/[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 121 by ph10, Mon Mar 12 12:12:47 2007 UTC revision 343 by ph10, Mon Apr 21 16:30:37 2008 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    #define isatty _isatty         /* This is what Windows calls them, I'm told */
75    #define fileno _fileno
76    
77  #else  #else
78  #include <sys/time.h>          /* These two includes are needed */  #include <sys/time.h>          /* These two includes are needed */
79  #include <sys/resource.h>      /* for setrlimit(). */  #include <sys/resource.h>      /* for setrlimit(). */
# Line 67  input mode under Windows. */ Line 82  input mode under Windows. */
82  #endif  #endif
83    
84    
85  #define PCRE_SPY        /* For Win32 build, import data, not export */  /* We have to include pcre_internal.h because we need the internal info for
86    displaying the results of pcre_study() and we also need to know about the
87  /* We include pcre_internal.h because we need the internal info for displaying  internal macros, structures, and other internal data values; pcretest has
88  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.
89  macros, structures, and other internal data values; pcretest has "inside  
90  information" compared to a program that strictly follows the PCRE API. */  Although pcre_internal.h does itself include pcre.h, we explicitly include it
91    here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
92    appropriately for an application, not for building PCRE. */
93    
94    #include "pcre.h"
95  #include "pcre_internal.h"  #include "pcre_internal.h"
96    
97  /* 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 87  symbols to prevent clashes. */ Line 105  symbols to prevent clashes. */
105  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
106  #define _pcre_utt              utt  #define _pcre_utt              utt
107  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
108    #define _pcre_utt_names        utt_names
109  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
110    
111  #include "pcre_tables.c"  #include "pcre_tables.c"
# Line 149  static int callout_count; Line 168  static int callout_count;
168  static int callout_extra;  static int callout_extra;
169  static int callout_fail_count;  static int callout_fail_count;
170  static int callout_fail_id;  static int callout_fail_id;
171    static int debug_lengths;
172  static int first_callout;  static int first_callout;
173  static int locale_set = 0;  static int locale_set = 0;
174  static int show_malloc;  static int show_malloc;
# Line 180  optimal way of handling this, but hey, t Line 200  optimal way of handling this, but hey, t
200  Arguments:  Arguments:
201    f            the file to read    f            the file to read
202    start        where in buffer to start (this *must* be within buffer)    start        where in buffer to start (this *must* be within buffer)
203      prompt       for stdin or readline()
204    
205  Returns:       pointer to the start of new data  Returns:       pointer to the start of new data
206                 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 208  Returns:       pointer to the start of n
208  */  */
209    
210  static uschar *  static uschar *
211  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, uschar *start, const char *prompt)
212  {  {
213  uschar *here = start;  uschar *here = start;
214    
# Line 198  for (;;) Line 219  for (;;)
219    if (rlen > 1000)    if (rlen > 1000)
220      {      {
221      int dlen;      int dlen;
222      if (fgets((char *)here, rlen,  f) == NULL)  
223        return (here == start)? NULL : start;      /* If libreadline support is required, use readline() to read a line if the
224        input is a terminal. Note that readline() removes the trailing newline, so
225        we must put it back again, to be compatible with fgets(). */
226    
227    #ifdef SUPPORT_LIBREADLINE
228        if (isatty(fileno(f)))
229          {
230          size_t len;
231          char *s = readline(prompt);
232          if (s == NULL) return (here == start)? NULL : start;
233          len = strlen(s);
234          if (len > 0) add_history(s);
235          if (len > rlen - 1) len = rlen - 1;
236          memcpy(here, s, len);
237          here[len] = '\n';
238          here[len+1] = 0;
239          free(s);
240          }
241        else
242    #endif
243    
244        /* Read the next line by normal means, prompting if the file is stdin. */
245    
246          {
247          if (f == stdin) printf(prompt);
248          if (fgets((char *)here, rlen,  f) == NULL)
249            return (here == start)? NULL : start;
250          }
251    
252      dlen = (int)strlen((char *)here);      dlen = (int)strlen((char *)here);
253      if (dlen > 0 && here[dlen - 1] == '\n') return start;      if (dlen > 0 && here[dlen - 1] == '\n') return start;
254      here += dlen;      here += dlen;
# Line 656  return count; Line 705  return count;
705    
706    
707  /*************************************************  /*************************************************
708    *         Case-independent strncmp() function    *
709    *************************************************/
710    
711    /*
712    Arguments:
713      s         first string
714      t         second string
715      n         number of characters to compare
716    
717    Returns:    < 0, = 0, or > 0, according to the comparison
718    */
719    
720    static int
721    strncmpic(uschar *s, uschar *t, int n)
722    {
723    while (n--)
724      {
725      int c = tolower(*s++) - tolower(*t++);
726      if (c) return c;
727      }
728    return 0;
729    }
730    
731    
732    
733    /*************************************************
734  *         Check newline indicator                *  *         Check newline indicator                *
735  *************************************************/  *************************************************/
736    
737  /* 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
738  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
739    no match.
