/[pcre]/code/tags/pcre-4.1/pcre.c

Diff of /code/tags/pcre-4.1/pcre.c

Parent Directory | Revision Log | View Patch Patch

-revision 39 by nigel,
Sat Feb 24 21:39:13 2007 UTC
+revision 43 by nigel,
Sat Feb 24 21:39:21 2007 UTC
 Line 9 
 the file Tech.Notes for some information
  Written by: Philip Hazel <ph10@cam.ac.uk>
-            Copyright (c) 1997-1999 University of Cambridge
+            Copyright (c) 1997-2000 University of Cambridge
  -----------------------------------------------------------------------------
  Permission is granted to anyone to use this software for any purpose on any
 Line 82 
 static const char *OP_names[] = {
    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",
    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",
    "*", "*?", "+", "+?", "?", "??", "{", "{",
-   "class", "Ref",
+   "class", "Ref", "Recurse",
    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",
    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cref",
    "Brazero", "Braminzero", "Bra"
 Line 107 
 static const short int escapes[] = {
 ,      0, -ESC_z                                            /* x - z */
  };
+ /* Tables of names of POSIX character classes and their lengths. The list is
+ terminated by a zero length entry. The first three must be alpha, upper, lower,
+ as this is assumed for handling case independence. */
+ static const char *posix_names[] = {
+   "alpha", "lower", "upper",
+   "alnum", "ascii", "cntrl", "digit", "graph",
+   "print", "punct", "space", "word",  "xdigit" };
+ static const uschar posix_name_lengths[] = {
+, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 };
+ /* Table of class bit maps for each POSIX class; up to three may be combined
+ to form the class. */
+ static const int posix_class_maps[] = {
+   cbit_lower, cbit_upper, -1,             /* alpha */
+   cbit_lower, -1,         -1,             /* lower */
+   cbit_upper, -1,         -1,             /* upper */
+   cbit_digit, cbit_lower, cbit_upper,     /* alnum */
+   cbit_print, cbit_cntrl, -1,             /* ascii */
+   cbit_cntrl, -1,         -1,             /* cntrl */
+   cbit_digit, -1,         -1,             /* digit */
+   cbit_graph, -1,         -1,             /* graph */
+   cbit_print, -1,         -1,             /* print */
+   cbit_punct, -1,         -1,             /* punct */
+   cbit_space, -1,         -1,             /* space */
+   cbit_word,  -1,         -1,             /* word */
+   cbit_xdigit,-1,         -1              /* xdigit */
+ };
  /* Definition to allow mutual recursion */
  static BOOL
-Line 161 
 return XSTRING(PCRE_MAJOR) "." XSTRING(P
+Line 193 
 return XSTRING(PCRE_MAJOR) "." XSTRING(P
  /*************************************************
- *       Return info about a compiled pattern     *
+ * (Obsolete) Return info about compiled pattern  *
  *************************************************/
- /* This function picks potentially useful data out of the private
+ /* This is the original "info" function. It picks potentially useful data out
- structure. The public options are passed back in an int - though the
+ of the private structure, but its interface was too rigid. It remains for
- re->options field has been expanded to a long int, all the public options
+ backwards compatibility. The public options are passed back in an int - though
+ the re->options field has been expanded to a long int, all the public options
  at the low end of it, and so even on 16-bit systems this will still be OK.
  Therefore, I haven't changed the API for pcre_info().
-Line 177 
 Arguments:
+Line 210 
 Arguments:
                  or -1 if multiline and all branches start ^,
                  or -2 otherwise
- Returns:        number of identifying extraction brackets
+ Returns:        number of capturing subpatterns
                  or negative values on error
