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

Diff of /code/trunk/pcretest.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 91 by nigel, Sat Feb 24 21:41:34 2007 UTC revision 236 by ph10, Tue Sep 11 12:57:06 2007 UTC
# Line 36  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  */  */
37    
38    
39    #ifdef HAVE_CONFIG_H
40    #include "config.h"
41    #endif
42    
43  #include <ctype.h>  #include <ctype.h>
44  #include <stdio.h>  #include <stdio.h>
45  #include <string.h>  #include <string.h>
# Line 44  POSSIBILITY OF SUCH DAMAGE. Line 48  POSSIBILITY OF SUCH DAMAGE.
48  #include <locale.h>  #include <locale.h>
49  #include <errno.h>  #include <errno.h>
50    
51  #ifndef _WIN32  
52  #include <sys/resource.h>  /* A number of things vary for Windows builds. Originally, pcretest opened its
53    input and output without "b"; then I was told that "b" was needed in some
54    environments, so it was added for release 5.0 to both the input and output. (It
55    makes no difference on Unix-like systems.) Later I was told that it is wrong
56    for the input on Windows. I've now abstracted the modes into two macros that
57    are set here, to make it easier to fiddle with them, and removed "b" from the
58    input mode under Windows. */
59    
60    #if defined(_WIN32) || defined(WIN32)
61    #include <io.h>                /* For _setmode() */
62    #include <fcntl.h>             /* For _O_BINARY */
63    #define INPUT_MODE   "r"
64    #define OUTPUT_MODE  "wb"
65    
66    #else
67    #include <sys/time.h>          /* These two includes are needed */
68    #include <sys/resource.h>      /* for setrlimit(). */
69    #define INPUT_MODE   "rb"
70    #define OUTPUT_MODE  "wb"
71  #endif  #endif
72    
 #define PCRE_SPY        /* For Win32 build, import data, not export */  
73    
74  /* 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
75  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
76  macros, structures, and other internal data values; pcretest has "inside  internal macros, structures, and other internal data values; pcretest has
77  information" compared to a program that strictly follows the PCRE API. */  "inside information" compared to a program that strictly follows the PCRE API.
78    
79    Although pcre_internal.h does itself include pcre.h, we explicitly include it
80    here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
81    appropriately for an application, not for building PCRE. */
82    
83    #include "pcre.h"
84  #include "pcre_internal.h"  #include "pcre_internal.h"
85    
86  /* We need access to the data tables that PCRE uses. So as not to have to keep  /* We need access to the data tables that PCRE uses. So as not to have to keep
# Line 74  symbols to prevent clashes. */ Line 100  symbols to prevent clashes. */
100    
101  /* We also need the pcre_printint() function for printing out compiled  /* We also need the pcre_printint() function for printing out compiled
102  patterns. This function is in a separate file so that it can be included in  patterns. This function is in a separate file so that it can be included in
103  pcre_compile.c when that module is compiled with debugging enabled. */  pcre_compile.c when that module is compiled with debugging enabled.
