setting default MIME charset not works with emacs 20.4

setting default MIME charset not works with emacs 20.4

[Vladimir Volovich]

Hi,

with emacs-20.3.6, i had

(set-language-environment "Cyrillic-KOI8")
;...
(coding-system-put
 (get-charset-property 'cyrillic-iso8859-5 'prefered-coding-system)
 'mime-charset 'koi8-r)

to set the default MIME charset for posting messages with cyrillic in
KOI8-R (instead of ISO-8859-5). This worked fine; but when i upgraded
to emacs 20.4, this no longer works (yes, i renamed
prefered-coding-system to preferred-coding-system): my messages are
sent in iso-8859-5 instead of koi8-r.

what should i do in emacs 20.4 to get things working? :-)

Thanks.

	Best regards, -- Vladimir.
-- 
Usage: fortune -P [] -a [xsz] [Q: [file]] [rKe9] -v6[+] dataspec ... inputdir

Re: setting default MIME charset not works with emacs 20.4

[Vladimir Volovich]

[-- Attachment #1: Type: text/plain, Size: 2678 bytes --]

"VV" == Vladimir Volovich writes:

 VV> Hi, with emacs-20.3.6, i had

 VV> (set-language-environment "Cyrillic-KOI8") ;...
 VV> (coding-system-put (get-charset-property 'cyrillic-iso8859-5
 VV> 'prefered-coding-system) 'mime-charset 'koi8-r)

 VV> to set the default MIME charset for posting messages with
 VV> cyrillic in KOI8-R (instead of ISO-8859-5). This worked fine; but
 VV> when i upgraded to emacs 20.4, this no longer works (yes, i
 VV> renamed prefered-coding-system to preferred-coding-system): my
 VV> messages are sent in iso-8859-5 instead of koi8-r.

 VV> what should i do in emacs 20.4 to get things working? :-)

(1) well, finally the following works:

(put-charset-property 'cyrillic-iso8859-5 'prefered-coding-system 'koi8-r)

i.e. instead of coding-system-put, use put-charset-property. i wonder,
why that change was needed, i.e. what changed between emacs-20.3.6 and
emacs-20.4?

(2) mm-util.el reads:

(defun mm-preferred-coding-system (charset)
  ;; A typo in some Emacs versions.
  (or (get-charset-property charset 'prefered-coding-system)
      (get-charset-property charset 'preffered-coding-system)))

in emacs-20.3, the name was prefered-coding-system, in emacs-20.4, it
was corrected: preferred-coding-system; but gnus does not check for
this name, but instead checks for preffered-coding-system, which seems
to be a typo in _gnus_, not in emacs (or is preffered-coding-system
really used somewhere?). [i.e. it seems that in mm-util.el, one should
either change preffered-coding-system to preferred-coding-system, or
add a check for preferred-coding-system].

(3) files in emacs/<version>/language/ define preferred mime charsets
    for coding systems, so gnus probably could extract that info from
    there instead of insisting to use such things as
    put-charset-property in .emacs file.

    i.e. (coding-system-get 'cyrillic-koi8 'mime-charset) returns
    koi8-r.

    Also, emacs/<version>/lisp/international/ contains some useful
    stuff for charsets, e.g. mule-cmds.el contains
    find-coding-systems-for-charsets which could be used in gnus.

(find-coding-systems-for-charsets (find-charset-string "test тест"))
(cyrillic-koi8 cyrillic-iso-8bit x-ctext iso-2022-7bit iso-2022-7bit-lock iso-2022-8bit-ss2 emacs-mule raw-text cyrillic-alternativnyj thai-tis620-with-esc tibetan-iso-8bit-with-esc lao-with-esc korean-iso-8bit-with-esc japanese-iso-8bit-with-esc in-is13194-devanagari-with-esc chinese-iso-8bit-with-esc iso-latin-5-with-esc hebrew-iso-8bit-with-esc greek-iso-8bit-with-esc cyrillic-iso-8bit-with-esc iso-latin-4-with-esc iso-latin-3-with-esc iso-latin-2-with-esc iso-latin-1-with-esc compound-text iso-2022-7bit-ss2)


[-- Attachment #2: Type: text/plain, Size: 563 bytes --]

