message-idna problem

message-idna problem

[Katsumi Yamaoka]

Hi,

When I send a mail containing the From header like

From: 山岡 克美 <yamaoka@jpl.org>

the *Messages* buffer collects the following warnings:

Unknown symbol: 山
Unknown symbol: 岡
Unknown symbol: 克
Unknown symbol: 美

I've never noticed this since I rarely use Japanese names in the
message header.  It arises if `message-use-idna' is non-nil and
there is a name of which the first letter is a non-ASCII character
in the header.

Those warnings are the ones that `ietf-drums-parse-addresses'
(a.k.a. `mail-header-parse-addresses') that
`message-idna-to-ascii-rhs-1' uses issues.  Assuming there will
potentially be non-ASCII names when `message-idna-to-ascii-rhs-1'
runs, how about using `mail-extract-address-components' rather
than an ietf-drums- function that does not support non-ASCII
characters?  This problem might not cause a real harm, though.

--- message.el~	2008-09-24 21:41:53 +0000
+++ message.el	2008-09-25 10:52:07 +0000
@@ -5633,7 +5633,8 @@
 		(mapcar (lambda (rhs) (or (cadr (split-string rhs "@")) ""))
 			(mapcar 'downcase
 				(mapcar
-				 'car (mail-header-parse-addresses field))))))
+				 'cadr
+				 (mail-extract-address-components field t))))))
 	(setq ace (if (string-match "\\`[[:ascii:]]+\\'" rhs)
 		      rhs
 		    (downcase (idna-to-ascii rhs))))

Regards,

Re: message-idna problem

[Reiner Steib]

On Thu, Sep 25 2008, Katsumi Yamaoka wrote:

> When I send a mail containing the From header like
>
> From: 山岡 克美 <yamaoka@jpl.org>
>
> the *Messages* buffer collects the following warnings:
>
> Unknown symbol: 山
> Unknown symbol: 岡
> Unknown symbol: 克
> Unknown symbol: 美
>
> I've never noticed this since I rarely use Japanese names in the
> message header.  It arises if `message-use-idna' is non-nil and
> there is a name of which the first letter is a non-ASCII character
> in the header.

I get "Unknown symbol: Ä" independent on the value of
`message-use-idna':

ELISP> (let ((message-use-idna nil))
	 (mail-header-parse-addresses "Äpfel Birne <foo@bar.invalid>"))
(("foo@bar.invalid" . "pfel Birne"))

ELISP> (let ((message-use-idna t))
	 (mail-header-parse-addresses "Äpfel Birne <foo@bar.invalid>"))
(("foo@bar.invalid" . "pfel Birne"))

> Those warnings are the ones that `ietf-drums-parse-addresses'
> (a.k.a. `mail-header-parse-addresses') that
> `message-idna-to-ascii-rhs-1' uses issues.  Assuming there will
> potentially be non-ASCII names when `message-idna-to-ascii-rhs-1'
> runs, how about using `mail-extract-address-components' rather
> than an ietf-drums- function that does not support non-ASCII
> characters?  This problem might not cause a real harm, though.

Fine with me.

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

Re: message-idna problem

[Katsumi Yamaoka]

>>>>> Reiner Steib <reinersteib+gmane@imap.cc> wrote:

> On Thu, Sep 25 2008, Katsumi Yamaoka wrote:

>> When I send a mail containing the From header like
>>
>> From: 山岡 克美 <yamaoka@jpl.org>
>>
>> the *Messages* buffer collects the following warnings:
>>
>> Unknown symbol: 山
>> Unknown symbol: 岡
>> Unknown symbol: 克
>> Unknown symbol: 美
>>
>> I've never noticed this since I rarely use Japanese names in the
>> message header.  It arises if `message-use-idna' is non-nil and
>> there is a name of which the first letter is a non-ASCII character
>> in the header.

> I get "Unknown symbol: Ä" independent on the value of
> `message-use-idna':

ELISP> (let ((message-use-idna nil))
>          (mail-header-parse-addresses "Äpfel Birne <foo@bar.invalid>"))
> (("foo@bar.invalid" . "pfel Birne"))

ELISP> (let ((message-use-idna t))
>          (mail-header-parse-addresses "Äpfel Birne <foo@bar.invalid>"))
> (("foo@bar.invalid" . "pfel Birne"))

>> Those warnings are the ones that `ietf-drums-parse-addresses'
>> (a.k.a. `mail-header-parse-addresses') that
>> `message-idna-to-ascii-rhs-1' uses issues.  Assuming there will
>> potentially be non-ASCII names when `message-idna-to-ascii-rhs-1'
>> runs, how about using `mail-extract-address-components' rather
>> than an ietf-drums- function that does not support non-ASCII
>> characters?  This problem might not cause a real harm, though.

> Fine with me.

Thanks.  I've modified `message-idna-to-ascii-rhs-1' so as to
use `mail-extract-address-components' rather than
`mail-header-parse-addresses'.

Regards,