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

Diff of /code/trunk/pcretest.c

Parent Directory | Revision Log | View Patch Patch

-revision 91 by nigel,
Sat Feb 24 21:41:34 2007 UTC
+revision 172 by ph10,
Tue Jun  5 10:40:13 2007 UTC
 Line 44 
 POSSIBILITY OF SUCH DAMAGE.
  #include <locale.h>
  #include <errno.h>
- #ifndef _WIN32
- #include <sys/resource.h>
+ /* A number of things vary for Windows builds. Originally, pcretest opened its
+ input and output without "b"; then I was told that "b" was needed in some
+ environments, so it was added for release 5.0 to both the input and output. (It
+ makes no difference on Unix-like systems.) Later I was told that it is wrong
+ for the input on Windows. I've now abstracted the modes into two macros that
+ are set here, to make it easier to fiddle with them, and removed "b" from the
+ input mode under Windows. */
+ #if defined(_WIN32) || defined(WIN32)
+ #include <io.h>                /* For _setmode() */
+ #include <fcntl.h>             /* For _O_BINARY */
+ #define INPUT_MODE   "r"
+ #define OUTPUT_MODE  "wb"
+ #else
+ #include <sys/time.h>          /* These two includes are needed */
+ #include <sys/resource.h>      /* for setrlimit(). */
+ #define INPUT_MODE   "rb"
+ #define OUTPUT_MODE  "wb"
  #endif
- #define PCRE_SPY        /* For Win32 build, import data, not export */
- /* We include pcre_internal.h because we need the internal info for displaying
+ /* We have to include pcre_internal.h because we need the internal info for
- the results of pcre_study() and we also need to know about the internal
+ displaying the results of pcre_study() and we also need to know about the
- macros, structures, and other internal data values; pcretest has "inside
+ internal macros, structures, and other internal data values; pcretest has
- information" compared to a program that strictly follows the PCRE API. */
+ "inside information" compared to a program that strictly follows the PCRE API.
+ Although pcre_internal.h does itself include pcre.h, we explicitly include it
+ here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
+ appropriately for an application, not for building PCRE. */
+ #include "pcre.h"
  #include "pcre_internal.h"
  /* We need access to the data tables that PCRE uses. So as not to have to keep
-Line 74 
 symbols to prevent clashes. */
+Line 96 
 symbols to prevent clashes. */
  /* 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. */
+ pcre_compile.c when that module is compiled with debugging enabled.
+ The definition of the macro PRINTABLE, which determines whether to print an
+ output character as-is or as a hex value when showing compiled patterns, is
+ contained in this file. We uses it here also, in cases when the locale has not
+ been explicitly changed, so as to get consistent output from systems that
+ differ in their output from isprint() even in the "C" locale. */
  #include "pcre_printint.src"
+ #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
  /* 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 87 
 Makefile. */
+Line 117 
 Makefile. */
  #include "pcreposix.h"
  #endif
- /* It is also possible, for the benefit of the version imported into Exim, to
+ /* It is also possible, for the benefit of the version currently imported into
- build pcretest without support for UTF8 (define NOUTF8), without the interface
+ Exim, to build pcretest without support for UTF8 (define NOUTF8), without the
- to the DFA matcher (NODFA), and without the doublecheck of the old "info"
+ interface to the DFA matcher (NODFA), and without the doublecheck of the old
- function (define NOINFOCHECK). */
+ "info" function (define NOINFOCHECK). In fact, we automatically cut out the
+ UTF8 support if PCRE is built without it. */
+ #ifndef SUPPORT_UTF8
+ #ifndef NOUTF8
+ #define NOUTF8
+ #endif
+ #endif
  /* Other parameters */
-Line 103 
 function (define NOINFOCHECK). */
+Line 140 
 function (define NOINFOCHECK). */
  #endif
  #endif
+ /* This is the default loop count for timing. */
  #define LOOPREPEAT 500000
  /* Static variables */
-Line 114 
 static int callout_extra;
+Line 153 
 static int callout_extra;
  static int callout_fail_count;
  static int callout_fail_id;
  static int first_callout;
+ static int locale_set = 0;
  static int show_malloc;
  static int use_utf8;
  static size_t gotten_store;
-Line 157 
 uschar *here = start;
+Line 197 
 uschar *here = start;
  for (;;)
    {
    int rlen = buffer_size - (here - buffer);
    if (rlen > 1000)
