Fix two obscure, but nasty bugs concerned with caseless matching with Unicode

property support.


git-svn-id: svn://vcs.exim.org/pcre/code/trunk@115 2f5784b3-3f2a-0410-8824-cb99058d5e15

6 files changed, 79 insertions, 3 deletions

diff --git a/ChangeLog b/ChangeLog
index cb2632e..b1adf50 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -70,6 +70,23 @@ Version 7.1 05-Mar-07
     pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter 
     case, it wraps all the #defines with #ifndefs. This script should be run 
     before "make dist".
+    
+12. Fixed two fairly obscure bugs concerned with quantified caseless matching
+    with Unicode property support.
+    
+    (a) For a maximizing quantifier, if the two different cases of the 
+        character were of different lengths in their UTF-8 codings (there are 
+        some cases like this - I found 11), and the matching function had to 
+        back up over a mixture of the two cases, it incorrectly assumed they
+        were both the same length.
+        
+    (b) When PCRE was configured to use the heap rather than the stack for 
+        recursion during matching, it was not correctly preserving the data for 
+        the other case of a UTF-8 character when checking ahead for a match 
+        while processing a minimizing repeat. If the check also involved 
+        matching a wide character, but failed, corruption could cause an
+        erroneous result when trying to check for a repeat of the original
+        character.
 
 
 Version 7.0 19-Dec-06
diff --git a/pcre_exec.c b/pcre_exec.c
index 8156fd2..0d80f2e 100644
--- a/pcre_exec.c
+++ b/pcre_exec.c
@@ -299,6 +299,8 @@ typedef struct heapframe {
   int Xprop_category;
   int Xprop_chartype;
   int Xprop_script;
+  int Xoclength; 
+  uschar Xocchars[8]; 
 #endif
 
   int Xctype;
@@ -441,6 +443,8 @@ HEAP_RECURSE:
 #define prop_category      frame->Xprop_category
 #define prop_chartype      frame->Xprop_chartype
 #define prop_script        frame->Xprop_script
+#define oclength           frame->Xoclength
+#define occhars            frame->Xocchars
 #endif
 
 #define ctype              frame->Xctype
@@ -494,6 +498,8 @@ int prop_fail_result;
 int prop_category;
 int prop_chartype;
 int prop_script;
+int oclength;
+uschar occhars[8];
 #endif
 
 int ctype;
@@ -2045,19 +2051,18 @@ for (;;)
 
       if (length > 1)
         {
-        int oclength = 0;
-        uschar occhars[8];
-
 #ifdef SUPPORT_UCP
         unsigned int othercase;
         if ((ims & PCRE_CASELESS) != 0 &&
             (othercase = _pcre_ucp_othercase(fc)) != NOTACHAR)
           oclength = _pcre_ord2utf8(othercase, occhars);
+        else oclength = 0;
 #endif  /* SUPPORT_UCP */
 
         for (i = 1; i <= min; i++)
           {
           if (memcmp(eptr, charptr, length) == 0) eptr += length;
+#ifdef SUPPORT_UCP          
           /* Need braces because of following else */
           else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
           else
@@ -2065,6 +2070,9 @@ for (;;)
             if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
             eptr += oclength;
             }
+#else   /* without SUPPORT_UCP */
+          else { RRETURN(MATCH_NOMATCH); }
+#endif  /* SUPPORT_UCP */                     
           }
 
         if (min == max) continue;
@@ -2077,6 +2085,7 @@ for (;;)
             if (rrc != MATCH_NOMATCH) RRETURN(rrc);
             if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
             if (memcmp(eptr, charptr, length) == 0) eptr += length;
+#ifdef SUPPORT_UCP            
             /* Need braces because of following else */
             else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
             else
@@ -2084,6 +2093,9 @@ for (;;)
               if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
               eptr += oclength;
               }
+#else   /* without SUPPORT_UCP */
+            else { RRETURN (MATCH_NOMATCH); }
+#endif  /* SUPPORT_UCP */
             }
           /* Control never gets here */
           }
@@ -2095,12 +2107,16 @@ for (;;)
             {
             if (eptr > md->end_subject - length) break;
             if (memcmp(eptr, charptr, length) == 0) eptr += length;
+#ifdef SUPPORT_UCP             
             else if (oclength == 0) break;
             else
               {
               if (memcmp(eptr, occhars, oclength) != 0) break;
               eptr += oclength;
               }
+#else   /* without SUPPORT_UCP */
+            else break;
+#endif  /* SUPPORT_UCP */                          
             }
 
           if (possessive) continue;
@@ -2108,7 +2124,12 @@ for (;;)
            {
            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
+#ifdef SUPPORT_UCP
+           eptr--;
+           BACKCHAR(eptr);
+#else   /* without SUPPORT_UCP */            
            eptr -= length;
+#endif  /* SUPPORT_UCP */            
            }
           RRETURN(MATCH_NOMATCH);
           }
diff --git a/testdata/testinput6 b/testdata/testinput6
index 5a541f3..a0de1f1 100644
--- a/testdata/testinput6
+++ b/testdata/testinput6
@@ -762,4 +762,14 @@ of case for anything other than the ASCII letters. /
 /^\p{Balinese}\p{Cuneiform}\p{Nko}\p{Phags_Pa}\p{Phoenician}/8
     \x{1b00}\x{12000}\x{7c0}\x{a840}\x{10900}
 
+/The next two are special cases where the lengths of the different cases of the 
+same character differ. The first went wrong with heap fram storage; the 2nd
+was broken in all cases./
+
+/^\x{023a}+?(\x{0130}+)/8i
+  \x{023a}\x{2c65}\x{0130}
+  
+/^\x{023a}+([^X])/8i
+  \x{023a}\x{2c65}X
+ 
 / End of testinput6 /
diff --git a/testdata/testinput9 b/testdata/testinput9
index 23d3d3e..e11dda1 100644
--- a/testdata/testinput9
+++ b/testdata/testinput9
@@ -813,4 +813,10 @@
      \x{1c5}XY
      AXY      
 
+/^\x{023a}+?(\x{0130}+)/8i
+  \x{023a}\x{2c65}\x{0130}
+  
+/^\x{023a}+([^X])/8i
+  \x{023a}\x{2c65}X
+ 
 / End / 
diff --git a/testdata/testoutput6 b/testdata/testoutput6
index a8bb0ad..3b907a4 100644
--- a/testdata/testoutput6
+++ b/testdata/testoutput6
@@ -1432,4 +1432,18 @@ of case for anything other than the ASCII letters. /
     \x{1b00}\x{12000}\x{7c0}\x{a840}\x{10900}
  0: \x{1b00}\x{12000}\x{7c0}\x{a840}\x{10900}
 
+/The next two are special cases where the lengths of the different cases of the 
+same character differ. The first went wrong with heap fram storage; the 2nd
+was broken in all cases./
+
+/^\x{023a}+?(\x{0130}+)/8i
+  \x{023a}\x{2c65}\x{0130}
+ 0: \x{23a}\x{2c65}\x{130}
+ 1: \x{130}
+  
+/^\x{023a}+([^X])/8i
+  \x{023a}\x{2c65}X
+ 0: \x{23a}\x{2c65}
+ 1: \x{2c65}
+ 
 / End of testinput6 /
diff --git a/testdata/testoutput9 b/testdata/testoutput9
index 000e2b9..2cc61ac 100644
--- a/testdata/testoutput9
+++ b/testdata/testoutput9
@@ -1624,4 +1624,12 @@ No match
      AXY      
 No match
 
+/^\x{023a}+?(\x{0130}+)/8i
+  \x{023a}\x{2c65}\x{0130}
+ 0: \x{23a}\x{2c65}\x{130}
+  
+/^\x{023a}+([^X])/8i
+  \x{023a}\x{2c65}X
+ 0: \x{23a}\x{2c65}
+ 
 / End /