Version: 1.5.9 (using KDE 3.1.9) Compiler: gcc version 3.3.1 20030710 (prerelease) OS: Linux (i686) release 2.6.0-test1 I've received a multipart/mixed e-mail containing two text/plain sections, each encoded with a different charset. The reason behind that is that mailman adds a footer to messages passing through a mailing list, and when the encoding of that footer is different than the encoding of the message body, it splits the message into two MIME sections. As far as I can tell, that's acceptable behaviour. And when displaying the message, KMail displays all non-ASCII characters correctly. There's even no distinction between the two sections. However, when hitting Reply, the quoted text shows up with undecoded UTF-8. If I set the encoding of the message to UTF-8, the body is displayed fine, but the footer isn't; the quoted text in the reply is properly decoded. If I set the encoding to ISO-8859-1, the footer is displayed ok, but neither the body nor the quoted text in the reply are correct.
Created attachment 2153 [details] The e-mail message that shows problems Here's the e-mail message that is showing the behaviour described above. As far as I can tell, the message is properly encoded.
Created attachment 2603 [details] Snapshot showing no raw UTF-8 sequences in the text As you can see, it works for me. My guess is that you have set the reader to force iso-8859-* instead of auto-detecting. Else, try to upgrade to current HEAD. I know that at one point in the 3.2 dev cycle, this was broken - but it's months ago.
See comment and screenshot above.
Will upgrade and report.
Nope, sorry. Using HEAD dated of 10 minutes ago and it is still not working. I don't think you actually got the problem. I've got a message with two text/plain parts, one encoded in UTF-8 and the other encoded in ISO-8859-1. When reading the message, it's displayed fine. When replying, it is not. I'm attaching a screenshot demonstrating the issue.
Created attachment 2605 [details] Snapshot showing the composer window with undecoded chars Here it goes. Note that you see two windows in this snapshot: the background window is KMail with the open e-mail. You'll see from the MIME tree that it has two "plain text" sections. Source code reveals one is UTF-8 and the other is ISO-8859-1 (as the e-mail in attachment #2153 [details]). There are no raw UTF-8 sequences in there, as your screenshot shows as well. however, the foreground window is a KMail composer window. As you'll see, it's a reply of the e-mail in the background (I just hit R). The UTF-8 sequences are there. This doesn't happen when the whole e-mail is UTF-8, but happens only for those in which there are two text sections.
Subject: kdepim/kmail CVS commit by mutz: When replying, don't overwrite the per-bodypart codec unless an override-codec is set or no per-part codec was found. Fixes #62140 CCMAIL: 62140-done@bugs.kde.org M +2 -1 kmmessage.cpp 1.427 --- kdepim/kmail/kmmessage.cpp #1.426:1.427 @@ -794,4 +794,5 @@ QCString KMMessage::asQuotedString( cons } + if ( mOverrideCodec || !codec ) codec = this->codec();
Thanks, confirmed.