/[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 142 by ph10, Fri Mar 30 15:55:18 2007 UTC revision 426 by ph10, Wed Aug 26 15:38:32 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      goto EXIT;      goto EXIT;
# 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 944  while (!done) Line 1046  while (!done)
1046    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1047    int do_study = 0;    int do_study = 0;
1048    int do_debug = debug;    int do_debug = debug;
   int debug_lengths = 1;  
1049    int do_G = 0;    int do_G = 0;
1050    int do_g = 0;    int do_g = 0;
1051    int do_showinfo = showinfo;    int do_showinfo = showinfo;
# Line 953  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 1055  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 1071  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 1164  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 1217  while (!done) Line 1325  while (!done)
1325  #endif  /* !defined NOPOSIX */  #endif  /* !defined NOPOSIX */
1326    
1327      {      {
1328        unsigned long int get_options;
1329    
1330      if (timeit > 0)      if (timeit > 0)
1331        {        {
1332        register int i;        register int i;
# Line 1246  while (!done) Line 1356  while (!done)
1356          {          {
1357          for (;;)          for (;;)
1358            {            {
1359            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1360              {              {
1361              done = 1;              done = 1;
1362              goto CONTINUE;              goto CONTINUE;
# Line 1260  while (!done) Line 1370  while (!done)
1370        goto CONTINUE;        goto CONTINUE;
1371        }        }
1372    
1373      /* Compilation succeeded; print data if required. There are now two      /* Compilation succeeded. It is now possible to set the UTF-8 option from
1374      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
1375      returns only limited data. Check that it agrees with the newer one. */      lines. */
1376    
1377        new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
1378        if ((get_options & PCRE_UTF8) != 0) use_utf8 = 1;
1379    
1380        /* Print information if required. There are now two info-returning
1381        functions. The old one has a limited interface and returns only limited
1382        data. Check that it agrees with the newer one. */
1383    
1384      if (log_store)      if (log_store)
1385        fprintf(outfile, "Memory allocation (code space): %d\n",        fprintf(outfile, "Memory allocation (code space): %d\n",
# Line 1309  while (!done) Line 1426  while (!done)
1426      if (do_flip)      if (do_flip)
1427        {        {
1428        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1429        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1430            byteflip(rre->magic_number, sizeof(rre->magic_number));
1431        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1432        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1433        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1434        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1435        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1436        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1437        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1438          rre->first_byte =
1439            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1440          rre->req_byte =
1441            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1442          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1443          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1444        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1445          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1446        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1447            sizeof(rre->name_count));
1448    
1449        if (extra != NULL)        if (extra != NULL)
1450          {          {
# Line 1340  while (!done) Line 1464  while (!done)
1464        pcre_printint(re, outfile, debug_lengths);        pcre_printint(re, outfile, debug_lengths);
1465        }        }
1466    
1467        /* We already have the options in get_options (see above) */
1468    
1469      if (do_showinfo)      if (do_showinfo)
1470        {        {
1471        unsigned long int get_options, all_options;        unsigned long int all_options;
1472  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1473        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1474  #endif  #endif
1475        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1476            hascrorlf;
1477        int nameentrysize, namecount;        int nameentrysize, namecount;
1478        const uschar *nametable;        const uschar *nametable;
1479    
       new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);  
1480        new_info(re, NULL, PCRE_INFO_SIZE, &size);        new_info(re, NULL, PCRE_INFO_SIZE, &size);
1481        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);        new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
1482        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);        new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax);
# Line 1359  while (!done) Line 1485  while (!done)
1485        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1486        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1487        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1488          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1489          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1490          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1491    
1492  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1493        old_count = pcre_info(re, &old_options, &old_first_char);        old_count = pcre_info(re, &old_options, &old_first_char);
# Line 1400  while (!done) Line 1529  while (!done)
1529            }            }
1530          }          }
1531    
1532        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1533        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1534    
1535        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1536        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");  
1537    
1538        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1539          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",
