/[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 123 by ph10, Mon Mar 12 15:19:06 2007 UTC revision 361 by ph10, Thu Jul 10 16:03:28 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 some of the data tables that PCRE uses. So as not to have
98  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
99  symbols to prevent clashes. */  external symbols to prevent clashes. */
100    
101    #define _pcre_ucp_gentype      ucp_gentype
102  #define _pcre_utf8_table1      utf8_table1  #define _pcre_utf8_table1      utf8_table1
103  #define _pcre_utf8_table1_size utf8_table1_size  #define _pcre_utf8_table1_size utf8_table1_size
104  #define _pcre_utf8_table2      utf8_table2  #define _pcre_utf8_table2      utf8_table2
# Line 87  symbols to prevent clashes. */ Line 106  symbols to prevent clashes. */
106  #define _pcre_utf8_table4      utf8_table4  #define _pcre_utf8_table4      utf8_table4
107  #define _pcre_utt              utt  #define _pcre_utt              utt
108  #define _pcre_utt_size         utt_size  #define _pcre_utt_size         utt_size
109    #define _pcre_utt_names        utt_names
110  #define _pcre_OP_lengths       OP_lengths  #define _pcre_OP_lengths       OP_lengths
111    
112  #include "pcre_tables.c"  #include "pcre_tables.c"
# Line 149  static int callout_count; Line 169  static int callout_count;
169  static int callout_extra;  static int callout_extra;
170  static int callout_fail_count;  static int callout_fail_count;
171  static int callout_fail_id;  static int callout_fail_id;
172    static int debug_lengths;
173  static int first_callout;  static int first_callout;
174  static int locale_set = 0;  static int locale_set = 0;
175  static int show_malloc;  static int show_malloc;
# Line 180  optimal way of handling this, but hey, t Line 201  optimal way of handling this, but hey, t
201  Arguments:  Arguments:
202    f            the file to read    f            the file to read
203    start        where in buffer to start (this *must* be within buffer)    start        where in buffer to start (this *must* be within buffer)
204      prompt       for stdin or readline()
205    
206  Returns:       pointer to the start of new data  Returns:       pointer to the start of new data
207                 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 209  Returns: pointer to the start of n
209  */  */
210    
211  static uschar *  static uschar *
212  extend_inputline(FILE *f, uschar *start)  extend_inputline(FILE *f, uschar *start, const char *prompt)
213  {  {
214  uschar *here = start;  uschar *here = start;
215    
# Line 198  for (;;) Line 220  for (;;)
220    if (rlen > 1000)    if (rlen > 1000)
221      {      {
222      int dlen;      int dlen;
223      if (fgets((char *)here, rlen,  f) == NULL)  
224        return (here == start)? NULL : start;      /* If libreadline support is required, use readline() to read a line if the
225        input is a terminal. Note that readline() removes the trailing newline, so
226        we must put it back again, to be compatible with fgets(). */
227    
228    #ifdef SUPPORT_LIBREADLINE
229        if (isatty(fileno(f)))
230          {
231          size_t len;
232          char *s = readline(prompt);
233          if (s == NULL) return (here == start)? NULL : start;
234          len = strlen(s);
235          if (len > 0) add_history(s);
236          if (len > rlen - 1) len = rlen - 1;
237          memcpy(here, s, len);
238          here[len] = '\n';
239          here[len+1] = 0;
240          free(s);
241          }
242        else
243    #endif
244    
245        /* Read the next line by normal means, prompting if the file is stdin. */
246    
247          {
248          if (f == stdin) printf(prompt);
249          if (fgets((char *)here, rlen,  f) == NULL)
250            return (here == start)? NULL : start;
251          }
252    
253      dlen = (int)strlen((char *)here);      dlen = (int)strlen((char *)here);
254      if (dlen > 0 && here[dlen - 1] == '\n') return start;      if (dlen > 0 && here[dlen - 1] == '\n') return start;
255      here += dlen;      here += dlen;
# Line 656  return count; Line 706  return count;
706    
707    
708  /*************************************************  /*************************************************
709    *         Case-independent strncmp() function    *
710    *************************************************/
711    
712    /*
713    Arguments:
714      s         first string
715      t         second string
716      n         number of characters to compare
717    
718    Returns:    < 0, = 0, or > 0, according to the comparison
719    */
720    
721    static int
722    strncmpic(uschar *s, uschar *t, int n)
723    {
724    while (n--)
725      {
726      int c = tolower(*s++) - tolower(*t++);
727      if (c) return c;
728      }
729    return 0;
730    }
731    
732    
733    
734    /*************************************************
735  *         Check newline indicator                *  *         Check newline indicator                *
736  *************************************************/  *************************************************/
737    
738  /* 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
739  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
740    no match.
