Decoding ISO8859-15 -- offending headers

Decoding ISO8859-15 -- offending headers

[Jonas Steverud]

I posted earlier about decoding iso8859-15 news articles and here is
an article where åäö is replaced by ? when I view it but they looked
nice when I did C-u g. The headers look alright to me. Anyone else who
has trouble with this article?

Path: news.chalmers.se!newsfeed.sunet.se!news01.sunet.se!logbridge.uoregon.edu!news.tele.dk!small.news.tele.dk!news-stob.telia.net!telia.net!194.22.194.4.MISMATCH!masternews.telia.net.!newsc.telia.net.POSTED!not-for-mail
From: Marcus <pictorNOSPAM@mac.com>
Subject: .Mac
Newsgroups: se.dator.sys.mac
Lines: 6
Organization: SUBLIMITETSAKADEMIEN
User-Agent: KNode/0.6.1
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-15
Content-Transfer-Encoding: 8Bit
Message-ID: <3ulZ8.47852$n4.11770780@newsc.telia.net>
Date: Wed, 17 Jul 2002 21:39:11 GMT
NNTP-Posting-Host: 217.210.56.30
X-Complaints-To: abuse@telia.com
X-Trace: newsc.telia.net 1026941951 217.210.56.30 (Wed, 17 Jul 2002 23:39:11 CEST)
NNTP-Posting-Date: Wed, 17 Jul 2002 23:39:11 CEST
Xref: news.chalmers.se se.dator.sys.mac:25740


-- 
(          www.dtek.chalmers.se/~d4jonas/         !     Wei Wu Wei     )
(        Meaning of U2 Lyrics, Roleplaying        !  To Do Without Do  )

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Jonas Steverud <d4jonas@dtek.chalmers.se> writes:

> I posted earlier about decoding iso8859-15 news articles and here is
> an article where åäö is replaced by ? when I view it 

This looks like a question mark.  Is this how it looks for you?  Or
was it an empty box?

The messages looked fine here, but I think you must have the proper
ISO-8859-15 fonts to display it.

> but they looked nice when I did C-u g. 

That's weird.  ÅÄÖ have the same code point in 15 as in 1, so if
somehow your emacs defaults to 1 this might happen, but it doesn't for
me.  When I C-u g an article containing non-ASCII I always get octal
sequences for the characters, which is what you should expect.  If you
C-u g this message, do you see ÅÄÖ as ÅÄÖ or \345\344\346?

Re: Decoding ISO8859-15 -- offending headers

[Jonas Steverud]

Simon Josefsson <jas@extundo.com> writes:

> Jonas Steverud <d4jonas@dtek.chalmers.se> writes:
>
>> I posted earlier about decoding iso8859-15 news articles and here is
>> an article where åäö is replaced by ? when I view it 
>
> This looks like a question mark.  Is this how it looks for you?  Or
> was it an empty box?

It's a questionmark. Sorry for not pointing that out. Should've known.

> The messages looked fine here, but I think you must have the proper
> ISO-8859-15 fonts to display it.

It might be it. Dunno how to check though.

>> but they looked nice when I did C-u g. 
>
> That's weird.  ÅÄÖ have the same code point in 15 as in 1, so if
> somehow your emacs defaults to 1 this might happen, but it doesn't for
> me.  When I C-u g an article containing non-ASCII I always get octal
> sequences for the characters, which is what you should expect.  If you
> C-u g this message, do you see ÅÄÖ as ÅÄÖ or \345\344\346?

As ÅÄÖ.

-- 
(          www.dtek.chalmers.se/~d4jonas/         !     Wei Wu Wei     )
(        Meaning of U2 Lyrics, Roleplaying        !  To Do Without Do  )

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Jonas Steverud <d4jonas@dtek.chalmers.se> writes:

>> That's weird.  ÅÄÖ have the same code point in 15 as in 1, so if
>> somehow your emacs defaults to 1 this might happen, but it doesn't for
>> me.  When I C-u g an article containing non-ASCII I always get octal
>> sequences for the characters, which is what you should expect.  If you
>> C-u g this message, do you see ÅÄÖ as ÅÄÖ or \345\344\346?
>
> As ÅÄÖ.

I'm not sure that's correct.  What kind of MULE configuration do you
have in .emacs?

