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

Diff of /code/trunk/pcre_compile.c

Parent Directory | Revision Log | View Patch Patch

-revision 97 by ph10,
Mon Mar  5 12:36:47 2007 UTC
+revision 172 by ph10,
Tue Jun  5 10:40:13 2007 UTC
 Line 6
  and semantics are as close as possible to those of the Perl 5 language.
                         Written by Philip Hazel
-            Copyright (c) 1997-2006 University of Cambridge
+            Copyright (c) 1997-2007 University of Cambridge
  -----------------------------------------------------------------------------
  Redistribution and use in source and binary forms, with or without
 Line 87 
 static const short int escapes[] = {
 ,      0,      0,      0,      0,      0,      0,      0,   /* 0 - 7 */
 ,      0,    ':',    ';',    '<',    '=',    '>',    '?',   /* 8 - ? */
     '@', -ESC_A, -ESC_B, -ESC_C, -ESC_D, -ESC_E,      0, -ESC_G,   /* @ - G */
-,      0,      0,      0,      0,      0,      0,      0,   /* H - O */
+,      0,      0, -ESC_K,      0,      0,      0,      0,   /* H - O */
  -ESC_P, -ESC_Q, -ESC_R, -ESC_S,      0,      0,      0, -ESC_W,   /* P - W */
  -ESC_X,      0, -ESC_Z,    '[',   '\\',    ']',    '^',    '_',   /* X - _ */
     '`',      7, -ESC_b,      0, -ESC_d,  ESC_e,  ESC_f,      0,   /* ` - g */
 Line 208 
 static const char *error_texts[] = {
    "malformed number or name after (?(",
    "conditional group contains more than two branches",
    "assertion expected after (?(",
-   "(?R or (?digits must be followed by )",
+   "(?R or (?[+-]digits must be followed by )",
    /* 30 */
    "unknown POSIX class name",
    "POSIX collating elements are not supported",
 Line 242 
 static const char *error_texts[] = {
    /* 55 */
    "repeating a DEFINE group is not allowed",
    "inconsistent NEWLINE options",
-   "\\g is not followed by an (optionally braced) non-zero number"
+   "\\g is not followed by a braced name or an optionally braced non-zero number",
+   "(?+ or (?- or (?(+ or (?(- must be followed by a non-zero number"
  };
-Line 452 
 else
+Line 453 
 else
      /* \g must be followed by a number, either plain or braced. If positive, it
      is an absolute backreference. If negative, it is a relative backreference.
-     This is a Perl 5.10 feature. */
+     This is a Perl 5.10 feature. Perl 5.10 also supports \g{name} as a
+     reference to a named group. This is part of Perl's movement towards a
+     unified syntax for back references. As this is synonymous with \k{name}, we
+     fudge it up by pretending it really was \k. */
      case 'g':
      if (ptr[1] == '{')
        {
+       const uschar *p;
+       for (p = ptr+2; *p != 0 && *p != '}'; p++)
+         if (*p != '-' && (digitab[*p] & ctype_digit) == 0) break;
+       if (*p != 0 && *p != '}')
+         {
+         c = -ESC_k;
+         break;
+         }
        braced = TRUE;
        ptr++;
        }
-Line 1246 
 for (;;)
+Line 1258 
 for (;;)
    else
      {
      code += _pcre_OP_lengths[c];
+ #ifdef SUPPORT_UTF8
      if (utf8) switch(c)
        {
        case OP_CHAR:
-Line 1266 
 for (;;)
+Line 1279 
 for (;;)
        if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f];
        break;
        }
+ #endif
      }
    }
  }
-Line 1309 
 for (;;)
+Line 1323 
 for (;;)
    else
      {
      code += _pcre_OP_lengths[c];
+ #ifdef SUPPORT_UTF8
      if (utf8) switch(c)
        {
        case OP_CHAR:
-Line 1329 
 for (;;)
+Line 1344 
 for (;;)
        if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f];
        break;
        }
+ #endif
      }
    }
  }
-Line 1366 
 for (code = first_significant_code(code
+Line 1382 
 for (code = first_significant_code(code
    c = *code;
+   /* Groups with zero repeats can of course be empty; skip them. */
+   if (c == OP_BRAZERO || c == OP_BRAMINZERO)
+     {
+     code += _pcre_OP_lengths[c];
+     do code += GET(code, 1); while (*code == OP_ALT);
+     c = *code;
+     continue;
+     }
+   /* For other groups, scan the branches. */