1540            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1541            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1542            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1543            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1544            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1545            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1546              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1547              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1548            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1549            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1550            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
# Line 1428  while (!done) Line 1553  while (!done)
1553            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1554            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1555    
1556          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1557    
1558        switch (get_options & PCRE_NEWLINE_BITS)        switch (get_options & PCRE_NEWLINE_BITS)
1559          {          {
1560          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
# Line 1442  while (!done) Line 1569  while (!done)
1569          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
1570          break;          break;
1571    
1572            case PCRE_NEWLINE_ANYCRLF:
1573            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1574            break;
1575    
1576          case PCRE_NEWLINE_ANY:          case PCRE_NEWLINE_ANY:
1577          fprintf(outfile, "Forced newline sequence: ANY\n");          fprintf(outfile, "Forced newline sequence: ANY\n");
1578          break;          break;
# Line 1546  while (!done) Line 1677  while (!done)
1677        else        else
1678          {          {
1679          uschar sbuf[8];          uschar sbuf[8];
1680          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
1681          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
1682          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
1683          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
1684    
1685          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
1686          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
1687          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
1688          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
1689    
1690          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
1691              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1591  while (!done) Line 1722  while (!done)
1722    for (;;)    for (;;)
1723      {      {
1724      uschar *q;      uschar *q;
1725      uschar *bptr = dbuffer;      uschar *bptr;
1726      int *use_offsets = offsets;      int *use_offsets = offsets;
1727      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
1728      int callout_data = 0;      int callout_data = 0;
1729      int callout_data_set = 0;      int callout_data_set = 0;
1730      int count, c;      int count, c;
1731      int copystrings = 0;      int copystrings = 0;
1732      int find_match_limit = 0;      int find_match_limit = default_find_match_limit;
1733      int getstrings = 0;      int getstrings = 0;
1734      int getlist = 0;      int getlist = 0;
1735      int gmatched = 0;      int gmatched = 0;
# Line 1628  while (!done) Line 1759  while (!done)
1759      len = 0;      len = 0;
1760      for (;;)      for (;;)
1761        {        {
1762        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
1763          {          {
1764          if (len > 0) break;          if (len > 0) break;
1765          done = 1;          done = 1;
# Line 1647  while (!done) Line 1777  while (!done)
1777      p = buffer;      p = buffer;
1778      while (isspace(*p)) p++;      while (isspace(*p)) p++;
1779    
1780      q = dbuffer;      bptr = q = dbuffer;
1781      while ((c = *p++) != 0)      while ((c = *p++) != 0)
1782        {        {
1783        int i = 0;        int i = 0;
# Line 1697  while (!done) Line 1827  while (!done)
1827              {              {
1828              unsigned char buff8[8];              unsigned char buff8[8];
1829              int ii, utn;              int ii, utn;
1830              utn = ord2utf8(c, buff8);              if (use_utf8)
1831              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
1832              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
1833                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
1834                  c = buff8[ii];   /* Last byte */
1835                  }
1836                else
1837                 {
1838                 if (c > 255)
1839                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
1840                     "UTF-8 mode is not enabled.\n"
1841                     "** Truncation will probably give the wrong result.\n", c);
1842                 }
1843              p = pt + 1;              p = pt + 1;
1844              break;              break;
1845              }              }
# Line 1842  while (!done) Line 1982  while (!done)
1982            if (offsets == NULL)            if (offsets == NULL)
1983              {              {
1984              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
1985                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
1986              yield = 1;              yield = 1;
1987              goto EXIT;              goto EXIT;
1988              }              }
# Line 1887  while (!done) Line 2027  while (!done)
2027          show_malloc = 1;          show_malloc = 1;
2028          continue;          continue;
2029    
2030            case 'Y':
2031            options |= PCRE_NO_START_OPTIMIZE;
2032            continue;
2033    
2034          case 'Z':          case 'Z':
2035          options |= PCRE_NOTEOL;          options |= PCRE_NOTEOL;
2036          continue;          continue;
# Line 1909  while (!done) Line 2053  while (!done)
2053      *q = 0;      *q = 0;
2054      len = q - dbuffer;      len = q - dbuffer;
2055    
2056        /* Move the data to the end of the buffer so that a read over the end of
2057        the buffer will be seen by valgrind, even if it doesn't cause a crash. If
2058        we are using the POSIX interface, we must include the terminating zero. */
2059    
2060    #if !defined NOPOSIX
2061        if (posix || do_posix)
2062          {
2063          memmove(bptr + buffer_size - len - 1, bptr, len + 1);
2064          bptr += buffer_size - len - 1;
2065          }
2066        else
2067    #endif
2068          {
2069          memmove(bptr + buffer_size - len, bptr, len);
2070          bptr += buffer_size - len;
2071          }
2072    
2073      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2074        {        {
2075        printf("**Match limit not relevant for DFA matching: ignored\n");        printf("**Match limit not relevant for DFA matching: ignored\n");
