Re: cvs: php-src(PHP_5_3) /ext/pdo pdo_stmt.c
| From: | Nuno Lopes | Date: | Thu, 11 Dec 2008 16:50:16 +0000 |
| Subject: | Re: cvs: php-src(PHP_5_3) /ext/pdo pdo_stmt.c | ||
| References: | 1 2 3 | Groups: | php.cvs |
| Request: | Send a blank email to php-cvs+get-54880@lists.php.net to get a copy of this message | ||
Weird.. Isn't that a bug in php_stream_printf() then? I would say it shouldn't read past the specified length.. otherwise the same bug may appear in other places.
What do you think?
Nuno
----- Original Message -----
The patch was already n 5.2, the issue is that the str (key) is not guaranteed to be NULL terminated (nor does it need to be), so when strlen() is attempted on top of it you could end up reading more data then necessary. On 11-Dec-08, at 10:43 AM, Nuno Lopes wrote:Ilia AlshanetskySorry for my ignorance, but isn't the new code exactly equivalent to the old one, albeit a bit slower? I can't really see how a strndup() can fix a corruption there.. If there's some problem, probably it's deeper than this.. NunoModified files: (Branch: PHP_5_3)/php-src/ext/pdo pdo_stmt.c Log: Fixed a possible corruption inside PDOStatement::debugDumpParams() http://cvs.php.net/viewvc.cgi/php-src/ext/pdo/pdo_stmt.c?r1=1.118.2.38.2.24.2.39&r2=1.118.2.38.2.24.2.40&diff_format=u Index: php-src/ext/pdo/pdo_stmt.c diff -u php-src/ext/pdo/pdo_stmt.c:1.118.2.38.2.24.2.39 php-src/ext/ pdo/pdo_stmt.c:1.118.2.38.2.24.2.40 @@ -2209,7 +2209,9 @@ if (res == HASH_KEY_IS_LONG) { php_stream_printf(out TSRMLS_CC, "Key: Position #%ld:\n", num); } else if (res == HASH_KEY_IS_STRING) { - php_stream_printf(out TSRMLS_CC, "Key: Name: [%d] %.*s\n", len, len, str); + char *s = estrndup(str, len); + php_stream_printf(out TSRMLS_CC, "Key: Name: [%d] %.*s\n", len, len, s); + efree(s); }