    if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE)
      {
      BOOL empty_branch;
-Line 1382 
 for (code = first_significant_code(code
+Line 1410 
 for (code = first_significant_code(code
        }
      while (*code == OP_ALT);
      if (!empty_branch) return FALSE;   /* All branches are non-empty */
+     c = *code;
-     /* Move past the KET and fudge things so that the increment in the "for"
-     above has no effect. */
-     c = OP_END;
-     code += 1 + LINK_SIZE - _pcre_OP_lengths[c];
      continue;
      }
-Line 2091 
 for (;; ptr++)
+Line 2114 
 for (;; ptr++)
    int class_lastchar;
    int newoptions;
    int recno;
+   int refsign;
    int skipbytes;
    int subreqbyte;
    int subfirstbyte;
-Line 3617 
 for (;; ptr++)
+Line 3641 
 for (;; ptr++)
          code[1+LINK_SIZE] = OP_CREF;
          skipbytes = 3;
+         refsign = -1;
          /* Check for a test for recursion in a named group. */
-Line 3640 
 for (;; ptr++)
+Line 3665 
 for (;; ptr++)
            terminator = '\'';
            ptr++;
            }
-         else terminator = 0;
+         else
+           {
+           terminator = 0;
+           if (ptr[1] == '-' || ptr[1] == '+') refsign = *(++ptr);
+           }
          /* We now expect to read a name; any thing else is an error */
-Line 3676 
 for (;; ptr++)
+Line 3705 
 for (;; ptr++)
          if (lengthptr != NULL) break;
          /* In the real compile we do the work of looking for the actual
-         reference. */
+         reference. If the string started with "+" or "-" we require the rest to
+         be digits, in which case recno will be set. */
+         if (refsign > 0)
+           {
+           if (recno <= 0)
+             {
+             *errorcodeptr = ERR58;
+             goto FAILED;
+             }
+           if (refsign == '-')
+             {
+             recno = cd->bracount - recno + 1;
+             if (recno <= 0)
+               {
+               *errorcodeptr = ERR15;
+               goto FAILED;
+               }
+             }
+           else recno += cd->bracount;
+           PUT2(code, 2+LINK_SIZE, recno);
+           break;
+           }
+         /* Otherwise (did not start with "+" or "-"), start by looking for the
+         name. */
          slot = cd->name_table;
          for (i = 0; i < cd->names_found; i++)
-Line 3995 
 for (;; ptr++)
+Line 4049 
 for (;; ptr++)
          /* ------------------------------------------------------------ */
+         case '-': case '+':
          case '0': case '1': case '2': case '3': case '4':   /* Recursion or */
          case '5': case '6': case '7': case '8': case '9':   /* subroutine */
            {
            const uschar *called;
+           if ((refsign = *ptr) == '+') ptr++;
+           else if (refsign == '-')
+             {
+             if ((digitab[ptr[1]] & ctype_digit) == 0)
+               goto OTHER_CHAR_AFTER_QUERY;
+             ptr++;
+             }
            recno = 0;
            while((digitab[*ptr] & ctype_digit) != 0)
              recno = recno * 10 + *ptr++ - '0';
            if (*ptr != ')')
              {
              *errorcodeptr = ERR29;
              goto FAILED;
              }
+           if (refsign == '-')
+             {
+             if (recno == 0)
+               {
+               *errorcodeptr = ERR58;
+               goto FAILED;
+               }
+             recno = cd->bracount - recno + 1;
+             if (recno <= 0)
+               {
+               *errorcodeptr = ERR15;
+               goto FAILED;
+               }
+             }
+           else if (refsign == '+')
+             {
+             if (recno == 0)
+               {
+               *errorcodeptr = ERR58;
+               goto FAILED;
+               }
+             recno += cd->bracount;
+             }
            /* Come here from code above that handles a named recursion */
            HANDLE_RECURSION:
-Line 4080 
 for (;; ptr++)
+Line 4169 
 for (;; ptr++)
          /* ------------------------------------------------------------ */