104    
105    The definition of the macro PRINTABLE, which determines whether to print an
106    output character as-is or as a hex value when showing compiled patterns, is
107    contained in this file. We uses it here also, in cases when the locale has not
108    been explicitly changed, so as to get consistent output from systems that
109    differ in their output from isprint() even in the "C" locale. */
110    
111  #include "pcre_printint.src"  #include "pcre_printint.src"
112    
113    #define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
114    
115    
116  /* It is possible to compile this test program without including support for  /* It is possible to compile this test program without including support for
117  testing the POSIX interface, though this is not available via the standard  testing the POSIX interface, though this is not available via the standard
# Line 87  Makefile. */ Line 121  Makefile. */
121  #include "pcreposix.h"  #include "pcreposix.h"
122  #endif  #endif
123    
124  /* 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
125  build pcretest without support for UTF8 (define NOUTF8), without the interface  Exim, to build pcretest without support for UTF8 (define NOUTF8), without the
126  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
127  function (define NOINFOCHECK). */  "info" function (define NOINFOCHECK). In fact, we automatically cut out the
128    UTF8 support if PCRE is built without it. */
129    
130    #ifndef SUPPORT_UTF8
131    #ifndef NOUTF8
132    #define NOUTF8
133    #endif
134    #endif
135    
136    
137  /* Other parameters */  /* Other parameters */
# Line 103  function (define NOINFOCHECK). */ Line 144  function (define NOINFOCHECK). */
144  #endif  #endif
145  #endif  #endif
146    
147    /* This is the default loop count for timing. */
148    
149  #define LOOPREPEAT 500000  #define LOOPREPEAT 500000
150    
151  /* Static variables */  /* Static variables */
# Line 113  static int callout_count; Line 156  static int callout_count;
156  static int callout_extra;  static int callout_extra;
157  static int callout_fail_count;  static int callout_fail_count;
158  static int callout_fail_id;  static int callout_fail_id;
159    static int debug_lengths;
160  static int first_callout;  static int first_callout;
161    static int locale_set = 0;
162  static int show_malloc;  static int show_malloc;
163  static int use_utf8;  static int use_utf8;
164  static size_t gotten_store;  static size_t gotten_store;
# Line 157  uschar *here = start; Line 202  uschar *here = start;
202  for (;;)  for (;;)
203    {    {
204    int rlen = buffer_size - (here - buffer);    int rlen = buffer_size - (here - buffer);
205    
206    if (rlen > 1000)    if (rlen > 1000)
207      {      {
208      int dlen;      int dlen;
# Line 213  return NULL;  /* Control never gets here Line 259  return NULL;  /* Control never gets here
259    
260  /* We don't use strtoul() because SunOS4 doesn't have it. Rather than mess  /* We don't use strtoul() because SunOS4 doesn't have it. Rather than mess
261  around with conditional compilation, just do the job by hand. It is only used  around with conditional compilation, just do the job by hand. It is only used
262  for unpicking the -o argument, so just keep it simple.  for unpicking arguments, so just keep it simple.
263    
264  Arguments:  Arguments:
265    str           string to be converted    str           string to be converted
# Line 311  Arguments: Line 357  Arguments:
357  Returns:     number of characters placed in the buffer  Returns:     number of characters placed in the buffer
358  */  */
359    
360    #if !defined NOUTF8
361    
362  static int  static int
363  ord2utf8(int cvalue, uschar *utf8bytes)  ord2utf8(int cvalue, uschar *utf8bytes)
364  {  {
# Line 327  for (j = i; j > 0; j--) Line 375  for (j = i; j > 0; j--)
375  return i + 1;  return i + 1;
376  }  }
377    
378    #endif
379    
380    
381    
382  /*************************************************  /*************************************************
# Line 353  while (length-- > 0) Line 403  while (length-- > 0)
403        {        {
404        length -= rc - 1;        length -= rc - 1;
405        p += rc;        p += rc;
406        if (c < 256 && isprint(c))        if (PRINTHEX(c))
407          {          {
408          if (f != NULL) fprintf(f, "%c", c);          if (f != NULL) fprintf(f, "%c", c);
409          yield++;          yield++;
410          }          }
411        else        else
412          {          {
413          int n;          int n = 4;
414          if (f != NULL) fprintf(f, "\\x{%02x}%n", c, &n);          if (f != NULL) fprintf(f, "\\x{%02x}", c);
415          yield += n;          yield += (n <= 0x000000ff)? 2 :
416                     (n <= 0x00000fff)? 3 :
417                     (n <= 0x0000ffff)? 4 :
418                     (n <= 0x000fffff)? 5 : 6;
419          }          }
420        continue;        continue;
421        }        }
# Line 371  while (length-- > 0) Line 424  while (length-- > 0)
424    
425     /* Not UTF-8, or malformed UTF-8  */     /* Not UTF-8, or malformed UTF-8  */
426    
427    if (isprint(c = *(p++)))    c = *p++;
428      if (PRINTHEX(c))
429      {      {
430      if (f != NULL) fprintf(f, "%c", c);      if (f != NULL) fprintf(f, "%c", c);
431      yield++;      yield++;
# Line 610  return count; Line 664  return count;
664    
665    
666  /*************************************************  /*************************************************
667    *         Case-independent strncmp() function    *
668    *************************************************/
669    
670    /*
671    Arguments:
672      s         first string
673      t         second string
674      n         number of characters to compare
675    
676    Returns:    < 0, = 0, or > 0, according to the comparison
677    */
678    
679    static int
680    strncmpic(uschar *s, uschar *t, int n)
681    {
682    while (n--)
683      {
684      int c = tolower(*s++) - tolower(*t++);
685      if (c) return c;
686      }
687    return 0;
688    }
689    
690    
691    
692    /*************************************************
693  *         Check newline indicator                *  *         Check newline indicator                *
694  *************************************************/  *************************************************/
695    
696  /* This is used both at compile and run-time to check for <xxx> escapes, where  /* This is used both at compile and run-time to check for <xxx> escapes, where
697  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
698    no match.
