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

Diff of /code/trunk/pcre_compile.c

Parent Directory | Revision Log | View Patch Patch

-revision 149 by ph10,
Mon Apr 16 15:28:08 2007 UTC
+revision 176 by ph10,
Mon Jun 11 13:48:37 2007 UTC
 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 373 
 static const unsigned char ebcdic_charta
+Line 374 
 static const unsigned char ebcdic_charta
  /* Definition to allow mutual recursion */
  static BOOL
-   compile_regex(int, int, uschar **, const uschar **, int *, BOOL, int, int *,
+   compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,
-     int *, branch_chain *, compile_data *, int *);
+     int *, int *, branch_chain *, compile_data *, int *);
-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 1370 
 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 1386 
 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 2110 
 for (;; ptr++)
    BOOL possessive_quantifier;
    BOOL is_quantifier;
    BOOL is_recurse;
+   BOOL reset_bracount;
    int class_charcount;
    int class_lastchar;
    int newoptions;
    int recno;
+   int refsign;
    int skipbytes;
    int subreqbyte;
    int subfirstbyte;
-Line 2659 
 for (;; ptr++)
+Line 2680 
 for (;; ptr++)
              unsigned int origd = d;
              while (get_othercase_range(&cc, origd, &occ, &ocd))
                {
-               if (occ >= c && ocd <= d) continue;  /* Skip embedded ranges */
+               if (occ >= (unsigned int)c &&
+                   ocd <= (unsigned int)d)
+                 continue;                          /* Skip embedded ranges */
-               if (occ < c  && ocd >= c - 1)        /* Extend the basic range */
+               if (occ < (unsigned int)c  &&
+                   ocd >= (unsigned int)c - 1)      /* Extend the basic range */
                  {                                  /* if there is overlap,   */
                  c = occ;                           /* noting that if occ < c */
                  continue;                          /* we can't have ocd > d  */
                  }                                  /* because a subrange is  */
-               if (ocd > d && occ <= d + 1)         /* always shorter than    */
+               if (ocd > (unsigned int)d &&
+                   occ <= (unsigned int)d + 1)      /* always shorter than    */
                  {                                  /* the basic range.       */
                  d = ocd;
                  continue;
-Line 3564 
 for (;; ptr++)
+Line 3589 
 for (;; ptr++)
      skipbytes = 0;
      bravalue = OP_CBRA;
      save_hwm = cd->hwm;
+     reset_bracount = FALSE;
      if (*(++ptr) == '?')
        {
-Line 3586 
 for (;; ptr++)
+Line 3612 
 for (;; ptr++)
          /* ------------------------------------------------------------ */
+         case '|':                 /* Reset capture count for each branch */
+         reset_bracount = TRUE;
+         /* Fall through */
+         /* ------------------------------------------------------------ */
          case ':':                 /* Non-capturing bracket */
          bravalue = OP_BRA;
          ptr++;
-Line 3621 
 for (;; ptr++)
+Line 3652 
 for (;; ptr++)
          code[1+LINK_SIZE] = OP_CREF;
          skipbytes = 3;
+         refsign = -1;
          /* Check for a test for recursion in a named group. */
-Line 3644 
 for (;; ptr++)
+Line 3676 
 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 3680 
 for (;; ptr++)
+Line 3716 
 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 3999 
 for (;; ptr++)
+Line 4060 
 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 4084 
 for (;; ptr++)
+Line 4180 
 for (;; ptr++)
          /* ------------------------------------------------------------ */
          default:              /* Other characters: check option setting */
+         OTHER_CHAR_AFTER_QUERY:
          set = unset = 0;
          optset = &set;
-Line 4218 
 for (;; ptr++)
+Line 4315 
 for (;; ptr++)
           errorcodeptr,                 /* Where to put an error message */
           (bravalue == OP_ASSERTBACK ||
            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */
+          reset_bracount,               /* True if (?| group */
           skipbytes,                    /* Skip over bracket number */
