bug: incorrect encoding of user-full-name

bug: incorrect encoding of user-full-name

[vvv]

Hi,

I've set my user-full-name to spell in Russian:

(setq user-full-name "Владимир Волович")

When i send my mails, the From: header looks like:

From: vvv@vvv.vsu.ru =?iso-8859-5?b?KLLb0NTY3NjgILLe297S2Ocp?=

which is incorrect, because the braces () are encoded with my name,
and =?iso-8859-5?b?KLLb0NTY3NjgILLe297S2Ocp?= looks like a recipient's
address. The correct would be something like

From: vvv@vvv.vsu.ru (=?iso-8859-5?b?...?=)

or, preferably,

From: =?iso-8859-5?b?...?= <vvv@vvv.vsu.ru>

(i wonder, why if i do not set user-full-name explicitly, and it
contains my finger data, which uses latin letters, gnus uses the
second form: "Vladimir Volovich <vvv@vvv.vsu.ru>", while when i set my
user-full-name value, gnus changed the format of From: header?)

	Best regards, -- Vladimir.

Re: bug: incorrect encoding of user-full-name

[Lars Magne Ingebrigtsen]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=us-ascii, Size: 978 bytes --]

vvv@vvv.vsu.ru.(²ÛÐÔØÜØà ²ÞÛÞÒØç) writes:

> When i send my mails, the From: header looks like:
> 
> From: vvv@vvv.vsu.ru =?iso-8859-5?b?KLLb0NTY3NjgILLe297S2Ocp?=

Yup.  Fix in Pterodactyl Gnus v0.30.

> From: vvv@vvv.vsu.ru (=?iso-8859-5?b?...?=)
> 
> or, preferably,
> 
> From: =?iso-8859-5?b?...?= <vvv@vvv.vsu.ru>
> 
> (i wonder, why if i do not set user-full-name explicitly, and it
> contains my finger data, which uses latin letters, gnus uses the
> second form: "Vladimir Volovich <vvv@vvv.vsu.ru>", while when i set my
> user-full-name value, gnus changed the format of From: header?)

It's because the From header is created before the header is encoded,
and the first form is used when the display name looks like it's not
going to fit in the second format.  Which is does only after encoding.

So to force the second form, set `message-from-style' to `angles'.

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

Re: bug: incorrect encoding of user-full-name

[Владимир Волович]

"LMI" == Lars Magne Ingebrigtsen writes:

 LMI> vvv@vvv.vsu.ru.( ŒèŒßŒâŒìŒíŒìŒÎ
 LMI> ŒçŒèŒçŒàŒìŒ³) writes:
 >> When i send my mails, the From: header looks like:
 >> 
 >> From: vvv@vvv.vsu.ru =?iso-8859-5?b?KLLb0NTY3NjgILLe297S2Ocp?=

 LMI> Yup.  Fix in Pterodactyl Gnus v0.30.

Thanks, it works.

 >> From: vvv@vvv.vsu.ru (=?iso-8859-5?b?...?=)
 >> 
 >> or, preferably,
 >> 
 >> From: =?iso-8859-5?b?...?= <vvv@vvv.vsu.ru>
 >> 
 >> (i wonder, why if i do not set user-full-name explicitly, and it
 >> contains my finger data, which uses latin letters, gnus uses the
 >> second form: "Vladimir Volovich <vvv@vvv.vsu.ru>", while when i
 >> set my user-full-name value, gnus changed the format of From:
 >> header?)

 LMI> It's because the From header is created before the header is
 LMI> encoded, and the first form is used when the display name looks
 LMI> like it's not going to fit in the second format.  Which is does
 LMI> only after encoding.

 LMI> So to force the second form, set `message-from-style' to
 LMI> `angles'.

Well, i think that the From header (and other headers) should be
created after encoding. For example, when i do
(setq message-from-style 'angles), my From header looks like:

From: "=?iso-8859-5?b?stvQ1Njc2OAgst7b3tLY5w==?=" <vvv@vvv.vsu.ru>

instead of perfectly legal (and better) form without quotes:

From: =?iso-8859-5?b?stvQ1Njc2OAgst7b3tLY5w==?= <vvv@vvv.vsu.ru>

without quotes.

For example, the following forms of headers are legal:

From: =?ISO-2022-JP?B?GyRCPGkyLBsoQiAbJEJDTkknGyhC?= <morioka@jaist.ac.jp>

From: =?ISO-2022-JP?B?GyRCPGkyLBsoQiAbJEJDTkknGyhC?= (MORIOKA Tomohiko) <morioka@jaist.ac.jp>

There is no need to quote the encoded-word imho, because it does not
contain any `dangerous' characters.

	Best regards, -- Vladimir.

Re: bug: incorrect encoding of user-full-name

[Lars Magne Ingebrigtsen]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=us-ascii, Size: 908 bytes --]

"²ÛÐÔØÜØà ²ÞÛÞÒØç" <vvv@vvv.vsu.ru> writes:

> Well, i think that the From header (and other headers) should be
> created after encoding. For example, when i do
> (setq message-from-style 'angles), my From header looks like:
> 
> From: "=?iso-8859-5?b?stvQ1Njc2OAgst7b3tLY5w==?=" <vvv@vvv.vsu.ru>

Uhm.  And this isn't even valid, since one can't have encoded-words
inside qouted-strings.  *sigh*

Er.

I really want to generate the headers first, and then encode then
afterwards; it makes things a whole lot easier.

Perhaps the encoder should treat qouted-strings as one word, and if it
contains something that is to be encoded, then it should remove the
quotes and then encode the rest?  Would that be correct under all
circumstances?  I guess \"-s inside the strings would have to be
undone first...

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

Re: bug: incorrect encoding of user-full-name

[Владимир Волович]

"LMI" == Lars Magne Ingebrigtsen writes:

 >> From: "=?iso-8859-5?b?stvQ1Njc2OAgst7b3tLY5w==?=" <vvv@vvv.vsu.ru>

 LMI> Uhm.  And this isn't even valid, since one can't have
 LMI> encoded-words inside qouted-strings.  *sigh*
[snip]
 LMI> Perhaps the encoder should treat qouted-strings as one word, and
 LMI> if it contains something that is to be encoded, then it should
 LMI> remove the quotes and then encode the rest?  Would that be
 LMI> correct under all circumstances?  I guess \"-s inside the
 LMI> strings would have to be undone first...

i dunno. semi seems to deal with these things pretty well.

	Best regards, -- Vladimir.

Re: bug: incorrect encoding of user-full-name

[François Pinard]

"Владимир Волович" <vvv@vvv.vsu.ru> writes:

> 	Best regards, -- Vladimir.

I do not know exactly why, but I love seeing foreign characters in the
author field of the summary buffer.  People have the right to their own
name, expressed according to their own culture.

P.S. - This message is also a test.  I stopped forcing Latin-1 in MIME
headers, curious to know what pgnus 0.31 will do about it :-)

P.P.S. - Oops!  I got "Can't encode messages with multiple charsets (yet)"
Let's try without my signature...