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

Diff of /code/branches/pcre16/pcre_compile.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 227 by ph10, Tue Aug 21 15:00:15 2007 UTC revision 230 by ph10, Mon Sep 10 13:23:56 2007 UTC
# Line 2975  for (;; ptr++) Line 2975  for (;; ptr++)
2975          }          }
2976    
2977        oldptr = ptr;        oldptr = ptr;
2978    
2979          /* Remember \r or \n */
2980    
2981          if (c == '\r' || c == '\n') cd->external_flags |= PCRE_HASCRORLF;
2982    
2983          /* Check for range */
2984    
2985        if (!inescq && ptr[1] == '-')        if (!inescq && ptr[1] == '-')
2986          {          {
# Line 3043  for (;; ptr++) Line 3049  for (;; ptr++)
3049    
3050          if (d == c) goto LONE_SINGLE_CHARACTER;  /* A few lines below */          if (d == c) goto LONE_SINGLE_CHARACTER;  /* A few lines below */
3051    
3052            /* Remember \r or \n */
3053    
3054            if (d == '\r' || d == '\n') cd->external_flags |= PCRE_HASCRORLF;
3055    
3056          /* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless          /* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless
3057          matching, we have to use an XCLASS with extra data items. Caseless          matching, we have to use an XCLASS with extra data items. Caseless
3058          matching for characters > 127 is available only if UCP support is          matching for characters > 127 is available only if UCP support is
# Line 3147  for (;; ptr++) Line 3157  for (;; ptr++)
3157        apparent range that isn't. */        apparent range that isn't. */
3158    
3159        LONE_SINGLE_CHARACTER:        LONE_SINGLE_CHARACTER:
3160    
3161        /* Handle a character that cannot go in the bit map */        /* Handle a character that cannot go in the bit map */
3162    
3163  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 3195  for (;; ptr++) Line 3205  for (;; ptr++)
3205        *errorcodeptr = ERR6;        *errorcodeptr = ERR6;
3206        goto FAILED;        goto FAILED;
3207        }        }
3208    
3209    
3210    /* This code has been disabled because it would mean that \s counts as
3211    an explicit \r or \n reference, and that's not really what is wanted. Now
3212    we set the flag only if there is a literal "\r" or "\n" in the class. */
3213    
3214    #if 0
3215      /* Remember whether \r or \n are in this class */      /* Remember whether \r or \n are in this class */
3216    
3217      if (negate_class)      if (negate_class)
3218        {        {
3219        if ((classbits[1] & 0x24) != 0x24) cd->external_options |= PCRE_HASCRORLF;        if ((classbits[1] & 0x24) != 0x24) cd->external_flags |= PCRE_HASCRORLF;
3220        }        }
3221      else      else
3222        {        {
3223        if ((classbits[1] & 0x24) != 0) cd->external_options |= PCRE_HASCRORLF;        if ((classbits[1] & 0x24) != 0) cd->external_flags |= PCRE_HASCRORLF;
3224        }        }
3225    #endif
3226    
3227    
3228      /* If class_charcount is 1, we saw precisely one character whose value is      /* If class_charcount is 1, we saw precisely one character whose value is
3229      less than 256. As long as there were no characters >= 128 and there was no      less than 256. As long as there were no characters >= 128 and there was no
# Line 3498  for (;; ptr++) Line 3516  for (;; ptr++)
3516        /* All real repeats make it impossible to handle partial matching (maybe        /* All real repeats make it impossible to handle partial matching (maybe
3517        one day we will be able to remove this restriction). */        one day we will be able to remove this restriction). */
3518    
3519        if (repeat_max != 1) cd->nopartial = TRUE;        if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL;
3520    
