/[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 107 by ph10, Wed Mar 7 11:02:28 2007 UTC revision 442 by ph10, Fri Sep 11 10:21:02 2009 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    #ifdef SUPPORT_LIBREADLINE
52    #ifdef HAVE_UNISTD_H
53    #include <unistd.h>
54    #endif
55    #include <readline/readline.h>
56    #include <readline/history.h>
57    #endif
58    
59    
60  /* A number of things vary for Windows builds. Originally, pcretest opened its  /* A number of things vary for Windows builds. Originally, pcretest opened its
61  input and output without "b"; then I was told that "b" was needed in some  input and output without "b"; then I was told that "b" was needed in some
# Line 59  input mode under Windows. */ Line 71  input mode under Windows. */
71  #define INPUT_MODE   "r"  #define INPUT_MODE   "r"
72  #define OUTPUT_MODE  "wb"  #define OUTPUT_MODE  "wb"
73    
74    #ifndef isatty
75    #define isatty _isatty         /* This is what Windows calls them, I'm told, */
76    #endif                         /* though in some environments they seem to   */
77                                   /* be already defined, hence the #ifndefs.    */
78    #ifndef fileno
79    #define fileno _fileno
80    #endif
81    
82  #else  #else
83  #include <sys/time.h>          /* These two includes are needed */  #include <sys/time.h>          /* These two includes are needed */
84  #include <sys/resource.h>      /* for setrlimit(). */  #include <sys/resource.h>      /* for setrlimit(). */
# Line 67  input mode under Windows. */ Line 87  input mode under Windows. */
87  #endif  #endif
88    
89    
90  #define PCRE_SPY        /* For Win32 build, import data, not export */  /* We have to include pcre_internal.h because we need the internal info for
91    displaying the results of pcre_study() and we also need to know about the
92  /* We include pcre_internal.h because we need the internal info for displaying  internal macros, structures, and other internal data values; pcretest has
93  the results of pcre_study() and we also need to know about the internal  "inside information" compared to a program that strictly follows the PCRE API.
94  macros, structures, and other internal data values; pcretest has "inside  
95  information" compared to a program that strictly follows the PCRE API. */  Although pcre_internal.h does itself include pcre.h, we explicitly include it
96    here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
97    appropriately for an application, not for building PCRE. */
98    
99    #include "pcre.h"
100  #include "pcre_internal.h"  #include "pcre_internal.h"
101    
102  /* We need access to the data tables that PCRE uses. So as not to have to keep  /* We need access to some of the data tables that PCRE uses. So as not to have
103  two copies, we include the source file here, changing the names of the external  to keep two copies, we include the source file here, changing the names of the
104  symbols to prevent clashes. */  external symbols to prevent clashes. */
105    
106    #define _pcre_ucp_gentype      ucp_gentype
107  #define _pcre_utf8_table1      utf8_table1  #define _pcre_utf8_table1      utf8_table1
108  #define _pcre_utf8_table1_size utf8_table1_size  #define _pcre_utf8_table1_size utf8_table1_size
109  #define _pcre_utf8_table2      utf8_table2  #define _pcre_utf8_table2      utf8_table2
# Line 87  symbols to prevent clashes. */ Line 111  symbols to prevent clashes. */
111  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
112  #define _pcre_utt              utt  #define _pcre_utt              utt
113  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
114    #define _pcre_utt_names        utt_names
115  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
116    
117  #include "pcre_tables.c"  #include "pcre_tables.c"
# Line 149  static int callout_count; Line 174  static int callout_count;
174  static int callout_extra;  static int callout_extra;
175  static int callout_fail_count;  static int callout_fail_count;
176  static int callout_fail_id;  static int callout_fail_id;
177    static int debug_lengths;
178  static int first_callout;  static int first_callout;
179  static int locale_set = 0;  static int locale_set = 0;
180  static int show_malloc;  static int show_malloc;
# Line 180  optimal way of handling this, but hey, t Line 206  optimal way of handling this, but hey, t
206  Arguments:  Arguments:
207    f            the file to read    f            the file to read
208    start        where in buffer to start (this *must* be within buffer)    start        where in buffer to start (this *must* be within buffer)
209      prompt       for stdin or readline()
210    
211  Returns:       pointer to the start of new data  Returns:       pointer to the start of new data
212                 could be a copy of start, or could be moved                 could be a copy of start, or could be moved
# Line 187  Returns:       pointer to the start of n Line 214  Returns:       pointer to the start of n
214  */  */
215    
216  static uschar *  static uschar *
217  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, uschar *start, const char *prompt)
218  {  {
219  uschar *here = start;  uschar *here = start;
220    
# Line 198  for (;;) Line 225  for (;;)
225    if (rlen > 1000)    if (rlen > 1000)
226      {      {
227      int dlen;      int dlen;
228      if (fgets((char *)here, rlen,  f) == NULL)  
229        return (here == start)? NULL : start;      /* If libreadline support is required, use readline() to read a line if the
230        input is a terminal. Note that readline() removes the trailing newline, so
231        we must put it back again, to be compatible with fgets(). */
232    
233    #ifdef SUPPORT_LIBREADLINE
234        if (isatty(fileno(f)))
235          {
236          size_t len;