      {
      int dlen;
-Line 213 
 return NULL;  /* Control never gets here
+Line 254 
 return NULL;  /* Control never gets here
  /* We don't use strtoul() because SunOS4 doesn't have it. Rather than mess
  around with conditional compilation, just do the job by hand. It is only used
- for unpicking the -o argument, so just keep it simple.
+ for unpicking arguments, so just keep it simple.
  Arguments:
    str           string to be converted
-Line 311 
 Arguments:
+Line 352 
 Arguments:
  Returns:     number of characters placed in the buffer
  */
+ #if !defined NOUTF8
  static int
  ord2utf8(int cvalue, uschar *utf8bytes)
  {
-Line 327 
 for (j = i; j > 0; j--)
+Line 370 
 for (j = i; j > 0; j--)
  return i + 1;
  }
+ #endif
  /*************************************************
-Line 353 
 while (length-- > 0)
+Line 398 
 while (length-- > 0)
        {
        length -= rc - 1;
        p += rc;
-       if (c < 256 && isprint(c))
+       if (PRINTHEX(c))
          {
          if (f != NULL) fprintf(f, "%c", c);
          yield++;
          }
        else
          {
-         int n;
+         int n = 4;
-         if (f != NULL) fprintf(f, "\\x{%02x}%n", c, &n);
+         if (f != NULL) fprintf(f, "\\x{%02x}", c);
-         yield += n;
+         yield += (n <= 0x000000ff)? 2 :
+                  (n <= 0x00000fff)? 3 :
+                  (n <= 0x0000ffff)? 4 :
+                  (n <= 0x000fffff)? 5 : 6;
          }
        continue;
        }
-Line 371 
 while (length-- > 0)
+Line 419 
 while (length-- > 0)
     /* Not UTF-8, or malformed UTF-8  */
-   if (isprint(c = *(p++)))
+   c = *p++;
+   if (PRINTHEX(c))
      {
      if (f != NULL) fprintf(f, "%c", c);
      yield++;
-Line 614 
 return count;
+Line 663 
 return count;
  *************************************************/
  /* This is used both at compile and run-time to check for <xxx> escapes, where
- xxx is LF, CR, or CRLF. Print a message and return 0 if there is no match.
+ xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is
+ no match.
  Arguments:
    p           points after the leading '<'
-Line 629 
 check_newline(uschar *p, FILE *f)
+Line 679 
 check_newline(uschar *p, FILE *f)
  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;
  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;
  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
+ if (strncmp((char *)p, "anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
+ if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY;
  fprintf(f, "Unknown newline type at: <%s\n", p);
  return 0;
  }
-Line 636 
 return 0;
+Line 688 
 return 0;
  /*************************************************
+ *             Usage function                     *
+ *************************************************/
+ static void
+ usage(void)
+ {
+ printf("Usage:     pcretest [options] [<input> [<output>]]\n");
+ printf("  -b       show compiled code (bytecode)\n");
+ printf("  -C       show PCRE compile-time options and exit\n");
+ printf("  -d       debug: show compiled code and information (-b and -i)\n");
+ #if !defined NODFA
+ printf("  -dfa     force DFA matching for all subjects\n");
+ #endif
+ printf("  -help    show usage information\n");
+ printf("  -i       show information about compiled patterns\n"
+        "  -m       output memory used information\n"
+        "  -o <n>   set size of offsets vector to <n>\n");
+ #if !defined NOPOSIX
+ printf("  -p       use POSIX interface\n");
+ #endif
+ printf("  -q       quiet: do not output PCRE version number at start\n");
+ printf("  -S <n>   set stack size to <n> megabytes\n");
+ printf("  -s       output store (memory) used information\n"
+        "  -t       time compilation and execution\n");
+ printf("  -t <n>   time compilation and execution, repeating <n> times\n");
+ printf("  -tm      time execution (matching) only\n");
+ printf("  -tm <n>  time execution (matching) only, repeating <n> times\n");
+ }
+ /*************************************************
  *                Main Program                    *
  *************************************************/
-Line 650 
 int options = 0;
+Line 734 
 int options = 0;
  int study_options = 0;
  int op = 1;
  int timeit = 0;
+ int timeitm = 0;
  int showinfo = 0;
  int showstore = 0;
  int quiet = 0;
-Line 681 
 buffer = (unsigned char *)malloc(buffer_
+Line 766 
 buffer = (unsigned char *)malloc(buffer_
  dbuffer = (unsigned char *)malloc(buffer_size);
  pbuffer = (unsigned char *)malloc(buffer_size);