           &subfirstbyte,                /* For possible first char */
           &subreqbyte,                  /* For possible last char */
-Line 4234 
 for (;; ptr++)
+Line 4332 
 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 4396 
 for (;; ptr++)
+Line 4496 
 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 4567 
 This function is used during the pre-com
+Line 4668 
 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
    ptrptr         -> the address of the current pattern pointer
    errorcodeptr   -> pointer to error code variable
    lookbehind     TRUE if this is a lookbehind assertion
+   reset_bracount TRUE to reset the count for each branch
    skipbytes      skip this many bytes at start (for brackets and OP_COND)
    firstbyteptr   place to put the first required character, or a negative number
    reqbyteptr     place to put the last required character, or a negative number
-Line 4587 
 Returns:         TRUE on success
+Line 4689 
 Returns:         TRUE on success
  static BOOL
  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,
-   int *errorcodeptr, BOOL lookbehind, int skipbytes, int *firstbyteptr,
+   int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
-   int *reqbyteptr, branch_chain *bcptr, compile_data *cd, int *lengthptr)
+   int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,
+   int *lengthptr)
  {
  const uschar *ptr = *ptrptr;
  uschar *code = *codeptr;
-Line 4598 
 uschar *reverse_count = NULL;
+Line 4701 
 uschar *reverse_count = NULL;
  int firstbyte, reqbyte;
  int branchfirstbyte, branchreqbyte;
  int length;
+ int orig_bracount;
+ int max_bracount;
  branch_chain bc;
  bc.outer = bcptr;
-Line 4626 
 code += 1 + LINK_SIZE + skipbytes;
+Line 4731 
 code += 1 + LINK_SIZE + skipbytes;
  /* Loop for each alternative branch */
+ orig_bracount = max_bracount = cd->bracount;
  for (;;)
    {
+   /* For a (?| group, reset the capturing bracket count so that each branch
+   uses the same numbers. */
+   if (reset_bracount) cd->bracount = orig_bracount;
    /* Handle a change of ims options at the start of the branch */
    if ((options & PCRE_IMS) != oldims)
-Line 4656 
 for (;;)
+Line 4767 
 for (;;)
      *ptrptr = ptr;
      return FALSE;
      }
+   /* Keep the highest bracket count in case (?| was used and some branch
+   has fewer than the rest. */
+   if (cd->bracount > max_bracount) max_bracount = cd->bracount;
    /* In the real compile phase, there is some post-processing to be done. */
-Line 4720 
 for (;;)
+Line 4836 
 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 4755 
 for (;;)
+Line 4874 
 for (;;)
        *code++ = oldims;
        length += 2;
        }
+     /* Retain the highest bracket number, in case resetting was used. */
+     cd->bracount = max_bracount;
      /* Set values to pass back */
-Line 4766 
 for (;;)
+Line 4889 
 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 5138 
 cd->cbits = tables + cbits_offset;
+Line 5273 
 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" and
  "anycrlf". */
  switch (options & (PCRE_NEWLINE_CRLF | PCRE_NEWLINE_ANY))
-Line 5149 
 switch (options & (PCRE_NEWLINE_CRLF | P
+Line 5284 
 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;
    }
-Line 5218 
 outside can help speed up starting point
+Line 5353 
 outside can help speed up starting point
  code = cworkspace;
  *code = OP_BRA;
  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,
-   &code, &ptr, &errorcode, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, &length);
+   &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd,
+   &length);
  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
-Line 5286 
 ptr = (const uschar *)pattern;
+Line 5422 
 ptr = (const uschar *)pattern;
  code = (uschar *)codestart;
  *code = OP_BRA;
  (void)compile_regex(re->options, re->options & PCRE_IMS, &code, &ptr,
-   &errorcode, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, NULL);
+   &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, NULL);
  re->top_bracket = cd->bracount;
  re->top_backref = cd->top_backref;

 Legend:



Removed from v.149
 


changed lines


 
Added in v.176
 Legend:



Removed from v.149
 


changed lines


 
Added in v.176
-Removed from v.149
+Added in v.176

webmaster@exim.org	ViewVC Help
Powered by ViewVC 1.1.12