699    
700  Arguments:  Arguments:
701    p           points after the leading '<'    p           points after the leading '<'
# Line 626  Returns:      appropriate PCRE_NEWLINE_x Line 707  Returns:      appropriate PCRE_NEWLINE_x
707  static int  static int
708  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
709  {  {
710  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
711  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
712  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
713    if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
714    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
715    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
716    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
717  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
718  return 0;  return 0;
719  }  }
# Line 636  return 0; Line 721  return 0;
721    
722    
723  /*************************************************  /*************************************************
724    *             Usage function                     *
725    *************************************************/
726    
727    static void
728    usage(void)
729    {
730    printf("Usage:     pcretest [options] [<input> [<output>]]\n");
731    printf("  -b       show compiled code (bytecode)\n");
732    printf("  -C       show PCRE compile-time options and exit\n");
733    printf("  -d       debug: show compiled code and information (-b and -i)\n");
734    #if !defined NODFA
735    printf("  -dfa     force DFA matching for all subjects\n");
736    #endif
737    printf("  -help    show usage information\n");
738    printf("  -i       show information about compiled patterns\n"
739           "  -m       output memory used information\n"
740           "  -o <n>   set size of offsets vector to <n>\n");
741    #if !defined NOPOSIX
742    printf("  -p       use POSIX interface\n");
743    #endif
744    printf("  -q       quiet: do not output PCRE version number at start\n");
745    printf("  -S <n>   set stack size to <n> megabytes\n");
746    printf("  -s       output store (memory) used information\n"
747           "  -t       time compilation and execution\n");
748    printf("  -t <n>   time compilation and execution, repeating <n> times\n");
749    printf("  -tm      time execution (matching) only\n");
750    printf("  -tm <n>  time execution (matching) only, repeating <n> times\n");
751    }
752    
753    
754    
755    /*************************************************
756  *                Main Program                    *  *                Main Program                    *
757  *************************************************/  *************************************************/
758    
# Line 650  int options = 0; Line 767  int options = 0;
767  int study_options = 0;  int study_options = 0;
768  int op = 1;  int op = 1;
769  int timeit = 0;  int timeit = 0;
770    int timeitm = 0;
771  int showinfo = 0;  int showinfo = 0;
772  int showstore = 0;  int showstore = 0;
773  int quiet = 0;  int quiet = 0;
# Line 681  buffer = (unsigned char *)malloc(buffer_ Line 799  buffer = (unsigned char *)malloc(buffer_
799  dbuffer = (unsigned char *)malloc(buffer_size);  dbuffer = (unsigned char *)malloc(buffer_size);
800  pbuffer = (unsigned char *)malloc(buffer_size);  pbuffer = (unsigned char *)malloc(buffer_size);
801    
802  /* 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) */  
803    
804  outfile = stdout;  outfile = stdout;
805    
806    /* The following  _setmode() stuff is some Windows magic that tells its runtime
807    library to translate CRLF into a single LF character. At least, that's what
808    I've been told: never having used Windows I take this all on trust. Originally
809    it set 0x8000, but then I was advised that _O_BINARY was better. */
810    
811    #if defined(_WIN32) || defined(WIN32)
812    _setmode( _fileno( stdout ), _O_BINARY );
813    #endif
814    
815  /* Scan options */  /* Scan options */
816    
817  while (argc > 1 && argv[op][0] == '-')  while (argc > 1 && argv[op][0] == '-')
# Line 699  while (argc > 1 && argv[op][0] == '-') Line 820  while (argc > 1 && argv[op][0] == '-')
820    
821    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)    if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)
822      showstore = 1;      showstore = 1;
   else if (strcmp(argv[op], "-t") == 0) timeit = 1;  
823    else if (strcmp(argv[op], "-q") == 0) quiet = 1;    else if (strcmp(argv[op], "-q") == 0) quiet = 1;
824      else if (strcmp(argv[op], "-b") == 0) debug = 1;
825    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
826    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
827  #if !defined NODFA  #if !defined NODFA
# Line 713  while (argc > 1 && argv[op][0] == '-') Line 834  while (argc > 1 && argv[op][0] == '-')
834      op++;      op++;
835      argc--;      argc--;
836      }      }
837      else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
838        {