- /* The outfile variable is static so that new_malloc can use it. The _setmode()
+ /* The outfile variable is static so that new_malloc can use it. */
- stuff is some magic that I don't understand, but which apparently does good
- things in Windows. It's related to line terminations.  */
- #if defined(_WIN32) || defined(WIN32)
- _setmode( _fileno( stdout ), 0x8000 );
- #endif  /* defined(_WIN32) || defined(WIN32) */
  outfile = stdout;
+ /* The following  _setmode() stuff is some Windows magic that tells its runtime
+ library to translate CRLF into a single LF character. At least, that's what
+ I've been told: never having used Windows I take this all on trust. Originally
+ it set 0x8000, but then I was advised that _O_BINARY was better. */
+ #if defined(_WIN32) || defined(WIN32)
+ _setmode( _fileno( stdout ), _O_BINARY );
+ #endif
  /* Scan options */
  while (argc > 1 && argv[op][0] == '-')
-Line 699 
 while (argc > 1 && argv[op][0] == '-')
+Line 787 
 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], "-b") == 0) debug = 1;
    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
  #if !defined NODFA
-Line 713 
 while (argc > 1 && argv[op][0] == '-')
+Line 801 
 while (argc > 1 && argv[op][0] == '-')
      op++;
      argc--;
      }
+   else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
+     {
+     int both = argv[op][2] == 0;
+     int temp;
+     if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),
+                      *endptr == 0))
+       {
+       timeitm = temp;
+       op++;
+       argc--;
+       }
+     else timeitm = LOOPREPEAT;
+     if (both) timeit = timeitm;
+     }
    else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
          *endptr == 0))
      {
- #ifdef _WIN32
+ #if defined(_WIN32) || defined(WIN32)
      printf("PCRE: -S not supported on this OS\n");
      exit(1);
  #else
-Line 749 
 while (argc > 1 && argv[op][0] == '-')
+Line 851 
 while (argc > 1 && argv[op][0] == '-')
      printf("  %sUnicode properties support\n", rc? "" : "No ");
      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :
-       (rc == '\n')? "LF" : "CRLF");
+       (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :
+       (rc == -2)? "ANYCRLF" :
+       (rc == -1)? "ANY" : "???");
      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
      printf("  Internal link size = %d\n", rc);
      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
-Line 760 
 while (argc > 1 && argv[op][0] == '-')
+Line 864 
 while (argc > 1 && argv[op][0] == '-')
      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);
+     goto EXIT;
+     }
+   else if (strcmp(argv[op], "-help") == 0 ||
+            strcmp(argv[op], "--help") == 0)
+     {
+     usage();
+     goto EXIT;
      }
    else
      {
      printf("** Unknown or malformed option %s\n", argv[op]);
-     printf("Usage:   pcretest [options] [<input> [<output>]]\n");
+     usage();
-     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");
- #if !defined NOPOSIX
-     printf("  -p     use POSIX interface\n");
- #endif
-     printf("  -S <n> set stack size to <n> megabytes\n");
-     printf("  -s     output store (memory) used information\n"
-            "  -t     time compilation and execution\n");
      yield = 1;
      goto EXIT;
      }
-Line 794 
 offsets = (int *)malloc(size_offsets_max
+Line 890 
 offsets = (int *)malloc(size_offsets_max
  if (offsets == NULL)
    {
    printf("** Failed to get %d bytes of memory for offsets vector\n",
-     size_offsets_max * sizeof(int));
+     (int)(size_offsets_max * sizeof(int)));
    yield = 1;
    goto EXIT;
    }
-Line 803 
 if (offsets == NULL)
+Line 899 
 if (offsets == NULL)
  if (argc > 1)
    {
-   infile = fopen(argv[op], "rb");
+   infile = fopen(argv[op], INPUT_MODE);
    if (infile == NULL)
      {
      printf("** Failed to open %s\n", argv[op]);
-Line 814 
 if (argc > 1)
+Line 910 
 if (argc > 1)
  if (argc > 2)
    {
-   outfile = fopen(argv[op+1], "wb");
+   outfile = fopen(argv[op+1], OUTPUT_MODE);
    if (outfile == NULL)
      {
      printf("** Failed to open %s\n", argv[op+1]);
-Line 854 
 while (!done)
+Line 950 
 while (!done)
    size_t size, regex_gotten_store;
    int do_study = 0;
    int do_debug = debug;
+   int debug_lengths = 1;
    int do_G = 0;
    int do_g = 0;
    int do_showinfo = showinfo;
    int do_showrest = 0;
    int do_flip = 0;
-   int erroroffset, len, delimiter;
+   int erroroffset, len, delimiter, poffset;