740    
741  Arguments:  Arguments:
742    p           points after the leading '<'    p           points after the leading '<'
# Line 672  Returns:      appropriate PCRE_NEWLINE_x Line 748  Returns:      appropriate PCRE_NEWLINE_x
748  static int  static int
749  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
750  {  {
751  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
752  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
753  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
754  if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
755    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
756    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
757    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
758  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
759  return 0;  return 0;
760  }  }
# Line 689  return 0; Line 768  return 0;
768  static void  static void
769  usage(void)  usage(void)
770  {  {
771  printf("Usage:     pcretest [options] [<input> [<output>]]\n");  printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
772    printf("Input and output default to stdin and stdout.\n");
773    #ifdef SUPPORT_LIBREADLINE
774    printf("If input is a terminal, readline() is used to read from it.\n");
775    #else
776    printf("This version of pcretest is not linked with readline().\n");
777    #endif
778    printf("\nOptions:\n");
779  printf("  -b       show compiled code (bytecode)\n");  printf("  -b       show compiled code (bytecode)\n");
780  printf("  -C       show PCRE compile-time options and exit\n");  printf("  -C       show PCRE compile-time options and exit\n");
781  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 847  while (argc > 1 && argv[op][0] == '-') Line 933  while (argc > 1 && argv[op][0] == '-')
933      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
934      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :
935        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :
936          (rc == -2)? "ANYCRLF" :
937        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
938        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
939        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
940                                         "all Unicode newlines");
941      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
942      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
943      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
# Line 884  offsets = (int *)malloc(size_offsets_max Line 974  offsets = (int *)malloc(size_offsets_max
974  if (offsets == NULL)  if (offsets == NULL)
975    {    {
976    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
977      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
978    yield = 1;    yield = 1;
979    goto EXIT;    goto EXIT;
980    }    }
# Line 944  while (!done) Line 1034  while (!done)
1034    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1035    int do_study = 0;    int do_study = 0;
1036    int do_debug = debug;    int do_debug = debug;
   int debug_lengths = 1;  
1037    int do_G = 0;    int do_G = 0;
1038    int do_g = 0;    int do_g = 0;
1039    int do_showinfo = showinfo;    int do_showinfo = showinfo;
# Line 953  while (!done) Line 1042  while (!done)
1042    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1043    
1044    use_utf8 = 0;    use_utf8 = 0;
1045      debug_lengths = 1;
1046    
1047    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1048    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1049    fflush(outfile);    fflush(outfile);
1050    
# Line 1055  while (!done) Line 1144  while (!done)
1144    
1145    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1146      {      {
1147      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1148      goto SKIP_DATA;      goto SKIP_DATA;
1149      }      }
1150    
# Line 1071  while (!done) Line 1160  while (!done)
1160        pp++;        pp++;
1161        }        }
1162      if (*pp != 0) break;      if (*pp != 0) break;
1163      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1164        {        {
1165        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1166        done = 1;        done = 1;
# Line 1135  while (!done) Line 1223  while (!done)
1223        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1224        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1225        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1226        case 'Z': debug_lengths = 0;        case 'Z': debug_lengths = 0; break;
1227        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1228        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1229    
# Line 1164  while (!done) Line 1252  while (!done)
1252    
1253        case '<':        case '<':
1254          {          {
1255          int x = check_newline(pp, outfile);          if (strncmp((char *)pp, "JS>", 3) == 0)
1256          if (x == 0) goto SKIP_DATA;            {
1257          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1258          while (*pp++ != '>');            pp += 3;
1259              }
1260            else
1261              {
1262              int x = check_newline(pp, outfile);
1263              if (x == 0) goto SKIP_DATA;
1264              options |= x;
1265              while (*pp++ != '>');
1266              }
1267          }          }
1268        break;        break;
1269    
# Line 1246  while (!done) Line 1342  while (!done)
1342          {          {
1343          for (;;)          for (;;)
1344            {            {
1345            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1346              {              {
1347              done = 1;              done = 1;
1348              goto CONTINUE;              goto CONTINUE;
# Line 1309  while (!done) Line 1405  while (!done)
1405      if (do_flip)      if (do_flip)
1406        {        {
1407        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1408        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1409            byteflip(rre->magic_number, sizeof(rre->magic_number));
1410        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1411        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1412        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1413        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1414        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1415        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1416        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1417          rre->first_byte =
1418            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1419          rre->req_byte =
1420            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1421          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1422          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1423        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1424          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1425        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1426            sizeof(rre->name_count));
1427    
1428        if (extra != NULL)        if (extra != NULL)
1429          {          {
# Line 1346  while (!done) Line 1449  while (!done)
1449  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1450        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1451  #endif  #endif
1452        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1453            hascrorlf;
1454        int nameentrysize, namecount;        int nameentrysize, namecount;
1455        const uschar *nametable;        const uschar *nametable;
1456    
