Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:34933
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain zend.com designates 212.25.124.162 as permitted sender)
Message-ID: <479A60BA.7030905@zend.com>
Date: Fri, 25 Jan 2008 14:20:42 -0800
Organization: Zend Technologies
User-Agent: Thunderbird 2.0.0.9 (Windows/20071031)
MIME-Version: 1.0
To: 'PHP Internals' <internals@lists.php.net>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: json_encode() bug
From: stas@zend.com (Stanislav Malyshev)

Hi!

Right now, if json_encode sees wrong UTF-8 data, it just cuts the string 
in the middle, no error returned, no message produced. Example:

var_dump(json_encode("ab\xE0"));
var_dump(json_encode("ab\xE0\""));

Both strings get cut at "ab". I think it's not a good idea to just 
silently cut the data. In fact, I think it is a bug caused by this code 
in ext/json/utf8_to_utf16.c:
         if (c < 0) {
             return UTF8_END ? the_index : UTF8_ERROR;
         }
which inherited this bug from code published on json.org. It should be:
         if (c < 0) {
             return (c == UTF8_END) ? the_index : UTF8_ERROR;
         }
Now this is an easy fix but would lead to bad strings silently converted 
to empty strings. The question is - should we have an error there? If 
so, which one - E_WARNING, E_NOTICE? I'm for E_WARNING.
Also filed as bug #43941.
Any comments?
-- 
Stanislav Malyshev, Zend Software Architect
stas@zend.com   http://www.zend.com/
(408)253-8829   MSN: stas@zend.com