/[pcre]/code/trunk/pcre_compile.c
ViewVC logotype

Diff of /code/trunk/pcre_compile.c

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

revision 1395 by ph10, Sat Nov 9 16:54:52 2013 UTC revision 1396 by ph10, Sun Nov 10 19:04:34 2013 UTC
# Line 4446  for (;; ptr++) Line 4446  for (;; ptr++)
4446    /* Get next character in the pattern */    /* Get next character in the pattern */
4447    
4448    c = *ptr;    c = *ptr;
4449    
4450    /* If we are at the end of a nested substitution, revert to the outer level    /* If we are at the end of a nested substitution, revert to the outer level
4451    string. Nesting only happens one level deep. */    string. Nesting only happens one level deep. */
4452    
# Line 4548  for (;; ptr++) Line 4548  for (;; ptr++)
4548          }          }
4549        goto NORMAL_CHAR;        goto NORMAL_CHAR;
4550        }        }
4551        /* Control does not reach here. */
4552        }
4553    
4554      /* In extended mode, skip white space and comments. We need a loop in order
4555      to check for more white space and more comments after a comment. */
4556    
4557      if ((options & PCRE_EXTENDED) != 0)
4558        {
4559        for (;;)
4560          {
4561          while (MAX_255(c) && (cd->ctypes[c] & ctype_space) != 0) c = *(++ptr);
4562          if (c != CHAR_NUMBER_SIGN) break;
4563          ptr++;
4564          while (*ptr != CHAR_NULL)
4565            {
4566            if (IS_NEWLINE(ptr))         /* For non-fixed-length newline cases, */
4567              {                          /* IS_NEWLINE sets cd->nllen. */
4568              ptr += cd->nllen;
4569              break;
4570              }
4571            ptr++;
4572    #ifdef SUPPORT_UTF
4573            if (utf) FORWARDCHAR(ptr);
4574    #endif
4575            }
4576          c = *ptr;     /* Either NULL or the char after a newline */
4577          }
4578      }      }
4579    
4580      /* See if the next thing is a quantifier. */
4581    
4582    is_quantifier =    is_quantifier =
4583      c == CHAR_ASTERISK || c == CHAR_PLUS || c == CHAR_QUESTION_MARK ||      c == CHAR_ASTERISK || c == CHAR_PLUS || c == CHAR_QUESTION_MARK ||
4584      (c == CHAR_LEFT_CURLY_BRACKET && is_counted_repeat(ptr+1));      (c == CHAR_LEFT_CURLY_BRACKET && is_counted_repeat(ptr+1));
# Line 4565  for (;; ptr++) Line 4594  for (;; ptr++)
4594      previous_callout = NULL;      previous_callout = NULL;
4595      }      }
4596    
4597    /* In extended mode, skip white space and comments. */    /* Create auto callout, except for quantifiers, or while processing property
4598      strings that are substituted for \w etc in UCP mode. */
   if ((options & PCRE_EXTENDED) != 0)  
     {  
     if (MAX_255(*ptr) && (cd->ctypes[c] & ctype_space) != 0) continue;  
     if (c == CHAR_NUMBER_SIGN)  
       {  
       ptr++;  
       while (*ptr != CHAR_NULL)  
         {  
         if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }  
         ptr++;  
 #ifdef SUPPORT_UTF  
         if (utf) FORWARDCHAR(ptr);  
 #endif  
         }  
       if (*ptr != CHAR_NULL) continue;  
   
       /* Else fall through to handle end of string */  
       c = 0;  
       }  
     }  
   
   /* No auto callout for quantifiers, or while processing property strings that  
   are substituted for \w etc in UCP mode. */  
4599    
4600    if ((options & PCRE_AUTO_CALLOUT) != 0 && !is_quantifier && nestptr == NULL)    if ((options & PCRE_AUTO_CALLOUT) != 0 && !is_quantifier && nestptr == NULL)
4601      {      {
4602      previous_callout = code;      previous_callout = code;
4603      code = auto_callout(code, ptr, cd);      code = auto_callout(code, ptr, cd);
4604      }      }
4605    
4606      /* Process the next pattern item. */
4607    
4608    switch(c)    switch(c)
4609      {      {
4610      /* ===================================================================*/      /* ===================================================================*/
4611      case 0:                        /* The branch terminates at string end */      case CHAR_NULL:                /* The branch terminates at string end */
4612      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
4613      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
4614      *firstcharptr = firstchar;      *firstcharptr = firstchar;
# Line 5445  for (;; ptr++) Line 5453  for (;; ptr++)
5453      insert something before it. */      insert something before it. */
5454    
5455      tempcode = previous;      tempcode = previous;
5456    
5457        /* Before checking for a possessive quantifier, we must skip over
5458        whitespace and comments in extended mode because Perl allows white space at
5459        this point. */
5460    
5461        if ((options & PCRE_EXTENDED) != 0)
5462          {
5463          const pcre_uchar *p = ptr + 1;
5464          for (;;)
5465            {
5466            while (MAX_255(*p) && (cd->ctypes[*p] & ctype_space) != 0) p++;
5467            if (*p != CHAR_NUMBER_SIGN) break;
5468            p++;
5469            while (*p != CHAR_NULL)
5470              {
5471              if (IS_NEWLINE(p))         /* For non-fixed-length newline cases, */
5472                {                        /* IS_NEWLINE sets cd->nllen. */
5473                p += cd->nllen;
5474                break;
5475                }
5476              p++;
5477    #ifdef SUPPORT_UTF
5478              if (utf) FORWARDCHAR(p);
5479    #endif
5480              }           /* Loop for comment characters */
5481            }             /* Loop for multiple comments */
5482          ptr = p - 1;    /* Character before the next significant one. */
5483          }
5484    
5485      /* If the next character is '+', we have a possessive quantifier. This      /* If the next character is '+', we have a possessive quantifier. This
5486      implies greediness, whatever the setting of the PCRE_UNGREEDY option.      implies greediness, whatever the setting of the PCRE_UNGREEDY option.
# Line 7752  for (;; ptr++) Line 7788  for (;; ptr++)
7788    
7789      /* ===================================================================*/      /* ===================================================================*/
7790      /* Handle a literal character. It is guaranteed not to be whitespace or #      /* Handle a literal character. It is guaranteed not to be whitespace or #
7791      when the extended flag is set. If we are in UTF-8 mode, it may be a      when the extended flag is set. If we are in a UTF mode, it may be a
7792      multi-byte literal character. */      multi-unit literal character. */
7793    
7794      default:      default:
7795      NORMAL_CHAR:      NORMAL_CHAR:
# Line 8899  else Line 8935  else
8935      cd->nl[0] = newline;      cd->nl[0] = newline;
8936      }      }
8937    }    }
8938    
8939  /* Maximum back reference and backref bitmap. The bitmap records up to 31 back  /* Maximum back reference and backref bitmap. The bitmap records up to 31 back
8940  references to help in deciding whether (.*) can be treated as anchored or not.  references to help in deciding whether (.*) can be treated as anchored or not.
8941  */  */
# Line 8952  outside can help speed up starting point Line 8988  outside can help speed up starting point
8988  ptr += skipatstart;  ptr += skipatstart;
8989  code = cworkspace;  code = cworkspace;
8990  *code = OP_BRA;  *code = OP_BRA;
8991    
8992  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
8993    FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,    FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,
8994    cd, &length);    cd, &length);
Legend:
Removed from v.1395  
changed lines
  Added in v.1396
  ViewVC Help
Powered by ViewVC 1.1.5