# Line 1359  while (!done) Line 1463  while (!done)
1463        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1464        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1465        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1466          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1467          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1468          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1469    
1470  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1471        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 1507  while (!done)
1507            }            }
1508          }          }
1509    
1510        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1511        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1512    
1513        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1514        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");  
1515    
1516        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1517          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",
1518            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1519            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1520            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1521            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1522            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1523            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1524              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1525              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1526            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1527            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1528            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
# Line 1428  while (!done) Line 1531  while (!done)
1531            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1532            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1533    
1534          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1535    
1536        switch (get_options & PCRE_NEWLINE_BITS)        switch (get_options & PCRE_NEWLINE_BITS)
1537          {          {
1538          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
# Line 1442  while (!done) Line 1547  while (!done)
1547          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
1548          break;          break;
1549    
1550            case PCRE_NEWLINE_ANYCRLF:
1551            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1552            break;
1553    
1554          case PCRE_NEWLINE_ANY:          case PCRE_NEWLINE_ANY:
1555          fprintf(outfile, "Forced newline sequence: ANY\n");          fprintf(outfile, "Forced newline sequence: ANY\n");
1556          break;          break;
# Line 1546  while (!done) Line 1655  while (!done)
1655        else        else
1656          {          {
1657          uschar sbuf[8];          uschar sbuf[8];
1658          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
1659          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
1660          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
1661          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
1662    
1663          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
1664          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
1665          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
1666          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
1667    
1668          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
1669              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1591  while (!done) Line 1700  while (!done)
1700    for (;;)    for (;;)
1701      {      {
1702      uschar *q;      uschar *q;
1703      uschar *bptr = dbuffer;      uschar *bptr;
1704      int *use_offsets = offsets;      int *use_offsets = offsets;
1705      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
1706      int callout_data = 0;      int callout_data = 0;
# Line 1628  while (!done) Line 1737  while (!done)
1737      len = 0;      len = 0;
1738      for (;;)      for (;;)
1739        {        {
1740        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
1741          {          {
1742          if (len > 0) break;          if (len > 0) break;
1743          done = 1;          done = 1;
# Line 1647  while (!done) Line 1755  while (!done)
1755      p = buffer;      p = buffer;
1756      while (isspace(*p)) p++;      while (isspace(*p)) p++;
1757    
1758      q = dbuffer;      bptr = q = dbuffer;
1759      while ((c = *p++) != 0)      while ((c = *p++) != 0)
1760        {        {
1761        int i = 0;        int i = 0;
# Line 1842  while (!done) Line 1950  while (!done)
1950            if (offsets == NULL)            if (offsets == NULL)
1951              {              {
1952              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
1953                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
1954              yield = 1;              yield = 1;
1955              goto EXIT;              goto EXIT;
1956              }              }
# Line 2211  while (!done) Line 2319  while (!done)
2319          }          }
2320    
2321        /* 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
2322        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
2323        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
2324        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2325        offset values to achieve this. We won't be at the end of the string -  
2326        that was checked before setting g_notempty. */        Complication arises in the case when the newline option is "any" or
2327          "anycrlf". If the previous match was at the end of a line terminated by
2328          CRLF, an advance of one character just passes the \r, whereas we should
2329          prefer the longer newline sequence, as does the code in pcre_exec().
2330          Fudge the offset value to achieve this.
2331    
2332          Otherwise, in the case of UTF-8 matching, the advance must be one
2333          character, not one byte. */
2334    
2335        else        else
2336          {          {
2337          if (g_notempty != 0)          if (g_notempty != 0)
2338            {            {
2339            int onechar = 1;            int onechar = 1;
2340              unsigned int obits = ((real_pcre *)re)->options;
2341            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2342            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2343                {
2344                int d;
2345                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2346                obits = (d == '\r')? PCRE_NEWLINE_CR :
2347                        (d == '\n')? PCRE_NEWLINE_LF :
2348                        (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :
2349                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2350                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2351                }
2352              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2353                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2354                  &&
2355                  start_offset < len - 1 &&
2356                  bptr[start_offset] == '\r' &&
2357                  bptr[start_offset+1] == '\n')
2358                onechar++;
2359              else if (use_utf8)
2360              {              {
2361              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2362                {                {
# Line 2258  while (!done) Line 2391  while (!done)
2391        character. */        character. */
2392    
2393        g_notempty = 0;        g_notempty = 0;
2394    
2395        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2396          {          {
2397          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;

Legend:
Removed from v.121  
changed lines
  Added in v.343

  ViewVC Help
Powered by ViewVC 1.1.5