Re: gnus should accept UTF8 even if UTF-8 is standard

[Katsumi Yamaoka <yamaoka@jpl.org>]

>>>>> In <E1LCJN1-0002Yy-00@quimby.gnus.org> (in gmane.emacs.gnus.cvs)
>>>>>	Reiner Steib wrote:
>     Date: Monday, December 15, 2008 @ 20:44:47
>   Author: cvs
>     Path: /usr/local/cvsroot/gnus/lisp

> Modified: ChangeLog mm-util.el

> (mm-charset-synonym-alist): Add bogus names "UTF8" and "ISO_8859-1".

This change potentially makes Gnus be a shameful newsreader that
sends messages with bogus charsets.  I must cope with it.

>>>>> In <b4mhc7dpf7w.fsf@jpl.org> Katsumi Yamaoka wrote:
> Type `C-c C-m p text/plain RET' in the message buffer and add
> charset=utf8 manually as follows:

> < #part type="text/plain" disposition=inline charset=utf8>

> In this case the message will be sent with charset=utf8 .

>>>>> In <873air44me.fsf@marauder.physik.uni-ulm.de>
>>>>>	Reiner Steib wrote:
>> I'm surprised that `mm-charset-synonym-alist' has any effect on
>> outgoing messages.  I though it would only be used when displaying an
>> article.  Could you please investigate this?

>>>>> In <b4m1vyaolxy.fsf@jpl.org> Katsumi Yamaoka wrote:
> That is due to `mml-generate-mime-1' that uses the charset specified
> as is (for encoding, it uses the valid coding system derived from
> that charset according to `mm-charset-synonym-alist', though).
> It's a bug, I have two solutions, and I like the second one:

> 1. Bind `mm-charset-synonym-alist' (and `mm-charset-eval-alist')
>    to nil when encoding.  And signal an error if the charset is
>    invalid.

> 2. Replace the charset that is an alias with the valid one that
>    Emacs knows.  Although it violates the idea that "the charset
>    aliases would be used only when displaying an article", it
>    will be convenience for users.

> WDYT?  The patches for 1. and 2. are below:

I've installed 2.  The code and the comment have been slightly
modified.

> Patch1:
> --- mml.el~	2008-10-03 05:47:11 +0000
> +++ mml.el	2008-10-20 23:58:29 +0000
> @@ -476,8 +476,11 @@
>  				 "application/octet-stream")
>  			   "text/plain")))
>  	       (charset (cdr (assq 'charset cont)))
> -	       (coding (mm-charset-to-coding-system charset))
> +	       (coding (let (mm-charset-eval-alist mm-charset-synonym-alist)
> +			 (mm-charset-to-coding-system charset)))
>  	       encoding flowed coded)
> +	  (unless coding
> +	    (error "Unknown charset: %s" charset))
>  	  (cond ((eq coding 'ascii)
>  		 (setq charset nil
>  		       coding nil))

> Patch2:
> --- mml.el~	2008-10-03 05:47:11 +0000
> +++ mml.el	2008-10-20 23:58:29 +0000
> @@ -482,7 +482,12 @@
>  		 (setq charset nil
>  		       coding nil))
>  		(charset
> -		 (setq charset (intern (downcase charset)))))
> +		 ;; `charset' might be an alias that `mm-charset-synonym-alist'
> +		 ;; provides and might not be in common use, so we prefer
> +		 ;; the one that Emacs knows for `coding'.
> +		 (setq charset (if coding
> +				   (mm-coding-system-to-mime-charset coding)
> +				 (intern (downcase charset))))))
>  	  (if (and (not raw)
>  		   (member (car (split-string type "/")) '("text" "message")))
>  	      (progn

Regards,