#18874 [Fbk]: Certain Japanese characters break PHP's $_GET and $_POST

ID: 18874 Updated by: wez@php.net Reported By: erica@simpli.biz Status: Feedback Bug Type: Output Control Operating System: Windows XP PHP Version: 4.1.2 New Comment: Also note that (for IE at least), you can force a more sensible encoding to be used by adding an ACCEPTCHARSET attribute to your form: <form acceptcharset="utf-8" ...> This will cause the data to be utf-8 encoded. And you should definitely be running with the mbstring module if you are dealing with Japanese text. Previous Comments: ------------------------------------------------------------------------ [2002-08-13 02:40:44] fujimoto@php.net The spec of Shift_JIS (== one of the japanese encodings) causes this problem. So, probably snapshots will not solve this. As mentioned in http://www.php.net/manual/en/ref.mbstring.php, using Shift_JIS as the script encoding or internal encoding is NOT recommended. Anyway, currently we have two workarounds. 1. use EUC-JP as script encoding 2. use snapshots w/ following settings: - enable zend engine to parse shift jis ./configure --enable-zend-multibyte [and so on] - enable encoding translation add following lines to your php.ini [mbstring] mbstring.internal_encoding=euc-jp mbstring.output_encoding=shift_jis mbstring.script_encoding=shift_jis mbstring.encoding_translation=1 if you have questions about japanese encoding, please ask me directly (because here is not suitable place to do that). ------------------------------------------------------------------------ [2002-08-13 02:15:32] rasmus@php.net Actually, he probably meant the latest snapshot from snaps.php.net ------------------------------------------------------------------------ [2002-08-13 02:07:13] erica@simpli.biz Still a bug in a fresh compile of 4.2.2 (on a friend's system.) His phpinfo(): http://redline.worldforge.org/en/info.php The script: http://redline.worldforge.org/hottie/contact2.php Thanks. ------------------------------------------------------------------------ [2002-08-13 01:05:28] kalowsky@php.net Does this still happen with a newer version of PHP? ------------------------------------------------------------------------ [2002-08-13 00:44:01] erica@simpli.biz I've put an example script here: http://clients.simpli.biz/contact.php and the .phps is http://clients.simpli.biz/contact.phps Basically, there is a specific Japanese character that translates to a [ in ASCII. Since this character contains a [, PHP treats this as the beginning of an array, and since it sees no ending delimiter, it ignores that form field. Any form input with that character then becomes broken in PHP. Since it is a fairly common character, this has a large impact. Go to the script above and change your character encoding, and you'll see exactly what I mean. When you hit Submit, the script just prints out all submitted variables. Any form inputs with a name containing that character do not submit at all. Thanks, Erica (SlashChick on irc.openprojects.net #php) ------------------------------------------------------------------------ -- Edit this bug report at http://bugs.php.net/?id=18874&edit=1