    use_utf8 = 0;
-Line 969 
 while (!done)
+Line 1066 
 while (!done)
      }
    pp = p;
+   poffset = p - buffer;
    for(;;)
      {
-Line 989 
 while (!done)
+Line 1087 
 while (!done)
      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
      }
+   /* The buffer may have moved while being extended; reset the start of data
+   pointer to the correct relative point in the buffer. */
+   p = buffer + poffset;
    /* If the first character after the delimiter is backslash, make
    the pattern end with backslash. This is purely to provide a way
    of testing for the error message when a pattern ends with backslash. */
-Line 1020 
 while (!done)
+Line 1123 
 while (!done)
        case '+': do_showrest = 1; break;
        case 'A': options |= PCRE_ANCHORED; break;
+       case 'B': do_debug = 1; break;
        case 'C': options |= PCRE_AUTO_CALLOUT; break;
        case 'D': do_debug = do_showinfo = 1; break;
        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
-Line 1037 
 while (!done)
+Line 1141 
 while (!done)
        case 'S': do_study = 1; break;
        case 'U': options |= PCRE_UNGREEDY; break;
        case 'X': options |= PCRE_EXTRA; break;
+       case 'Z': debug_lengths = 0; break;
        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
        case '?': options |= PCRE_NO_UTF8_CHECK; break;
        case 'L':
        ppp = pp;
-       /* The '\r' test here is so that it works on Windows */
+       /* The '\r' test here is so that it works on Windows. */
-       while (*ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
+       /* The '0' test is just in case this is an unterminated line. */
+       while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
        *ppp = 0;
        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
          {
          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
          goto SKIP_DATA;
          }
+       locale_set = 1;
        tables = pcre_maketables();
        pp = ppp;
        break;
-Line 1116 
 while (!done)
+Line 1223 
 while (!done)
  #endif  /* !defined NOPOSIX */
      {
-     if (timeit)
+     if (timeit > 0)
        {
        register int i;
        clock_t time_taken;
        clock_t start_time = clock();
-       for (i = 0; i < LOOPREPEAT; i++)
+       for (i = 0; i < timeit; i++)
          {
          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
          if (re != NULL) free(re);
          }
        time_taken = clock() - start_time;
-       fprintf(outfile, "Compile time %.3f milliseconds\n",
+       fprintf(outfile, "Compile time %.4f milliseconds\n",
-         (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /
+         (((double)time_taken * 1000.0) / (double)timeit) /
            (double)CLOCKS_PER_SEC);
        }
-Line 1180 
 while (!done)
+Line 1287 
 while (!done)
      if (do_study)
        {
-       if (timeit)
+       if (timeit > 0)
          {
          register int i;
          clock_t time_taken;
          clock_t start_time = clock();
-         for (i = 0; i < LOOPREPEAT; i++)
+         for (i = 0; i < timeit; i++)
            extra = pcre_study(re, study_options, &error);
          time_taken = clock() - start_time;
          if (extra != NULL) free(extra);
-         fprintf(outfile, "  Study time %.3f milliseconds\n",
+         fprintf(outfile, "  Study time %.4f milliseconds\n",
-           (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /
+           (((double)time_taken * 1000.0) / (double)timeit) /
              (double)CLOCKS_PER_SEC);
          }
        extra = pcre_study(re, study_options, &error);
-Line 1233 
 while (!done)
+Line 1340 
 while (!done)
      SHOW_INFO:
+     if (do_debug)
+       {
+       fprintf(outfile, "------------------------------------------------------------------\n");
+       pcre_printint(re, outfile, debug_lengths);
+       }
      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 count, backrefmax, first_char, need_char, okpartial, jchanged;
        int nameentrysize, namecount;
        const uschar *nametable;
-       if (do_debug)
-         {
-         fprintf(outfile, "------------------------------------------------------------------\n");
-         pcre_printint(re, outfile);
-         }
        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
        new_info(re, NULL, PCRE_INFO_SIZE, &size);
        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
-Line 1258 
 while (!done)
+Line 1365 
 while (!done)
