Coding system for outgoing mail

Coding system for outgoing mail

[konrad.hinsen]

This bug report will be sent to the Free Software Foundation,
not to your local site managers!
Please write in English, because the Emacs maintainers do not have
translators to read other languages for them.

Your bug report will be posted to the bug-gnu-emacs@gnu.org mailing list,
and to the gnu.emacs.bug news group.

In GNU Emacs 21.2.1 (i386-redhat-linux-gnu, X toolkit, Xaw3d scroll bars)
 of 2002-04-09 on porky.devel.redhat.com
configured using `configure  i386-redhat-linux --prefix=/usr --exec-prefix=/usr --bindir=/usr/bin --sbindir=/usr/sbin --sysconfdir=/etc --datadir=/usr/share --includedir=/usr/include --libdir=/usr/lib --libexecdir=/usr/libexec --localstatedir=/var --sharedstatedir=/usr/com --mandir=/usr/share/man --infodir=/usr/share/info --with-gcc --with-pop --with-sound'
Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: nil
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: en_US
  locale-coding-system: iso-latin-1
  default-enable-multibyte-characters: t

Please describe exactly what actions triggered the bug
and the precise symptoms of the bug:

Since I set up my Emacs for international character set support,
every time I send a mail with non-ASCII characters I have to specify
the latin-1 coding system manually. According to all documentation
I could find, it should be sufficient to set sendmail-coding-system
to latin-1 to have outgoing mail sent as Latin-1 by default, but that
doesn't work.

After a lot of searching, my hypothesis is that something sets the
local value of buffer-file-coding-system to binary for every new mail
buffer. So I added the following to my .emacs:

 (add-hook 'mail-mode-hook
	   (function (lambda ()
		       (setq buffer-file-coding-system 'iso-latin-1))))

but that makes no difference either.

Recent input:
<return> ) C-a C-y \ <backspace> <backspace> <right> 
<backspace> ) <down-mouse-1> <mouse-1> <tab> <down> 
<tab> C-e <left> <left> <left> <left> <left> <left> 
<left> <left> <left> <left> <left> <right> M-d C-d 
C-d C-d l a t i n 0 1 <backspace> <backspace> - 1 C-x 
C-s <down-mouse-1> <mouse-1> C-a C-SPC <down> <down> 
<down> C-w C-x C-s <down-mouse-1> <mouse-1> <down-mouse-1> 
<mouse-1> <backspace> ) ) C-x C-s M-x b u <tab> g <tab> 
<backspace> <backspace> <backspace> s e n <tab> <backspace> 
<backspace> <backspace> <backspace> <backspace> <backspace> 
<backspace> <backspace> <backspace> <backspace> <backspace> 
<backspace> <backspace> <backspace> r e p <tab> o <tab> 
r <tab> <return>

Recent messages:
Undo! [3 times]
Wrote /home/hinsen/.emacs
Mark set [2 times]
Wrote /home/hinsen/.emacs [2 times]
Mark set
Wrote /home/hinsen/.emacs
Mark set
Wrote /home/hinsen/.emacs [2 times]
Making completion list... [3 times]
Loading emacsbug...done


-- 
-------------------------------------------------------------------------------
Konrad Hinsen                            | E-Mail: hinsen@cnrs-orleans.fr
Centre de Biophysique Moleculaire (CNRS) | Tel.: +33-2.38.25.56.24
Rue Charles Sadron                       | Fax:  +33-2.38.63.15.17
45071 Orleans Cedex 2                    | Deutsch/Esperanto/English/
France                                   | Nederlands/Francais
-------------------------------------------------------------------------------

Re: Coding system for outgoing mail

[Kenichi Handa]

In article <200208202142.g7KLgOO19478@localhost.localdomain>, konrad.hinsen@laposte.net writes:
> Since I set up my Emacs for international character set support,
> every time I send a mail with non-ASCII characters I have to specify
> the latin-1 coding system manually. According to all documentation
> I could find, it should be sufficient to set sendmail-coding-system
> to latin-1 to have outgoing mail sent as Latin-1 by default, but that
> doesn't work.

> After a lot of searching, my hypothesis is that something sets the
> local value of buffer-file-coding-system to binary for every new mail
> buffer. So I added the following to my .emacs:

>  (add-hook 'mail-mode-hook
> 	   (function (lambda ()
> 		       (setq buffer-file-coding-system 'iso-latin-1))))

> but that makes no difference either.