839        int both = argv[op][2] == 0;
840        int temp;
841        if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),
842                         *endptr == 0))
843          {
844          timeitm = temp;
845          op++;
846          argc--;
847          }
848        else timeitm = LOOPREPEAT;
849        if (both) timeit = timeitm;
850        }
851    else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&    else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
852        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),        ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
853          *endptr == 0))          *endptr == 0))
854      {      {
855  #ifdef _WIN32  #if defined(_WIN32) || defined(WIN32)
856      printf("PCRE: -S not supported on this OS\n");      printf("PCRE: -S not supported on this OS\n");
857      exit(1);      exit(1);
858  #else  #else
# Line 749  while (argc > 1 && argv[op][0] == '-') Line 884  while (argc > 1 && argv[op][0] == '-')
884      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
885      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
886      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :
887        (rc == '\n')? "LF" : "CRLF");        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :
888          (rc == -2)? "ANYCRLF" :
889          (rc == -1)? "ANY" : "???");
890        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
891        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
892                                         "all Unicode newlines");
893      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
894      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
895      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
# Line 760  while (argc > 1 && argv[op][0] == '-') Line 900  while (argc > 1 && argv[op][0] == '-')
900      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %d\n", rc);
901      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
902      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
903      exit(0);      goto EXIT;
904        }
905      else if (strcmp(argv[op], "-help") == 0 ||
906               strcmp(argv[op], "--help") == 0)
907        {
908        usage();
909        goto EXIT;
910      }      }
911    else    else
912      {      {
913      printf("** Unknown or malformed option %s\n", argv[op]);      printf("** Unknown or malformed option %s\n", argv[op]);
914      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");  
915      yield = 1;      yield = 1;
916      goto EXIT;      goto EXIT;
917      }      }
# Line 794  offsets = (int *)malloc(size_offsets_max Line 926  offsets = (int *)malloc(size_offsets_max
926  if (offsets == NULL)  if (offsets == NULL)
927    {    {
928    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
929      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
930    yield = 1;    yield = 1;
931    goto EXIT;    goto EXIT;
932    }    }
# Line 803  if (offsets == NULL) Line 935  if (offsets == NULL)
935    
936  if (argc > 1)  if (argc > 1)
937    {    {
938    infile = fopen(argv[op], "rb");    infile = fopen(argv[op], INPUT_MODE);
939    if (infile == NULL)    if (infile == NULL)
940      {      {
941      printf("** Failed to open %s\n", argv[op]);      printf("** Failed to open %s\n", argv[op]);
# Line 814  if (argc > 1) Line 946  if (argc > 1)
946    
947  if (argc > 2)  if (argc > 2)
948    {    {
949    outfile = fopen(argv[op+1], "wb");    outfile = fopen(argv[op+1], OUTPUT_MODE);
950    if (outfile == NULL)    if (outfile == NULL)
951      {      {
952      printf("** Failed to open %s\n", argv[op+1]);      printf("** Failed to open %s\n", argv[op+1]);
# Line 859  while (!done) Line 991  while (!done)
991    int do_showinfo = showinfo;    int do_showinfo = showinfo;
992    int do_showrest = 0;    int do_showrest = 0;
993    int do_flip = 0;    int do_flip = 0;
994    int erroroffset, len, delimiter;    int erroroffset, len, delimiter, poffset;
995    
996    use_utf8 = 0;    use_utf8 = 0;
997      debug_lengths = 1;
998    
999    if (infile == stdin) printf("  re> ");    if (infile == stdin) printf("  re> ");
1000    if (extend_inputline(infile, buffer) == NULL) break;    if (extend_inputline(infile, buffer) == NULL) break;
# Line 969  while (!done) Line 1102  while (!done)
1102      }      }
1103    
1104    pp = p;    pp = p;
1105      poffset = p - buffer;
1106    
1107    for(;;)    for(;;)
1108      {      {
# Line 989  while (!done) Line 1123  while (!done)
1123      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);      if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
1124      }      }
1125    
1126      /* The buffer may have moved while being extended; reset the start of data
1127      pointer to the correct relative point in the buffer. */
1128    
1129      p = buffer + poffset;
1130    
1131    /* If the first character after the delimiter is backslash, make    /* If the first character after the delimiter is backslash, make