237          char *s = readline(prompt);
238          if (s == NULL) return (here == start)? NULL : start;
239          len = strlen(s);
240          if (len > 0) add_history(s);
241          if (len > rlen - 1) len = rlen - 1;
242          memcpy(here, s, len);
243          here[len] = '\n';
244          here[len+1] = 0;
245          free(s);
246          }
247        else
248    #endif
249    
250        /* Read the next line by normal means, prompting if the file is stdin. */
251    
252          {
253          if (f == stdin) printf(prompt);
254          if (fgets((char *)here, rlen,  f) == NULL)
255            return (here == start)? NULL : start;
256          }
257    
258      dlen = (int)strlen((char *)here);      dlen = (int)strlen((char *)here);
259      if (dlen > 0 && here[dlen - 1] == '\n') return start;      if (dlen > 0 && here[dlen - 1] == '\n') return start;
260      here += dlen;      here += dlen;
# Line 656  return count; Line 711  return count;
711    
712    
713  /*************************************************  /*************************************************
714    *         Case-independent strncmp() function    *
715    *************************************************/
716    
717    /*
718    Arguments:
719      s         first string
720      t         second string
721      n         number of characters to compare
722    
723    Returns:    < 0, = 0, or > 0, according to the comparison
724    */
725    
726    static int
727    strncmpic(uschar *s, uschar *t, int n)
728    {
729    while (n--)
730      {
731      int c = tolower(*s++) - tolower(*t++);
732      if (c) return c;
733      }
734    return 0;
735    }
736    
737    
738    
739    /*************************************************
740  *         Check newline indicator                *  *         Check newline indicator                *
741  *************************************************/  *************************************************/
742    
743  /* 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
744  xxx is LF, CR, CRLF, or ANY. 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
745    no match.
746    
747  Arguments:  Arguments:
748    p           points after the leading '<'    p           points after the leading '<'
# Line 672  Returns:      appropriate PCRE_NEWLINE_x Line 754  Returns:      appropriate PCRE_NEWLINE_x
754  static int  static int
755  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
756  {  {
757  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
758  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
759  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
760  if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
761    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
762    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
763    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
764  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
765  return 0;  return 0;
766  }  }
# Line 689  return 0; Line 774  return 0;
774  static void  static void
775  usage(void)  usage(void)
776  {  {
777  printf("Usage:     pcretest [options] [<input> [<output>]]\n");  printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
778    printf("Input and output default to stdin and stdout.\n");
779    #ifdef SUPPORT_LIBREADLINE
780    printf("If input is a terminal, readline() is used to read from it.\n");
781    #else
782    printf("This version of pcretest is not linked with readline().\n");
783    #endif
784    printf("\nOptions:\n");
785  printf("  -b       show compiled code (bytecode)\n");  printf("  -b       show compiled code (bytecode)\n");
786  printf("  -C       show PCRE compile-time options and exit\n");  printf("  -C       show PCRE compile-time options and exit\n");
787  printf("  -d       debug: show compiled code and information (-b and -i)\n");  printf("  -d       debug: show compiled code and information (-b and -i)\n");
# Line 698  printf("  -dfa     force DFA matching fo Line 790  printf("  -dfa     force DFA matching fo
790  #endif  #endif
791  printf("  -help    show usage information\n");  printf("  -help    show usage information\n");
792  printf("  -i       show information about compiled patterns\n"  printf("  -i       show information about compiled patterns\n"
793           "  -M       find MATCH_LIMIT minimum for each subject\n"
794         "  -m       output memory used information\n"         "  -m       output memory used information\n"
795         "  -o <n>   set size of offsets vector to <n>\n");         "  -o <n>   set size of offsets vector to <n>\n");
796  #if !defined NOPOSIX  #if !defined NOPOSIX
# Line 727  int main(int argc, char **argv) Line 820  int main(int argc, char **argv)
820  FILE *infile = stdin;  FILE *infile = stdin;
821  int options = 0;  int options = 0;
822  int study_options = 0;  int study_options = 0;
823    int default_find_match_limit = FALSE;
824  int op = 1;  int op = 1;
825  int timeit = 0;  int timeit = 0;
826  int timeitm = 0;  int timeitm = 0;
# Line 786  while (argc > 1 && argv[op][0] == '-') Line 880  while (argc > 1 && argv[op][0] == '-')
880    else if (strcmp(argv[op], "-b") == 0) debug = 1;    else if (strcmp(argv[op], "-b") == 0) debug = 1;
881    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;    else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
882    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;    else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
883      else if (strcmp(argv[op], "-M") == 0) default_find_match_limit = TRUE;
884  #if !defined NODFA  #if !defined NODFA
885    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;    else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1;
886  #endif  #endif
