On 6/17/06, Zvezdan Petkovic <[EMAIL PROTECTED]> wrote:
OpenBSD works fine with UTF-8 despite not having support for locales. As already described in my previous message, Vim, XTerm, and KMail work just fine with these messages. It's a mutt port issue.
Vim has its own functions to display multibyte characters. So vim works fine with UTF-8. But, if every program had such functions, things would be very complicated. Instead, mutt simply relies on ncurses.
> So, we have two ways to use UTF-8 on mutt: > (1) Use mutt's wc-funcs. i.e. build it with --without-wc-funcs. As shown in my examples this solves one big part of the problem. The message can be read correctly most of the time, and the subject in UTF-8 is accepted correctly. Can we get this into port?
I am not the maintainer, but I love to have this option in port.
However, this still leaves an issue of a strange behaviour with mixed Latin (ASCII, probably) and non-Latin (non-ASCII) characters. Also, the input of subject line in UTF-8 used to be visible as you type it, as opposed to the current port where it looks mangled until Enter is pressed. This _worked_ in mutt before. I don't see why it doesn't any more. Perhaps something else should be declared --without-...
I guess your two problems are in ncurses. Mutt just uses ncurses' addstr or other functions. But ncurses 5.2 doesn't support UTF-8. IIRC, only libncursesw of ncurses 5.3 or higher supports it.
Notice that the version of mutt did not change (except security patches) between the time when it worked and now. Thus, it's all in configuration of mutt. I'll continue to experiment with configuration
The trigger is that OpenBSD implemented wc-funcs which doesn't support UTF-8. So I'm afraid that you can do nothing on mutt other than the --without-wc-funcs option. -- tamo
