UTF-8 breakage (was: Re: Gnus 5.9.0 - Arithmetic error)

[pinard@iro.umontreal.ca (François Pinard)]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=utf-8, Size: 2724 bytes --]

[Matthias Andree]

> pinard@iro.umontreal.ca (François Pinard) writes:

> >> pinard@iro.umontreal.ca (Fran篩s Pinard) writes:
>                                 ^
> I can smell utf-8 breakage here.... looks like some Chinese symbol.

There was some breakage indeed.

I'm quoting below the edited copy I go for my own outgoing message,
after having replaced non-ASCII characters by `[0x...]' to make very
clear what they were.  From the headers, it seem that the message left
the place in quoted-printable.  Was it quoted-printable over UTF-8?
I cannot say.  Conversion back to 8 bits occurred automatically, and this
might have produced some Latin-1 characters.  I guess one needs a lot of
administrative power along the way to really trace down all this mess. :-)

The abnormality I see is that the message encoding is said to be UTF-8,
but the c-cedilla in the second line of attribution was Latin-1.  Surely,
it should have been UTF-8 there.  However, in the same message, the
c-cedilla in the signature seems correctly written in UTF-8.

Besides the various MTAs, it might be an Emacs or a Gnus problem.  Emacs is
surely a complex program for the poor little me, but Mule is an area within
Emacs where I feel particularily lost 1/2 :-).  I feel overwhelmed! :-)

---------------------------------------------------------------------->
X-From-Line: pinard@IRO.UMontreal.CA  Sun Nov 25 17:12:21 2002
[...]
Sender: pinard@sram.qc.ca
To: Jesper Harder <harder@ifa.au.dk>
Cc: ding@gnus.org
Subject: Re: Gnus 5.9.0 - Arithmetic error
[...]
From: pinard@IRO.UMontreal.CA (=?iso-8859-1?q?Fran=E7ois?= Pinard)
Date: 25 Nov 2002 17:12:21 -0500
In-Reply-To: <m37kf1qta8.fsf@defun.localdomain>
[...]
Message-ID: <oqel99ti2i.fsf@carouge.sram.qc.ca>
User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2.92
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by mercure.iro.umontreal.ca id gAPMCLO8020161
[...]

[Jesper Harder]

> pinard@iro.umontreal.ca (Fran[0xE7]ois Pinard) writes:

> > At first glance, I would believe that `M P a' should mark all articles,
> > as its name implies.

> `M P b' does (nearly) the same thing.  It is faster and generally
> better unless you want to decode some multipart binary.  Does `M P b'
> also trigger the error?

`M P b' does not trigger the error on the offending article, and does
indeed mark all articles.  Thanks for the suggestion.

-- 
Fran[0xC3A7]ois Pinard   http://www.iro.umontreal.ca/~pinard
----------------------------------------------------------------------<

-- 
François Pinard   http://www.iro.umontreal.ca/~pinard