utf-translate-cjk-mode question

utf-translate-cjk-mode question

[Kai Großjohann]

I have some questions about utf-translate-cjk-mode -- I think it
might not be working, so I'm trying to understand it.  I probably
b0rked it when I changed it to use define-minor-mode.  It's rather
embarrassing :-/

In my ~/.emacs file, I have this piece of code:

(when (fboundp 'utf-translate-cjk-mode)
  (let ((current-language-environment "Chinese-GB"))
    (utf-translate-cjk-mode 1)))

I start Emacs.  I open a new file.  I say C-u C-\ chinese-py RET.  I
then type "nihao" into the buffer.  I say C-x RET f utf-8 RET.  Then
I try to save the buffer.

Emacs tells me that it can't encode the Chinese characters, and it
tried iso-latin-1 and utf-8 as encodings, and now it's suggesting
gb2312 as the encoding to use.

Is utf-translate-cjk-mode supposed to enable me to store the file as
utf-8?



The second question contains the code itself.  I tried to just copy
and paste it from the original defcustom.  But maybe I did something
wrong.  After loading all the subst-foo files, there is a let
statement that binds the symbol `table'.  But after this, nothing is
done with that symbol -- it does not appear to be accessed after
being populated.  I also looked in version 1.27 (the last version
before my change), and there, too, the symbol `table' does not appear
to be accessed.

What is going on here?  Is there some kind of magic?

I have the nagging feeling that the `table' needs to be passed to
define-translation-table somehow, but I can't find the code that used
to do that.

-- 
file-error; Data: (Opening input file no such file or directory ~/.signature)

Re: utf-translate-cjk-mode question

[Kai Großjohann]

kai.grossjohann@gmx.net (Kai Großjohann) writes:

> In my ~/.emacs file, I have this piece of code:
>
> (when (fboundp 'utf-translate-cjk-mode)
>   (let ((current-language-environment "Chinese-GB"))
>     (utf-translate-cjk-mode 1)))
>
> I start Emacs.  I open a new file.  I say C-u C-\ chinese-py RET.  I
> then type "nihao" into the buffer.  I say C-x RET f utf-8 RET.  Then
> I try to save the buffer.
>
> Emacs tells me that it can't encode the Chinese characters, and it
> tried iso-latin-1 and utf-8 as encodings, and now it's suggesting
> gb2312 as the encoding to use.

I've now tried (encode-coding-region (point-min) (point-max) 'utf-8)
on that buffer.  This produces \344\275\240\345\245\275 which I guess
might be right (but I didn't check).

So why can Emacs encode the buffer into UTF-8 but can't save the
buffer in UTF-8 encoding?

coding-category-list contains coding-category-utf-8 as the last
element.

I'm quite confused now.
-- 
file-error; Data: (Opening input file no such file or directory ~/.signature)

Re: utf-translate-cjk-mode question

[Kenichi Handa]

Sorry for the late response on this matter.

In article <84adej4tmj.fsf@lucy.is.informatik.uni-duisburg.de>, kai.grossjohann@gmx.net (Kai Großjohann) writes:
> In my ~/.emacs file, I have this piece of code:

> (when (fboundp 'utf-translate-cjk-mode)
>   (let ((current-language-environment "Chinese-GB"))
>     (utf-translate-cjk-mode 1)))

> I start Emacs.  I open a new file.  I say C-u C-\ chinese-py RET.  I
> then type "nihao" into the buffer.  I say C-x RET f utf-8 RET.  Then
> I try to save the buffer.

> Emacs tells me that it can't encode the Chinese characters, and it
> tried iso-latin-1 and utf-8 as encodings, and now it's suggesting
> gb2312 as the encoding to use.

> Is utf-translate-cjk-mode supposed to enable me to store the file as
> utf-8?

Yes, and this bug is my fault.  My previous change to
find-coding-systems-region-internal was incomplete.  I've
just installed a fix.  Please try again with the latest
HEAD.

> The second question contains the code itself.  I tried to just copy
> and paste it from the original defcustom.  But maybe I did something
> wrong.  After loading all the subst-foo files, there is a let
> statement that binds the symbol `table'.  But after this, nothing is
> done with that symbol -- it does not appear to be accessed after
> being populated.  I also looked in version 1.27 (the last version
> before my change), and there, too, the symbol `table' does not appear
> to be accessed.

This is also my fault.  When I made register-char-codings
obsolete, I forgot to delete this unnecessary code.

---
Ken'ichi HANDA
handa@m17n.org

Re: utf-translate-cjk-mode question

[Kai Großjohann]

Kenichi Handa <handa@m17n.org> writes:

> In article <84adej4tmj.fsf@lucy.is.informatik.uni-duisburg.de>, kai.grossjohann@gmx.net (Kai Großjohann) writes:
>
>> Emacs tells me that it can't encode the Chinese characters, and it
>> tried iso-latin-1 and utf-8 as encodings, and now it's suggesting
>> gb2312 as the encoding to use.
>
>> Is utf-translate-cjk-mode supposed to enable me to store the file as
>> utf-8?
>
> Yes, and this bug is my fault.  My previous change to
> find-coding-systems-region-internal was incomplete.  I've
> just installed a fix.  Please try again with the latest
> HEAD.

I waited a bit until head would compile again.  This morning, it did,
and I tested, and it works!

Thanks a lot.

>> The second question contains the code itself.  I tried to just copy
>> and paste it from the original defcustom.  But maybe I did something
>> wrong.  After loading all the subst-foo files, there is a let
>> statement that binds the symbol `table'.  But after this, nothing is
>> done with that symbol -- it does not appear to be accessed after
>> being populated.  I also looked in version 1.27 (the last version
>> before my change), and there, too, the symbol `table' does not appear
>> to be accessed.
>
> This is also my fault.  When I made register-char-codings
> obsolete, I forgot to delete this unnecessary code.

Ah, it's gone now.

And it works to turn utf-translate-cjk-mode on, and it also works to
turn it off.  That's good.

I'm sorry to cause you so much work; I wish I grokked the Mule code
better to fix it myself.
-- 
file-error; Data: (Opening input file no such file or directory ~/.signature)