Re: Decoding ISO8859-15 -- offending headers

[Jonas Steverud]

Simon Josefsson <jas@extundo.com> writes:

> Jonas Steverud <d4jonas@dtek.chalmers.se> writes:
[...]
>> As ÅÄÖ.
>
> I'm not sure that's correct.  What kind of MULE configuration do you
> have in .emacs?

Huh?

Is it 
(set-language-environment "Latin-1")
you talk of?

But set-language-environment is maybe not needed anylonger? I have
never entirely understood how Emacs handles diffrent input methods and
how to make it understand me so set-language-environment might be outdated.

GNU Emacs 21.1.2 (sparc-sun-solaris2.8, X toolkit, Xaw3d scroll bars)
of 2001-10-21 on licia.dtek.chalmers.se

-- 
(          www.dtek.chalmers.se/~d4jonas/         !     Wei Wu Wei     )
(        Meaning of U2 Lyrics, Roleplaying        !  To Do Without Do  )

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Jonas Steverud <d4jonas@dtek.chalmers.se> writes:

> Simon Josefsson <jas@extundo.com> writes:
>
>> Jonas Steverud <d4jonas@dtek.chalmers.se> writes:
> [...]
>>> As ÅÄÖ.
>>
>> I'm not sure that's correct.  What kind of MULE configuration do you
>> have in .emacs?
>
> Huh?
>
> Is it 
> (set-language-environment "Latin-1")
> you talk of?

Yes, and other similar stuff.  Evaluating that expression doesn't
change the behaviour for me though, I still get octal sequences in C-u
g buffers.  Pinpointing the code that changes this for you could be a
lead in solving the original problem.  What do other people see?
Perhaps it is my configuration..

> But set-language-environment is maybe not needed anylonger? 

The C locale is used to guess the language environment, I think, so
usually it is not needed.

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Simon Josefsson <jas@extundo.com> writes:

>>>> As ÅÄÖ.
>>>
>>> I'm not sure that's correct.  What kind of MULE configuration do you
>>> have in .emacs?
>>
>> Huh?
>>
>> Is it 
>> (set-language-environment "Latin-1")
>> you talk of?

(standard-display-european 1) has the effect you mention.  If you have
that in .emacs, consider removing it.  I'll see if I can fix it so
that s-d-e doesn't break Gnus though.

Re: Decoding ISO8859-15 -- offending headers

[Jonas Steverud]

Simon Josefsson <jas@extundo.com> writes:

> Jonas Steverud <d4jonas@dtek.chalmers.se> writes:
[...]
>> Is it 
>> (set-language-environment "Latin-1")
>> you talk of?
>
> Yes, and other similar stuff.

As far as I know and can find, s-l-e is the only function I call
except for a (require 'latin-1).
  
> Perhaps it is my configuration..

My problem is that I have very many files loaded in my .emacs so it is
hard to debug but I don't think I have anything more then the above. My
locale is set to:

LANG=
LC_CTYPE=iso_8859_1
LC_NUMERIC="C"
LC_TIME="C"
LC_COLLATE="C"
LC_MONETARY="C"
LC_MESSAGES="C"
LC_ALL=

I ssh from Mac OS X running OroborOSX 0.8preview2 as window manager to
a Solaris system. But that's hardly the problem.

There are very few messages that looks like this so it is not a big
problem, but it would be nice to fix it anyway.

-- 
(          www.dtek.chalmers.se/~d4jonas/         !     Wei Wu Wei     )
(        Meaning of U2 Lyrics, Roleplaying        !  To Do Without Do  )

Re: Decoding ISO8859-15 -- offending headers

[Reiner Steib]

On Sat, Jul 27 2002, Simon Josefsson wrote:

> I still get octal sequences in C-u g buffers. Pinpointing the code
> that changes this for you could be a lead in solving the original
> problem. What do other people see?

I don't see any octal sequences. I see the normal Latin-1 characters
(`M-x describe-char-after RET' gives `eight-bit-graphic' in the raw
article, not `latin-iso8859-1').