# Line 1928  while (!done) Line 2089  while (!done)
2089          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);          pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets);
2090        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;        if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL;
2091        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;        if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL;
2092          if ((options & PCRE_NOTEMPTY) != 0) eflags |= REG_NOTEMPTY;
2093    
2094        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);        rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags);
2095    
# Line 1972  while (!done) Line 2134  while (!done)
2134    
2135      for (;; gmatched++)    /* Loop for /g or /G */      for (;; gmatched++)    /* Loop for /g or /G */
2136        {        {
       int gany_fudge;  
2137        if (timeitm > 0)        if (timeitm > 0)
2138          {          {
2139          register int i;          register int i;
# Line 2202  while (!done) Line 2363  while (!done)
2363        else if (count == PCRE_ERROR_PARTIAL)        else if (count == PCRE_ERROR_PARTIAL)
2364          {          {
2365          fprintf(outfile, "Partial match");          fprintf(outfile, "Partial match");
2366  #if !defined NODFA          if (use_size_offsets > 1)
2367          if ((all_use_dfa || use_dfa) && use_size_offsets > 2)            {
2368            fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0],            fprintf(outfile, ": ");
2369              bptr + use_offsets[0]);            pchars(bptr + use_offsets[0], use_offsets[1] - use_offsets[0],
2370  #endif              outfile);
2371              }
2372          fprintf(outfile, "\n");          fprintf(outfile, "\n");
2373          break;  /* Out of the /g loop */          break;  /* Out of the /g loop */
2374          }          }
2375    
2376        /* 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
2377        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
2378        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
2379        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2380        offset values to achieve this. We won't be at the end of the string -  
2381        that was checked before setting g_notempty. */        Complication arises in the case when the newline option is "any" or
2382          "anycrlf". If the previous match was at the end of a line terminated by
2383          CRLF, an advance of one character just passes the \r, whereas we should
2384          prefer the longer newline sequence, as does the code in pcre_exec().
2385          Fudge the offset value to achieve this.
2386    
2387          Otherwise, in the case of UTF-8 matching, the advance must be one
2388          character, not one byte. */
2389    
2390        else        else
2391          {          {
2392          if (g_notempty != 0)          if (g_notempty != 0)
2393            {            {
2394            int onechar = 1;            int onechar = 1;
2395              unsigned int obits = ((real_pcre *)re)->options;
2396            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2397            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2398                {
2399                int d;
2400                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2401                /* Note that these values are always the ASCII ones, even in
2402                EBCDIC environments. CR = 13, NL = 10. */
2403                obits = (d == 13)? PCRE_NEWLINE_CR :
2404                        (d == 10)? PCRE_NEWLINE_LF :
2405                        (d == (13<<8 | 10))? PCRE_NEWLINE_CRLF :
2406                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2407                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2408                }
2409              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2410                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2411                  &&
2412                  start_offset < len - 1 &&
2413                  bptr[start_offset] == '\r' &&
2414                  bptr[start_offset+1] == '\n')
2415                onechar++;
2416              else if (use_utf8)
2417              {              {
2418              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2419                {                {
# Line 2256  while (!done) Line 2445  while (!done)
2445        what Perl's /g options does. This turns out to be rather cunning. First        what Perl's /g options does. This turns out to be rather cunning. First
2446        we set PCRE_NOTEMPTY and PCRE_ANCHORED and try the match again at the        we set PCRE_NOTEMPTY and PCRE_ANCHORED and try the match again at the
2447        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
2448        character.        character. */
   
       Yet more complication arises in the case when the newline option is  
       "any" and a pattern in multiline mode has to match at the start of a  
       line. If a previous match was at the end of a line, and advance of one  
       character just passes the \r, whereas we should prefer the longer newline  
       sequence, as does the code in pcre_exec(). So we fudge it. */  
2449    
2450        g_notempty = 0;        g_notempty = 0;
       gany_fudge = 0;  
2451    
2452        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2453          {          {
2454          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
2455          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;          g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED;
         if ((((real_pcre *)re)->options & PCRE_STARTLINE) != 0 &&  
             (((real_pcre *)re)->options & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY &&  
             use_offsets[0] < len - 1 &&  
             bptr[use_offsets[0]] == '\r' &&  
             bptr[use_offsets[0]+1] == '\n')  
           gany_fudge = 1;  
2456          }          }
2457    
2458        /* For /g, update the start offset, leaving the rest alone */        /* For /g, update the start offset, leaving the rest alone */
2459    
2460        if (do_g) start_offset = use_offsets[1] + gany_fudge;        if (do_g) start_offset = use_offsets[1];
2461    
2462        /* For /G, update the pointer and length */        /* For /G, update the pointer and length */
2463    
2464        else        else
2465          {          {
2466          bptr += use_offsets[1] + gany_fudge;          bptr += use_offsets[1];
2467          len -= use_offsets[1] + gany_fudge;          len -= use_offsets[1];
2468          }          }
2469        }  /* End of loop for /g and /G */        }  /* End of loop for /g and /G */
2470    

Legend:
Removed from v.142  
changed lines
  Added in v.426

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12