/[pcre]/code/trunk/pcretest.c

Diff of /code/trunk/pcretest.c

Parent Directory | Revision Log | View Patch Patch

-revision 77 by nigel,
Sat Feb 24 21:40:45 2007 UTC
+revision 87 by nigel,
Sat Feb 24 21:41:21 2007 UTC
 Line 46 
 POSSIBILITY OF SUCH DAMAGE.
  #define PCRE_SPY        /* For Win32 build, import data, not export */
- /* We need the internal info for displaying the results of pcre_study() and
+ /* We include pcre_internal.h because we need the internal info for displaying
- other internal data; pcretest also uses some of the fixed tables, and generally
+ the results of pcre_study() and we also need to know about the internal
- has "inside information" compared to a program that strictly follows the PCRE
+ macros, structures, and other internal data values; pcretest has "inside
- API. */
+ information" compared to a program that strictly follows the PCRE API. */
  #include "pcre_internal.h"
+ /* We need access to the data tables that PCRE uses. So as not to have to keep
+ two copies, we include the source file here, changing the names of the external
+ symbols to prevent clashes. */
+ #define _pcre_utf8_table1      utf8_table1
+ #define _pcre_utf8_table1_size utf8_table1_size
+ #define _pcre_utf8_table2      utf8_table2
+ #define _pcre_utf8_table3      utf8_table3
+ #define _pcre_utf8_table4      utf8_table4
+ #define _pcre_utt              utt
+ #define _pcre_utt_size         utt_size
+ #define _pcre_OP_lengths       OP_lengths
+ #include "pcre_tables.c"
+ /* We also need the pcre_printint() function for printing out compiled
+ patterns. This function is in a separate file so that it can be included in
+ pcre_compile.c when that module is compiled with debugging enabled. */
+ #include "pcre_printint.src"
  /* It is possible to compile this test program without including support for
  testing the POSIX interface, though this is not available via the standard
-Line 62 
 Makefile. */
+Line 83 
 Makefile. */
  #include "pcreposix.h"
  #endif
+ /* It is also possible, for the benefit of the version imported into Exim, to
+ build pcretest without support for UTF8 (define NOUTF8), without the interface
+ to the DFA matcher (NODFA), and without the doublecheck of the old "info"
+ function (define NOINFOCHECK). */
+ /* Other parameters */
  #ifndef CLOCKS_PER_SEC
  #ifdef CLK_TCK
  #define CLOCKS_PER_SEC CLK_TCK
-Line 77 
 Makefile. */
+Line 106 
 Makefile. */
  #define DBUFFER_SIZE BUFFER_SIZE
+ /* Static variables */
  static FILE *outfile;
  static int log_store = 0;
  static int callout_count;
-Line 135 
 Returns:   >  0 => the number of bytes c
+Line 166 
 Returns:   >  0 => the number of bytes c
             -6 to 0 => malformed UTF-8 character at offset = (-return)
  */
+ #if !defined NOUTF8
  static int
  utf82ord(unsigned char *buffer, int *vptr)
  {
-Line 154 
 if (i == 0 || i == 6) return 0;        /
+Line 187 
 if (i == 0 || i == 6) return 0;        /
  /* i now has a value in the range 1-5 */
  s = 6*i;
- d = (c & _pcre_utf8_table3[i]) << s;
+ d = (c & utf8_table3[i]) << s;
  for (j = 0; j < i; j++)
    {
-Line 166 
 for (j = 0; j < i; j++)
+Line 199 
 for (j = 0; j < i; j++)
  /* Check that encoding was the correct unique one */
- for (j = 0; j < _pcre_utf8_table1_size; j++)
+ for (j = 0; j < utf8_table1_size; j++)
-   if (d <= _pcre_utf8_table1[j]) break;
+   if (d <= utf8_table1[j]) break;
  if (j != i) return -(i+1);
  /* Valid value */
-Line 176 
 if (j != i) return -(i+1);
+Line 209 
 if (j != i) return -(i+1);
  return i+1;
  }
+ #endif
+ /*************************************************
+ *       Convert character value to UTF-8         *
+ *************************************************/
+ /* This function takes an integer value in the range 0 - 0x7fffffff
+ and encodes it as a UTF-8 character in 0 to 6 bytes.
+ Arguments:
+   cvalue     the character value
+   buffer     pointer to buffer for result - at least 6 bytes long
+ Returns:     number of characters placed in the buffer
+ */
+ static int
+ ord2utf8(int cvalue, uschar *buffer)
+ {
+ register int i, j;
+ for (i = 0; i < utf8_table1_size; i++)
+   if (cvalue <= utf8_table1[i]) break;
+ buffer += i;
+ for (j = i; j > 0; j--)
+  {
+  *buffer-- = 0x80 | (cvalue & 0x3f);
+  cvalue >>= 6;
+  }
+ *buffer = utf8_table2[i] | cvalue;