Oort Gnus v0.07, Emacs/21.1, (set-language-environment "Latin-1")

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo--- PGP key available via WWW   http://rsteib.home.pages.de/

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Reiner Steib <4uce.02.r.steib@gmx.net> writes:

> On Sat, Jul 27 2002, Simon Josefsson wrote:
>
>> I still get octal sequences in C-u g buffers. Pinpointing the code
>> that changes this for you could be a lead in solving the original
>> problem. What do other people see?
>
> I don't see any octal sequences. I see the normal Latin-1 characters
> (`M-x describe-char-after RET' gives `eight-bit-graphic' in the raw
> article, not `latin-iso8859-1').
>
> Oort Gnus v0.07, Emacs/21.1, (set-language-environment "Latin-1")

Interesting, I get the below.  However, the Ä is a octal sequence in
the *Help* buffer, but a proper glyph when I cut'n'paste it into this
buffer.

  character: Ä (0304, 196, 0xc4)
    charset: eight-bit-graphic (8-bit graphic char (0xA0..0xFF))
 code point: 196
     syntax:   	which means: whitespace
   category:
buffer code: 0xC4
  file code: not encodable by coding system nil
       font: -Adobe-Courier-Medium-R-Normal--17-120-100-100-M-100-ISO8859-1

Re: Decoding ISO8859-15 -- offending headers

[Reiner Steib]

On Mon, Jul 29 2002, Simon Josefsson wrote:

> Reiner Steib <4uce.02.r.steib@gmx.net> writes:
[...]
>> I don't see any octal sequences. I see the normal Latin-1 characters
>> (`M-x describe-char-after RET' gives `eight-bit-graphic' in the raw
>> article, not `latin-iso8859-1').
>>
>> Oort Gnus v0.07, Emacs/21.1, (set-language-environment "Latin-1")

I found that my locale is responsible for this (namely
LC_CTYPE=en_US.ISO_8859-1, all other are C or POSIX). With LC_CTYPE=C,
I see \304 too.

Emacs 21.1 (--no-site-file, minimal ~/.{emacs,gnus}), Oort:
- LC_CTYPE=C                ==> displayed as \304
- LC_CTYPE=en_US.ISO_8859-1 ==> displayed as Ä
(eight-bit-graphic in both cases)

> Interesting, I get the below.  However, the Ä is a octal sequence in
> the *Help* buffer, but a proper glyph when I cut'n'paste it into this
> buffer.

Same here, and in the message buffer, I have »charset: latin-iso8859-1«.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo--- PGP key available via WWW   http://rsteib.home.pages.de/

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Reiner Steib <4uce.02.r.steib@gmx.net> writes:

> On Mon, Jul 29 2002, Simon Josefsson wrote:
>
>> Reiner Steib <4uce.02.r.steib@gmx.net> writes:
> [...]
>>> I don't see any octal sequences. I see the normal Latin-1 characters
>>> (`M-x describe-char-after RET' gives `eight-bit-graphic' in the raw
>>> article, not `latin-iso8859-1').
>>>
>>> Oort Gnus v0.07, Emacs/21.1, (set-language-environment "Latin-1")
>
> I found that my locale is responsible for this (namely
> LC_CTYPE=en_US.ISO_8859-1, all other are C or POSIX). With LC_CTYPE=C,
> I see \304 too.
>
> Emacs 21.1 (--no-site-file, minimal ~/.{emacs,gnus}), Oort:
> - LC_CTYPE=C                ==> displayed as \304
> - LC_CTYPE=en_US.ISO_8859-1 ==> displayed as Ä
> (eight-bit-graphic in both cases)

Ok.  Jonas, does using a non-ISO8859-1 locale modify the behaviour for
your original problem?  What locale settings do you use?

Re: Decoding ISO8859-15 -- offending headers

[Jonas Steverud]

Simon Josefsson <jas@extundo.com> writes:

[...]
> Ok.  Jonas, does using a non-ISO8859-1 locale modify the behaviour for
> your original problem?  What locale settings do you use?

LANG=
LC_CTYPE=iso_8859_1
LC_NUMERIC="C"
LC_TIME="C"
LC_COLLATE="C"
LC_MONETARY="C"
LC_MESSAGES="C"
LC_ALL=

I will nopt be able to access my email from saturday and about a week
forward and I do not have time to check before, I think. I'll put it
on my to-do-list.

-- 
(          www.dtek.chalmers.se/~d4jonas/         !     Wei Wu Wei     )
(        Meaning of U2 Lyrics, Roleplaying        !  To Do Without Do  )

Re: Decoding ISO8859-15 -- offending headers

[Jonas Steverud]

Simon Josefsson <jas@extundo.com> writes:

[...]
> Ok.  Jonas, does using a non-ISO8859-1 locale modify the behaviour for
> your original problem?  What locale settings do you use?

I don't remeber what I have answered and the mailing list archive's
Glimps index "was not found" when I searched att gnus.org so I can't
check which article (the message-id) is was that offended Gnus.