741    
742  Arguments:  Arguments:
743    p           points after the leading '<'    p           points after the leading '<'
# Line 672  Returns: appropriate PCRE_NEWLINE_x Line 749  Returns: appropriate PCRE_NEWLINE_x
749  static int  static int
750  check_newline(uschar *p, FILE *f)  check_newline(uschar *p, FILE *f)
751  {  {
752  if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR;  if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR;
753  if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF;  if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF;
754  if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;  if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF;
755  if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY;  if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF;
756    if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY;
757    if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF;
758    if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE;
759  fprintf(f, "Unknown newline type at: <%s\n", p);  fprintf(f, "Unknown newline type at: <%s\n", p);
760  return 0;  return 0;
761  }  }
# Line 689  return 0; Line 769  return 0;
769  static void  static void
770  usage(void)  usage(void)
771  {  {
772  printf("Usage:     pcretest [options] [<input> [<output>]]\n");  printf("Usage:     pcretest [options] [<input file> [<output file>]]\n\n");
773    printf("Input and output default to stdin and stdout.\n");
774    #ifdef SUPPORT_LIBREADLINE
775    printf("If input is a terminal, readline() is used to read from it.\n");
776    #else
777    printf("This version of pcretest is not linked with readline().\n");
778    #endif
779    printf("\nOptions:\n");
780  printf("  -b       show compiled code (bytecode)\n");  printf("  -b       show compiled code (bytecode)\n");
781  printf("  -C       show PCRE compile-time options and exit\n");  printf("  -C       show PCRE compile-time options and exit\n");
782  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 934  while (argc > 1 && argv[op][0] == '-')
934      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);      (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
935      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :      printf("  Newline sequence is %s\n", (rc == '\r')? "CR" :
936        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :        (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :
937          (rc == -2)? "ANYCRLF" :
938        (rc == -1)? "ANY" : "???");        (rc == -1)? "ANY" : "???");
939        (void)pcre_config(PCRE_CONFIG_BSR, &rc);
940        printf("  \\R matches %s\n", rc? "CR, LF, or CRLF only" :
941                                         "all Unicode newlines");
942      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);      (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
943      printf("  Internal link size = %d\n", rc);      printf("  Internal link size = %d\n", rc);
944      (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 975  offsets = (int *)malloc(size_offsets_max
975  if (offsets == NULL)  if (offsets == NULL)
976    {    {
977    printf("** Failed to get %d bytes of memory for offsets vector\n",    printf("** Failed to get %d bytes of memory for offsets vector\n",
978      size_offsets_max * sizeof(int));      (int)(size_offsets_max * sizeof(int)));
979    yield = 1;    yield = 1;
980    goto EXIT;    goto EXIT;
981    }    }
# Line 944  while (!done) Line 1035  while (!done)
1035    size_t size, regex_gotten_store;    size_t size, regex_gotten_store;
1036    int do_study = 0;    int do_study = 0;
1037    int do_debug = debug;    int do_debug = debug;
   int debug_lengths = 1;  
1038    int do_G = 0;    int do_G = 0;
1039    int do_g = 0;    int do_g = 0;
1040    int do_showinfo = showinfo;    int do_showinfo = showinfo;
# Line 953  while (!done) Line 1043  while (!done)
1043    int erroroffset, len, delimiter, poffset;    int erroroffset, len, delimiter, poffset;
1044    
1045    use_utf8 = 0;    use_utf8 = 0;
1046      debug_lengths = 1;
1047    
1048    if (infile == stdin) printf("  re> ");    if (extend_inputline(infile, buffer, "  re> ") == NULL) break;
   if (extend_inputline(infile, buffer) == NULL) break;  
1049    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);    if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
1050    fflush(outfile);    fflush(outfile);
1051    
# Line 1055  while (!done) Line 1145  while (!done)
1145    
1146    if (isalnum(delimiter) || delimiter == '\\')    if (isalnum(delimiter) || delimiter == '\\')
1147      {      {
1148      fprintf(outfile, "** Delimiter must not be alphameric or \\\n");      fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n");
1149      goto SKIP_DATA;      goto SKIP_DATA;
1150      }      }
1151    
# Line 1071  while (!done) Line 1161  while (!done)
1161        pp++;        pp++;
1162        }        }
1163      if (*pp != 0) break;      if (*pp != 0) break;
1164      if (infile == stdin) printf("    > ");      if ((pp = extend_inputline(infile, pp, "    > ")) == NULL)
     if ((pp = extend_inputline(infile, pp)) == NULL)  
1165        {        {
1166        fprintf(outfile, "** Unexpected EOF\n");        fprintf(outfile, "** Unexpected EOF\n");
1167        done = 1;        done = 1;
# Line 1135  while (!done) Line 1224  while (!done)
1224        case 'S': do_study = 1; break;        case 'S': do_study = 1; break;