# Line 838  while (argc > 1 && argv[op][0] == '-') Line 933  while (argc > 1 && argv[op][0] == '-')
933    else if (strcmp(argv[op], "-C") == 0)    else if (strcmp(argv[op], "-C") == 0)
934      {      {
935      int rc;      int rc;
936        unsigned long int lrc;
937      printf("PCRE version %s\n", pcre_version());      printf("PCRE version %s\n", pcre_version());
938      printf("Compiled with\n");      printf("Compiled with\n");
939      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);      (void)pcre_config(PCRE_CONFIG_UTF8, &rc);
# Line 845  while (argc > 1 && argv[op][0] == '-') Line 941  while (argc > 1 && argv[op][0] == '-')
941      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);      (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
942      printf("  %sUnicode properties support\n", rc? "" : "No ");      printf("  %sUnicode properties support\n", rc? "" : "No ");
943      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
944      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      /* Note that these values are always the ASCII values, even
945        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :      in EBCDIC environments. CR is 13 and NL is 10. */
946        printf("  Newline sequence is %s\n", (rc == 13)? "CR" :
947          (rc == 10)? "LF" : (rc == (13<<8 | 10))? "CRLF" :
948          (rc == -2)? "ANYCRLF" :
949        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
950        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
951        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
952                                         "all Unicode newlines");
953      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
954      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
955      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
956      printf("  POSIX malloc threshold = %d\n", rc);      printf("  POSIX malloc threshold = %d\n", rc);
957      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &lrc);
958      printf("  Default match limit = %d\n", rc);      printf("  Default match limit = %ld\n", lrc);
959      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);      (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &lrc);
960      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %ld\n", lrc);
961      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
962      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
963      exit(0);      goto EXIT;
964      }      }
965    else if (strcmp(argv[op], "-help") == 0 ||    else if (strcmp(argv[op], "-help") == 0 ||
966             strcmp(argv[op], "--help") == 0)             strcmp(argv[op], "--help") == 0)
# Line 884  offsets = (int *)malloc(size_offsets_max Line 986  offsets = (int *)malloc(size_offsets_max
986  if (offsets == NULL)  if (offsets == NULL)
987    {    {
988    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
989      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
990    yield = 1;    yield = 1;
991    goto EXIT;    goto EXIT;
992    }    }
# Line 952  while (!done) Line 1054  while (!done)
1054    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1055    
1056    use_utf8 = 0;    use_utf8 = 0;
1057      debug_lengths = 1;
1058    
1059    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1060    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1061    fflush(outfile);    fflush(outfile);
1062    
# Line 1054  while (!done) Line 1156  while (!done)
1156    
1157    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1158      {      {
1159      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1160      goto SKIP_DATA;      goto SKIP_DATA;
1161      }      }
1162    
# Line 1070  while (!done) Line 1172  while (!done)
1172        pp++;        pp++;
1173        }        }
1174      if (*pp != 0) break;      if (*pp != 0) break;
1175      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1176        {        {
1177        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1178        done = 1;        done = 1;
# Line 1134  while (!done) Line 1235  while (!done)
1235        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1236        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1237        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1238          case 'Z': debug_lengths = 0; break;
1239        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1240        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1241    
# Line 1162  while (!done) Line 1264  while (!done)
1264    
1265        case '<':        case '<':
1266          {          {
1267          int x = check_newline(pp, outfile);          if (strncmp((char *)pp, "JS>", 3) == 0)
1268          if (x == 0) goto SKIP_DATA;            {
1269          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1270          while (*pp++ != '>');            pp += 3;
1271              }
1272            else
1273              {
1274              int x = check_newline(pp, outfile);
1275              if (x == 0) goto SKIP_DATA;
1276              options |= x;
1277              while (*pp++ != '>');
1278              }
1279          }          }
1280        break;        break;
1281    
# Line 1195  while (!done) Line 1305  while (!done)
1305      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;      if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
1306      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;      if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
1307      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;      if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
1308        if ((options & PCRE_UNGREEDY) != 0) cflags |= REG_UNGREEDY;
1309    
1310      rc = regcomp(&preg, (char *)p, cflags);      rc = regcomp(&preg, (char *)p, cflags);
1311    
# Line 1215  while (!done) Line 1326  while (!done)
1326  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1327    
