315                 {
 316                   finish.tv_nsec -= start.tv_nsec - 1000000000;
 317                   finish.tv_sec -= 1 + start.tv_sec;
 318                 }
 319               else
 320                 {
 321                   finish.tv_nsec -= start.tv_nsec;
 322                   finish.tv_sec -= start.tv_sec;
 323                 }
 324               if (finish.tv_sec < mintime.tv_sec
 325                   || (finish.tv_sec == mintime.tv_sec
 326                       && finish.tv_nsec < mintime.tv_nsec))
 327                 mintime = finish;
 328             }
 329         }
 330       printf ("elapsed time: %jd.%09jd sec\n",
 331               (intmax_t) mintime.tv_sec, (intmax_t) mintime.tv_nsec);
 332     }
 333 #endif
 334
 335   /* Return an error if the number of matches found is not match we
 336      expect.  */
 337   return cnt != expected;
 338 }
 339
 340
 341 static int
 342 run_test_backwards (const char *expr, const char *mem, size_t memlen,
 343                     int icase, int expected)
 344 {
 345 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 346   struct timespec start;
 347   struct timespec finish;
 348 #endif
 349   struct re_pattern_buffer re;
 350   const char *err;
 351   size_t offset;
 352   int cnt;
 353
 354 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 355   if (use_clock && !timing)
 356     use_clock = clock_gettime (cl, &start) == 0;
 357 #endif
 358
 359   re_set_syntax ((RE_SYNTAX_POSIX_BASIC & ~RE_DOT_NEWLINE)
 360                  | RE_HAT_LISTS_NOT_NEWLINE
 361                  | (icase ? RE_ICASE : 0));
 362
 363   memset (&re, 0, sizeof (re));
 364   re.fastmap = malloc (256);
 365   if (re.fastmap == NULL)
 366     error (EXIT_FAILURE, errno, "cannot allocate fastmap");
 367
 368   err = re_compile_pattern (expr, strlen (expr), &re);
 369   if (err != NULL)
 370     error (EXIT_FAILURE, 0, "cannot compile expression: %s", err);
 371
 372   if (re_compile_fastmap (&re))
 373     error (EXIT_FAILURE, 0, "couldn't compile fastmap");
 374
 375   cnt = 0;
 376   offset = memlen;
 377   assert (mem[memlen] == '\0');
 378   while (offset <= memlen)
 379     {
 380       int start;
 381       const char *sp;
 382       const char *ep;
 383
 384       start = re_search (&re, mem, memlen, offset, -offset, NULL);
 385       if (start == -1)
 386         break;
 387
 388       if (start == -2)
 389         error (EXIT_FAILURE, 0, "internal error in re_search");
 390
 391       sp = mem + start;
 392       while (sp > mem && sp[-1] != '\n')
 393         --sp;
 394
 395       ep = mem + start;
 396       while (*ep != '\0' && *ep != '\n')
 397         ++ep;
 398
 399       printf ("match %d: \"%.*s\"\n", ++cnt, (int) (ep - sp), sp);
 400
 401       offset = sp - 1 - mem;
 402     }
 403
 404   regfree (&re);
 405
 406 #if defined _POSIX_CPUTIME && _POSIX_CPUTIME >= 0
 407   if (use_clock && !timing)
 408     {
 409       use_clock = clock_gettime (cl, &finish) == 0;
 410       if (use_clock)
 411         {
 412           if (finish.tv_nsec < start.tv_nsec)
 413             {
 414               finish.tv_nsec -= start.tv_nsec - 1000000000;
 415               finish.tv_sec -= 1 + start.tv_sec;
 416             }
 417           else
 418             {
 419               finish.tv_nsec -= start.tv_nsec;
 420               finish.tv_sec -= start.tv_sec;
 421             }
 422
 423           printf ("elapsed time: %jd.%09jd sec\n",
 424                   (intmax_t) finish.tv_sec, (intmax_t) finish.tv_nsec);
 425         }
 426     }
 427
 428   if (use_clock && timing)
 429     {
 430       struct timespec mintime = { .tv_sec = 24 * 60 * 60 };
 431
 432       for (int i = 0; i < 10; ++i)
 433         {
 434           offset = memlen;
 435           use_clock = clock_gettime (cl, &start) == 0;
 436
 437           if (!use_clock)
 438             continue;
 439
 440           memset (&re, 0, sizeof (re));
 441           re.fastmap = malloc (256);
 442           if (re.fastmap == NULL)
 443             continue;
 444
 445           err = re_compile_pattern (expr, strlen (expr), &re);
 446           if (err != NULL)
 447             continue;
 448
 449           if (re_compile_fastmap (&re))
 450             {
 451               regfree (&re);
 452               continue;
 453             }
 454
 455           while (offset <= memlen)
 456             {
 457               int start;
 458               const char *sp;
 459
 460               start = re_search (&re, mem, memlen, offset, -offset, NULL);
 461               if (start < -1)
 462                 break;
 463
 464               sp = mem + start;
 465               while (sp > mem && sp[-1] != '\n')
 466                 --sp;
 467
 468               offset = sp - 1 - mem;
 469             }
 470
 471           regfree (&re);
 472
 473           use_clock = clock_gettime (cl, &finish) == 0;
 474           if (use_clock)
 475             {
 476               if (finish.tv_nsec < start.tv_nsec)
 477                 {
 478                   finish.tv_nsec -= start.tv_nsec - 1000000000;
 479                   finish.tv_sec -= 1 + start.tv_sec;
 480                 }
 481               else
 482                 {
 483                   finish.tv_nsec -= start.tv_nsec;
 484                   finish.tv_sec -= start.tv_sec;
 485                 }
 486               if (finish.tv_sec < mintime.tv_sec
 487                   || (finish.tv_sec == mintime.tv_sec
 488                       && finish.tv_nsec < mintime.tv_nsec))
 489                 mintime = finish;
 490             }
 491         }
 492       printf ("elapsed time: %jd.%09jd sec\n",
 493               (intmax_t) mintime.tv_sec, (intmax_t) mintime.tv_nsec);
 494     }
 495 #endif
 496
 497   /* Return an error if the number of matches found is not match we
 498      expect.  */
 499   return cnt != expected;
 500 }
 501
 502 /* If --timing is used we will need a larger timout.  */
 503 #define TIMEOUT 50
 504 #define CMDLINE_OPTIONS \
 505    {"timing", no_argument, &timing, 1 },
 506 #define TEST_FUNCTION do_test ()
 507 #include "../test-skeleton.c"