Thank you for the bug report, but I can't reproduce that
bug.  Could you give me the following information?

o Can you reproduce the bug when you start Emacs with "-q
  --no-site-file"?

o What do you mean by `non-ASCII characters'?   Does't it
  contain non-Latin-1 characters?

o How do you type (or insert) those non-ASCII characters?

o Does Emacs also asks coding system on saving that mail
  buffer into a file?

o Please show me the contents of *Warning* buffer shown when
  Emacs asks you to select a coding system.

o How do you create that mail buffer?  By typing C-x m, from
  RMAIL, from Gnus, or else?

o What is the actual value of `buffer-file-coding-system'
  when you don't setup that mail-mode-hook.

o What is the value of `mail-user-agent'?

o What is the value of `default-sendmail-coding-system'?

---
Ken'ichi HANDA
handa@etl.go.jp

Re: Coding system for outgoing mail

[konrad.hinsen]

> o Can you reproduce the bug when you start Emacs with "-q
>   --no-site-file"?

No. I went through my .emacs line by line to find out what the
cause is, and it turns out to be the line

   (setq mail-signature t)

Then I disabled that line and checked what happens when I insert
the signature by hand (C-c C-w). Before inserting the signature,
describe-coding-system says "none, use default". After inserting
the signature, it says "no conversion (binary)".

My signature file contains only ASCII characters (see below), so that
cannot be the problem.

> o What do you mean by `non-ASCII characters'?   Does't it
>   contain non-Latin-1 characters?

No, just French text.

> o How do you type (or insert) those non-ASCII characters?

Directly from a French keyboard.

> o Does Emacs also asks coding system on saving that mail
>   buffer into a file?

No, only when sending it.

> o Please show me the contents of *Warning* buffer shown when
>   Emacs asks you to select a coding system.

   These default coding systems were tried:
     no-conversion iso-latin-1
   no-conversion safely encodes the target text,
   but it is not recommended for encoding text in this context,
   e.g., for sending an email message.

   Select the above, or one of the following safe coding systems:
     iso-8859-1 utf-8 iso-2022-jp-2 x-ctext mac-roman iso-2022-7bit
     raw-text emacs-mule no-conversion ctext-no-compositions
     iso-2022-8bit-ss2 iso-2022-7bit-lock iso-2022-7bit-ss2
     tibetan-iso-8bit-with-esc thai-tis620-with-esc lao-with-esc
     korean-iso-8bit-with-esc hebrew-iso-8bit-with-esc
     greek-iso-8bit-with-esc iso-latin-9-with-esc iso-latin-8-with-esc
     iso-latin-5-with-esc iso-latin-4-with-esc iso-latin-3-with-esc
     iso-latin-2-with-esc iso-latin-1-with-esc
     in-is13194-devanagari-with-esc cyrillic-iso-8bit-with-esc
     chinese-iso-8bit-with-esc japanese-iso-8bit-with-esc

> o How do you create that mail buffer?  By typing C-x m, from
>   RMAIL, from Gnus, or else?

Either C-x m or by replying in RMAIL, the effect is the same.

> o What is the actual value of `buffer-file-coding-system'
>   when you don't setup that mail-mode-hook.

With automatic signature:

   buffer-file-coding-system's value is no-conversion
   Local in buffer *mail*; global value is iso-latin-1

Without:

   buffer-file-coding-system's value is iso-latin-1

> o What is the value of `mail-user-agent'?

   mail-user-agent's value is sendmail-user-agent

> o What is the value of `default-sendmail-coding-system'?

  default-sendmail-coding-system's value is iso-latin-1

Thanks for your help!

Konrad.
-- 
-------------------------------------------------------------------------------
Konrad Hinsen                            | E-Mail: hinsen@cnrs-orleans.fr
Centre de Biophysique Moleculaire (CNRS) | Tel.: +33-2.38.25.56.24
Rue Charles Sadron                       | Fax:  +33-2.38.63.15.17
45071 Orleans Cedex 2                    | Deutsch/Esperanto/English/
France                                   | Nederlands/Francais
-------------------------------------------------------------------------------

Re: Coding system for outgoing mail

[Kenichi Handa]

In article <200208220818.g7M8IuA02321@localhost.localdomain>, konrad.hinsen@laposte.net writes:
>>  o Can you reproduce the bug when you start Emacs with "-q
>>    --no-site-file"?

> No. I went through my .emacs line by line to find out what the
> cause is, and it turns out to be the line

>    (setq mail-signature t)

> Then I disabled that line and checked what happens when I insert
> the signature by hand (C-c C-w). Before inserting the signature,
> describe-coding-system says "none, use default". After inserting
> the signature, it says "no conversion (binary)".

> My signature file contains only ASCII characters (see below), so that
> cannot be the problem.

Thank you for tracking down the problem to here.  It's a
very strange phenomenon.  I tried the same thing with my
ASCII-only signagure file, but, of course, I can't reproduce
it.  Could you please send your signagure file by uuencoding
it?

