|Dependency Graph||View Issue Relation Graph Vertical|
View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0011562||phpList 3 application||Internationalization (l18n)||public||21-09-07 05:01||17-05-11 15:00|
|Target Version||Fixed in Version||2.11.6|
|Summary||0011562: Random Character Encoding Bug in SHIFT-JIS Japanese emails body & Subject|
I have been using PHPLIST now for 3 years, from version 2.8 something, to send mail magazines in Japanese.
Japanese email clients have a poor support of UTF-8, sadly, so at the setup/test stage, I had to abandon this first obvious solution in favor of the SHIFT-JIS encoding.
This works well, except for a very problematic bug: some characters (only some, on an apparently random fashion) in the email body and subject are badly converted at some point when saving or sending the email. Instead of the real character,
This forces me to triple check all emails before I can actually send them, and to use alternative expressions when a given character bugs.
I never took the time to make tests to see if the same characters always bugged, on much it depended on context, etc... (I WANT TO USE THE SOFT, NOT DEBUG IT MYSELF !!!).
If someone could dig into this problem and solve it, it would be sooooo cool.
Thanks in advance to all,
PS: if you need something from my install, tell me which files and database tables in details and I'll provide them.
|Additional Information||Examples from my last (yesterday) mailing:|
アレクサンドル・ソクーロフ becomes アレクサンドル・ャNーロフ
( ソク => ャN )
来日を予定しています becomes 来日を嵐閧ｵています
( 予定 => 嵐閧ｵ )
|Tags||No tags attached.|
||One possible explanation for this apparent random behavior might be this: Some SHIFT-JIS and BIG5 encoded characters contain backslashes which may get stripped when saving the message, thus breaking the character's encoding. See http://mantis.phplist.com/view.php?id=15551#c51125|