Version: (using Devel) Installed from: Compiled sources Compiler: gcc (GCC) 4.1.3 20070929 (prerelease) (Ubuntu 4.1.2-16ubuntu2) OS: Linux The To, From, Cc and Bcc fields are encoded incorrectly, if contatin non-Latin chars. Only the display name must be encoded, while the mail address itself should left as-is. For example, my work email address, with my name in Russian, gets encoded as =?utf-8?B?0KHQtdGA0LPQtdC5INCR0LXQudC70LjQvSA8c2JlaWxpbkBsYWItbS5ydT4=?= while must be (domain name hidden to avoid spam) "=?utf-8?b?0KHQtdGA0LPQtdC5INCR0LXQudC70LjQvQ==?= <sbeilin@MYWORK.ru>" This error leads to the following message (kmmessage.cpp:2251) kmail(4815) KMMessage::splitAddrField: Error in address splitting: parseAddressList returned false! and messages stay in Outbox. Reverting kmmsgbase.h and kmmsgbase.cpp to revision r788000 fixes the problem. It seems that the following commit r788330 | tmcguire | 2008-03-21 15:36:46 +0400 (Fri, 21 Mar 2008) | 2 lines Use decode/encodeRFC2047String from KMime instead of rolling our own. is the source.
Created attachment 24114 [details] a patch approach The following code seems to be working. Hope it helps.
SVN commit 792011 by tmcguire: Revert parts of: >SVN commit 788330 by tmcguire: >Use decode/encodeRFC2047String from KMime instead of rolling our own. The reason is that our encodeRFC2047String does not encode quoted words (e.g. the address part of a mailbox), while the encodeRFC2047String from KMime does, and I couldn't find a method in KMime which does not encode quoted words. BUG: 159995 M +92 -5 kmmsgbase.cpp M +3 -0 kmmsgbase.h WebSVN link: http://websvn.kde.org/?view=rev&revision=792011
> a patch approach > The following code seems to be working. Hope it helps. Thanks for the bug report. I didn't use your patch, because it was hardcoded to use utf-8. Instead I just reverted my earlier commit which broke things.
Hi Thomas! > Thanks for the bug report. I didn't use your patch, because it was hardcoded to use utf-8. Instead I just reverted my earlier commit which broke things. Oops! My fault, a copy-paste from a wrong window ;) You can change "utf-8" constant to variable 'cset', it works for me. -- Best regards, Sergei A. Beilin Jabber ID: SAABeilin@jabber.ru