---
Ken'ichi HANDA
handa@etl.go.jp

Re: Coding system for outgoing mail

[konrad.hinsen]

> ASCII-only signagure file, but, of course, I can't reproduce
> it.  Could you please send your signagure file by uuencoding
> it?

Sure:

begin 644 signature-file
M+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM
M+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+0I+;VYR860@2&EN
M<V5N("`@("`@("`@("`@("`@("`@("`@("`@("`@('P@12U-86EL.B!H:6YS
M96Y`8VYR<RUO<FQE86YS+F9R"D-E;G1R92!D92!":6]P:'ES:7%U92!-;VQE
M8W5L86ER92`H0TY24RD@?"!496PN.B`K,S,M,BXS."XR-2XU-BXR-`I2=64@
M0VAA<FQE<R!3861R;VX@("`@("`@("`@("`@("`@("`@("`@('P@1F%X.B`@
M*S,S+3(N,S@N-C,N,34N,3<*-#4P-S$@3W)L96%N<R!#961E>"`R("`@("`@
M("`@("`@("`@("`@("!\($1E=71S8V@O17-P97)A;G1O+T5N9VQI<V@O"D9R
M86YC92`@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@?"!.961E
M<FQA;F1S+T9R86YC86ES"BTM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM
M+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM+2TM
&+2TM+2T*
`
end

-- 
-------------------------------------------------------------------------------
Konrad Hinsen                            | E-Mail: hinsen@cnrs-orleans.fr
Centre de Biophysique Moleculaire (CNRS) | Tel.: +33-2.38.25.56.24
Rue Charles Sadron                       | Fax:  +33-2.38.63.15.17
45071 Orleans Cedex 2                    | Deutsch/Esperanto/English/
France                                   | Nederlands/Francais
-------------------------------------------------------------------------------

Re: Coding system for outgoing mail

[Kenichi Handa]

In article <200208220946.g7M9kGr28137@chinon.cnrs-orleans.fr>, konrad.hinsen@laposte.net writes:
>>  ASCII-only signagure file, but, of course, I can't reproduce
>>  it.  Could you please send your signagure file by uuencoding
>>  it?

> Sure:

> begin 644 signature-file

Thank you, but even with that file, I can't reproduce the bug.

Is "signature-file" the actual file name?

Please show me the value of these variables:
	file-coding-system-alist
	auto-coding-alist
	auto-coding-regexp-alist

---
Ken'ichi HANDA
handa@etl.go.jp

Re: Coding system for outgoing mail

[konrad.hinsen]

> Thank you, but even with that file, I can't reproduce the bug.
> 
> Is "signature-file" the actual file name?

No, it is ".signature", but I considered it safer not to send it under
that reserved name.

> Please show me the value of these variables:
> 	file-coding-system-alist

file-coding-system-alist's value is 
(("\\.po[tx]?\\'\\|\\.po\\." . po-find-file-coding-system)
 ("\\.elc\\'" emacs-mule . emacs-mule)
 ("\\(\\`\\|/\\)loaddefs.el\\'" raw-text . raw-text-unix)
 ("\\.tar\\'" no-conversion . no-conversion)
 ("" undecided))

> 	auto-coding-alist

auto-coding-alist's value is 
(("\\.\\(arc\\|zip\\|lzh\\|zoo\\|jar\\|tar\\|tgz\\)\\'" . no-conversion)
 ("\\.\\(gz\\|Z\\|bz\\|bz2\\|gpg\\)\\'" . no-conversion))

> 	auto-coding-regexp-alist

auto-coding-regexp-alist's value is 
(("^BABYL OPTIONS:[ 	]*-\\*-[ 	]*rmail[ 	]*-\\*-" . no-conversion))


--
-------------------------------------------------------------------------------
Konrad Hinsen                            | E-Mail: hinsen@cnrs-orleans.fr
Centre de Biophysique Moleculaire (CNRS) | Tel.: +33-2.38.25.56.24
Rue Charles Sadron                       | Fax:  +33-2.38.63.15.17
45071 Orleans Cedex 2                    | Deutsch/Esperanto/English/
France                                   | Nederlands/Francais
-------------------------------------------------------------------------------

Re: Coding system for outgoing mail

[Kenichi Handa]

In article <200208221941.g7MJfoo01752@localhost.localdomain>, konrad.hinsen@laposte.net writes:
>>  Thank you, but even with that file, I can't reproduce the bug.
>>  
>>  Is "signature-file" the actual file name?

> No, it is ".signature", but I considered it safer not to send it under
> that reserved name.

Thank you for considering that.  But, as file name sometimes
does affect how it is read, I wanted to know the exact file
name.

But, as far as seeing these values in your emacs:

>>  	file-coding-system-alist
>>  	auto-coding-alist
>>  	auto-coding-regexp-alist

I see no reason why inserting that file makes the buffer
unibyte.  I still can't reproduce it with setting those
values exactly as yours.