          default:              /* Other characters: check option setting */
+         OTHER_CHAR_AFTER_QUERY:
          set = unset = 0;
          optset = &set;
-Line 4230 
 for (;; ptr++)
+Line 4320 
 for (;; ptr++)
      is on the bracket. */
      /* If this is a conditional bracket, check that there are no more than
-     two branches in the group, or just one if it's a DEFINE group. */
+     two branches in the group, or just one if it's a DEFINE group. We do this
+     in the real compile phase, not in the pre-pass, where the whole group may
+     not be available. */
-     if (bravalue == OP_COND)
+     if (bravalue == OP_COND && lengthptr == NULL)
        {
        uschar *tc = code;
        int condcount = 0;
-Line 4392 
 for (;; ptr++)
+Line 4484 
 for (;; ptr++)
        zerofirstbyte = firstbyte;
        zeroreqbyte = reqbyte;
-       /* \k<name> or \k'name' is a back reference by name (Perl syntax) */
+       /* \k<name> or \k'name' is a back reference by name (Perl syntax).
+       We also support \k{name} (.NET syntax) */
-       if (-c == ESC_k && (ptr[1] == '<' || ptr[1] == '\''))
+       if (-c == ESC_k && (ptr[1] == '<' || ptr[1] == '\'' || ptr[1] == '{'))
          {
          is_recurse = FALSE;
-         terminator = (*(++ptr) == '<')? '>' : '\'';
+         terminator = (*(++ptr) == '<')? '>' : (*ptr == '\'')? '\'' : '}';
          goto NAMED_REF_OR_RECURSE;
          }
-Line 4563 
 This function is used during the pre-com
+Line 4656 
 This function is used during the pre-com
  out the amount of memory needed, as well as during the real compile phase. The
  value of lengthptr distinguishes the two phases.
- Argument:
+ Arguments:
    options        option bits, including any changes for this subpattern
    oldims         previous settings of ims option bits
    codeptr        -> the address of the current code pointer
-Line 4716 
 for (;;)
+Line 4809 
 for (;;)
        }
      }
-   /* Reached end of expression, either ')' or end of pattern. Go back through
+   /* Reached end of expression, either ')' or end of pattern. In the real
-   the alternative branches and reverse the chain of offsets, with the field in
+   compile phase, go back through the alternative branches and reverse the chain
-   the BRA item now becoming an offset to the first alternative. If there are
+   of offsets, with the field in the BRA item now becoming an offset to the
-   no alternatives, it points to the end of the group. The length in the
+   first alternative. If there are no alternatives, it points to the end of the
-   terminating ket is always the length of the whole bracketed item. If any of
+   group. The length in the terminating ket is always the length of the whole
-   the ims options were changed inside the group, compile a resetting op-code
+   bracketed item. If any of the ims options were changed inside the group,
-   following, except at the very end of the pattern. Return leaving the pointer
+   compile a resetting op-code following, except at the very end of the pattern.
-   at the terminating char. */
+   Return leaving the pointer at the terminating char. */
    if (*ptr != '|')
      {
-     int branch_length = code - last_branch;
+     if (lengthptr == NULL)
-     do
        {
-       int prev_length = GET(last_branch, 1);
+       int branch_length = code - last_branch;
-       PUT(last_branch, 1, branch_length);
+       do
-       branch_length = prev_length;
+         {
-       last_branch -= branch_length;
+         int prev_length = GET(last_branch, 1);
+         PUT(last_branch, 1, branch_length);
+         branch_length = prev_length;
+         last_branch -= branch_length;
+         }
+       while (branch_length > 0);
        }
-     while (branch_length > 0);
      /* Fill in the ket */
-Line 4762 
 for (;;)
+Line 4858 
 for (;;)
      return TRUE;
      }
-   /* Another branch follows; insert an "or" node. Its length field points back
+   /* Another branch follows. In the pre-compile phase, we can move the code
