Re: UTF-8 reading broken?

Re: UTF-8 reading broken?

[Jakub Martyński]

The following message is a courtesy copy of an article
that has been posted to gnu.emacs.gnus as well.

* Person Sergei Pokrovsky <pok@nbsp.nsk.su> wrote:

`lo.

I have another question: i want make "iso-8859-2" default coding for
my posts and mails, but when i reply to "utf-8" i got charset "utf-8"
in my headers (like in that post). How set recoding all "utf-8" posts
to iso-8859-2 in Gnus?

Gnus from cvs, Mule-UCS, Emacs 20.7.

-- 
Jakub Martyński
jakub@martynski.pl

Re: UTF-8 reading broken?

[Jakub Martyński]

The following message is a courtesy copy of an article
that has been posted to gnu.emacs.gnus as well.

* Person jakub@martynski.pl (Jakub Martyński) wrote:

> `lo.  I have another question: i want make "iso-8859-2" default
> coding for my posts and mails, but when i reply to "utf-8" i got
> charset "utf-8" in my headers (like in that post). How set recoding
> all "utf-8" posts to iso-8859-2 in Gnus?  Gnus from cvs, Mule-UCS,
> Emacs 20.7.

Oh s...t, now i got good header... Sorry for disturbing...

-- 
Jakub Martyński
jakub@martynski.pl

Re: UTF-8 reading broken?

[Toby Speight]

[-- Attachment #1: Type: text/plain, Size: 1239 bytes --]

0> In <URL:news:m3d7b47917.fsf@wazabi.tpi.pl>,
0> Jakub Marty <URL:mailto:jakub@martynski.pl> ("Jakub") wrote:

Jakub> I have another question: i want make "iso-8859-2" default
Jakub> coding for my posts and mails, but when i reply to "utf-8" i
Jakub> got charset "utf-8" in my headers (like in that post). How set
Jakub> recoding all "utf-8" posts to iso-8859-2 in Gnus?

I think this is a side-effect of how MULE works - it doesn't really
understand that a "á" character (for example) is in Latin-2 as well as
Latin-1, so when it reads "á", it puts a Latin-1 "á" in the message,
rather than a Latin-2 "á"[*].  If you quote this, and add some text of
your own with a Latin-2 "á" (or if the quote contains a character like
"ń" which isn't in Latin-1), then Mule thinks you've mixed two different
encodings.  That's why it uses UTF-8.

The fix isn't easy - though it may be possible to patch mule-ucs to
prefer Latin-2 encoding when reading.  I wouldn't know how to do this,
unfortunately.

[*] I've distinguished the two characters during composition, but this
    will be sent as UTF-8, so you won't be able tell them apart.
    Perhaps if I divide the message by hand, they'll get coded
    seperately:



[-- Attachment #2: Type: text/plain, Size: 217 bytes --]

> I think this is a side-effect of how MULE works - it doesn't really
> understand that a "á" character (for example) is in Latin-2 as well as
> Latin-1, so when it reads "á", it puts a Latin-1 "á" in the message,

[-- Attachment #3: Type: text/plain, Size: 265 bytes --]

> rather than a Latin-2 "á"[*].  If you quote this, and add some text of
> your own with a Latin-2 "á" (or if the quote contains a character like
> "ń" which isn't in Latin-1), then Mule thinks you've mixed two different
> encodings.  That's why it uses UTF-8.