1328      {      {
1329        unsigned long int get_options;
1330    
1331      if (timeit > 0)      if (timeit > 0)
1332        {        {
1333        register int i;        register int i;
# Line 1244  while (!done) Line 1357  while (!done)
1357          {          {
1358          for (;;)          for (;;)
1359            {            {
1360            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1361              {              {
1362              done = 1;              done = 1;
1363              goto CONTINUE;              goto CONTINUE;
# Line 1258  while (!done) Line 1371  while (!done)
1371        goto CONTINUE;        goto CONTINUE;
1372        }        }
1373    
1374      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1375      info-returning functions. The old one has a limited interface and      within the regex; check for this so that we know how to process the data
1376      returns only limited data. Check that it agrees with the newer one. */      lines. */
1377    
1378        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1379        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1380    
1381        /* Print information if required. There are now two info-returning
1382        functions. The old one has a limited interface and returns only limited
1383        data. Check that it agrees with the newer one. */
1384    
1385      if (log_store)      if (log_store)
1386        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1307  while (!done) Line 1427  while (!done)
1427      if (do_flip)      if (do_flip)
1428        {        {
1429        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1430        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1431            byteflip(rre->magic_number, sizeof(rre->magic_number));
1432        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1433        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1434        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1435        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1436        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1437        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1438        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1439          rre->first_byte =
1440            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1441          rre->req_byte =
1442            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1443          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1444          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1445        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1446          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1447        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1448            sizeof(rre->name_count));
1449    
1450        if (extra != NULL)        if (extra != NULL)
1451          {          {
# Line 1335  while (!done) Line 1462  while (!done)
1462      if (do_debug)      if (do_debug)
1463        {        {
1464        fprintf(outfile, "------------------------------------------------------------------\n");        fprintf(outfile, "------------------------------------------------------------------\n");
1465        pcre_printint(re, outfile);        pcre_printint(re, outfile, debug_lengths);
1466        }        }
1467    
1468        /* We already have the options in get_options (see above) */
1469    
1470      if (do_showinfo)      if (do_showinfo)
1471        {        {
1472        unsigned long int get_options, all_options;        unsigned long int all_options;
1473  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1474        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1475  #endif  #endif
1476        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1477            hascrorlf;
1478        int nameentrysize, namecount;        int nameentrysize, namecount;
1479        const uschar *nametable;        const uschar *nametable;
1480    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1481        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1482        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1483        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1357  while (!done) Line 1486  while (!done)
1486        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1487        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1488        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1489          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1490          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1491          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1492    
1493  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1494        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
# Line 1398  while (!done) Line 1530  while (!done)
1530            }            }
1531          }          }
1532    
1533        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1534        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1535    
1536        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1537        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");  
1538    
1539        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1540          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",
1541            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1542            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1543            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1544            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1545            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1546            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1547              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1548              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1549            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1550            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1551            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
# Line 1426  while (!done) Line 1554  while (!done)
1554            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1555            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1556    
1557          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1558    
1559        switch (get_options & PCRE_NEWLINE_BITS)        switch (get_options & PCRE_NEWLINE_BITS)
1560          {          {
1561          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
# Line 1440  while (!done) Line 1570  while (!done)
1570          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
1571          break;          break;
1572    
1573            case PCRE_NEWLINE_ANYCRLF:
1574            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1575            break;
1576    
1577          case PCRE_NEWLINE_ANY:          case PCRE_NEWLINE_ANY:
1578          fprintf(outfile, "Forced newline sequence: ANY\n");          fprintf(outfile, "Forced newline sequence: ANY\n");
1579          break;          break;
# Line 1544  while (!done) Line 1678  while (!done)
1678        else        else
1679          {          {
1680          uschar sbuf[8];          uschar sbuf[8];
1681          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
1682          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
1683          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
1684          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
1685    
1686          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
1687          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
1688          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
1689          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
1690    
1691          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
1692              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1589  while (!done) Line 1723  while (!done)
1723    for (;;)    for (;;)
1724      {      {
1725      uschar *q;      uschar *q;
1726      uschar *bptr = dbuffer;      uschar *bptr;
1727      int *use_offsets = offsets;      int *use_offsets = offsets;
1728      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
1729      int callout_data = 0;      int callout_data = 0;
1730      int callout_data_set = 0;      int callout_data_set = 0;
1731      int count, c;      int count, c;
1732      int copystrings = 0;      int copystrings = 0;
1733      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
1734      int getstrings = 0;      int getstrings = 0;
1735      int getlist = 0;      int getlist = 0;
1736      int gmatched = 0;      int gmatched = 0;
# Line 1626  while (!done) Line 1760  while (!done)
1760      len = 0;      len = 0;
1761      for (;;)      for (;;)
1762        {        {
1763        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
1764          {          {
1765          if (len > 0) break;          if (len > 0) break;
1766          done = 1;          done = 1;
# Line 1645  while (!done) Line 1778  while (!done)
1778      p = buffer;      p = buffer;
1779      while (isspace(*p)) p++;      while (isspace(*p)) p++;
1780    
1781      q = dbuffer;      bptr = q = dbuffer;
1782      while ((c = *p++) != 0)      while ((c = *p++) != 0)
1783        {        {
1784        int i = 0;        int i = 0;
# Line 1695  while (!done) Line 1828  while (!done)
1828              {              {
1829              unsigned char buff8[8];              unsigned char buff8[8];
1830              int ii, utn;              int ii, utn;
1831              utn = ord2utf8(c, buff8);              if (use_utf8)
1832              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
1833              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
1834                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
1835                  c = buff8[ii];   /* Last byte */
1836                  }
1837                else
1838                 {
1839                 if (c > 255)
1840                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
1841                     "UTF-8 mode is not enabled.\n"
1842                     "** Truncation will probably give the wrong result.\n", c);
1843                 }
1844              p = pt + 1;              p = pt + 1;
1845              break;              break;
1846              }              }
# Line 1827  while (!done) Line 1970  while (!done)
1970          continue;          continue;
1971    
1972          case 'N':          case 'N':
1973          options |= PCRE_NOTEMPTY;          if ((options & PCRE_NOTEMPTY) != 0)
1974              options = (options & ~PCRE_NOTEMPTY) | PCRE_NOTEMPTY_ATSTART;
1975            else
1976              options |= PCRE_NOTEMPTY;
1977          continue;          continue;
1978    
1979          case 'O':          case 'O':
# Line 1840  while (!done) Line 1986  while (!done)
1986            if (offsets == NULL)            if (offsets == NULL)
1987              {              {
1988              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
1989                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
1990              yield = 1;              yield = 1;
1991              goto EXIT;              goto EXIT;
1992              }              }
# Line 1850  while (!done) Line 1996  while (!done)
1996          continue;          continue;
1997    
1998          case 'P':          case 'P':
1999          options |= PCRE_PARTIAL;          options |= ((options & PCRE_PARTIAL_SOFT) == 0)?
