pGnus 0.13 & archive

pGnus 0.13 & archive

[Jean-Yves Perrier]

Hello,

I have a strange behaviour in pGnus 0.13:

When I post to myself a message with accentued chars (iso 8859-1),
the autodetect works (By the way should it send it using 8bit encoding,
or quoted-printable, rather than 7bit ?). Good.

When I receive it, the translation works. Very Good. (seems to work with
other charsets too: I see square as I don't have other fonts installed)

But when I open it in my nnfolder:archive group, i see \201é and so on (MULE
feature...) 

I have Emacs 20.3, pGnus0.13 & no other stuff to do MIME coding (no TM,
nothing...)
Note : I have (standard-display-european t) in my .emacs

Lars, thank you for the good work.
-- 
Jean-Yves Perrier

Re: pGnus 0.13 & archive

[Lars Magne Ingebrigtsen]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="us-ascii", Size: 544 bytes --]

Jean-Yves Perrier <perrier@nagra-kudelski.ch> writes:

> But when I open it in my nnfolder:archive group, i see \201é and so on (MULE
> feature...) 

Hm.  I'm unable to reproduce this bug.

> I have Emacs 20.3, pGnus0.13 & no other stuff to do MIME coding (no TM,
> nothing...)
> Note : I have (standard-display-european t) in my .emacs

I use the same.

Is it possible that your Emacs does some extra de/encoding of your
nnfolder folders?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: pGnus 0.13 & archive

[Jean-Yves Perrier]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Hm.  I'm unable to reproduce this bug.

OK, I will try to precise...

> Is it possible that your Emacs does some extra de/encoding of your
> nnfolder folders?

I did try the following:

I entered the article with C-u g.
Accents were perfect.

I emptied my gnus-article-display-hook, then I added the different hooks I
had one by one and I found that one and only one of them triggers the bug:

gnus-article-decode-charset


More, there is two different cases: one triggers the problem, the other not.

Case 1: -- The problem does not appear
- start configuration: I have no *Article* buffer (I just kill it) OR the
                       *article* buffer is in ISO-8859-1 ( '-1:--' on the
                       mode line.
- I enter this message: everything fine.

Case 2: -- The problem does appear
- start configuration: I have an *Article* buffer with '-J:--' on the mode
                       line.
- I enter this message: \201 in front of accents (éèà...)

I noticed also the following: this happen with message that have:
Content-Transfer-Encoding: 7bit

This does not happen on messages received on this mailing-list as I receive
them with:
Content-Transfer-Encoding: quoted-printable

I was never able to trigger the bug with quoted-printable.

Note: the original message (the one which does not show correctly) was
made with message-mode and pGnus 0.13

Hope it helps you,
-- 
Jean-Yves Perrier

Re: pGnus 0.13 & archive

[Lars Magne Ingebrigtsen]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="us-ascii", Size: 411 bytes --]

Jean-Yves Perrier <perrier@nagra-kudelski.ch> writes:

> Case 2: -- The problem does appear
> - start configuration: I have an *Article* buffer with '-J:--' on the mode
>                        line.
> - I enter this message: \201 in front of accents (éèà...)

How do you get the article buffer to be "J:"?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: pGnus 0.13 & archive

[Jean-Yves Perrier]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> How do you get the article buffer to be "J:"?

I read the article from SL Baur which where posted on this mailing-list
one or two days ago.

I always see this when the article buffer is "-:" by reading my last
message on this mailing-list.

It never happens if I read the accentued characters first. (The buffer
is then "-1:"

Hope it helps,
-- 
Jean-Yves Perrier

Re: pGnus 0.13 & archive

[Lars Magne Ingebrigtsen]

Jean-Yves Perrier <perrier@nagra-kudelski.ch> writes:

> > How do you get the article buffer to be "J:"?
> 
> I read the article from SL Baur which where posted on this mailing-list
> one or two days ago.

