/[pcre]/code/branches/pcre16/pcretest.c
ViewVC logotype

Diff of /code/branches/pcre16/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 259 by ph10, Wed Sep 19 09:52:44 2007 UTC
# Line 36  POSSIBILITY OF SUCH DAMAGE. Line 36  POSSIBILITY OF SUCH DAMAGE.
36  */  */
37    
38    
39    #ifdef HAVE_CONFIG_H
40    #include "config.h"
41    #endif
42    
43  #include <ctype.h>  #include <ctype.h>
44  #include <stdio.h>  #include <stdio.h>
45  #include <string.h>  #include <string.h>
# Line 67  input mode under Windows. */ Line 71  input mode under Windows. */
71  #endif  #endif
72    
73    
74  #define PCRE_SPY        /* For Win32 build, import data, not export */  /* We have to include pcre_internal.h because we need the internal info for
75    displaying the results of pcre_study() and we also need to know about the
76  /* We include pcre_internal.h because we need the internal info for displaying  internal macros, structures, and other internal data values; pcretest has
77  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.
78  macros, structures, and other internal data values; pcretest has "inside  
79  information" compared to a program that strictly follows the PCRE API. */  Although pcre_internal.h does itself include pcre.h, we explicitly include it
80    here before pcre_internal.h so that the PCRE_EXP_xxx macros get set
81    appropriately for an application, not for building PCRE. */
82    
83    #include "pcre.h"
84  #include "pcre_internal.h"  #include "pcre_internal.h"
85    
86  /* We need access to the data tables that PCRE uses. So as not to have to keep  /* We need access to the data tables that PCRE uses. So as not to have to keep
# Line 87  symbols to prevent clashes. */ Line 94  symbols to prevent clashes. */
94  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
95  #define _pcre_utt              utt  #define _pcre_utt              utt
96  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
97    #define _pcre_utt_names        utt_names
98  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
99    
100  #include "pcre_tables.c"  #include "pcre_tables.c"
# Line 149  static int callout_count; Line 157  static int callout_count;
157  static int callout_extra;  static int callout_extra;
158  static int callout_fail_count;  static int callout_fail_count;
159  static int callout_fail_id;  static int callout_fail_id;
160    static int debug_lengths;
161  static int first_callout;  static int first_callout;
162  static int locale_set = 0;  static int locale_set = 0;
163  static int show_malloc;  static int show_malloc;
# Line 656  return count; Line 665  return count;
665    
666    
667  /*************************************************  /*************************************************
668    *         Case-independent strncmp() function    *
669    *************************************************/
670    
671    /*
672    Arguments:
673      s         first string
674      t         second string
675      n         number of characters to compare
676    
677    Returns:    < 0, = 0, or > 0, according to the comparison
678    */
679    
680    static int
681    strncmpic(uschar *s, uschar *t, int n)
682    {
683    while (n--)
684      {
685      int c = tolower(*s++) - tolower(*t++);
686      if (c) return c;
687      }
688    return 0;
689    }
690    
691    
692    
693    /*************************************************
694  *         Check newline indicator                *  *         Check newline indicator                *
695  *************************************************/  *************************************************/
696    
697  /* 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
698  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
699    no match.
700    
701  Arguments:  Arguments:
702    p           points after the leading '<'    p           points after the leading '<'
# Line 672  Returns: appropriate PCRE_NEWLINE_x Line 708  Returns: appropriate PCRE_NEWLINE_x
708  static int  static int
709  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
710  {  {
711  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
712  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
713  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
714  if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
715    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
716    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
717    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
718  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
719  return 0;  return 0;
720  }  }
# Line 847  while (argc > 1 && argv[op][0] == '-') Line 886  while (argc > 1 && argv[op][0] == '-')
886      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
887      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :
888        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :
889          (rc == -2)? "ANYCRLF" :
890        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
891        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
892        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
893                                         "all Unicode newlines");
894      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
895      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
896      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);      (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
# Line 858  while (argc > 1 && argv[op][0] == '-') Line 901  while (argc > 1 && argv[op][0] == '-')
901      printf("  Default recursion depth limit = %d\n", rc);      printf("  Default recursion depth limit = %d\n", rc);
902      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);      (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
903      printf("  Match recursion uses %s\n", rc? "stack" : "heap");      printf("  Match recursion uses %s\n", rc? "stack" : "heap");
904      exit(0);      goto EXIT;
905      }      }
906    else if (strcmp(argv[op], "-help") == 0 ||    else if (strcmp(argv[op], "-help") == 0 ||
907             strcmp(argv[op], "--help") == 0)             strcmp(argv[op], "--help") == 0)
# Line 884  offsets = (int *)malloc(size_offsets_max Line 927  offsets = (int *)malloc(size_offsets_max
927  if (offsets == NULL)  if (offsets == NULL)
928    {    {
929    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
930      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
931    yield = 1;    yield = 1;
932    goto EXIT;    goto EXIT;
933    }    }
# Line 952  while (!done) Line 995  while (!done)
995    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
996    
997    use_utf8 = 0;    use_utf8 = 0;
998      debug_lengths = 1;
999    
1000    if (infile == stdin) printf("  re> ");    if (infile == stdin) printf("  re> ");
1001    if (extend_inputline(infile, buffer) == NULL) break;    if (extend_inputline(infile, buffer) == NULL) break;
# Line 1134  while (!done) Line 1178  while (!done)
1178        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1179        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1180        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1181          case 'Z': debug_lengths = 0; break;
1182        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1183        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1184    
# Line 1307  while (!done) Line 1352  while (!done)
1352      if (do_flip)      if (do_flip)
1353        {        {
1354        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1355        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1356            byteflip(rre->magic_number, sizeof(rre->magic_number));
1357        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1358        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1359        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1360        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1361        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1362        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1363        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1364          rre->first_byte =
1365            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1366          rre->req_byte =
1367            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1368          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1369          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1370        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1371          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1372        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1373            sizeof(rre->name_count));
1374    
1375        if (extra != NULL)        if (extra != NULL)