1132    the pattern end with backslash. This is purely to provide a way    the pattern end with backslash. This is purely to provide a way
1133    of testing for the error message when a pattern ends with backslash. */    of testing for the error message when a pattern ends with backslash. */
# Line 1020  while (!done) Line 1159  while (!done)
1159    
1160        case '+': do_showrest = 1; break;        case '+': do_showrest = 1; break;
1161        case 'A': options |= PCRE_ANCHORED; break;        case 'A': options |= PCRE_ANCHORED; break;
1162          case 'B': do_debug = 1; break;
1163        case 'C': options |= PCRE_AUTO_CALLOUT; break;        case 'C': options |= PCRE_AUTO_CALLOUT; break;
1164        case 'D': do_debug = do_showinfo = 1; break;        case 'D': do_debug = do_showinfo = 1; break;
1165        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;        case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
# Line 1037  while (!done) Line 1177  while (!done)
1177        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1178        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1179        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1180          case 'Z': debug_lengths = 0; break;
1181        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1182        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1183    
1184        case 'L':        case 'L':
1185        ppp = pp;        ppp = pp;
1186        /* The '\r' test here is so that it works on Windows */        /* The '\r' test here is so that it works on Windows. */
1187        while (*ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;        /* The '0' test is just in case this is an unterminated line. */
1188          while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
1189        *ppp = 0;        *ppp = 0;
1190        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)        if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
1191          {          {
1192          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);          fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
1193          goto SKIP_DATA;          goto SKIP_DATA;
1194          }          }
1195          locale_set = 1;
1196        tables = pcre_maketables();        tables = pcre_maketables();
1197        pp = ppp;        pp = ppp;
1198        break;        break;
# Line 1116  while (!done) Line 1259  while (!done)
1259  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1260    
1261      {      {
1262      if (timeit)      if (timeit > 0)
1263        {        {
1264        register int i;        register int i;
1265        clock_t time_taken;        clock_t time_taken;
1266        clock_t start_time = clock();        clock_t start_time = clock();
1267        for (i = 0; i < LOOPREPEAT; i++)        for (i = 0; i < timeit; i++)
1268          {          {
1269          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);          re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
1270          if (re != NULL) free(re);          if (re != NULL) free(re);
1271          }          }
1272        time_taken = clock() - start_time;        time_taken = clock() - start_time;
1273        fprintf(outfile, "Compile time %.3f milliseconds\n",        fprintf(outfile, "Compile time %.4f milliseconds\n",
1274          (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /          (((double)time_taken * 1000.0) / (double)timeit) /
1275            (double)CLOCKS_PER_SEC);            (double)CLOCKS_PER_SEC);
1276        }        }
1277    
# Line 1180  while (!done) Line 1323  while (!done)
1323    
1324      if (do_study)      if (do_study)
1325        {        {
1326        if (timeit)        if (timeit > 0)
1327          {          {
1328          register int i;          register int i;
1329          clock_t time_taken;          clock_t time_taken;
1330          clock_t start_time = clock();          clock_t start_time = clock();
1331          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeit; i++)
1332            extra = pcre_study(re, study_options, &error);            extra = pcre_study(re, study_options, &error);
1333          time_taken = clock() - start_time;          time_taken = clock() - start_time;
1334          if (extra != NULL) free(extra);          if (extra != NULL) free(extra);
1335          fprintf(outfile, "  Study time %.3f milliseconds\n",          fprintf(outfile, "  Study time %.4f milliseconds\n",
1336            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeit) /
1337              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
1338          }          }
1339        extra = pcre_study(re, study_options, &error);        extra = pcre_study(re, study_options, &error);
# Line 1211  while (!done) Line 1354  while (!done)
1354        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));