+ return i + 1;
+ }
  /*************************************************
-Line 188 
 chars without printing. */
+Line 255 
 chars without printing. */
  static int pchars(unsigned char *p, int length, FILE *f)
  {
- int c;
+ int c = 0;
  int yield = 0;
  while (length-- > 0)
    {
+ #if !defined NOUTF8
    if (use_utf8)
      {
      int rc = utf82ord(p, &c);
-Line 215 
 while (length-- > 0)
+Line 283 
 while (length-- > 0)
        continue;
        }
      }
+ #endif
     /* Not UTF-8, or malformed UTF-8  */
-Line 406 
 return ((value & 0x000000ff) << 24) |
+Line 475 
 return ((value & 0x000000ff) << 24) |
  /*************************************************
+ *        Check match or recursion limit          *
+ *************************************************/
+ static int
+ check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len,
+   int start_offset, int options, int *use_offsets, int use_size_offsets,
+   int flag, unsigned long int *limit, int errnumber, const char *msg)
+ {
+ int count;
+ int min = 0;
+ int mid = 64;
+ int max = -1;
+ extra->flags |= flag;
+ for (;;)
+   {
+   *limit = mid;
+   count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options,
+     use_offsets, use_size_offsets);
+   if (count == errnumber)
+     {
+     /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
+     min = mid;
+     mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;
+     }
+   else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||
+                          count == PCRE_ERROR_PARTIAL)
+     {
+     if (mid == min + 1)
+       {
+       fprintf(outfile, "Minimum %s limit = %d\n", msg, mid);
+       break;
+       }
+     /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
+     max = mid;
+     mid = (min + mid)/2;
+     }
+   else break;    /* Some other error */
+   }
+ extra->flags &= ~flag;
+ return count;
+ }
+ /*************************************************
  *                Main Program                    *
  *************************************************/
-Line 422 
 int op = 1;
+Line 542 
 int op = 1;
  int timeit = 0;
  int showinfo = 0;
  int showstore = 0;
+ int quiet = 0;
  int size_offsets = 45;
  int size_offsets_max;
  int *offsets = NULL;
-Line 462 
 while (argc > 1 && argv[op][0] == '-')
+Line 583 
 while (argc > 1 && argv[op][0] == '-')
    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)
      showstore = 1;
    else if (strcmp(argv[op], "-t") == 0) timeit = 1;
+   else if (strcmp(argv[op], "-q") == 0) quiet = 1;
    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
+ #if !defined NODFA
    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
+ #endif
    else if (strcmp(argv[op], "-o") == 0 && argc > 2 &&
        ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)),
          *endptr == 0))
-Line 492 
 while (argc > 1 && argv[op][0] == '-')
+Line 616 
 while (argc > 1 && argv[op][0] == '-')
      printf("  POSIX malloc threshold = %d\n", rc);
      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);
      printf("  Default match limit = %d\n", rc);
+     (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);
+     printf("  Default recursion depth limit = %d\n", rc);
      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
      exit(0);
-Line 502 
 while (argc > 1 && argv[op][0] == '-')
+Line 628 
 while (argc > 1 && argv[op][0] == '-')
      printf("Usage:   pcretest [-d] [-i] [-o <n>] [-p] [-s] [-t] [<input> [<output>]]\n");
      printf("  -C     show PCRE compile-time options and exit\n");
      printf("  -d     debug: show compiled code; implies -i\n");
+ #if !defined NODFA
      printf("  -dfa   force DFA matching for all subjects\n");
+ #endif
      printf("  -i     show information about compiled pattern\n"
             "  -m     output memory used information\n"
             "  -o <n> set size of offsets vector to <n>\n");
-Line 561 
 pcre_free = new_free;
+Line 689 
 pcre_free = new_free;
  pcre_stack_malloc = stack_malloc;
  pcre_stack_free = stack_free;
- /* Heading line, then prompt for first regex if stdin */
+ /* Heading line unless quiet, then prompt for first regex if stdin */
- fprintf(outfile, "PCRE version %s\n\n", pcre_version());
+ if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version());