  */
-Line 196 
 return re->top_bracket;
+Line 229 
 return re->top_bracket;
+ /*************************************************
+ *        Return info about compiled pattern      *
+ *************************************************/
+ /* This is a newer "info" function which has an extensible interface so
+ that additional items can be added compatibly.
+ Arguments:
+   external_re      points to compiled code
+   external_study   points to study data, or NULL
+   what             what information is required
+   where            where to put the information
+ Returns:           0 if data returned, negative on error
+ */
+ int
+ pcre_fullinfo(const pcre *external_re, const pcre_extra *study_data, int what,
+   void *where)
+ {
+ const real_pcre *re = (const real_pcre *)external_re;
+ const real_pcre_extra *study = (const real_pcre_extra *)study_data;
+ if (re == NULL || where == NULL) return PCRE_ERROR_NULL;
+ if (re->magic_number != MAGIC_NUMBER) return PCRE_ERROR_BADMAGIC;
+ switch (what)
+   {
+   case PCRE_INFO_OPTIONS:
+   *((unsigned long int *)where) = re->options & PUBLIC_OPTIONS;
+   break;
+   case PCRE_INFO_SIZE:
+   *((size_t *)where) = re->size;
+   break;
+   case PCRE_INFO_CAPTURECOUNT:
+   *((int *)where) = re->top_bracket;
+   break;
+   case PCRE_INFO_BACKREFMAX:
+   *((int *)where) = re->top_backref;
+   break;
+   case PCRE_INFO_FIRSTCHAR:
+   *((int *)where) =
+     ((re->options & PCRE_FIRSTSET) != 0)? re->first_char :
+     ((re->options & PCRE_STARTLINE) != 0)? -1 : -2;
+   break;
+   case PCRE_INFO_FIRSTTABLE:
+   *((const uschar **)where) =
+     (study != NULL && (study->options & PCRE_STUDY_MAPPED) != 0)?
+       study->start_bits : NULL;
+   break;
+   case PCRE_INFO_LASTLITERAL:
+   *((int *)where) =
+     ((re->options & PCRE_REQCHSET) != 0)? re->req_char : -1;
+   break;
+   default: return PCRE_ERROR_BADOPTION;
+   }
+ return 0;
+ }
  #ifdef DEBUG
  /*************************************************
-Line 255 
 check_escape(const uschar **ptrptr, cons
+Line 356 
 check_escape(const uschar **ptrptr, cons
    int options, BOOL isclass, compile_data *cd)
  {
  const uschar *ptr = *ptrptr;
- int c = *(++ptr) & 255;   /* Ensure > 0 on signed-char systems */
+ int c, i;
- int i;
+ c = *(++ptr) & 255;   /* Ensure > 0 on signed-char systems */
  if (c == 0) *errorptr = ERR1;
  /* Digits or letters may have special meaning; all others are literals. */
-Line 622 
 for (;;)
+Line 723 
 for (;;)
  /*************************************************
+ *           Check for POSIX class syntax         *
+ *************************************************/
+ /* This function is called when the sequence "[:" or "[." or "[=" is
+ encountered in a character class. It checks whether this is followed by an
+ optional ^ and then a sequence of letters, terminated by a matching ":]" or
+ ".]" or "=]".
+ Argument:
+   ptr      pointer to the initial [
+   endptr   where to return the end pointer
+   cd       pointer to compile data
+ Returns:   TRUE or FALSE
+ */
+ static BOOL
+ check_posix_syntax(const uschar *ptr, const uschar **endptr, compile_data *cd)
+ {
+ int terminator;          /* Don't combine these lines; the Solaris cc */
+ terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
+ if (*(++ptr) == '^') ptr++;
+ while ((cd->ctypes[*ptr] & ctype_letter) != 0) ptr++;
+ if (*ptr == terminator && ptr[1] == ']')
+   {
+   *endptr = ptr;
+   return TRUE;
+   }
+ return FALSE;
+ }
+ /*************************************************
+ *          Check POSIX class name                *
+ *************************************************/
+ /* This function is called to check the name given in a POSIX-style class entry
+ such as [:alnum:].
+ Arguments:
+   ptr        points to the first letter
+   len        the length of the name
