[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: mutt and asian fonts

Hash: SHA1


    Did iconv of glibc integrated big5<->utf8<->gb2312 mapping of
libhz already ? That's very cool, indeed! My problem here is, map of
the libiconv from GNU is just a crap ;) The results is useless
converted big5/utf8/gb2312 characters. How can we tell FSF libiconv
people to use the correct one ?

    BTW, the patchset included some modification of mutt internal
pager. For example, how could you accept that a Chinese character is
wrapp'd by space instead of by margin ?

    ex: a very long line like 
中文中文中文中文中文中文中文中文中文中文 中文中文中文中文中文中文中文中文中文中文中文中文
    (suppose this comes from a random mail)

    mutt's default behavior will wrap this line like this:
    (This is manually wrapped. mutt's original pager will wrap it like this.)

    But, I think that's bad. It should be like this:
中文中文中文中文中文中文中文中文中文中文 中文中文中文中文中文中文中文中文中文中
    (This is manually wrapped, the correct one)

    And, about the line wrapping... I've seen people on this list said
that, mutt can not wrap Chinese correctly. It may cut a word into 2
pieces, and the result is decreasing joy of reading. It's interesting
that Linux should already have full featured wcrtomb(), mbrtowc(),
iswprint() and wcwidth(). Correct line wrapping should not be a

    uhm, the patchset lacks of documentation, I think I have to
explain more. BBS is used extensively around China, HongKong, and
Taiwan. Many of them can not do anything about QP or MIME encoded
mails. If I want to use mutt for every mail stuff, mutt must have the
ability to send mails as-is. But this violates RFC ;) Thus I have to
apply a local patch to provide this option, or mutt can not be used in
my daily life.

    Yet another tiny thing is, mutt developers refuses to decode
attachment filenames. They told us that this violates RFC
(again!). hm... well, that's very bad if I got a file named "財務報表.xls" or
"期中作業報告.doc". Without decoding the filename, the
attachment becomes crap. See this in parse.c of the patchset.

    Mutt has its own problem, and it definitely needs extra patch for
daily use ;) Any tiny glitch is ... err, no good ;)


On Sun, Sep 15, 2002 at 04:00:27PM +0800, Anthony Fok wrote:
> Hello Clive,
> Just to let you know that it appears mutt on Linux does not need any extra
> patches.  It automatically takes advantage of glibc's iconv functions and
> does the necessary conversions when required.  It is not perfect yet, but it
> already handles big5<->gb2312 and utf-8<->big5 (in zh_TW.Big5 locale) and
> utf-8<->gb2312(in zh_CN.GB2312) locale.  When mutt is running under the
> zh_TW.Big5 locale, it may not properly display charset=gb18030 mail, but
> that problem should be fixed by adding a new gconv module in glibc instead.
Version: GnuPG v1.0.7 (FreeBSD)


Reply to: