Re: message-forward-subject-name-subject sensitive to non-ascii

Re: message-forward-subject-name-subject sensitive to non-ascii

[Lars Magne Ingebrigtsen]

Maciej Matysiak <phoner.ding@blah.pl> writes:

> (raw: From: Witold Bar=?ISO-8859-2?B?s/M=?=g <witek68#poczta.onet.pl>)

That's not a valid RFC2047 encoding.  

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

Re: message-forward-subject-name-subject sensitive to non-ascii

[Jesper Harder]

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

> Maciej Matysiak <phoner.ding@blah.pl> writes:
>
>> (raw: From: Witold Bar=?ISO-8859-2?B?s/M=?=g <witek68#poczta.onet.pl>)
>
> That's not a valid RFC2047 encoding.  

Why not?

(rfc2047-decode-string 
 "From: Witold Bar=?ISO-8859-2?B?s/M=?=g <witek68#poczta.onet.pl>")
=> "From: Witold Barłóg <witek68#poczta.onet.pl>"

I think the problem Maciej is mentioning is real, i.e. that
`mail-header-parse-address' fails in
`message-forward-subject-name-subject' because it's used on a From
line which hasn't been RFC 2047-decoded yet.

Re: message-forward-subject-name-subject sensitive to non-ascii

[Lars Magne Ingebrigtsen]

Jesper Harder <harder@myrealbox.com> writes:

>>> (raw: From: Witold Bar=?ISO-8859-2?B?s/M=?=g <witek68#poczta.onet.pl>)
>>
>> That's not a valid RFC2047 encoding.  
>
> Why not?

Encoded-words can't appear in the middle of other words.  They can
only be intruduced on boundaries (CLWS, parentheses, etc).

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

Re: message-forward-subject-name-subject sensitive to non-ascii

[Jesper Blåbærgrød Harder]

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

> Jesper Harder <harder@myrealbox.com> writes:
>
>>>> (raw: From: Witold Bar=?ISO-8859-2?B?s/M=?=g <witek68#poczta.onet.pl>)
>>>
>>> That's not a valid RFC2047 encoding.  
>>
>> Why not?
>
> Encoded-words can't appear in the middle of other words.  They can
> only be intruduced on boundaries (CLWS, parentheses, etc).

Ah, right.

But I still suspect `message-forward-subject-name-subject' is broken.
`mail-header-parse-address' doesn't work on RFC2047-encoded strings,
they must be decoded first.

Lets see ... if I forward a message with this From line:

From: Jesper =?iso-8859-1?q?Bl=E5b=E6rgr=F8d?= Harder <harder@ifa.au.dk>

I get this:

Subject: [Jesper =?iso-8859-1?q?Bl =E5b =E6rgr =F8d? = Harder] Da da da

which isn't right.

Re: message-forward-subject-name-subject sensitive to non-ascii

[Lars Magne Ingebrigtsen]

Jesper Blåbærgrød Harder <harder@myrealbox.com> writes:

> Lets see ... if I forward a message with this From line:
>
> From: Jesper =?iso-8859-1?q?Bl=E5b=E6rgr=F8d?= Harder <harder@ifa.au.dk>

I tried `S O m', and I got this:

Subject: [Jesper Blåbærgrød Harder] message-forward-subject-name-subject sensitive to non-ascii

What command are you using?

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

Re: message-forward-subject-name-subject sensitive to non-ascii

[Jesper Harder]

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

> Jesper Blåbærgrød Harder <harder@myrealbox.com> writes:
>
>> Lets see ... if I forward a message with this From line:
>>
>> From: Jesper =?iso-8859-1?q?Bl=E5b=E6rgr=F8d?= Harder <harder@ifa.au.dk>
>
> I tried `S O m', and I got this:
>
> Subject: [Jesper Blåbærgrød Harder] message-forward-subject-name-subject sensitive to non-ascii
>
> What command are you using?

`C-c C-f'.

Re: message-forward-subject-name-subject sensitive to non-ascii

[Lars Magne Ingebrigtsen]

Jesper Harder <harder@myrealbox.com> writes:

> `C-c C-f'.