My locales are accordingly:
On the Solaris where Emacs lives:
> locale
LANG=
LC_CTYPE=iso_8859_1
LC_NUMERIC="C"
LC_TIME="C"
LC_COLLATE="C"
LC_MONETARY="C"
LC_MESSAGES="C"
LC_ALL=

I ssh from a Mac OS X running XFree86 with OrorborOSX 0.8b2 as WM and
there I have:
> env | egrep '^L[A,C]'
LANG=sv_SE
LC_CTYPE=iso_8859_1

Since I don't know the offending article I cannot (yet) check how the
locale affects Gnus. I do plan to drop the issue since it is not a
great matter - *very* few articles are a problem.

What is a greater problem (as a side note) is the \201 that is printed
in the Groups buffer for 0.06 and 0.07. (See other mails regarding
this.) They prevents me from upgrading... :-/ (Since they are so ugly.)

-- 
(          www.dtek.chalmers.se/~d4jonas/         !     Wei Wu Wei     )
(        Meaning of U2 Lyrics, Roleplaying        !  To Do Without Do  )

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Jonas Steverud <d4jonas@dtek.chalmers.se> writes:

> Since I don't know the offending article I cannot (yet) check how the
> locale affects Gnus. I do plan to drop the issue since it is not a
> great matter - *very* few articles are a problem.

I haven't been able to reproduce it yet, but I'll try with your locale
settings.

> What is a greater problem (as a side note) is the \201 that is printed
> in the Groups buffer for 0.06 and 0.07. (See other mails regarding
> this.) They prevents me from upgrading... :-/ (Since they are so ugly.)

I see them too, but I don't know where to start looking.  It would be
a big help if you could do a binary search between working and
non-working CVS until you find the buggy patch.  ("cvs -D ...")

Re: Decoding ISO8859-15 -- offending headers

[Jonas Steverud]

Simon Josefsson <jas@extundo.com> writes:

> Jonas Steverud <d4jonas@dtek.chalmers.se> writes:
[...]
>> What is a greater problem (as a side note) is the \201 that is printed
>> in the Groups buffer for 0.06 and 0.07. (See other mails regarding
>> this.) They prevents me from upgrading... :-/ (Since they are so ugly.)
>
> I see them too, but I don't know where to start looking.  It would be
> a big help if you could do a binary search between working and
> non-working CVS until you find the buggy patch.  ("cvs -D ...")

:-D

I've been thinking on doing a diff between the 0.05 and 0.06 dirs but
I expect them to be very large diffs. I do not have cvs set up and I
think it is a little too much learning on my behalf for thay approch
to be effective.

Hmm... I maybe should do that directory diff after all...

What is the function that prints the contents to the Group buffer
named? It would help me somewhat to debug this.

-- 
(          www.dtek.chalmers.se/~d4jonas/         !     Wei Wu Wei     )
(        Meaning of U2 Lyrics, Roleplaying        !  To Do Without Do  )

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Jonas Steverud <d4jonas@dtek.chalmers.se> writes:

> What is the function that prints the contents to the Group buffer
> named? It would help me somewhat to debug this.

I think I found the patch.  I'll boot XEmacs to see what is breaking.
Masatoshi, do you remember the details?

	* gnus-group.el (gnus-group-name-decode): Don't test
	multibyte-string, because it breaks XEmacs.

Index: gnus-group.el
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/gnus-group.el,v
retrieving revision 6.68
retrieving revision 6.69
diff -u -p -r6.68 -r6.69
--- gnus-group.el	2002/02/20 00:15:31	6.68
+++ gnus-group.el	2002/02/20 16:15:00	6.69
@@ -1030,8 +1030,7 @@ The following commands are available:
       result)))
 
 (defun gnus-group-name-decode (string charset)
-  (if (and string charset (featurep 'mule)
-	   (not (mm-multibyte-string-p string)))
+  (if (and string charset (featurep 'mule))
       (mm-decode-coding-string string charset)
     string))
 

Re: Decoding ISO8859-15 -- offending headers

[Jonas Steverud]

Simon Josefsson <jas@extundo.com> writes:

> Jonas Steverud <d4jonas@dtek.chalmers.se> writes:
>
>> What is the function that prints the contents to the Group buffer
>> named? It would help me somewhat to debug this.
>
> I think I found the patch.

So do I.

