Re: Misformatted replies to UTF-8 messages

[ShengHuo ZHU]

Eli Zaretskii <address@hidden> writes:

> This is off the top of my head, before I had time to step through the 
> code.
>
> On Mon, 25 Feb 2002, ShengHuo ZHU wrote:
>
>> I received a bug report in Emacs 21.1.95 which is because
>> find-coding-systems-region prefer ctext-no-compositions to mule-utf-8,
>> e.g.  the following lines return ctext-no-compositions.
>> 
>>   (car (find-coding-systems-string
>>        (concat (char-to-string 2276) (char-to-string 2367))))
>
> Why is this regarded as a bug?  UTF-8 is generally very low on the 
> priority list of the possible encodings, so IIRC the user needs to invoke 
> prefer-coding-system to get UTF.  If she doesn't, almost any other 
> encoding will be preferred to UTF-8.

Because I saw it before in the Emacs main branch and later it returns
mule-utf-8.

> Is this some language environment that is already supposed to prefer 
> UTF-8?  If so, could the user who saw the problem please submit a full
> bug  report with the necessary info supplied by "M-x report-emacs-bug"?

I've asked the user to submit the bug report.

>> I think it has been fixed in the main branch. Could the fix be ported
>> to EMACS_21_1_RC?
>
> If the problem doesn't exist in CVS head, it's by sheer luck, because 
> IIRC the definition of ctext-no-compositions is identical in both 
> branches.

But the definition of utf-8 is changed, probably in Dave Love's
2001-12-18, -15, -07 changes.

ShengHuo