3521        /* Combine the op_type with the repeat_type */        /* Combine the op_type with the repeat_type */
3522    
# Line 3648  for (;; ptr++) Line 3666  for (;; ptr++)
3666        /* All real repeats make it impossible to handle partial matching (maybe        /* All real repeats make it impossible to handle partial matching (maybe
3667        one day we will be able to remove this restriction). */        one day we will be able to remove this restriction). */
3668    
3669        if (repeat_max != 1) cd->nopartial = TRUE;        if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL;
3670    
3671        if (repeat_min == 0 && repeat_max == -1)        if (repeat_min == 0 && repeat_max == -1)
3672          *code++ = OP_CRSTAR + repeat_type;          *code++ = OP_CRSTAR + repeat_type;
# Line 4645  for (;; ptr++) Line 4663  for (;; ptr++)
4663    
4664              case 'J':    /* Record that it changed in the external options */              case 'J':    /* Record that it changed in the external options */
4665              *optset |= PCRE_DUPNAMES;              *optset |= PCRE_DUPNAMES;
4666              cd->external_options |= PCRE_JCHANGED;              cd->external_flags |= PCRE_JCHANGED;
4667              break;              break;
4668    
4669              case 'i': *optset |= PCRE_CASELESS; break;              case 'i': *optset |= PCRE_CASELESS; break;
# Line 5065  for (;; ptr++) Line 5083  for (;; ptr++)
5083      /* Remember if \r or \n were seen */      /* Remember if \r or \n were seen */
5084    
5085      if (mcbuffer[0] == '\r' || mcbuffer[0] == '\n')      if (mcbuffer[0] == '\r' || mcbuffer[0] == '\n')
5086        cd->external_options |= PCRE_HASCRORLF;        cd->external_flags |= PCRE_HASCRORLF;
5087    
5088      /* Set the first and required bytes appropriately. If no previous first      /* Set the first and required bytes appropriately. If no previous first
5089      byte, set it from this character, but revert to none on a zero repeat.      byte, set it from this character, but revert to none on a zero repeat.
# Line 5834  cd->hwm = cworkspace; Line 5852  cd->hwm = cworkspace;
5852  cd->start_pattern = (const uschar *)pattern;  cd->start_pattern = (const uschar *)pattern;
5853  cd->end_pattern = (const uschar *)(pattern + strlen(pattern));  cd->end_pattern = (const uschar *)(pattern + strlen(pattern));
5854  cd->req_varyopt = 0;  cd->req_varyopt = 0;
 cd->nopartial = FALSE;  
5855  cd->external_options = options;  cd->external_options = options;
5856    cd->external_flags = 0;
5857    
5858  /* Now do the pre-compile. On error, errorcode will be set non-zero, so we  /* Now do the pre-compile. On error, errorcode will be set non-zero, so we
5859  don't need to look at the result of the function here. The initial options have  don't need to look at the result of the function here. The initial options have
# Line 5874  if (re == NULL) Line 5892  if (re == NULL)
5892    goto PCRE_EARLY_ERROR_RETURN;    goto PCRE_EARLY_ERROR_RETURN;
5893    }    }
5894    
5895  /* Put in the magic number, and save the sizes, initial options, and character  /* Put in the magic number, and save the sizes, initial options, internal
5896  table pointer. NULL is used for the default character tables. The nullpad field  flags, and character table pointer. NULL is used for the default character
5897  is at the end; it's there to help in the case when a regex compiled on a system  tables. The nullpad field is at the end; it's there to help in the case when a
5898  with 4-byte pointers is run on another with 8-byte pointers. */  regex compiled on a system with 4-byte pointers is run on another with 8-byte
5899    pointers. */
5900    
5901  re->magic_number = MAGIC_NUMBER;  re->magic_number = MAGIC_NUMBER;
5902  re->size = size;  re->size = size;
5903  re->options = cd->external_options;  re->options = cd->external_options;
5904    re->flags = cd->external_flags;