  /* Main loop */
-Line 823 
 while (!done)
+Line 951 
 while (!done)
      if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;
      if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;
      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
+     if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
+     if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
      rc = regcomp(&preg, (char *)p, cflags);
      /* Compilation failed; go back for another re, skipping to blank line
-Line 962 
 while (!done)
+Line 1093 
 while (!done)
      if (do_showinfo)
        {
        unsigned long int get_options, all_options;
+ #if !defined NOINFOCHECK
        int old_first_char, old_options, old_count;
+ #endif
        int count, backrefmax, first_char, need_char;
        int nameentrysize, namecount;
        const uschar *nametable;
-Line 970 
 while (!done)
+Line 1103 
 while (!done)
        if (do_debug)
          {
          fprintf(outfile, "------------------------------------------------------------------\n");
-         _pcre_printint(re, outfile);
+         pcre_printint(re, outfile);
          }
        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
-Line 983 
 while (!done)
+Line 1116 
 while (!done)
        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
+ #if !defined NOINFOCHECK
        old_count = pcre_info(re, &old_options, &old_first_char);
        if (count < 0) fprintf(outfile,
          "Error %d from pcre_info()\n", count);
-Line 1000 
 while (!done)
+Line 1134 
 while (!done)
            "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n",
              get_options, old_options);
          }
+ #endif
        if (size != regex_gotten_store) fprintf(outfile,
          "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n",
-Line 1034 
 while (!done)
+Line 1169 
 while (!done)
          fprintf(outfile, "Partial matching not supported\n");
        if (get_options == 0) fprintf(outfile, "No options\n");
-         else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s\n",
+         else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s\n",
            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
-Line 1044 
 while (!done)
+Line 1179 
 while (!done)
            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
+           ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
            ((get_options & PCRE_UTF8) != 0)? " utf8" : "",
            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "");
-Line 1189 
 while (!done)
+Line 1325 
 while (!done)
    for (;;)
      {
-     unsigned char *q;
+     uschar *q;
-     unsigned char *bptr = dbuffer;
+     uschar *bptr = dbuffer;
      int *use_offsets = offsets;
      int use_size_offsets = size_offsets;
      int callout_data = 0;
-Line 1259 
 while (!done)
+Line 1395 
 while (!done)
          /* Handle \x{..} specially - new Perl thing for utf8 */
+ #if !defined NOUTF8
          if (*p == '{')
            {
            unsigned char *pt = p;
-Line 1269 
 while (!done)
+Line 1406 
 while (!done)
              {
              unsigned char buff8[8];
              int ii, utn;
-             utn = _pcre_ord2utf8(c, buff8);
+             utn = ord2utf8(c, buff8);
              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
              c = buff8[ii];   /* Last byte */
              p = pt + 1;
-Line 1277 
 while (!done)
+Line 1414 
 while (!done)
              }
            /* Not correct form; fall through */
            }
+ #endif
          /* Ordinary \x */
-Line 1357 
 while (!done)
+Line 1495 
 while (!done)
            }
          continue;
+ #if !defined NODFA
          case 'D':
+ #if !defined NOPOSIX
          if (posix || do_posix)
            printf("** Can't use dfa matching in POSIX mode: \\D ignored\n");
          else
+ #endif
            use_dfa = 1;
          continue;
          case 'F':
          options |= PCRE_DFA_SHORTEST;
          continue;
+ #endif
          case 'G':
          if (isdigit(*p))
-Line 1422 
 while (!done)
+Line 1564 
 while (!done)
          options |= PCRE_PARTIAL;
          continue;
+ #if !defined NODFA
          case 'R':
          options |= PCRE_DFA_RESTART;
          continue;
+ #endif
          case 'S':
