From: Tomas K. <to...@us...> - 2006-01-12 17:51:35
|
> Squirrelmail 1.4.5, Courier on Fedora Core 3, in case it matters. > My default charset is set to utf-8 > > > I have some users who are Japanese who have noticed that since compose > switches the charset (to iso-2022-jp), unless the recipient (also on > squirrelmail), is displaying in Japanese, the characters are garbled. > > If I send japanese from my English (custom charset), everyone can see the > characters fine, regardless of their language setting. > > What I need is some way to make it so that Japanese composes in UTF-8 so > that recipients of any charset can display the messages properly. > > Is it as simple as editing the i18n.php and setting the charset for > Japanese to utf-8. I did that on a test box and it seems to resolve the > issue, but I am wondering what ramifications this will have on other > aspects of squirrelmail. > > Any ideas? Japanese translation is very delicate thing. Translation is designed to use euc-jp internally and send emails in iso-2022-jp. I've warned Fedora devels that they must be careful when switching Japanese to utf-8. First thing that you must do, is to disable XTRA_CODE key in Japanese language description. It removes main Japanese specific functions. Second thing - grep through SquirrelMail source for 'ja_JP' and disable character set specific Japanese workarounds. I think functions/i18n.php, functions/page_header.php, functions/strings.php and some functions/decode/ files need fixes. You need 1.4.6-rc1 and php with mbstring support in order to create correct utf-8 email headers. And you will need modified extra decoding library in order to read other Japanese emails. Which Japanese email client can't read utf-8? I know only that some unnamed email software works only with iso-2022-jp. -- Tomas |