5905  re->dummy1 = 0;  re->dummy1 = 0;
5906  re->first_byte = 0;  re->first_byte = 0;
5907  re->req_byte = 0;  re->req_byte = 0;
# Line 5906  codestart = cd->name_table + re->name_en Line 5926  codestart = cd->name_table + re->name_en
5926  cd->start_code = codestart;  cd->start_code = codestart;
5927  cd->hwm = cworkspace;  cd->hwm = cworkspace;
5928  cd->req_varyopt = 0;  cd->req_varyopt = 0;
 cd->nopartial = FALSE;  
5929  cd->had_accept = FALSE;  cd->had_accept = FALSE;
5930    
5931  /* Set up a starting, non-extracting bracket, then compile the expression. On  /* Set up a starting, non-extracting bracket, then compile the expression. On
# Line 5920  code = (uschar *)codestart; Line 5939  code = (uschar *)codestart;
5939    &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, NULL);    &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, NULL);
5940  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
5941  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
5942    re->flags = cd->external_flags;
5943    
 if (cd->nopartial) re->options |= PCRE_NOPARTIAL;  
5944  if (cd->had_accept) reqbyte = -1;   /* Must disable after (*ACCEPT) */  if (cd->had_accept) reqbyte = -1;   /* Must disable after (*ACCEPT) */
5945    
5946  /* If not reached end of pattern on success, there's an excess bracket. */  /* If not reached end of pattern on success, there's an excess bracket. */
# Line 5993  if ((re->options & PCRE_ANCHORED) == 0) Line 6012  if ((re->options & PCRE_ANCHORED) == 0)
6012        int ch = firstbyte & 255;        int ch = firstbyte & 255;
6013        re->first_byte = ((firstbyte & REQ_CASELESS) != 0 &&        re->first_byte = ((firstbyte & REQ_CASELESS) != 0 &&
6014           cd->fcc[ch] == ch)? ch : firstbyte;           cd->fcc[ch] == ch)? ch : firstbyte;
6015        re->options |= PCRE_FIRSTSET;        re->flags |= PCRE_FIRSTSET;
6016        }        }
6017      else if (is_startline(codestart, 0, cd->backref_map))      else if (is_startline(codestart, 0, cd->backref_map))
6018        re->options |= PCRE_STARTLINE;        re->flags |= PCRE_STARTLINE;
6019      }      }
6020    }    }
6021    
# Line 6010  if (reqbyte >= 0 && Line 6029  if (reqbyte >= 0 &&
6029    int ch = reqbyte & 255;    int ch = reqbyte & 255;
6030    re->req_byte = ((reqbyte & REQ_CASELESS) != 0 &&    re->req_byte = ((reqbyte & REQ_CASELESS) != 0 &&
6031      cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte;      cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte;
6032    re->options |= PCRE_REQCHSET;    re->flags |= PCRE_REQCHSET;
6033    }    }
6034    
6035  /* Print out the compiled data if debugging is enabled. This is never the  /* Print out the compiled data if debugging is enabled. This is never the
# Line 6023  printf("Length = %d top_bracket = %d top Line 6042  printf("Length = %d top_bracket = %d top
6042    
6043  printf("Options=%08x\n", re->options);  printf("Options=%08x\n", re->options);
6044    
6045  if ((re->options & PCRE_FIRSTSET) != 0)  if ((re->flags & PCRE_FIRSTSET) != 0)
6046    {    {
6047    int ch = re->first_byte & 255;    int ch = re->first_byte & 255;
6048    const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)?    const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)?
# Line 6032  if ((re->options & PCRE_FIRSTSET) != 0) Line 6051  if ((re->options & PCRE_FIRSTSET) != 0)
6051      else printf("First char = \\x%02x%s\n", ch, caseless);      else printf("First char = \\x%02x%s\n", ch, caseless);
6052    }    }
6053    
6054  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6055    {    {
6056    int ch = re->req_byte & 255;    int ch = re->req_byte & 255;
6057    const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)?    const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)?

Legend:
Removed from v.227  
changed lines
  Added in v.230

  ViewVC Help
Powered by ViewVC 1.1.5