+ Returns:     a value representing the name, or -1 if unknown
+ */
+ static int
+ check_posix_name(const uschar *ptr, int len)
+ {
+ register int yield = 0;
+ while (posix_name_lengths[yield] != 0)
+   {
+   if (len == posix_name_lengths[yield] &&
+     strncmp((const char *)ptr, posix_names[yield], len) == 0) return yield;
+   yield++;
+   }
+ return -1;
+ }
+ /*************************************************
  *           Compile one branch                   *
  *************************************************/
-Line 764 
 for (;; ptr++)
+Line 930 
 for (;; ptr++)
          goto FAILED;
          }
+       /* Handle POSIX class names. Perl allows a negation extension of the
+       form [:^name]. A square bracket that doesn't match the syntax is
+       treated as a literal. We also recognize the POSIX constructions
+       [.ch.] and [=ch=] ("collating elements") and fault them, as Perl
+.6 does. */
+       if (c == '[' &&
+           (ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') &&
+           check_posix_syntax(ptr, &tempptr, cd))
+         {
+         BOOL local_negate = FALSE;
+         int posix_class, i;
+         register const uschar *cbits = cd->cbits;
+         if (ptr[1] != ':')
+           {
+           *errorptr = ERR31;
+           goto FAILED;
+           }
+         ptr += 2;
+         if (*ptr == '^')
+           {
+           local_negate = TRUE;
+           ptr++;
+           }
+         posix_class = check_posix_name(ptr, tempptr - ptr);
+         if (posix_class < 0)
+           {
+           *errorptr = ERR30;
+           goto FAILED;
+           }
+         /* If matching is caseless, upper and lower are converted to
+         alpha. This relies on the fact that the class table starts with
+         alpha, lower, upper as the first 3 entries. */
+         if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
+           posix_class = 0;
+         /* Or into the map we are building up to 3 of the static class
+         tables, or their negations. */
+         posix_class *= 3;
+         for (i = 0; i < 3; i++)
+           {
+           int taboffset = posix_class_maps[posix_class + i];
+           if (taboffset < 0) break;
+           if (local_negate)
+             for (c = 0; c < 32; c++) class[c] |= ~cbits[c+taboffset];
+           else
+             for (c = 0; c < 32; c++) class[c] |= cbits[c+taboffset];
+           }
+         ptr = tempptr + 1;
+         class_charcount = 10;  /* Set > 1; assumes more than 1 per class */
+         continue;
+         }
        /* Backslash may introduce a single character, or it may introduce one
        of the specials, which just set a flag. Escaped items are checked for
        validity in the pre-compiling pass. The sequence \b is a special case.
-Line 791 
 for (;; ptr++)
+Line 1017 
 for (;; ptr++)
              continue;
              case ESC_w:
-             for (c = 0; c < 32; c++)
+             for (c = 0; c < 32; c++) class[c] |= cbits[c+cbit_word];
-               class[c] |= (cbits[c+cbit_digit] | cbits[c+cbit_word]);
              continue;
              case ESC_W:
-             for (c = 0; c < 32; c++)
+             for (c = 0; c < 32; c++) class[c] |= ~cbits[c+cbit_word];
-               class[c] |= ~(cbits[c+cbit_digit] | cbits[c+cbit_word]);
              continue;
              case ESC_s:
-Line 1360 
 for (;; ptr++)
+Line 1584 
 for (;; ptr++)
          ptr++;
          break;
+         case 'R':                 /* Pattern recursion */
+         *code++ = OP_RECURSE;
+         ptr++;
+         continue;
          default:                  /* Option setting */
          set = unset = 0;
          optset = &set;
-Line 2015 
 pcre_compile(const char *pattern, int op
+Line 2244 
 pcre_compile(const char *pattern, int op
  real_pcre *re;
  int length = 3;      /* For initial BRA plus length */
  int runlength;
