Very Computer

OK, I surrender and I need help. I can not configure mutt
to work with our czech locales on hp-ux 10.20.
Some facts:
- mutt 1.4i
- hp-ux 10.20
- unix locale cs_CZ.IBM-852
- ncurses (5.2 ?)
- compiler: gcc 2.95.2

./configure --prefix=/opt/mutt --with-curses=/opt/ncurses
   --enable-pop --enable-imap --enable-exact-address --enable-debug
make
make install

I do not see czech chars (in both subject and message body),
even after trying numerous env (LC_*, LANG, charset) settings.
Some national chars seem to be 8bit in some other coding, some
are rewritten as \octal.

What should I do to see czech chars in mutt ? What to set, unset,
check ? Does anybody use mutt on hp-ux 10.20 with other then
english character set ?

Thanks,
                Michal

--
--------------------------------------------------------------

--------------------------------------------------------------

We just had a thread here WRT Nordic 8-bit characters in the headers.
The answer is no, about that, except you could try parsing your mail as
it comes in to do any translations.

--

  or put "not-spam" or "/dev/rob0" in Subject header to reply

depends on if your xterm is able to display it.
I can even display chinese charactere but with another the same xterm
that support chinese charset.

--
William.

I use terminal emulation (text mode) under win32. Yes, the
emulator can display czech characters correctly.

When I send mail from myself (from mutt) to myself with czech characters
both in subject and body:
- subject is correctly encoded and displayed when read in mutt
   (displayed correctly also in mozilla in windows),
    but SOME characters even can not be typed
- body (written in external editor) is correctly encoded
    (displayed OK in mozilla in windows),
    but SOME characters are displayed as numbers (eg. \233)
    in mutt
- mail written in other czech coding (than IBM852, which we use on HP)
   are not displayed correctly. Mutt does decoding(s) from 7bit coding
   to 8bit BUT it seems it does not translate code pages.

It seems mutt CAN display (most of) czech characters and can do
decoding both in subject and body. BUT there is problem in character
table definition(s) and/or code tables translating.

I can believe that translating between code pages is essential
thing and I believe that mutt can do that.

Michal

--
--------------------------------------------------------------

--------------------------------------------------------------

Hello Michal,

Quote:> I can not configure mutt to work with our czech locales on hp-ux
> 10.20. [...] unix locale cs_CZ.IBM-852

    Which libiconv do you use? The very new libiconv 1.8 needs to be
configured with "--enable-extra-encodings" to support CP852. Without it,
Mutt will show you raw chars as they are, without conversion. BTW I
tried here, and the spelling "IBM-852" was unknown to my iconv. It knows
this charset as 852 and aliases:

$ iconv -l | grep 852
852 CP852 IBM852 CSPCP852

Quote:> Some national chars [...] are rewritten as \octal.

    Check your locale cs_CZ.IBM-852 really exist as declared, and works.
Perhaps using the little program checklocale.c found near end of page on
<URL:http://www.uni-ulm.de/~s_smasch/Locale/>. Especially the ??Setting
LC_CTYPE to its own value failed/succeeded.?? and ??Testing LC_CTYPE
with isprint():?? sections. In CP852, you should not have a single "#"
in the second half of the charset.

Quote:> What should I do to see czech chars in mutt ?

    Normal settings could be:

set charset="cp852"   # your terminal's charset, must be known by iconv
set locale="$LC_TIME" # or cs_CZ.IBM-852, or C, for localized or not date

    For $send_charset I don't know: what do you use as interchange
charset? Latin-2? Windows-1250? Something else?

Quote:> Does anybody use mutt on hp-ux 10.20

    Not me, but I hope something here will help you. A setup not too far
from yours works well here, but in CP850: Just to say it can be done.

Quote:> I use terminal emulation (text mode) under win32. Yes, the emulator
> can display czech characters correctly.

    Which emulator exactly? Could it be replaced by PuTTY 0.52?

Quote:> subject is correctly encoded [...] but SOME characters even can not be
> typed

    Replaced by dots or beeps? Typical of bad locales.

Quote:> SOME characters are displayed as numbers (eg. \233) in mutt

    Bad locale. \233 is not a printable char in us-ascii or any ISO
locales. It is printable in CP852 though, and Mutt will print it once
your locale will say it to libc's isprint() function.

Quote:> Mutt does decoding(s) from 7bit coding to 8bit BUT it seems it does
> not translate code pages.

    Typical of iconv problem. Either mail's charset or $charset in
muttrc unknown to iconv. Or also non MIME messages: AFAIK Mutt never
translates them.

    Well: I'm not sure, but I'd say you have 2 problems. :-(

Bye!    Alain.