From: Anatol Belski <ab@php.net>
Date: Tue, 30 Dec 2014 15:50:22 +0000 (+0100)
Subject: Fixed bug #66679 Alignment Bug in PCRE 8.34 upstream
X-Git-Tag: php-5.6.5RC1~26^2
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=d5123415f680b9c70ec44309f0bce3f0a512ee52;p=php

Fixed bug #66679 Alignment Bug in PCRE 8.34 upstream
---

diff --git a/NEWS b/NEWS
index d0d3c0f8b3..9b2fd10b8f 100644
--- a/NEWS
+++ b/NEWS
@@ -43,6 +43,10 @@ PHP                                                                        NEWS
   . Fixed bug #60509 (pcntl_signal doesn't decrease ref-count of old handler
     when setting SIG_DFL). (Julien)
 
+- PCRE:
+  . Fixed bug #66679 (Alignment Bug in PCRE 8.34 upstream).
+    (Rainer Jung, Anatol Belski)
+
 - PDO_mysql
   . Fixed bug #68424 (Add new PDO mysql connection attr to control multi
     statements option). (peter dot wolanin at acquia dot com)
diff --git a/ext/pcre/pcrelib/pcre_compile.c b/ext/pcre/pcrelib/pcre_compile.c
index 853fb24793..1075aaec86 100644
--- a/ext/pcre/pcrelib/pcre_compile.c
+++ b/ext/pcre/pcrelib/pcre_compile.c
@@ -3071,7 +3071,7 @@ const pcre_uint32 *ochr_ptr;
 const pcre_uint32 *list_ptr;
 const pcre_uchar *next_code;
 const pcre_uint8 *class_bitset;
-const pcre_uint32 *set1, *set2, *set_end;
+const pcre_uint8 *set1, *set2, *set_end;
 pcre_uint32 chr;
 BOOL accepted, invert_bits;
 
@@ -3202,12 +3202,12 @@ for(;;)
     if (base_list[0] == OP_CLASS)
 #endif
       {
-      set1 = (pcre_uint32 *)(base_end - base_list[2]);
+      set1 = (pcre_uint8 *)(base_end - base_list[2]);
       list_ptr = list;
       }
     else
       {
-      set1 = (pcre_uint32 *)(code - list[2]);
+      set1 = (pcre_uint8 *)(code - list[2]);
       list_ptr = base_list;
       }
 
@@ -3216,41 +3216,38 @@ for(;;)
       {
       case OP_CLASS:
       case OP_NCLASS:
-      set2 = (pcre_uint32 *)
+      set2 = (pcre_uint8 *)
         ((list_ptr == list ? code : base_end) - list_ptr[2]);
       break;
 
-      /* OP_XCLASS cannot be supported here, because its bitset
-      is not necessarily complete. E.g: [a-\0x{200}] is stored
-      as a character range, and the appropriate bits are not set. */
-
       case OP_NOT_DIGIT:
-        invert_bits = TRUE;
-        /* Fall through */
+      invert_bits = TRUE;
+      /* Fall through */
       case OP_DIGIT:
-        set2 = (pcre_uint32 *)(cd->cbits + cbit_digit);
-        break;
+      set2 = (pcre_uint8 *)(cd->cbits + cbit_digit);
+      break;
 
       case OP_NOT_WHITESPACE:
-        invert_bits = TRUE;
-        /* Fall through */
+      invert_bits = TRUE;
+      /* Fall through */
       case OP_WHITESPACE:
-        set2 = (pcre_uint32 *)(cd->cbits + cbit_space);
-        break;
+      set2 = (pcre_uint8 *)(cd->cbits + cbit_space);
+      break;
 
       case OP_NOT_WORDCHAR:
-        invert_bits = TRUE;
-        /* Fall through */
+      invert_bits = TRUE;
+      /* Fall through */
       case OP_WORDCHAR:
-        set2 = (pcre_uint32 *)(cd->cbits + cbit_word);
-        break;
+      set2 = (pcre_uint8 *)(cd->cbits + cbit_word);
+      break;
 
       default:
       return FALSE;
       }
 
-    /* Compare 4 bytes to improve speed. */
-    set_end = set1 + (32 / 4);
+    /* Because the sets are unaligned, we need
+    to perform byte comparison here. */
+    set_end = set1 + 32;
     if (invert_bits)
       {
       do