1376          {          {
# Line 1335  while (!done) Line 1387  while (!done)
1387      if (do_debug)      if (do_debug)
1388        {        {
1389        fprintf(outfile, "------------------------------------------------------------------\n");        fprintf(outfile, "------------------------------------------------------------------\n");
1390        pcre_printint(re, outfile);        pcre_printint(re, outfile, debug_lengths);
1391        }        }
1392    
1393      if (do_showinfo)      if (do_showinfo)
# Line 1344  while (!done) Line 1396  while (!done)
1396  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1397        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1398  #endif  #endif
1399        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1400            hascrorlf;
1401        int nameentrysize, namecount;        int nameentrysize, namecount;
1402        const uschar *nametable;        const uschar *nametable;
1403    
# Line 1357  while (!done) Line 1410  while (!done)
1410        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1411        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1412        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1413          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1414          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1415          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1416    
1417  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1418        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 1454  while (!done)
1454            }            }
1455          }          }
1456    
1457        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1458        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1459    
1460        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1461        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");  
1462    
1463        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1464          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",
1465            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1466            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1467            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1468            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1469            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1470            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1471              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1472              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1473            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1474            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1475            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
# Line 1426  while (!done) Line 1478  while (!done)
1478            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1479            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1480    
1481          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1482    
1483        switch (get_options & PCRE_NEWLINE_BITS)        switch (get_options & PCRE_NEWLINE_BITS)
1484          {          {
1485          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
# Line 1440  while (!done) Line 1494  while (!done)
1494          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
1495          break;          break;
1496    
1497            case PCRE_NEWLINE_ANYCRLF:
1498            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1499            break;
1500    
1501          case PCRE_NEWLINE_ANY:          case PCRE_NEWLINE_ANY:
1502          fprintf(outfile, "Forced newline sequence: ANY\n");          fprintf(outfile, "Forced newline sequence: ANY\n");
1503          break;          break;
# Line 1544  while (!done) Line 1602  while (!done)
1602        else        else
1603          {          {
1604          uschar sbuf[8];          uschar sbuf[8];
1605          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
1606          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
1607          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
1608          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
1609    
1610          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
1611          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
1612          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
1613          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
1614    
1615          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
1616              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1589  while (!done) Line 1647  while (!done)
1647    for (;;)    for (;;)
1648      {      {
1649      uschar *q;      uschar *q;
1650      uschar *bptr = dbuffer;      uschar *bptr;
1651      int *use_offsets = offsets;      int *use_offsets = offsets;
1652      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
1653      int callout_data = 0;      int callout_data = 0;
# Line 1645  while (!done) Line 1703  while (!done)
1703      p = buffer;      p = buffer;
1704      while (isspace(*p)) p++;      while (isspace(*p)) p++;
1705    
1706      q = dbuffer;      bptr = q = dbuffer;
1707      while ((c = *p++) != 0)      while ((c = *p++) != 0)
1708        {        {
1709        int i = 0;        int i = 0;
# Line 1840  while (!done) Line 1898  while (!done)
1898            if (offsets == NULL)            if (offsets == NULL)
1899              {              {
1900              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
1901                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
1902              yield = 1;              yield = 1;
1903              goto EXIT;              goto EXIT;
1904              }              }
# Line 2209  while (!done) Line 2267  while (!done)
2267          }          }
2268    
2269        /* 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
2270        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
2271        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
2272        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2273        offset values to achieve this. We won't be at the end of the string -  
2274        that was checked before setting g_notempty. */        Complication arises in the case when the newline option is "any" or
2275          "anycrlf". If the previous match was at the end of a line terminated by
2276          CRLF, an advance of one character just passes the \r, whereas we should
2277          prefer the longer newline sequence, as does the code in pcre_exec().
2278          Fudge the offset value to achieve this.
2279    
2280          Otherwise, in the case of UTF-8 matching, the advance must be one
2281          character, not one byte. */
2282    
2283        else        else
2284          {          {
2285          if (g_notempty != 0)          if (g_notempty != 0)
2286            {            {
2287            int onechar = 1;            int onechar = 1;
2288              unsigned int obits = ((real_pcre *)re)->options;
2289            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2290            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2291                {
2292                int d;
2293                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2294                obits = (d == '\r')? PCRE_NEWLINE_CR :
2295                        (d == '\n')? PCRE_NEWLINE_LF :
2296                        (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :
2297                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2298                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2299                }
2300              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2301                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2302                  &&
2303                  start_offset < len - 1 &&
2304                  bptr[start_offset] == '\r' &&
2305                  bptr[start_offset+1] == '\n')
2306                onechar++;
2307              else if (use_utf8)
2308              {              {
2309              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2310                {                {
# Line 2256  while (!done) Line 2339  while (!done)
2339        character. */        character. */
2340    
2341        g_notempty = 0;        g_notempty = 0;
2342    
2343        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2344          {          {
2345          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;
Legend:
Removed from v.107  
changed lines
  Added in v.259
webmaster@exim.org
 ViewVC Help
Powered by ViewVC 1.1.12