Non-encoded character in the subject (was: Emacs 21/22: french ç and œ)

Non-encoded character in the subject (was: Emacs 21/22: french ç and œ)

[Reiner Steib]

On Mon, Oct 27 2008, Uwe Brauer wrote:

>>>>>> "Miles" == Miles Bader <miles@gnu.org> writes:
>
>    > Hmm, why did gnus (message-mode) only encode the first non-ascii
>    > character in the Subject: header?  Odd...  Lemme try again...

The `œ' isn't encode, but sent as a raw iso-2022-7bit character
(character U+001B replaced with <ESC>):

| Subject: Re: Emacs 21/22:  french =?iso-8859-1?q?=E7?= and <ESC>$(D)M<ESC>(B

I cannot reproduce this, neither with Emacs 22.1 nor with current
Emacs trunk (2008-10-25):

ELISP> (let ((mm-coding-system-priorities '(iso-8859-1 iso-2022-jp-2 utf-8)))
	 (rfc2047-encode-string "french ç and œ oe"))
"french =?iso-8859-1?Q?=E7?= and =?iso-2022-jp-2?B?GyQoRClNGyhC?= oe"

ELISP> (let ((mm-coding-system-priorities '(iso-2022-jp-2 utf-8)))
	 (rfc2047-encode-string "french ç and œ oe"))
"french =?iso-2022-jp-2?B?GyQoRCsuGyhC?= and =?iso-2022-jp-2?B?GyQoRClNGyhC?=\n oe"
ELISP> (let ((mm-coding-system-priorities nil))
	 (rfc2047-encode-string "french ç and œ oe"))
"french =?utf-8?Q?=C3=A7?= and =?utf-8?Q?=C5=93?= oe"

> My mistake, I presume, 

No, your message was encoded correctly, AFAICS:

| Subject: Emacs 21/22:  french =?iso-8859-1?q?=E7?= and =?iso-8859-15?q?=BD?=

> but I will ask in the gnus list, I send this mail using xemacs 21.4
> and maybe the Mule support is flawed in gnus. (Or better said in
> xemacs 21.4)
>
> I should have used GNU 21/22 or Xemacs 21.5.
>
> Alas so many emacs(en) around.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Non-encoded character in the subject

[Katsumi Yamaoka]

>>>>> Reiner Steib <reinersteib+gmane@imap.cc> wrote:
> On Mon, Oct 27 2008, Uwe Brauer wrote:
>>>>>>> "Miles" == Miles Bader <miles@gnu.org> writes:
>>
>>> Hmm, why did gnus (message-mode) only encode the first non-ascii
>>> character in the Subject: header?  Odd...  Lemme try again...

> The `œ' isn't encode, but sent as a raw iso-2022-7bit character
> (character U+001B replaced with <ESC>):

>| Subject: Re: Emacs 21/22:  french =?iso-8859-1?q?=E7?= and <ESC>$(D)M<ESC>(B

> I cannot reproduce this, neither with Emacs 22.1 nor with current
> Emacs trunk (2008-10-25):

If the subject is encoded correctly when performing

C-u C-c C-m P

in the message buffer, the culprit is not Gnus in all likelihood.
It might be MTA's doing.  Though it doesn't come under the case,
Reiner's message I received from the MS Exchange POP server is
broken as follows:

Subject:
 =?Windows-1252?Q?Non-encoded_character_in_the_subject_(was:_Emacs_21/22:_?=
 =?Windows-1252?Q?_french_=E7_and_=C5=93)?=
i.e.,
Subject:
 Non-encoded character in the subject (was: Emacs 21/22:  french ç and Å“)

But no problem in the same message having reached Gmane.

Regards,

Re: Non-encoded character in the subject

[Reiner Steib]

On Tue, Oct 28 2008, Katsumi Yamaoka wrote:

> If the subject is encoded correctly when performing
> C-u C-c C-m P
> in the message buffer, the culprit is not Gnus in all likelihood.
> It might be MTA's doing.  Though it doesn't come under the case,
> Reiner's message I received from the MS Exchange POP server is
> broken as follows:
>
> Subject:
>  =?Windows-1252?Q?Non-encoded_character_in_the_subject_(was:_Emacs_21/22:_?=
>  =?Windows-1252?Q?_french_=E7_and_=C5=93)?=
> i.e.,
> Subject:
>  Non-encoded character in the subject (was: Emacs 21/22:  french ç and Å“)
>
> But no problem in the same message having reached Gmane.

Uh, this is the first time I hear that MTAs are changing the MIME
encoding of mail headers (I know about MTAs converting the
Content-Transfer-Encoding, e.g. qp <-> 8bit, of the body).

In fact, the article on Gmane has:

,----[ http://article.gmane.org/gmane.emacs.devel/105053/raw ]
| Subject: Non-encoded character in the subject (was: Emacs 21/22:  french
|  =?iso-8859-1?q?=E7?= and =?utf-8?b?xZMp?=
`----

My gcc-ed copy has:

,----[ Gcc ]
| Subject: Non-encoded character in the subject (was: Emacs 21/22:  french
|  =?iso-8859-1?Q?=E7?= and =?utf-8?Q?=C5=93=29?=
`----

Note the difference: binary vs. QP for the second encoded word.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Non-encoded character in the subject

[Katsumi Yamaoka]

>>>>> Reiner Steib <reinersteib+gmane@imap.cc> wrote:
> Uh, this is the first time I hear that MTAs are changing the MIME
> encoding of mail headers (I know about MTAs converting the
> Content-Transfer-Encoding, e.g. qp <-> 8bit, of the body).

The MS Exchange server changes not only mail headers but also
the MIME structure in a message body.  For instance, all inline
parts in a message body are turned into attachments.  Body
encoded with utf-8 and quoted-printable are turned into base64,
etc.  That's really boneheaded.  Sigh.

Re: Non-encoded character in the subject

[Miles Bader]

Katsumi Yamaoka <yamaoka@jpl.org> writes:
> The MS Exchange server changes not only mail headers but also
> the MIME structure in a message body.  For instance, all inline
> parts in a message body are turned into attachments.  Body
> encoded with utf-8 and quoted-printable are turned into base64,
> etc.  That's really boneheaded.  Sigh.

Or as my friend-who-works-at-microsoft says, "that's so MS."

-Miles

-- 
Opposition, n. In politics the party that prevents the Goverment from running
amok by hamstringing it.

Re: Non-encoded character in the subject

[Miles Bader]

Reiner Steib <reinersteib+gmane@imap.cc> writes:
> Uh, this is the first time I hear that MTAs are changing the MIME
> encoding of mail headers (I know about MTAs converting the
> Content-Transfer-Encoding, e.g. qp <-> 8bit, of the body).

My "original" (written by FCC: into a local file here) message has this:

   Subject: Re: Emacs 21/22:  french =?iso-8859-1?Q?=E7?= and
    =?iso-2022-jp-2?B?GyQoRClNGyhC?=

but the copy of that same message I get via NNTP from Gmane has this:

   Subject: Re: Emacs 21/22:  french =?iso-8859-1?q?=E7?= and ^[$(D)M^[(B

where I've replaced literal ESC characters with the two-character
sequence ^[ in this reply.

-Miles

-- 
Road, n. A strip of land along which one may pass from where it is too
tiresome to be to where it is futile to go.