1355        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1356        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1357          rre->flags = byteflip(rre->flags, sizeof(rre->flags));
1358        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1359        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));
1360        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));
# Line 1233  while (!done) Line 1377  while (!done)
1377    
1378      SHOW_INFO:      SHOW_INFO:
1379    
1380        if (do_debug)
1381          {
1382          fprintf(outfile, "------------------------------------------------------------------\n");
1383          pcre_printint(re, outfile, debug_lengths);
1384          }
1385    
1386      if (do_showinfo)      if (do_showinfo)
1387        {        {
1388        unsigned long int get_options, all_options;        unsigned long int get_options, all_options;
1389  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1390        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1391  #endif  #endif
1392        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1393            hascrorlf;
1394        int nameentrysize, namecount;        int nameentrysize, namecount;
1395        const uschar *nametable;        const uschar *nametable;
1396    
       if (do_debug)  
         {  
         fprintf(outfile, "------------------------------------------------------------------\n");  
         pcre_printint(re, outfile);  
         }  
   
1397        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1398        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1399        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
# Line 1258  while (!done) Line 1403  while (!done)
1403        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1404        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1405        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1406          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1407          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1408          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1409    
1410  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1411        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
# Line 1299  while (!done) Line 1447  while (!done)
1447            }            }
1448          }          }
1449    
1450        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1451        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1452    
1453        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1454        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");  
1455    
1456        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1457          else fprintf(outfile, "Options:%s%s%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%s%s%s\n",
1458            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1459            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1460            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1461            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1462            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1463            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1464              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1465              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1466            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1467            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1468            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
# Line 1327  while (!done) Line 1471  while (!done)
1471            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1472            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1473    
1474        switch (get_options & PCRE_NEWLINE_CRLF)        if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1475    
1476          switch (get_options & PCRE_NEWLINE_BITS)
1477          {          {
1478          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
1479          fprintf(outfile, "Forced newline sequence: CR\n");          fprintf(outfile, "Forced newline sequence: CR\n");
# Line 1341  while (!done) Line 1487  while (!done)
1487          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
1488          break;          break;
1489    
1490            case PCRE_NEWLINE_ANYCRLF:
1491            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1492            break;
1493    
1494            case PCRE_NEWLINE_ANY:
1495            fprintf(outfile, "Forced newline sequence: ANY\n");
1496            break;
1497    
1498          default:          default:
1499          break;          break;
1500          }          }
# Line 1358  while (!done) Line 1512  while (!done)
1512          int ch = first_char & 255;          int ch = first_char & 255;
1513          const char *caseless = ((first_char & REQ_CASELESS) == 0)?          const char *caseless = ((first_char & REQ_CASELESS) == 0)?