2000              PCRE_PARTIAL_SOFT : PCRE_PARTIAL_HARD;
2001          continue;          continue;
2002    
2003          case 'Q':          case 'Q':
# Line 1885  while (!done) Line 2032  while (!done)
2032          show_malloc = 1;          show_malloc = 1;
2033          continue;          continue;
2034    
2035            case 'Y':
2036            options |= PCRE_NO_START_OPTIMIZE;
2037            continue;
2038    
2039          case 'Z':          case 'Z':
2040          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2041          continue;          continue;
# Line 1907  while (!done) Line 2058  while (!done)
2058      *q = 0;      *q = 0;
2059      len = q - dbuffer;      len = q - dbuffer;
2060    
2061        /* Move the data to the end of the buffer so that a read over the end of
2062        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2063        we are using the POSIX interface, we must include the terminating zero. */
2064    
2065    #if !defined NOPOSIX
2066        if (posix || do_posix)
2067          {
2068          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2069          bptr += buffer_size - len - 1;
2070          }
2071        else
2072    #endif
2073          {
2074          memmove(bptr + buffer_size - len, bptr, len);
2075          bptr += buffer_size - len;
2076          }
2077    
2078      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2079        {        {
2080        printf("**Match limit not relevant for DFA matching: ignored\n");        printf("**Match limit not relevant for DFA matching: ignored\n");
# Line 1926  while (!done) Line 2094  while (!done)
2094          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2095        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2096        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2097          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2098    
2099        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2100    
# Line 2199  while (!done) Line 2368  while (!done)
2368        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2369          {          {
2370          fprintf(outfile, "Partial match");          fprintf(outfile, "Partial match");
2371  #if !defined NODFA          if (use_size_offsets > 1)
2372          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)            {
2373            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            fprintf(outfile, ": ");
2374              bptr + use_offsets[0]);            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2375  #endif              outfile);
2376              }
2377          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2378          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2379          }          }
2380    
2381        /* 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
2382        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
2383        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
2384        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2385        offset values to achieve this. We won't be at the end of the string -  
2386        that was checked before setting g_notempty. */        Complication arises in the case when the newline option is "any" or
2387          "anycrlf". If the previous match was at the end of a line terminated by
2388          CRLF, an advance of one character just passes the \r, whereas we should
2389          prefer the longer newline sequence, as does the code in pcre_exec().
2390          Fudge the offset value to achieve this.
2391    
2392          Otherwise, in the case of UTF-8 matching, the advance must be one
2393          character, not one byte. */
2394    
2395        else        else
2396          {          {
2397          if (g_notempty != 0)          if (g_notempty != 0)
2398            {            {
2399            int onechar = 1;            int onechar = 1;
2400              unsigned int obits = ((real_pcre *)re)->options;
2401            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2402            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2403                {
2404                int d;
2405                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2406                /* Note that these values are always the ASCII ones, even in
2407                EBCDIC environments. CR = 13, NL = 10. */
2408                obits = (d == 13)? PCRE_NEWLINE_CR :
2409                        (d == 10)? PCRE_NEWLINE_LF :
2410                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2411                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2412                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2413                }
2414              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2415                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2416                  &&
2417                  start_offset < len - 1 &&
2418                  bptr[start_offset] == '\r' &&
2419                  bptr[start_offset+1] == '\n')
2420                onechar++;
2421              else if (use_utf8)
2422              {              {
2423              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2424                {                {
# Line 2249  while (!done) Line 2446  while (!done)
2446        if (!do_g && !do_G) break;        if (!do_g && !do_G) break;
2447    
2448        /* If we have matched an empty string, first check to see if we are at        /* If we have matched an empty string, first check to see if we are at
2449        the end of the subject. If so, the /g loop is over. Otherwise, mimic        the end of the subject. If so, the /g loop is over. Otherwise, mimic what
2450        what Perl's /g options does. This turns out to be rather cunning. First        Perl's /g options does. This turns out to be rather cunning. First we set
2451        we set PCRE_NOTEMPTY and PCRE_ANCHORED and try the match again at the        PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED and try the match again at the
2452        same point. If this fails (picked up above) we advance to the next        same point. If this fails (picked up above) we advance to the next
2453        character. */        character. */
2454    
2455        g_notempty = 0;        g_notempty = 0;
2456    
2457        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2458          {          {
2459          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2460          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY_ATSTART | PCRE_ANCHORED;
2461          }          }
2462    
2463        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */

Legend:
Removed from v.107  
changed lines
  Added in v.442

  ViewVC Help
Powered by ViewVC 1.1.5