        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
+       new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
+       new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
  #if !defined NOINFOCHECK
        old_count = pcre_info(re, &old_options, &old_first_char);
-Line 1299 
 while (!done)
+Line 1408 
 while (!done)
            }
          }
-       /* The NOPARTIAL bit is a private bit in the options, so we have
+       if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
-       to fish it out via out back door */
        all_options = ((real_pcre *)re)->options;
-       if (do_flip)
+       if (do_flip) all_options = byteflip(all_options, sizeof(all_options));
-         {
-         all_options = byteflip(all_options, sizeof(all_options));
-          }
-       if ((all_options & PCRE_NOPARTIAL) != 0)
-         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%s%s\n",
-Line 1327 
 while (!done)
+Line 1429 
 while (!done)
            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
-       switch (get_options & PCRE_NEWLINE_CRLF)
+       if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
+       switch (get_options & PCRE_NEWLINE_BITS)
          {
          case PCRE_NEWLINE_CR:
          fprintf(outfile, "Forced newline sequence: CR\n");
-Line 1341 
 while (!done)
+Line 1445 
 while (!done)
          fprintf(outfile, "Forced newline sequence: CRLF\n");
          break;
+         case PCRE_NEWLINE_ANYCRLF:
+         fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
+         break;
+         case PCRE_NEWLINE_ANY:
+         fprintf(outfile, "Forced newline sequence: ANY\n");
+         break;
          default:
          break;
          }
-Line 1358 
 while (!done)
+Line 1470 
 while (!done)
          int ch = first_char & 255;
          const char *caseless = ((first_char & REQ_CASELESS) == 0)?
            "" : " (caseless)";
-         if (isprint(ch))
+         if (PRINTHEX(ch))
            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);
          else
            fprintf(outfile, "First char = %d%s\n", ch, caseless);
-Line 1373 
 while (!done)
+Line 1485 
 while (!done)
          int ch = need_char & 255;
          const char *caseless = ((need_char & REQ_CASELESS) == 0)?
            "" : " (caseless)";
-         if (isprint(ch))
+         if (PRINTHEX(ch))
            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);
          else
            fprintf(outfile, "Need char = %d%s\n", ch, caseless);
-Line 1409 
 while (!done)
+Line 1521 
 while (!done)
                    fprintf(outfile, "\n  ");
                    c = 2;
                    }
-                 if (isprint(i) && i != ' ')
+                 if (PRINTHEX(i) && i != ' ')
                    {
                    fprintf(outfile, "%c ", i);
                    c += 2;
-Line 1468 
 while (!done)
+Line 1580 
 while (!done)
                  strerror(errno));
                }
              else fprintf(outfile, "Study data written to %s\n", to_file);
              }
            }
          fclose(f);
-Line 1485 
 while (!done)
+Line 1598 
 while (!done)
    for (;;)
      {
      uschar *q;
-     uschar *bptr = dbuffer;
+     uschar *bptr;
      int *use_offsets = offsets;
      int use_size_offsets = size_offsets;
      int callout_data = 0;
-Line 1541 
 while (!done)
+Line 1654 
 while (!done)
      p = buffer;
      while (isspace(*p)) p++;
-     q = dbuffer;
+     bptr = q = dbuffer;
      while ((c = *p++) != 0)
        {
        int i = 0;
-Line 1736 
 while (!done)
+Line 1849 
 while (!done)
            if (offsets == NULL)
              {
              printf("** Failed to get %d bytes of memory for offsets vector\n",
-               size_offsets_max * sizeof(int));
+               (int)(size_offsets_max * sizeof(int)));
              yield = 1;
              goto EXIT;
              }
-Line 1866 
 while (!done)
+Line 1979 
 while (!done)
      for (;; gmatched++)    /* Loop for /g or /G */
        {
-       if (timeit)
+       if (timeitm > 0)
          {
          register int i;
          clock_t time_taken;
-Line 1876 
 while (!done)
+Line 1989 
 while (!done)
          if (all_use_dfa || use_dfa)
            {
            int workspace[1000];
-           for (i = 0; i < LOOPREPEAT; i++)
+           for (i = 0; i < timeitm; i++)
              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,
                options | g_notempty, use_offsets, use_size_offsets, workspace,