Yup.  Fix in Oort Gnus v0.23 (i. e., CVS).

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

Re: message-forward-subject-name-subject sensitive to non-ascii

[Jesper Harder]

Jesper Harder <harder@myrealbox.com> writes:

> But I still suspect `message-forward-subject-name-subject' is broken.
> `mail-header-parse-address' doesn't work on RFC2047-encoded strings,
> they must be decoded first.

Oops, that was a bogus suggestion.  

`mail-header-parse-address' didn't work because "=" had the wrong
syntax, not because the string should rfc2047-decoded first (it
shouldn't).

Re: message-forward-subject-name-subject sensitive to non-ascii

[Lars Magne Ingebrigtsen]

Jesper Harder <harder@myrealbox.com> writes:

> I think the problem Maciej is mentioning is real, i.e. that
> `mail-header-parse-address' fails in
> `message-forward-subject-name-subject' because it's used on a From
> line which hasn't been RFC 2047-decoded yet.

That's possible.  A test case that had valid input would be nicer,
though.  :-)

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

Re: message-forward-subject-name-subject sensitive to non-ascii

[Lars Magne Ingebrigtsen]

Maciej Matysiak <phoner.ding@blah.pl> writes:

>  From: =?ISO-8859-2?Q?Rafa=B3_Komorowski?= <komor#mac,com>
> (addresses malformed to avoid harvesting).
> the names are decoded properly in the summary buffer.
>
> at the moment, gnus almost does the job. there's no more encoding garbage,
> just a tilde in a place of latin2 character. is that fixable any further?

A tilde?  That's odd.  Are you sure it's really a tilde and not a
font problem?  (following-char) should tell you.  

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

Re: message-forward-subject-name-subject sensitive to non-ascii

[Lars Magne Ingebrigtsen]

Maciej żółty Matysiak <phoner.ding@blah.pl> writes:

> i've added a word with latin2 characters in my `from:', you can try
> forwarding this article.

Things worked for me, but perhaps this is the result over things in
rfc2047.el not quite working as they should in XEmacs?  I think this
might have been fixed by now...

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

Re: message-forward-subject-name-subject sensitive to non-ascii

[Jesper Harder]

Maciej Matysiak <phoner.ding@blah.pl> writes:

> sorry, but that bug still exists :( iso-8859-2 characters in people's names
> are changed into tildes (literally) in the subject line of forwarded message.
> can be, that it's xemacs specific. i don't have fsf emacs here, so i can't
> check.
>
>  m.m. (cvs gnus updated few minutes ago)

Try to update again.  It now works for me in Gnu Emacs (it didn't
previously with your test message).

Re: message-forward-subject-name-subject sensitive to non-ascii

[山岡克美]

>>>>> In <m3he86w7na.fsf@defun.localdomain>
>>>>>	Jesper Harder <harder@myrealbox.com> wrote:

> Try to update again.  It now works for me in Gnu Emacs (it didn't
> previously with your test message).

Both Emacs and XEmacs signal an error to Japanese From.  This
message is just an example. :)
-- 
Katsumi Yamaoka <yamaoka@jpl.org>

Re: message-forward-subject-name-subject sensitive to non-ascii

[Jesper Harder]

山岡克美 <yamaoka@jpl.org> writes:

>>>>>> In <m3he86w7na.fsf@defun.localdomain>
>>>>>>	Jesper Harder <harder@myrealbox.com> wrote:
>
>> Try to update again.  It now works for me in Gnu Emacs (it didn't
>> previously with your test message).
>
> Both Emacs and XEmacs signal an error to Japanese From.  This
> message is just an example. :)

Weird ... it works fine for me in both Emacs 21.3.50 and XEmacs 21.4.
What does the backtrace look like?

Re: message-forward-subject-name-subject sensitive to non-ascii

[Katsumi Yamaoka]

>>>>> In <m34r455wcs.fsf@defun.localdomain> 
>>>>>	Jesper Harder <harder@myrealbox.com> wrote:

>> Both Emacs and XEmacs signal an error to Japanese From.  This
>> message is just an example. :)

> Weird ... it works fine for me in both Emacs 21.3.50 and XEmacs 21.4.
> What does the backtrace look like?

