--- code/trunk/pcre.c	2007/02/24 21:38:21	13
+++ code/trunk/pcre.c	2007/02/24 21:38:33	19
@@ -9,7 +9,7 @@
 
 Written by: Philip Hazel <ph10@cam.ac.uk>
 
-           Copyright (c) 1997 University of Cambridge
+           Copyright (c) 1998 University of Cambridge
 
 -----------------------------------------------------------------------------
 Permission is granted to anyone to use this software for any purpose on any
@@ -49,10 +49,17 @@
 #include "internal.h"
 
 
+/* Allow compilation as C++ source code, should anybody want to do that. */
+
+#ifdef __cplusplus
+#define class pcre_class
+#endif
+
+
 /* Min and max values for the common repeats; for the maxima, 0 => infinity */
 
-static char rep_min[] = { 0, 0, 1, 1, 0, 0 };
-static char rep_max[] = { 0, 0, 0, 0, 1, 1 };
+static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };
+static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };
 
 /* Text forms of OP_ values and things, for debugging (not all used) */
 
@@ -76,7 +83,7 @@
 on. Zero means further processing is needed (for things like \x), or the escape
 is invalid. */
 
-static short int escapes[] = {
+static const short int escapes[] = {
     0,      0,      0,      0,      0,      0,      0,      0,   /* 0 - 7 */
     0,      0,    ':',    ';',    '<',    '=',    '>',    '?',   /* 8 - ? */
   '@', -ESC_A, -ESC_B,      0, -ESC_D,      0,      0,      0,   /* @ - G */
@@ -257,6 +264,13 @@
       case OP_KETRMIN:
       return TRUE;
 
+      /* Skip over entire bracket groups with zero lower bound */
+
+      case OP_BRAZERO:
+      case OP_BRAMINZERO:
+      cc++;
+      /* Fall through */
+
       /* Skip over assertive subpatterns */
 
       case OP_ASSERT:
@@ -271,8 +285,6 @@
       case OP_EOD:
       case OP_CIRC:
       case OP_DOLL:
-      case OP_BRAZERO:
-      case OP_BRAMINZERO:
       case OP_NOT_WORD_BOUNDARY:
       case OP_WORD_BOUNDARY:
       cc++;
@@ -611,6 +623,7 @@
 int repeat_type, op_type;
 int repeat_min, repeat_max;
 int bravalue, length;
+int greedy_default, greedy_non_default;
 register int c;
 register uschar *code = *codeptr;
 const uschar *ptr = *ptrptr;
@@ -618,6 +631,11 @@
 uschar *previous = NULL;
 uschar class[32];
 
+/* Set up the default and non-default settings for greediness */
+
+greedy_default = ((options & PCRE_UNGREEDY) != 0);
+greedy_non_default = greedy_default ^ 1;
+
 /* Switch on next character until the end of the branch */
 
 for (;; ptr++)
@@ -895,10 +913,13 @@
       goto FAILED;
       }
 
-    /* If the next character is '?' this is a minimizing repeat. Advance to the
+    /* If the next character is '?' this is a minimizing repeat, by default,
+    but if PCRE_UNGREEDY is set, it works the other way round. Advance to the
     next character. */
 
-    if (ptr[1] == '?') { repeat_type = 1; ptr++; } else repeat_type = 0;
+    if (ptr[1] == '?')
+      { repeat_type = greedy_non_default; ptr++; }
+    else repeat_type = greedy_default;
 
     /* If the maximum is zero then the minimum must also be zero; Perl allows
     this case, so we do too - by simply omitting the item altogether. */
@@ -1137,6 +1158,8 @@
         case 'm':
         case 's':
         case 'x':
+        case 'U':
+        case 'X':
         ptr++;
         while (*ptr != ')') ptr++;
         previous = NULL;
@@ -1296,7 +1319,7 @@
     the next state. */
 
     previous[1] = length;
-    ptr--;
+    if (length < 255) ptr--;
     break;
     }
   }                   /* end of big loop */
@@ -1740,7 +1763,7 @@
         ptr += 2;
         break;
         }
-      /* Else fall thourh */
+      /* Else fall through */
 
       /* Else loop setting valid options until ) is met. Anything else is an
       error. */
@@ -1770,6 +1793,16 @@
           length -= spaces;          /* Already counted spaces */
           continue;
           }
+        else if (c == 'X')
+          {
+          options |= PCRE_EXTRA;
+          continue;
+          }
+        else if (c == 'U')
+          {
+          options |= PCRE_UNGREEDY;
+          continue;
+          }
         else if (c == ')') break;
 
         *errorptr = ERR12;
@@ -1975,14 +2008,15 @@
 
 if (re->options != 0)
   {
-  printf("%s%s%s%s%s%s%s\n",
+  printf("%s%s%s%s%s%s%s%s\n",
     ((re->options & PCRE_ANCHORED) != 0)? "anchored " : "",
     ((re->options & PCRE_CASELESS) != 0)? "caseless " : "",
     ((re->options & PCRE_EXTENDED) != 0)? "extended " : "",
     ((re->options & PCRE_MULTILINE) != 0)? "multiline " : "",
     ((re->options & PCRE_DOTALL) != 0)? "dotall " : "",
     ((re->options & PCRE_DOLLAR_ENDONLY) != 0)? "endonly " : "",
-    ((re->options & PCRE_EXTRA) != 0)? "extra " : "");
+    ((re->options & PCRE_EXTRA) != 0)? "extra " : "",
+    ((re->options & PCRE_UNGREEDY) != 0)? "ungreedy " : "");
   }
 
 if ((re->options & PCRE_FIRSTSET) != 0)
@@ -3450,7 +3484,7 @@
 if (re->top_backref > 0 && re->top_backref >= ocount/2)
   {
   ocount = re->top_backref * 2 + 2;
-  match_block.offset_vector = (pcre_malloc)(ocount * sizeof(int));
+  match_block.offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));
   if (match_block.offset_vector == NULL) return PCRE_ERROR_NOMEMORY;
   using_temporary_offsets = TRUE;
   DPRINTF(("Got memory to hold back references\n"));