1514            "" : " (caseless)";            "" : " (caseless)";
1515          if (isprint(ch))          if (PRINTHEX(ch))
1516            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);
1517          else          else
1518            fprintf(outfile, "First char = %d%s\n", ch, caseless);            fprintf(outfile, "First char = %d%s\n", ch, caseless);
# Line 1373  while (!done) Line 1527  while (!done)
1527          int ch = need_char & 255;          int ch = need_char & 255;
1528          const char *caseless = ((need_char & REQ_CASELESS) == 0)?          const char *caseless = ((need_char & REQ_CASELESS) == 0)?
1529            "" : " (caseless)";            "" : " (caseless)";
1530          if (isprint(ch))          if (PRINTHEX(ch))
1531            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);            fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);
1532          else          else
1533            fprintf(outfile, "Need char = %d%s\n", ch, caseless);            fprintf(outfile, "Need char = %d%s\n", ch, caseless);
# Line 1409  while (!done) Line 1563  while (!done)
1563                    fprintf(outfile, "\n  ");                    fprintf(outfile, "\n  ");
1564                    c = 2;                    c = 2;
1565                    }                    }
1566                  if (isprint(i) && i != ' ')                  if (PRINTHEX(i) && i != ' ')
1567                    {                    {
1568                    fprintf(outfile, "%c ", i);                    fprintf(outfile, "%c ", i);
1569                    c += 2;                    c += 2;
# Line 1468  while (!done) Line 1622  while (!done)
1622                  strerror(errno));                  strerror(errno));
1623                }                }
1624              else fprintf(outfile, "Study data written to %s\n", to_file);              else fprintf(outfile, "Study data written to %s\n", to_file);
1625    
1626              }              }
1627            }            }
1628          fclose(f);          fclose(f);
# Line 1485  while (!done) Line 1640  while (!done)
1640    for (;;)    for (;;)
1641      {      {
1642      uschar *q;      uschar *q;
1643      uschar *bptr = dbuffer;      uschar *bptr;
1644      int *use_offsets = offsets;      int *use_offsets = offsets;
1645      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
1646      int callout_data = 0;      int callout_data = 0;
# Line 1541  while (!done) Line 1696  while (!done)
1696      p = buffer;      p = buffer;
1697      while (isspace(*p)) p++;      while (isspace(*p)) p++;
1698    
1699      q = dbuffer;      bptr = q = dbuffer;
1700      while ((c = *p++) != 0)      while ((c = *p++) != 0)
1701        {        {
1702        int i = 0;        int i = 0;
# Line 1736  while (!done) Line 1891  while (!done)
1891            if (offsets == NULL)            if (offsets == NULL)
1892              {              {
1893              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
1894                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
1895              yield = 1;              yield = 1;
1896              goto EXIT;              goto EXIT;
1897              }              }
# Line 1866  while (!done) Line 2021  while (!done)
2021    
2022      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2023        {        {
2024        if (timeit)        if (timeitm > 0)
2025          {          {
2026          register int i;          register int i;
2027          clock_t time_taken;          clock_t time_taken;
# Line 1876  while (!done) Line 2031  while (!done)
2031          if (all_use_dfa || use_dfa)          if (all_use_dfa || use_dfa)
2032            {            {
2033            int workspace[1000];            int workspace[1000];
2034            for (i = 0; i < LOOPREPEAT; i++)            for (i = 0; i < timeitm; i++)
2035              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,              count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,
2036                options | g_notempty, use_offsets, use_size_offsets, workspace,                options | g_notempty, use_offsets, use_size_offsets, workspace,
2037                sizeof(workspace)/sizeof(int));                sizeof(workspace)/sizeof(int));
# Line 1884  while (!done) Line 2039  while (!done)
2039          else          else
2040  #endif  #endif
2041    
2042          for (i = 0; i < LOOPREPEAT; i++)          for (i = 0; i < timeitm; i++)
2043            count = pcre_exec(re, extra, (char *)bptr, len,            count = pcre_exec(re, extra, (char *)bptr, len,
2044              start_offset, options | g_notempty, use_offsets, use_size_offsets);              start_offset, options | g_notempty, use_offsets, use_size_offsets);
2045    
2046          time_taken = clock() - start_time;          time_taken = clock() - start_time;