1225        case 'U': options |= PCRE_UNGREEDY; break;        case 'U': options |= PCRE_UNGREEDY; break;
1226        case 'X': options |= PCRE_EXTRA; break;        case 'X': options |= PCRE_EXTRA; break;
1227        case 'Z': debug_lengths = 0;        case 'Z': debug_lengths = 0; break;
1228        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;        case '8': options |= PCRE_UTF8; use_utf8 = 1; break;
1229        case '?': options |= PCRE_NO_UTF8_CHECK; break;        case '?': options |= PCRE_NO_UTF8_CHECK; break;
1230    
# Line 1164  while (!done) Line 1253  while (!done)
1253    
1254        case '<':        case '<':
1255          {          {
1256          int x = check_newline(pp, outfile);          if (strncmp((char *)pp, "JS>", 3) == 0)
1257          if (x == 0) goto SKIP_DATA;            {
1258          options |= x;            options |= PCRE_JAVASCRIPT_COMPAT;
1259          while (*pp++ != '>');            pp += 3;
1260              }
1261            else
1262              {
1263              int x = check_newline(pp, outfile);
1264              if (x == 0) goto SKIP_DATA;
1265              options |= x;
1266              while (*pp++ != '>');
1267              }
1268          }          }
1269        break;        break;
1270    
# Line 1246  while (!done) Line 1343  while (!done)
1343          {          {
1344          for (;;)          for (;;)
1345            {            {
1346            if (extend_inputline(infile, buffer) == NULL)            if (extend_inputline(infile, buffer, NULL) == NULL)
1347              {              {
1348              done = 1;              done = 1;
1349              goto CONTINUE;              goto CONTINUE;
# Line 1309  while (!done) Line 1406  while (!done)
1406      if (do_flip)      if (do_flip)
1407        {        {
1408        real_pcre *rre = (real_pcre *)re;        real_pcre *rre = (real_pcre *)re;
1409        rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number));        rre->magic_number =
1410            byteflip(rre->magic_number, sizeof(rre->magic_number));
1411        rre->size = byteflip(rre->size, sizeof(rre->size));        rre->size = byteflip(rre->size, sizeof(rre->size));
1412        rre->options = byteflip(rre->options, sizeof(rre->options));        rre->options = byteflip(rre->options, sizeof(rre->options));
1413        rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket));        rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags));
1414        rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref));        rre->top_bracket =
1415        rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte));          (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket));
1416        rre->req_byte = byteflip(rre->req_byte, sizeof(rre->req_byte));        rre->top_backref =
1417        rre->name_table_offset = byteflip(rre->name_table_offset,          (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref));
1418          rre->first_byte =
1419            (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte));
1420          rre->req_byte =
1421            (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte));
1422          rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset,
1423          sizeof(rre->name_table_offset));          sizeof(rre->name_table_offset));
1424        rre->name_entry_size = byteflip(rre->name_entry_size,        rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size,
1425          sizeof(rre->name_entry_size));          sizeof(rre->name_entry_size));
1426        rre->name_count = byteflip(rre->name_count, sizeof(rre->name_count));        rre->name_count = (pcre_uint16)byteflip(rre->name_count,
1427            sizeof(rre->name_count));
1428    
1429        if (extra != NULL)        if (extra != NULL)
1430          {          {
# Line 1346  while (!done) Line 1450  while (!done)
1450  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1451        int old_first_char, old_options, old_count;        int old_first_char, old_options, old_count;
1452  #endif  #endif
1453        int count, backrefmax, first_char, need_char;        int count, backrefmax, first_char, need_char, okpartial, jchanged,
1454            hascrorlf;
1455        int nameentrysize, namecount;        int nameentrysize, namecount;
1456        const uschar *nametable;        const uschar *nametable;
1457    
# Line 1359  while (!done) Line 1464  while (!done)
1464        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);        new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize);
1465        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);        new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount);
1466        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);        new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable);
1467          new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial);
1468          new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1469          new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf);
1470    
1471  #if !defined NOINFOCHECK  #if !defined NOINFOCHECK
1472        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 1508  while (!done)
1508            }            }
1509          }          }
1510    
1511        /* The NOPARTIAL bit is a private bit in the options, so we have        if (!okpartial) fprintf(outfile, "Partial matching not supported\n");