Hm.  This doesn't seem to happen to me.  Could someone post a message
that should make my article buffer be "J:"?  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: pGnus 0.13 & archive

[Jean-Yves Perrier]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Hm.  This doesn't seem to happen to me.  Could someone post a message
> that should make my article buffer be "J:"?  :-)

This should be enough: 「日本語」

I played around with such a message (killing the *Article* buffer, then entering
a japanese message first, or an iso-8859-1 msg first, or an ascii msg first) I was
able to see the same japanese encoded message:

with --:-- and square for japanese (which is correct as jap fonts are not installed)
     -1:-- and japanese replaced by accentued chars !
     -J:-- and square for japanese

That's really odd... A message display shouldn't depend of the previous display :-(

(I begin to thing that MULE has some problems...)

Hope it helps,
-- 
Jean-Yves Perrier

Re: pGnus 0.13 & archive

[Lars Magne Ingebrigtsen]

Jean-Yves Perrier <perrier@nagra-kudelski.ch> writes:

> This should be enough: ^[$B!VF|K\8l!W^[(B
> 
> I played around with such a message (killing the *Article* buffer, then entering
> a japanese message first, or an iso-8859-1 msg first, or an ascii msg first) I was
> able to see the same japanese encoded message:
> 
> with --:-- and square for japanese (which is correct as jap fonts are not installed)
>      -1:-- and japanese replaced by accentued chars !
>      -J:-- and square for japanese

Argh.

I always get "1:", so I've now removed

	(set-language-environment "Latin-1")

from .emacs.  And now I get "-:" all the time.  No matter what I do.
And if I call

(encode-coding-region (point-min) (point-max) 'iso-2022-jp)

in a buffer with

"This should be enough: ^[$B!VF|K\8l!W^[(B"

in it, I get

"This should be enough: ^[$B!VF|K\8l!W^[(B"

(i.e., absolutely no change whatsoever.)

Argh.

I really need to read the preprint of the new manual, I guess.  Or is
this a bug?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: pGnus 0.13 & archive

[Jon Kvebaek]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
| 
| I always get "1:", so I've now removed
| 
| 	(set-language-environment "Latin-1")

I always got -:, so I added just that :)
 
| from .emacs.  And now I get "-:" all the time.  No matter what I do.

And I get...hey, not in thîs¹ message buffer.

| And if I call
| 
| (encode-coding-region (point-min) (point-max) 'iso-2022-jp)
| "This should be enough: ^[$B!VF|K\8l
| (i.e., absolutely no change whatsoever.)

Exactly the same here, so I don't think set-language-environment does
much in that particular case.

| I really need to read the preprint of the new manual, I guess.  Or
| is this a bug?

`I wish I had never come here, and I don't want to see no more magic,'
he said, and fell silent.

¹Let's see...
-- 
Jon Kvebaek <jkv@sys.sol.no>
Wrinkled earlobes are a sign of impending heart attacks.

Re: pGnus 0.13 & archive

[Lars Magne Ingebrigtsen]

Jon Kvebaek <jkv@a.sol.no> writes:

> ¹Let's see...

I've now added explicit checks to `mm-body-encoding' to see whether we
really have 8bit chars.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: pGnus 0.13 & archive

[Lars Balker Rasmussen]

Jon Kvebaek <jkv@a.sol.no> writes:
> | "This should be enough: 

Heh.  I just get a screen full of assorted far eastern glyphs in the
quoted article, from the quoted and onwards.  Can't post it, as Emacs
complains:  Can't encode messages with multiple charsets (yet).

But still 1:.

Weird.
-- 
Lars Balker Rasmussen                                        "Woo hoo!?"
.

Re: pGnus 0.13 & archive

[Jon Kvebaek]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

| Jon Kvebaek <jkv@a.sol.no> writes:
| 
| > ¹Let's see...
| 
| I've now added explicit checks to `mm-body-encoding' to see whether we
| really have 8bit chars.

This seems like an immediate¹ success - the headers have been correct
so far.