[...]
>  (defun gnus-group-name-decode (string charset)
> -  (if (and string charset (featurep 'mule)
> -	   (not (mm-multibyte-string-p string)))

I reverted that change in 0.07 and now it works! Great!

I consider the bug to be found.

-- 
(          www.dtek.chalmers.se/~d4jonas/         !     Wei Wu Wei     )
(        Meaning of U2 Lyrics, Roleplaying        !  To Do Without Do  )

Re: Decoding ISO8859-15 -- offending headers

[Daiki Ueno]

>>>>> In <iluptwg2zlg.fsf@latte.josefsson.org> 
>>>>>	Simon Josefsson <jas@extundo.com> wrote:

> I think I found the patch.  I'll boot XEmacs to see what is breaking.
> Masatoshi, do you remember the details?

> 	* gnus-group.el (gnus-group-name-decode): Don't test
> 	multibyte-string, because it breaks XEmacs.

I should comment on this because I suggested him to remove the multibyte
check.  And, for my real intention, I can't read any Japanese newsgroups
whose names are encoded in UTF-8[1] without the patch.

I'll show you why the old group name encoder/decoder went wrong for
both GNU Emacs and XEmacs-MULE.

XEmacs-MULE: Without the patch, all the group names will be decoded
twice because the current implementation of `mm-multibyte-string-p'
for XEmacs is just an alias to `ignore'.  Unfortunately, we can't
implement working multibyte-string-p for XEmacs-MULE because
XEmacs-MULE uses a Latin-1 character so as to represent a byte in the
range from 160 to 255.  See the last paragraph of
"(internals)Character Sets" for details.

GNU Emacs: nntp-server-buffer is marked as multibyte.  A string picked
from the buffer is always regarded as multibyte, unless it doesn't
contain a non-ASCII character.

Footnotes: 
[1]  Section 5.5. in USEFOR draft:
<http://www.ietf.org/internet-drafts/draft-ietf-usefor-article-07.txt>
-- 
Daiki Ueno

Re: Decoding ISO8859-15 -- offending headers

[Jonas Steverud]

Daiki Ueno <ueno@unixuser.org> writes:

[...]
> I'll show you why the old group name encoder/decoder went wrong for
> both GNU Emacs and XEmacs-MULE.

I.e. it seems like the code has to check for which version of Emacs
that is running. A not very beatiful solution IMNSHO. The reason I
brought up the issue is that nnfolder groups is attacked by \201 in
the name so is it a better solution to do the mutlibyte process in the
nnfolder code before the group names are sent to the higher levels?

Gnus shall expect to find non-A-Z in the group names of the
mailfolders and process it accordingly. From a data arbritation (sp?)
point of view, I would say that the best place to process the various
characters are in the lower levels and then have all backends return
multibyte (or whatever) and send it to the buffer where the buffer
display code converts it to whatever the user can handle.

Please consider that I'm no expert of the internal functionality of Gnus.

-- 
(          www.dtek.chalmers.se/~d4jonas/         !     Wei Wu Wei     )
(        Meaning of U2 Lyrics, Roleplaying        !  To Do Without Do  )

Re: Decoding ISO8859-15 -- offending headers

[Kai Großjohann]

Jonas Steverud <d4jonas@dtek.chalmers.se> writes:

> Daiki Ueno <ueno@unixuser.org> writes:
>
> [...]
>> I'll show you why the old group name encoder/decoder went wrong for
>> both GNU Emacs and XEmacs-MULE.
>
> I.e. it seems like the code has to check for which version of Emacs
> that is running.

I wonder if Dave's Gnus changes in the Emacs CVS branch emacs-unicode
address this problem.  He said that he tried to make Gnus less
dependent on the internal encoding used.

kai
-- 
A large number of young women don't trust men with beards.  (BFBS Radio)

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> I wonder if Dave's Gnus changes in the Emacs CVS branch emacs-unicode
> address this problem.  He said that he tried to make Gnus less
> dependent on the internal encoding used.

The functions involved look the same in the unicode branch, at least
when I looked quickly.

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Daiki Ueno <ueno@unixuser.org> writes:

>>>>>> In <iluptwg2zlg.fsf@latte.josefsson.org> 
>>>>>>	Simon Josefsson <jas@extundo.com> wrote:
>
>> I think I found the patch.  I'll boot XEmacs to see what is breaking.
>> Masatoshi, do you remember the details?
>
>> 	* gnus-group.el (gnus-group-name-decode): Don't test
>> 	multibyte-string, because it breaks XEmacs.
>
> I should comment on this because I suggested him to remove the multibyte
> check.  And, for my real intention, I can't read any Japanese newsgroups
> whose names are encoded in UTF-8[1] without the patch.

