Version: 1.5.4 (using KDE KDE 3.1.4) Installed from: Gentoo Packages Compiler: GNU C Compiler 3.2.3 OS: Linux When I receive an ISO-8859-15 encoded message and my UNIX locale is set to use the Unicode encoding, KMail displays the body of the message (in the message pane) using ISO-8859-15, but it uses Unicode to display the subject (in the messages list pane and the message pane), without first converting the subject string from ISO-8859-15 to Unicode, resulting in garbled characters. Interestingly, if I choose to show all headers in the message pane, KMail seems to use the same encoding for the headers and the body, so the subject line appears correctly. I would expect KMail to assume the subject line uses the same encoding as the body, and to convert between this encoding and Unicode before displaying the line. Here is an example of what happens : MandrakeSoft r?ompens?pour le Club et la 9.1 (subject line with garbled accented characters) MandrakeSoft r
are you using IMAP? Can you attach an example mail?
Another question: what do you mean by "Unicode locale"? A locale like fr_CA.UTF-8?
Created attachment 2733 [details] Example ISO encoded message whose subject line contains no ASCII chars I use a POP3 account. My POSIX locale is set to fr_CA.UTF-8, as is my XFree86 locale. An example mail is attached. As you can appreciate, all non-ascii characters get garbled since they are ISO encoded but interpreted as if they were Unicode encoded.
Typo in comment #3 : ... contains non-ASCII chars
The e-mail is invalid. Please have the source fix their mailer. KMail is not at fault here. Your Subject line comes in this header in your e-mail: Subject: Mandrake Newsletter Flash: MandrakeSoft r
The reason for using ::fromLocal8Bit() is that it works for most Asian people. A config option to choose the fallback charset for headers as well as bodies would be nice, but it's a wish, not a bug.
An option to configure a violation of the standards? The subject is supposed to be US-ASCII and KMail already is graceful enough to tolerate those violations.
well, it's already used for the body, why exclude the headers from it? People won't stop creating broken mails too soon ;( [bugs.kde.org is just another bad example, btw]
Subject: Re: Incorrect encoding is assumed when displaying messages subjects The body is allowed to contain non-US-ASCII characters (at least if cte is 8bit). The headers OTOH must never contain anything else than US-ASCII characters (0x01 - 0x7F).
I know the mail violates the RFC. What I pointed out is that since KMail detects (or falls back) to a correct encoding, it might as well do so for the entire mail rather than only for the body. Even if the subject line is not supposed to include non-ASCII characters, many clients do make this mistake. I understand it's not the developers job to find workarounds for mistakes in other people's applications, but since the feature is already there, it would be nice and perhaps not too much work to extend it a little. In any case, it would make KMail behave the way the end-user expects it. An average user is not going to suppose the message is incorrectly formed, he is going to assume KMail is buggy, and that would be sad, since it's not the case. The best would certainly be to have a default configurable fallback encoding that would apply both to the Subject and From fields and to the body.
Unless it is made configurable, you're asking us to replace a feature that already violates the standards with another feature that still violates them and would break support for (also broken) asian-language e-mails. In any event, be sure to warn the source about their problem. A less forgiving MTA or MDA could simply decide to drop those characters that were not well protected, as was my case with my ISP in France. So, it's up to the KMail developers to decide whether they will implement this feature, but note that it should not come before KDE 3.2 (we're in feature-list freeze already).
I have the same problem, too. I live in Austria and many clients seem to transmit "umlaute" as 8-bit-characters in the headers, especially auto-generated mails from web services (like EBay and my Chello newsletter etc.). Strange thing is, before I updated to SuSE 9.1 from 8.0, those characters were shown correctly. But maybe this is due to the change of the default charset for SuSE 9.1 to UTF-8? Is there something I can do about it?
Since KMail 1.8 it's possible to define a fallback character encoding that's used for messages which are lacking an explicitly mentioned character encoding. This fallback should also be used for the subject.
The fallback encoding (or manually set encoding, for that matter) does not seem to be applied to the subject of a message. Was this ever implemented?
Thank you for your feature request. Kmail1 is currently unmaintained so we are closing all wishes. Please feel free to reopen a feature request for Kmail2 if it has not already been implemented. Thank you for your understanding.
Instead of creating a new feature request, please confirm here if the wishlist is still valid for kmail2.