1512        to fish it out via out back door */        if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n");
1513    
1514        all_options = ((real_pcre *)re)->options;        all_options = ((real_pcre *)re)->options;
1515        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");  
1516    
1517        if (get_options == 0) fprintf(outfile, "No options\n");        if (get_options == 0) fprintf(outfile, "No options\n");
1518          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",
1519            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",            ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
1520            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",            ((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
1521            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",            ((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
1522            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",            ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "",
1523            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",            ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "",
1524            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",            ((get_options & PCRE_DOTALL) != 0)? " dotall" : "",
1525              ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "",
1526              ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "",
1527            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",            ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
1528            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",            ((get_options & PCRE_EXTRA) != 0)? " extra" : "",
1529            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",            ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
# Line 1428  while (!done) Line 1532  while (!done)
1532            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",            ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
1533            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");            ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
1534    
1535          if (jchanged) fprintf(outfile, "Duplicate name status changes\n");
1536    
1537        switch (get_options & PCRE_NEWLINE_BITS)        switch (get_options & PCRE_NEWLINE_BITS)
1538          {          {
1539          case PCRE_NEWLINE_CR:          case PCRE_NEWLINE_CR:
# Line 1442  while (!done) Line 1548  while (!done)
1548          fprintf(outfile, "Forced newline sequence: CRLF\n");          fprintf(outfile, "Forced newline sequence: CRLF\n");
1549          break;          break;
1550    
1551            case PCRE_NEWLINE_ANYCRLF:
1552            fprintf(outfile, "Forced newline sequence: ANYCRLF\n");
1553            break;
1554    
1555          case PCRE_NEWLINE_ANY:          case PCRE_NEWLINE_ANY:
1556          fprintf(outfile, "Forced newline sequence: ANY\n");          fprintf(outfile, "Forced newline sequence: ANY\n");
1557          break;          break;
# Line 1546  while (!done) Line 1656  while (!done)
1656        else        else
1657          {          {
1658          uschar sbuf[8];          uschar sbuf[8];
1659          sbuf[0] = (true_size >> 24)  & 255;          sbuf[0] = (uschar)((true_size >> 24) & 255);
1660          sbuf[1] = (true_size >> 16)  & 255;          sbuf[1] = (uschar)((true_size >> 16) & 255);
1661          sbuf[2] = (true_size >>  8)  & 255;          sbuf[2] = (uschar)((true_size >>  8) & 255);
1662          sbuf[3] = (true_size)  & 255;          sbuf[3] = (uschar)((true_size) & 255);
1663    
1664          sbuf[4] = (true_study_size >> 24)  & 255;          sbuf[4] = (uschar)((true_study_size >> 24) & 255);
1665          sbuf[5] = (true_study_size >> 16)  & 255;          sbuf[5] = (uschar)((true_study_size >> 16) & 255);
1666          sbuf[6] = (true_study_size >>  8)  & 255;          sbuf[6] = (uschar)((true_study_size >>  8) & 255);
1667          sbuf[7] = (true_study_size)  & 255;          sbuf[7] = (uschar)((true_study_size) & 255);
1668    
1669          if (fwrite(sbuf, 1, 8, f) < 8 ||          if (fwrite(sbuf, 1, 8, f) < 8 ||
1670              fwrite(re, 1, true_size, f) < true_size)              fwrite(re, 1, true_size, f) < true_size)
# Line 1591  while (!done) Line 1701  while (!done)
1701    for (;;)    for (;;)
1702      {      {
1703      uschar *q;      uschar *q;
1704      uschar *bptr = dbuffer;      uschar *bptr;
1705      int *use_offsets = offsets;      int *use_offsets = offsets;
1706      int use_size_offsets = size_offsets;      int use_size_offsets = size_offsets;
1707      int callout_data = 0;      int callout_data = 0;
# Line 1628  while (!done) Line 1738  while (!done)
1738      len = 0;      len = 0;
1739      for (;;)      for (;;)
1740        {        {
1741        if (infile == stdin) printf("data> ");        if (extend_inputline(infile, buffer + len, "data> ") == NULL)
       if (extend_inputline(infile, buffer + len) == NULL)  
1742          {          {
1743          if (len > 0) break;          if (len > 0) break;
1744          done = 1;          done = 1;
# Line 1647  while (!done) Line 1756  while (!done)
1756      p = buffer;      p = buffer;
1757      while (isspace(*p)) p++;      while (isspace(*p)) p++;
1758    
1759      q = dbuffer;      bptr = q = dbuffer;