+   pointer back to where it was for the start of the first branch. (That is,
+   pretend that each branch is the only one.)
+   In the real compile phase, insert an ALT node. Its length field points back
    to the previous branch while the bracket remains open. At the end the chain
    is reversed. It's done like this so that the start of the bracket has a
    zero offset until it is closed, making it possible to detect recursion. */
-   *code = OP_ALT;
+   if (lengthptr != NULL)
-   PUT(code, 1, code - last_branch);
+     {
-   bc.current = last_branch = code;
+     code = *codeptr + 1 + LINK_SIZE + skipbytes;
-   code += 1 + LINK_SIZE;
+     length += 1 + LINK_SIZE;
+     }
+   else
+     {
+     *code = OP_ALT;
+     PUT(code, 1, code - last_branch);
+     bc.current = last_branch = code;
+     code += 1 + LINK_SIZE;
+     }
    ptr++;
-   length += 1 + LINK_SIZE;
    }
  /* Control never reaches here */
  }
-Line 5039 
 Returns:        pointer to compiled data
+Line 5147 
 Returns:        pointer to compiled data
                  with errorptr and erroroffset set
  */
- PCRE_DATA_SCOPE pcre *
+ PCRE_EXP_DEFN pcre *
  pcre_compile(const char *pattern, int options, const char **errorptr,
    int *erroroffset, const unsigned char *tables)
  {
-Line 5047 
 return pcre_compile2(pattern, options, N
+Line 5155 
 return pcre_compile2(pattern, options, N
  }
- PCRE_DATA_SCOPE pcre *
+ PCRE_EXP_DEFN pcre *
  pcre_compile2(const char *pattern, int options, int *errorcodeptr,
    const char **errorptr, int *erroroffset, const unsigned char *tables)
  {
-Line 5096 
 if (errorcodeptr != NULL) *errorcodeptr
+Line 5204 
 if (errorcodeptr != NULL) *errorcodeptr
  if (erroroffset == NULL)
    {
    errorcode = ERR16;
-   goto PCRE_EARLY_ERROR_RETURN;
+   goto PCRE_EARLY_ERROR_RETURN2;
    }
  *erroroffset = 0;
-Line 5109 
 if (utf8 && (options & PCRE_NO_UTF8_CHEC
+Line 5217 
 if (utf8 && (options & PCRE_NO_UTF8_CHEC
       (*erroroffset = _pcre_valid_utf8((uschar *)pattern, -1)) >= 0)
    {
    errorcode = ERR44;
-   goto PCRE_UTF8_ERROR_RETURN;
+   goto PCRE_EARLY_ERROR_RETURN2;
    }
  #else
  if ((options & PCRE_UTF8) != 0)
-Line 5134 
 cd->cbits = tables + cbits_offset;
+Line 5242 
 cd->cbits = tables + cbits_offset;
  cd->ctypes = tables + ctypes_offset;
  /* Handle different types of newline. The three bits give seven cases. The
- current code allows for fixed one- or two-byte sequences, plus "any". */
+ current code allows for fixed one- or two-byte sequences, plus "any" and
+ "anycrlf". */
  switch (options & (PCRE_NEWLINE_CRLF | PCRE_NEWLINE_ANY))
    {
-Line 5144 
 switch (options & (PCRE_NEWLINE_CRLF | P
+Line 5253 
 switch (options & (PCRE_NEWLINE_CRLF | P
    case PCRE_NEWLINE_CR+
         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;
    case PCRE_NEWLINE_ANY: newline = -1; break;
+   case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
    default: errorcode = ERR56; goto PCRE_EARLY_ERROR_RETURN;
    }
- if (newline < 0)
+ if (newline == -2)
+   {
+   cd->nltype = NLTYPE_ANYCRLF;
+   }
+ else if (newline < 0)
    {
    cd->nltype = NLTYPE_ANY;
    }
-Line 5321 
 if (errorcode != 0)
+Line 5435 
 if (errorcode != 0)
    (pcre_free)(re);
    PCRE_EARLY_ERROR_RETURN:
    *erroroffset = ptr - (const uschar *)pattern;
- #ifdef SUPPORT_UTF8
+   PCRE_EARLY_ERROR_RETURN2:
-   PCRE_UTF8_ERROR_RETURN:
- #endif
    *errorptr = error_texts[errorcode];
    if (errorcodeptr != NULL) *errorcodeptr = errorcode;
    return NULL;
-Line 5413 
 if ((re->options & PCRE_REQCHSET) != 0)
+Line 5525 
 if ((re->options & PCRE_REQCHSET) != 0)
      else printf("Req char = \\x%02x%s\n", ch, caseless);
    }
- pcre_printint(re, stdout);
+ pcre_printint(re, stdout, TRUE);
  /* This check is done here in the debugging case so that the code that
  was compiled can be seen. */

 Legend:



Removed from v.97
 


changed lines


 
Added in v.172
 Legend:



Removed from v.97
 


changed lines


 
Added in v.172
-Removed from v.97
+Added in v.172

webmaster@exim.org	ViewVC Help
Powered by ViewVC 1.1.12