          show_malloc = 1;
-Line 1470 
 while (!done)
+Line 1614 
 while (!done)
          (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);
          fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);
          }
+       else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE)
+               != 0)
+         {
+         fprintf(outfile, "Matched with REG_NOSUB\n");
+         }
        else
          {
          size_t i;
-Line 1507 
 while (!done)
+Line 1656 
 while (!done)
          clock_t time_taken;
          clock_t start_time = clock();
+ #if !defined NODFA
          if (all_use_dfa || use_dfa)
            {
            int workspace[1000];
-Line 1516 
 while (!done)
+Line 1666 
 while (!done)
                sizeof(workspace)/sizeof(int));
            }
          else
+ #endif
          for (i = 0; i < LOOPREPEAT; i++)
            count = pcre_exec(re, extra, (char *)bptr, len,
-Line 1528 
 while (!done)
+Line 1679 
 while (!done)
          }
        /* If find_match_limit is set, we want to do repeated matches with
-       varying limits in order to find the minimum value. */
+       varying limits in order to find the minimum value for the match limit and
+       for the recursion limit. */
        if (find_match_limit)
          {
-         int min = 0;
-         int mid = 64;
-         int max = -1;
          if (extra == NULL)
            {
            extra = (pcre_extra *)malloc(sizeof(pcre_extra));
            extra->flags = 0;
            }
-         extra->flags |= PCRE_EXTRA_MATCH_LIMIT;
-         for (;;)
-           {
-           extra->match_limit = mid;
-           count = pcre_exec(re, extra, (char *)bptr, len, start_offset,
-             options | g_notempty, use_offsets, use_size_offsets);
-           if (count == PCRE_ERROR_MATCHLIMIT)
-             {
-             /* fprintf(outfile, "Testing match limit = %d\n", mid); */
-             min = mid;
-             mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;
-             }
-           else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||
-                                  count == PCRE_ERROR_PARTIAL)
-             {
-             if (mid == min + 1)
-               {
-               fprintf(outfile, "Minimum match limit = %d\n", mid);
-               break;
-               }
-             /* fprintf(outfile, "Testing match limit = %d\n", mid); */
-             max = mid;
-             mid = (min + mid)/2;
-             }
-           else break;    /* Some other error */
-           }
-         extra->flags &= ~PCRE_EXTRA_MATCH_LIMIT;
+         count = check_match_limit(re, extra, bptr, len, start_offset,
+           options|g_notempty, use_offsets, use_size_offsets,
+           PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
+           PCRE_ERROR_MATCHLIMIT, "match()");
+         count = check_match_limit(re, extra, bptr, len, start_offset,
+           options|g_notempty, use_offsets, use_size_offsets,
+           PCRE_EXTRA_MATCH_LIMIT_RECURSION, &(extra->match_limit_recursion),
+           PCRE_ERROR_RECURSIONLIMIT, "match() recursion");
          }
        /* If callout_data is set, use the interface with additional data */
-Line 1591 
 while (!done)
+Line 1720 
 while (!done)
        /* The normal case is just to do the match once, with the default
        value of match_limit. */
+ #if !defined NODFA
        else if (all_use_dfa || use_dfa)
          {
          int workspace[1000];
-Line 1603 
 while (!done)
+Line 1733 
 while (!done)
            count = use_size_offsets/2;
            }
          }
+ #endif
        else
          {
-Line 1699 
 while (!done)
+Line 1830 
 while (!done)
        else if (count == PCRE_ERROR_PARTIAL)
          {
          fprintf(outfile, "Partial match");
+ #if !defined NODFA
          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)
            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],
              bptr + use_offsets[0]);
+ #endif
          fprintf(outfile, "\n");
          break;  /* Out of the /g loop */
          }

 Legend:



Removed from v.77
 


changed lines


 
Added in v.87
 Legend:



Removed from v.77
 


changed lines


 
Added in v.87
-Removed from v.77
+Added in v.87

	ViewVC Help
Powered by ViewVC 1.1.5