¹Note: immediate :)
-- 
Jon Kvebaek <jkv@sys.sol.no>
Wrinkled earlobes are a sign of impending heart attacks.

Re: pGnus 0.13 & archive

[Lars Magne Ingebrigtsen]

Lars Balker Rasmussen <gnort@daimi.aau.dk> writes:

> > | "This should be enough: 
> 
> Heh.  I just get a screen full of assorted far eastern glyphs in the
> quoted article, from the quoted and onwards.

That's if you `C-u W M c iso-2202-jp' the article, I take it?  Surely
not automatically?

> But still 1:.

I can't get it to be anything but "1:".

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: pGnus 0.13 & archive

[Lars Balker Rasmussen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Lars Balker Rasmussen <gnort@daimi.aau.dk> writes:
> > > | "This should be enough: 
> > 
> > Heh.  I just get a screen full of assorted far eastern glyphs in the
> > quoted article, from the quoted and onwards.
> 
> That's if you `C-u W M c iso-2202-jp' the article, I take it?  Surely
> not automatically?

Automatically (with 0.13 and Emacs 20.2!).

I've just tried it with 0.14, and now it's not that way, as I can't get
it to show japanese in articles claiming to be iso-8859-1 or us-ascii.

Jean-Yves' article <yf4hfypoeyq.fsf@tom.nagra-kudelski.ch> claims to be
iso-2022-jp and shows ^[$(B!VF|K\8l!W^[(B in japanese, as it should.
-- 
Lars Balker Rasmussen, Software Engineer, Mjolner Informatics ApS
lbr@mjolner.dk

Re: pGnus 0.13 & archive

[Lars Balker Rasmussen]

Jon Kvebaek <jkv@a.sol.no> writes:
> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> | I've now added explicit checks to `mm-body-encoding' to see whether we
> | really have 8bit chars.
> 
> This seems like an immediate¹ success - the headers have been correct
> so far.
> 
> ¹Note: immediate :)

I just tried something.  Stripped of boring details, this is what I
posted to a test group.  Same thing with Emacs 20.3.  This is with
iso-insert.

----------------------------------------------------------------------
Subject: Tæst
Lines: 3
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 8bit
X-Newsreader: Pterodactyl Gnus v0.14/Emacs 20.2

Tæst
----------------------------------------------------------------------

What's going wrong to get this abomination?

What about encoding iso-8859 in headers?
-- 
Lars Balker Rasmussen, Software Engineer, Mjolner Informatics ApS
lbr@mjolner.dk

Re: pGnus 0.13 & archive

[Lars Magne Ingebrigtsen]

Lars Balker Rasmussen <lbr@mjolner.dk> writes:

> I've just tried it with 0.14, and now it's not that way, as I can't get
> it to show japanese in articles claiming to be iso-8859-1 or us-ascii.
> 
> Jean-Yves' article <yf4hfypoeyq.fsf@tom.nagra-kudelski.ch> claims to be
> iso-2022-jp and shows ^[$(B!VF|K\8l!W^[(B in japanese, as it should.

If I `C-u W M c iso-2202-jp' this article under Emacs 20.3, I get the
expected Japanese glyphs...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: pGnus 0.13 & archive

[Lars Magne Ingebrigtsen]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="us-ascii", Size: 808 bytes --]

Lars Balker Rasmussen <lbr@mjolner.dk> writes:

> I just tried something.  Stripped of boring details, this is what I
> posted to a test group.  Same thing with Emacs 20.3.  This is with
> iso-insert.
> 
> ----------------------------------------------------------------------
> Subject: Tæst
> Lines: 3
> Mime-Version: 1.0
> Content-Type: text/plain; charset="us-ascii"
> Content-Transfer-Encoding: 8bit
> X-Newsreader: Pterodactyl Gnus v0.14/Emacs 20.2
> 
> Tæst
> ----------------------------------------------------------------------
> 
> What's going wrong to get this abomination?

Eek.  The iso-insert functions are seriously bogus.  I've reported
this as a bug to RMS.  We'll see what he says...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen