[Python-checkins] r83395 - python/branches/py3k/Objects/unicodeobject.c

[georg.brandl]

Author: georg.brandl
Date: Sun Aug  1 10:49:18 2010
New Revision: 83395

Log:
#8821: do not rely on Unicode strings being terminated with a \u0000, rather explicitly check range before looking for a second surrogate character.

Modified:
   python/branches/py3k/Objects/unicodeobject.c

Modified: python/branches/py3k/Objects/unicodeobject.c
==============================================================================
--- python/branches/py3k/Objects/unicodeobject.c	(original)
+++ python/branches/py3k/Objects/unicodeobject.c	Sun Aug  1 10:49:18 2010
@@ -3734,7 +3734,7 @@
 
             ch2 = *s++;
             size--;
-            if (ch2 >= 0xDC00 && ch2 <= 0xDFFF) {
+            if (ch2 >= 0xDC00 && ch2 <= 0xDFFF && size) {
                 ucs = (((ch & 0x03FF) << 10) | (ch2 & 0x03FF)) + 0x00010000;
                 *p++ = '\\';
                 *p++ = 'U';
@@ -3976,7 +3976,7 @@
 
                 ch2 = *s++;
                 size--;
-                if (ch2 >= 0xDC00 && ch2 <= 0xDFFF) {
+                if (ch2 >= 0xDC00 && ch2 <= 0xDFFF && size) {
                     ucs = (((ch & 0x03FF) << 10) | (ch2 & 0x03FF)) + 0x00010000;
                     *p++ = '\\';
                     *p++ = 'U';