                sizeof(workspace)/sizeof(int));
-Line 1884 
 while (!done)
+Line 1997 
 while (!done)
          else
  #endif
-         for (i = 0; i < LOOPREPEAT; i++)
+         for (i = 0; i < timeitm; i++)
            count = pcre_exec(re, extra, (char *)bptr, len,
              start_offset, options | g_notempty, use_offsets, use_size_offsets);
          time_taken = clock() - start_time;
-         fprintf(outfile, "Execute time %.3f milliseconds\n",
+         fprintf(outfile, "Execute time %.4f milliseconds\n",
-           (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /
+           (((double)time_taken * 1000.0) / (double)timeitm) /
              (double)CLOCKS_PER_SEC);
          }
-Line 1966 
 while (!done)
+Line 2079 
 while (!done)
        if (count >= 0)
          {
-         int i;
+         int i, maxcount;
+ #if !defined NODFA
+         if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
+ #endif
+           maxcount = use_size_offsets/3;
+         /* This is a check against a lunatic return value. */
+         if (count > maxcount)
+           {
+           fprintf(outfile,
+             "** PCRE error: returned count %d is too big for offset size %d\n",
+             count, use_size_offsets);
+           count = use_size_offsets/3;
+           if (do_g || do_G)
+             {
+             fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
+             do_g = do_G = FALSE;        /* Break g/G loop */
+             }
+           }
          for (i = 0; i < count * 2; i += 2)
            {
            if (use_offsets[i] < 0)
-Line 2084 
 while (!done)
+Line 2218 
 while (!done)
          }
        /* Failed to match. If this is a /g or /G loop and we previously set
-       g_notempty after a null match, this is not necessarily the end.
+       g_notempty after a null match, this is not necessarily the end. We want
-       We want to advance the start offset, and continue. In the case of UTF-8
+       to advance the start offset, and continue. We won't be at the end of the
-       matching, the advance must be one character, not one byte. Fudge the
+       string - that was checked before setting g_notempty.
-       offset values to achieve this. We won't be at the end of the string -
-       that was checked before setting g_notempty. */
+       Complication arises in the case when the newline option is "any" or
+       "anycrlf". If the previous match was at the end of a line terminated by
+       CRLF, an advance of one character just passes the \r, whereas we should
+       prefer the longer newline sequence, as does the code in pcre_exec().
+       Fudge the offset value to achieve this.
+       Otherwise, in the case of UTF-8 matching, the advance must be one
+       character, not one byte. */
        else
          {
          if (g_notempty != 0)
            {
            int onechar = 1;
+           unsigned int obits = ((real_pcre *)re)->options;
            use_offsets[0] = start_offset;
-           if (use_utf8)
+           if ((obits & PCRE_NEWLINE_BITS) == 0)
+             {
+             int d;
+             (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
+             obits = (d == '\r')? PCRE_NEWLINE_CR :
+                     (d == '\n')? PCRE_NEWLINE_LF :
+                     (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :
+                     (d == -2)? PCRE_NEWLINE_ANYCRLF :
+                     (d == -1)? PCRE_NEWLINE_ANY : 0;
+             }
+           if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
+                (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
+               &&
+               start_offset < len - 1 &&
+               bptr[start_offset] == '\r' &&
+               bptr[start_offset+1] == '\n')
+             onechar++;
+           else if (use_utf8)
              {
              while (start_offset + onechar < len)
                {
-Line 2131 
 while (!done)
+Line 2290 
 while (!done)
        character. */
        g_notempty = 0;
        if (use_offsets[0] == use_offsets[1])
          {
          if (use_offsets[0] == len) break;
-Line 2165 
 while (!done)
+Line 2325 
 while (!done)
      {
      new_free((void *)tables);
      setlocale(LC_CTYPE, "C");
+     locale_set = 0;
      }
    }

 Legend:



Removed from v.91
 


changed lines


 
Added in v.172
 Legend:



Removed from v.91
 


changed lines


 
Added in v.172
-Removed from v.91
+Added in v.172

webmaster@exim.org	ViewVC Help
Powered by ViewVC 1.1.12