You wrote:
>>  o Can you reproduce the bug when you start Emacs with "-q
>>    --no-site-file"?

> No. I went through my .emacs line by line to find out what the
> cause is, and it turns out to be the line

>    (setq mail-signature t)

> Then I disabled that line and checked what happens when I insert
> the signature by hand (C-c C-w). Before inserting the signature,
> describe-coding-system says "none, use default". After inserting
> the signature, it says "no conversion (binary)".

What happens if you insert the signature file by this?
	C-x i ~/.signature RET

And what happens if you start emacs with -q --no-site-file
and type C-c C-w to insert the signature file?

Do they also change buffer-file-coding-system?

Do they also change buffer-file-coding-system when you
insert it not in *mail* buffer, but *scratch* buffer?

And, to which function C-c C-w is bound, and in which file
that function is defined?  In my case, it is
`mail-signature' defined in lisp/mail/sendmail.el.

---
Ken'ichi HANDA
handa@etl.go.jp

Re: Coding system for outgoing mail

[konrad.hinsen]

> > Then I disabled that line and checked what happens when I insert
> > the signature by hand (C-c C-w). Before inserting the signature,
> > describe-coding-system says "none, use default". After inserting
> > the signature, it says "no conversion (binary)".
> 
> What happens if you insert the signature file by this?
> 	C-x i ~/.signature RET

Same effect, the coding system is changed to "no conversion (binary)".

> And what happens if you start emacs with -q --no-site-file
> and type C-c C-w to insert the signature file?

Before inserting, there is no buffer-specific coding system. After
inserting, it is "undecided-unix".

The line in .emacs that causes the change from "undecided-unix"
to "no conversion (binary)" is

  (require 'crypt)

However, the file .signature is neither compressed nor encrypted in any
way, so crypt mode should not become active.

> Do they also change buffer-file-coding-system?

Yes.

> Do they also change buffer-file-coding-system when you
> insert it not in *mail* buffer, but *scratch* buffer?

Yes.

> And, to which function C-c C-w is bound, and in which file
> that function is defined?  In my case, it is
> `mail-signature' defined in lisp/mail/sendmail.el.

On my system it is the same.

Konrad.
--
-------------------------------------------------------------------------------
Konrad Hinsen                            | E-Mail: hinsen@cnrs-orleans.fr
Centre de Biophysique Moleculaire (CNRS) | Tel.: +33-2.38.25.56.24
Rue Charles Sadron                       | Fax:  +33-2.38.63.15.17
45071 Orleans Cedex 2                    | Deutsch/Esperanto/English/
France                                   | Nederlands/Francais
-------------------------------------------------------------------------------

Re: Coding system for outgoing mail

[Kenichi Handa]

I'm re-sending this because the previous one to
bug-gnu-emacs@gnu.org was returned as error somehow.

In article <200208260859.g7Q8xtB02375@localhost.localdomain>, konrad.hinsen@laposte.net writes:
> The line in .emacs that causes the change from "undecided-unix"
> to "no conversion (binary)" is

>   (require 'crypt)

Do you mean the current bug doesn't appear if you delete the
above line in your .emacs?  Then, we should check what crypt
does instead of collecting circumstantial evidence.

I've just downloaded the latest crypt package (Ver.2.91,
actually the file name is now crypt++.el, not crypt.el) and
loaded it, but I still can't reproduce the bug.

Which version are you using?  If you are not using the
latest version, please update it and try again.

As crypt++.el has this line:
	(add-hook 'find-file-hooks 'crypt-find-file-hook)
there's a possibility that the old version had a bug in
crypt-find-file-hook.

---
Ken'ichi HANDA
handa@etl.go.jp

Re: Coding system for outgoing mail

[konrad.hinsen]

> Do you mean the current bug doesn't appear if you delete the
> above line in your .emacs?  Then, we should check what crypt
> does instead of collecting circumstantial evidence.
> 
> I've just downloaded the latest crypt package (Ver.2.91,
> actually the file name is now crypt++.el, not crypt.el) and
> loaded it, but I still can't reproduce the bug.

Sorry for taking so long, my computer needed some repair.

I updated to crypt 2.91, and the problem seems to have disappeared.
So perhaps it was a bug in crypt, I was using version 1.3.
Thanks for your help!

Konrad.

Re: Coding system for outgoing mail

[Kenichi Handa]

In article <200209030819.g838J4u13500@localhost.localdomain>, konrad.hinsen@laposte.net writes:
> I updated to crypt 2.91, and the problem seems to have disappeared.
> So perhaps it was a bug in crypt, I was using version 1.3.

I see.   Thank you for confirming that.

---
Ken'ichi HANDA
handa@etl.go.jp