2047          fprintf(outfile, "Execute time %.3f milliseconds\n",          fprintf(outfile, "Execute time %.4f milliseconds\n",
2048            (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /            (((double)time_taken * 1000.0) / (double)timeitm) /
2049              (double)CLOCKS_PER_SEC);              (double)CLOCKS_PER_SEC);
2050          }          }
2051    
# Line 1966  while (!done) Line 2121  while (!done)
2121    
2122        if (count >= 0)        if (count >= 0)
2123          {          {
2124          int i;          int i, maxcount;
2125    
2126    #if !defined NODFA
2127            if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
2128    #endif
2129              maxcount = use_size_offsets/3;
2130    
2131            /* This is a check against a lunatic return value. */
2132    
2133            if (count > maxcount)
2134              {
2135              fprintf(outfile,
2136                "** PCRE error: returned count %d is too big for offset size %d\n",
2137                count, use_size_offsets);
2138              count = use_size_offsets/3;
2139              if (do_g || do_G)
2140                {
2141                fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
2142                do_g = do_G = FALSE;        /* Break g/G loop */
2143                }
2144              }
2145    
2146          for (i = 0; i < count * 2; i += 2)          for (i = 0; i < count * 2; i += 2)
2147            {            {
2148            if (use_offsets[i] < 0)            if (use_offsets[i] < 0)
# Line 2084  while (!done) Line 2260  while (!done)
2260          }          }
2261    
2262        /* Failed to match. If this is a /g or /G loop and we previously set        /* Failed to match. If this is a /g or /G loop and we previously set
2263        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
2264        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
2265        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2266        offset values to achieve this. We won't be at the end of the string -  
2267        that was checked before setting g_notempty. */        Complication arises in the case when the newline option is "any" or
2268          "anycrlf". If the previous match was at the end of a line terminated by
2269          CRLF, an advance of one character just passes the \r, whereas we should
2270          prefer the longer newline sequence, as does the code in pcre_exec().
2271          Fudge the offset value to achieve this.
2272    
2273          Otherwise, in the case of UTF-8 matching, the advance must be one
2274          character, not one byte. */
2275    
2276        else        else
2277          {          {
2278          if (g_notempty != 0)          if (g_notempty != 0)
2279            {            {
2280            int onechar = 1;            int onechar = 1;
2281              unsigned int obits = ((real_pcre *)re)->options;
2282            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2283            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2284                {
2285                int d;
2286                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2287                obits = (d == '\r')? PCRE_NEWLINE_CR :
2288                        (d == '\n')? PCRE_NEWLINE_LF :
2289                        (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :
2290                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2291                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2292                }
2293              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2294                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2295                  &&
2296                  start_offset < len - 1 &&
2297                  bptr[start_offset] == '\r' &&
2298                  bptr[start_offset+1] == '\n')
2299                onechar++;
2300              else if (use_utf8)
2301              {              {
2302              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2303                {                {
# Line 2131  while (!done) Line 2332  while (!done)
2332        character. */        character. */
2333    
2334        g_notempty = 0;        g_notempty = 0;
2335    
2336        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2337          {          {
2338          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
# Line 2165  while (!done) Line 2367  while (!done)
2367      {      {
2368      new_free((void *)tables);      new_free((void *)tables);
2369      setlocale(LC_CTYPE, "C");      setlocale(LC_CTYPE, "C");
2370        locale_set = 0;
2371      }      }
2372    }    }
2373    

Legend:
Removed from v.91  
changed lines
  Added in v.236

  ViewVC Help
Powered by ViewVC 1.1.5