[Evolution] does evolution respect standards of web ?

[hulin <hulin thibaud wanadoo fr>]

Hi!

Sorry for this provocative subject and thanks for your work.
I use evolution on Debian testing and I'm subscribed at a national
mailinglist (for professionals of libraries) that use sympa mailing list
server (www.sympa.org). However, I can't read the subject of that
mailing list, because evolution use the bad encodage to show headings.

I checked that evolution use the ISO-8859-1 (accentued caracters for
Ouest Europe)
in preferences menu / edition and view messages. Curiously, the
subject is unreadable in the list frame, but not in view message frame.

By example, when the unreadable subject give in the blue/grey frame :
Sujet: March=?ISO-8859-1?B?6SA=?=- Ville de Betton

I can read in the white frame :
Objet : Marché public - Fourniture de documents  imprimés pour la
médiathèque de BETTON

I wrote at the evolution mailing list about this problem, but I had
noone answer. So, I wrote at the list administrators. They forwarded my
message to sympa administrators. According to them, the problem is not
from sympa, but evolution. In fact, thunderbird and claws-mail can
perfectly show headings of this mailing list. Sympa's developper who
replied me think conclue that evolution don't respect standards of web,
but I don't think so. Then, what's the problem with evolution ?

Thanks very much,
Thibaud.

PS : that's the reply of sympa list with technical precisions. Sorry,
it's a traduction from the original french message...


Sympa don't touch with the encodage of messages any more. Those what it
generate are in UTF-8, not in ISO-8859-1. Thus, the subject concerned
was not generated by sympa.
Then :
- the heading of the heading of provided message is perfectly valid from
the
point of view of the RFC 2047 (encoding of the characters of the
headings). It is thus the customer mail of your user who seems at
fault here; - All the messages containing of the accents which I saw
in the list biblio-Fr (a fortnight, all dating from October) have this
behaviour. Their author is... "Moderateur BIBLIO-FR"... ;) it is your
customer mail (what "Microsoft-entourage" with the fact, I never did
intend some to speak?) who creates these headings; but, once again,
they are perfectly valid; - the messages in question are tolerated
perfectly by Thunderbird.

To return in detail, which could interest your user or the developers
of evolution: The headings of messages (of which the subject forms
part) are encodés only in ASCII. One cannot lay at it stresses,
protocol smtp is thus defined. As one wants nevertheless to write
accents, they should be coded in one way or another. Basic syntax is:
=?*charset*?*encoding*?*result *? = charset: the encoding employed
(ISO-8859-1 in what concerns us) encoding: B (bases 64) or Q (quoted
printable) the way in which the characters are coded result: the code
of the character itself.

As follows: "Subject: How =?ISO-8859-1?Q?=E7a_va_=3F? = " will be
interpreted by a prone field in your customer mail containing: "How
that is?"

You see that: 1 in the encoded part (= E7a_va_=3F), spaces are
represented by "_". 2 whole is not encodé. "How" not containing
anything with encoder, it is quite simply left in-outside zone of
encoding.

It is this second point which plants evolution. I compared with
messages generated by Thunderbird: this one included all the subject
in a zone of encoding, even what does not have to be encodé. It would
have thus encodé "How that goes" in "Subject:
=?ISO-8859-1?Q?Comment_=E7a_va_=3F? = ". Such messages are read
without problem by evolution. What it does not seem to be able to
make, it is to cover a subject of which a part is encoded and another
is not it.

It is in that that it is wrong.

Here. No the final solution with this problem, therefore. It is
necessary that evolution... evolves/moves to respect the standards of
the Web;)