1760      while ((c = *p++) != 0)      while ((c = *p++) != 0)
1761        {        {
1762        int i = 0;        int i = 0;
# Line 1697  while (!done) Line 1806  while (!done)
1806              {              {
1807              unsigned char buff8[8];              unsigned char buff8[8];
1808              int ii, utn;              int ii, utn;
1809              utn = ord2utf8(c, buff8);              if (use_utf8)
1810              for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];                {
1811              c = buff8[ii];   /* Last byte */                utn = ord2utf8(c, buff8);
1812                  for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
1813                  c = buff8[ii];   /* Last byte */
1814                  }
1815                else
1816                 {
1817                 if (c > 255)
1818                   fprintf(outfile, "** Character \\x{%x} is greater than 255 and "
1819                     "UTF-8 mode is not enabled.\n"
1820                     "** Truncation will probably give the wrong result.\n", c);
1821                 }
1822              p = pt + 1;              p = pt + 1;
1823              break;              break;
1824              }              }
# Line 1842  while (!done) Line 1961  while (!done)
1961            if (offsets == NULL)            if (offsets == NULL)
1962              {              {
1963              printf("** Failed to get %d bytes of memory for offsets vector\n",              printf("** Failed to get %d bytes of memory for offsets vector\n",
1964                size_offsets_max * sizeof(int));                (int)(size_offsets_max * sizeof(int)));
1965              yield = 1;              yield = 1;
1966              goto EXIT;              goto EXIT;
1967              }              }
# Line 1908  while (!done) Line 2027  while (!done)
2027        }        }
2028      *q = 0;      *q = 0;
2029      len = q - dbuffer;      len = q - dbuffer;
2030    
2031        /* Move the data to the end of the buffer so that a read over the end of
2032        the buffer will be seen by valgrind, even if it doesn't cause a crash. */
2033    
2034        memmove(bptr + buffer_size - len, bptr, len);
2035        bptr += buffer_size - len;
2036    
2037      if ((all_use_dfa || use_dfa) && find_match_limit)      if ((all_use_dfa || use_dfa) && find_match_limit)
2038        {        {
# Line 2211  while (!done) Line 2336  while (!done)
2336          }          }
2337    
2338        /* 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
2339        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
2340        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
2341        matching, the advance must be one character, not one byte. Fudge the        string - that was checked before setting g_notempty.
2342        offset values to achieve this. We won't be at the end of the string -  
2343        that was checked before setting g_notempty. */        Complication arises in the case when the newline option is "any" or
2344          "anycrlf". If the previous match was at the end of a line terminated by
2345          CRLF, an advance of one character just passes the \r, whereas we should
2346          prefer the longer newline sequence, as does the code in pcre_exec().
2347          Fudge the offset value to achieve this.
2348    
2349          Otherwise, in the case of UTF-8 matching, the advance must be one
2350          character, not one byte. */
2351    
2352        else        else
2353          {          {
2354          if (g_notempty != 0)          if (g_notempty != 0)
2355            {            {
2356            int onechar = 1;            int onechar = 1;
2357              unsigned int obits = ((real_pcre *)re)->options;
2358            use_offsets[0] = start_offset;            use_offsets[0] = start_offset;
2359            if (use_utf8)            if ((obits & PCRE_NEWLINE_BITS) == 0)
2360                {
2361                int d;
2362                (void)pcre_config(PCRE_CONFIG_NEWLINE, &d);
2363                obits = (d == '\r')? PCRE_NEWLINE_CR :
2364                        (d == '\n')? PCRE_NEWLINE_LF :
2365                        (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF :
2366                        (d == -2)? PCRE_NEWLINE_ANYCRLF :
2367                        (d == -1)? PCRE_NEWLINE_ANY : 0;
2368                }
2369              if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY ||
2370                   (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF)
2371                  &&
2372                  start_offset < len - 1 &&
2373                  bptr[start_offset] == '\r' &&
2374                  bptr[start_offset+1] == '\n')
2375                onechar++;
2376              else if (use_utf8)
2377              {              {
2378              while (start_offset + onechar < len)              while (start_offset + onechar < len)
2379                {                {
# Line 2258  while (!done) Line 2408  while (!done)
2408        character. */        character. */
2409    
2410        g_notempty = 0;        g_notempty = 0;
2411    
2412        if (use_offsets[0] == use_offsets[1])        if (use_offsets[0] == use_offsets[1])
2413          {          {
2414          if (use_offsets[0] == len) break;          if (use_offsets[0] == len) break;

Legend:
Removed from v.123  
changed lines
  Added in v.361

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12