- int c, size, reqchar, countlits;
+ int c, reqchar, countlits;
  int bracount = 0;
  int top_backref = 0;
  int branch_extra = 0;
  int branch_newextra;
  unsigned int brastackptr = 0;
+ size_t size;
  uschar *code;
  const uschar *ptr;
  compile_data compile_block;
-Line 2248 
 while ((c = *(++ptr)) != 0)
+Line 2478 
 while ((c = *(++ptr)) != 0)
          ptr += 2;
          break;
+         /* A recursive call to the regex is an extension, to provide the
+         facility which can be obtained by $(?p{perl-code}) in Perl 5.6. */
+         case 'R':
+         if (ptr[3] != ')')
+           {
+           *errorptr = ERR29;
+           goto PCRE_ERROR_RETURN;
+           }
+         ptr += 3;
+         length += 1;
+         break;
          /* Lookbehinds are in Perl from version 5.005 */
          case '<':
-Line 2550 
 if (re == NULL)
+Line 2793 
 if (re == NULL)
    return NULL;
    }
- /* Put in the magic number and the options. */
+ /* Put in the magic number, and save the size, options, and table pointer */
  re->magic_number = MAGIC_NUMBER;
+ re->size = size;
  re->options = options;
  re->tables = tables;
-Line 3147 
 for (;;)
+Line 3391 
 for (;;)
      ecode += 3;
      break;
+     /* Recursion matches the current regex, nested. If there are any capturing
+     brackets started but not finished, we have to save their starting points
+     and reinstate them after the recursion. However, we don't know how many
+     such there are (offset_top records the completed total) so we just have
+     to save all the potential data. There may be up to 99 such values, which
+     is a bit large to put on the stack, but using malloc for small numbers
+     seems expensive. As a compromise, the stack is used when there are fewer
+     than 16 values to store; otherwise malloc is used. A problem is what to do
+     if the malloc fails ... there is no way of returning to the top level with
+     an error. Save the top 15 values on the stack, and accept that the rest
+     may be wrong. */
+     case OP_RECURSE:
+       {
+       BOOL rc;
+       int *save;
+       int stacksave[15];
+       c = md->offset_max;
+       if (c < 16) save = stacksave; else
+         {
+         save = (int *)(pcre_malloc)((c+1) * sizeof(int));
+         if (save == NULL)
+           {
+           save = stacksave;
+           c = 15;
+           }
+         }
+       for (i = 1; i <= c; i++)
+         save[i] = md->offset_vector[md->offset_end - i];
+       rc = match(eptr, md->start_pattern, offset_top, md, ims, FALSE, eptrb);
+       for (i = 1; i <= c; i++)
+         md->offset_vector[md->offset_end - i] = save[i];
+       if (save != stacksave) (pcre_free)(save);
+       if (!rc) return FALSE;
+       /* In case the recursion has set more capturing values, save the final
+       number, then move along the subject till after the recursive match,
+       and advance one byte in the pattern code. */
+       offset_top = md->end_offset_top;
+       eptr = md->end_match_ptr;
+       ecode++;
+       }
+     break;
      /* "Once" brackets are like assertion brackets except that after a match,
      the point in the subject string is not moved back. Thus there can never be
-Line 4216 
 if (re == NULL || subject == NULL ||
+Line 4507 
 if (re == NULL || subject == NULL ||
     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;
  if (re->magic_number != MAGIC_NUMBER) return PCRE_ERROR_BADMAGIC;
+ match_block.start_pattern = re->code;
  match_block.start_subject = (const uschar *)subject;
  match_block.end_subject = match_block.start_subject + length;
  end_subject = match_block.end_subject;

 Legend:



Removed from v.39
 


changed lines


 
Added in v.43
 Legend:



Removed from v.39
 


changed lines


 
Added in v.43
-Removed from v.39
+Added in v.43

webmaster@exim.org	ViewVC Help
Powered by ViewVC 1.1.12