Right, it seem to break my UTF-8 nntp groups too.

It also makes the .newsrc.eld contain different codings of the group
name depending on if I use Emacs or XEmacs.  Not good.

> I'll show you why the old group name encoder/decoder went wrong for
> both GNU Emacs and XEmacs-MULE.
>
> XEmacs-MULE: Without the patch, all the group names will be decoded
> twice because the current implementation of `mm-multibyte-string-p'
> for XEmacs is just an alias to `ignore'.  Unfortunately, we can't
> implement working multibyte-string-p for XEmacs-MULE because
> XEmacs-MULE uses a Latin-1 character so as to represent a byte in the
> range from 160 to 255.  See the last paragraph of
> "(internals)Character Sets" for details.
>
> GNU Emacs: nntp-server-buffer is marked as multibyte.  A string picked
> from the buffer is always regarded as multibyte, unless it doesn't
> contain a non-ASCII character.

Ok.  What would the solution be?  Make all the backends return
multibyte group strings?  Right now it seems nnfolder (and nnimap)
uses unibyte, and nntp uses multibyte, which won't work.

Re: Decoding ISO8859-15 -- offending headers

[Kai Großjohann]

Simon Josefsson <jas@extundo.com> writes:

> It also makes the .newsrc.eld contain different codings of the group
> name depending on if I use Emacs or XEmacs.  Not good.

I think it would be good if Gnus always used emacs-mule as the
encoding for .newsrc.eld.  What do you think?

kai
-- 
A large number of young women don't trust men with beards.  (BFBS Radio)

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> Simon Josefsson <jas@extundo.com> writes:
>
>> It also makes the .newsrc.eld contain different codings of the group
>> name depending on if I use Emacs or XEmacs.  Not good.
>
> I think it would be good if Gnus always used emacs-mule as the
> encoding for .newsrc.eld.  

It has a emacs-lisp cookie, so I think Emacs does this (or doesn't
emacs store elisp in emacs-mule encoding?).

> What do you think?

Won't work when I use XEmacs without MULE.

Saving and reading .newsrc.eld in an interoperable encoding,
preferably even standardized, is the only acceptable solution, I
think, but the set to chose from is empty and will be for many years
too.  I'd love to be mistaken here though.

Btw, how about making Gnus put all of the recipient addresses in To:
instead of messing up the headers like it does in this mail?  It looks
as if I'm addressing Daiki here (or is that Ueno?).

Re: Decoding ISO8859-15 -- offending headers

[Kai Großjohann]

Simon Josefsson <jas@extundo.com> writes:

> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
>
>> Simon Josefsson <jas@extundo.com> writes:
>>
>>> It also makes the .newsrc.eld contain different codings of the group
>>> name depending on if I use Emacs or XEmacs.  Not good.
>>
>> I think it would be good if Gnus always used emacs-mule as the
>> encoding for .newsrc.eld.  
>
> It has a emacs-lisp cookie, so I think Emacs does this (or doesn't
> emacs store elisp in emacs-mule encoding?).

Hm?  Why should the mode imply any encoding?  I used to frob the
encoding of ~/.gnus between latin-1, utf-8, and latin-9, for example.

kai
-- 
A large number of young women don't trust men with beards.  (BFBS Radio)

Re: Decoding ISO8859-15 -- offending headers

[Simon Josefsson]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> Simon Josefsson <jas@extundo.com> writes:
>
>> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
>>
>>> Simon Josefsson <jas@extundo.com> writes:
>>>
>>>> It also makes the .newsrc.eld contain different codings of the group
>>>> name depending on if I use Emacs or XEmacs.  Not good.
>>>
>>> I think it would be good if Gnus always used emacs-mule as the
>>> encoding for .newsrc.eld.  
>>
>> It has a emacs-lisp cookie, so I think Emacs does this (or doesn't
>> emacs store elisp in emacs-mule encoding?).
>
> Hm?  Why should the mode imply any encoding?  I used to frob the
> encoding of ~/.gnus between latin-1, utf-8, and latin-9, for example.

You are right, I was thinking of *.elc (which is loaded as emacs-mule)
and confused things, sorry.