Oops!  There's no problem in my home computer.  Maybe there's
something wrong in the office.  I need to investigate and report
it tomorrow.  I'm sorry for my fault.
-- 
Katsumi@Yamaoka.cc

Re: message-forward-subject-name-subject sensitive to non-ascii

[Katsumi Yamaoka]

>>>>> In <yotlptmtsaxp.fsf@jpl.org>
>>>>>	Katsumi Yamaoka <yamaoka@jpl.org> wrote:

>> Weird ... it works fine for me in both Emacs 21.3.50 and XEmacs 21.4.
>> What does the backtrace look like?

> Oops!  There's no problem in my home computer.

I was silly to have forgotten to update Gnus in the new linux
machine.  Beat me (if your whip reaches to Japan).
-- 
Katsumi Yamaoka <yamaoka@jpl.org>

Re: message-forward-subject-name-subject sensitive to non-ascii

[Simon Josefsson]

Maciej Matysiak <phoner.ding@blah.pl> writes:

> 2. there's still problem with compiling ssl.el with xemacs:
>
> Compiling /home/phoner/Elisp/current-gnus/lisp/ssl.el...
> While compiling open-ssl-stream in file /home/phoner/Elisp/current-gnus/lisp/ssl.el:
>   ** function open-ssl-stream used to take 0+ arguments, now takes 4
> Wrote /home/phoner/Elisp/current-gnus/lisp/ssl.elc

contrib/ssl.el is present as a convenience for people that don't have
the file, it is not supposed to be used (or byte compiled) if you
already have SSL functions installed (which XEmacs does).

Since we have a free replacement in tls.el + GnuTLS now, I think we
should remove ssl.el.

Re: message-forward-subject-name-subject sensitive to non-ascii

[Dave Love]

Simon Josefsson <jas@extundo.com> writes:

> Since we have a free replacement in tls.el + GnuTLS now, I think we
> should remove ssl.el.

What's non-free about ssl.el?

Re: message-forward-subject-name-subject sensitive to non-ascii

[Simon Josefsson]

Dave Love <d.love@dl.ac.uk> writes:

> Simon Josefsson <jas@extundo.com> writes:
>
>> Since we have a free replacement in tls.el + GnuTLS now, I think we
>> should remove ssl.el.
>
> What's non-free about ssl.el?

It uses OpenSSL, which makes the manual refer to non-free code.  I
recall trying to get ssl.el included in Emacs, but finally ended up
with tls.el that Richard accepted.

Re: message-forward-subject-name-subject sensitive to non-ascii

[Dave Love]

Simon Josefsson <jas@extundo.com> writes:

> It uses OpenSSL, which makes the manual refer to non-free code.

I don't understand.  Do you mean openssl is non-free?  It presumably
satisfies the Debian guidelines, even if it's not GPL-compatible.  [I
ask because ssl.el comes from W3 and actually should have been
installed in Emacs long ago.]

Re: message-forward-subject-name-subject sensitive to non-ascii

[Simon Josefsson]

(NB, mail-followup-to set to emacs-devel@gnu.org.)

Dave Love <d.love@dl.ac.uk> writes:

> Simon Josefsson <jas@extundo.com> writes:
>
>> It uses OpenSSL, which makes the manual refer to non-free code.
>
> I don't understand.  Do you mean openssl is non-free?  It presumably
> satisfies the Debian guidelines, even if it's not GPL-compatible. [I
> ask because ssl.el comes from W3 and actually should have been
> installed in Emacs long ago.]

I don't claim to understand it either; but I recall having tryied to
get ssl.el included in Emacs but failed with a reason I didn't
understand, but attributed to some license issue.  It was some time
ago, so perhaps someone would like to try again.  OTOH, we now have
tls.el and it might be better to use and support that instead so any
enhancements end up in GnuTLS instead of OpenSSL.