(find-coding-systems-for-charsets (find-charset-string "test êãõëå"))
(iso-latin-1 iso-2022-jp-2 x-ctext iso-2022-7bit iso-2022-7bit-lock iso-2022-8bit-ss2 emacs-mule raw-text thai-tis620-with-esc tibetan-iso-8bit-with-esc lao-with-esc korean-iso-8bit-with-esc japanese-iso-8bit-with-esc in-is13194-devanagari-with-esc chinese-iso-8bit-with-esc iso-latin-5-with-esc hebrew-iso-8bit-with-esc greek-iso-8bit-with-esc cyrillic-iso-8bit-with-esc iso-latin-4-with-esc iso-latin-3-with-esc iso-latin-2-with-esc iso-latin-1-with-esc compound-text iso-2022-7bit-ss2)


[-- Attachment #3: Type: text/plain, Size: 66 bytes --]

(find-coding-systems-for-charsets (find-charset-string "test тест 

[-- Attachment #4: Type: text/plain, Size: 781 bytes --]

êãõëå"))
(x-ctext iso-2022-7bit iso-2022-7bit-lock iso-2022-8bit-ss2 emacs-mule raw-text thai-tis620-with-esc tibetan-iso-8bit-with-esc lao-with-esc korean-iso-8bit-with-esc japanese-iso-8bit-with-esc in-is13194-devanagari-with-esc chinese-iso-8bit-with-esc iso-latin-5-with-esc hebrew-iso-8bit-with-esc greek-iso-8bit-with-esc cyrillic-iso-8bit-with-esc iso-latin-4-with-esc iso-latin-3-with-esc iso-latin-2-with-esc iso-latin-1-with-esc compound-text iso-2022-7bit-ss2)

i.e. gnus should find-charset-string for the message part body, then
call find-coding-systems-for-charsets, and finally select the
appropriate charset from the returned list; it could then extract the
mime-charset to use for encoding that charset via coding-system-get.

	Best regards, -- Vladimir.

Re: setting default MIME charset not works with emacs 20.4

[Lars Magne Ingebrigtsen]

Vladimir Volovich <vvv@vvv.vsu.ru> writes:

> (defun mm-preferred-coding-system (charset)
>   ;; A typo in some Emacs versions.
>   (or (get-charset-property charset 'prefered-coding-system)
>       (get-charset-property charset 'preffered-coding-system)))

Yup.  Fix in Pterodactyl Gnus v0.97.

> (3) files in emacs/<version>/language/ define preferred mime charsets
>     for coding systems, so gnus probably could extract that info from
>     there instead of insisting to use such things as
>     put-charset-property in .emacs file.
> 
>     i.e. (coding-system-get 'cyrillic-koi8 'mime-charset) returns
>     koi8-r.

There is code in mm to use that stuff, and it seems to work for me.
(Although I don't write Cyrillic, so I haven't really tested. :-)
Under what circumstances does it fail?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: setting default MIME charset not works with emacs 20.4

[Vladimir Volovich]

"LMI" == Lars Magne Ingebrigtsen writes:

 >> (3) files in emacs/<version>/language/ define preferred mime
 >> charsets for coding systems, so gnus probably could extract that
 >> info from there instead of insisting to use such things as
 >> put-charset-property in .emacs file.
 >> 
 >> i.e. (coding-system-get 'cyrillic-koi8 'mime-charset) returns
 >> koi8-r.

 LMI> There is code in mm to use that stuff, and it seems to work for
 LMI> me.  (Although I don't write Cyrillic, so I haven't really
 LMI> tested. :-) Under what circumstances does it fail?

Well, i set

(set-language-environment "Cyrillic-KOI8")

and i expect that in this language environment gnus should prefer
koi8-r when sending cyrillic letters. but gnus still uses iso-8859-5.
so, it seems that gnus does not extract that info from coding system
properties (or elsewhere), to guess the preferred charset to use.

	Best regards, -- Vladimir.

Re: setting default MIME charset not works with emacs 20.4

[Lars Magne Ingebrigtsen]

Vladimir Volovich <vvv@vvv.vsu.ru> writes:

> Well, i set
> 
> (set-language-environment "Cyrillic-KOI8")
> 
> and i expect that in this language environment gnus should prefer
> koi8-r when sending cyrillic letters. but gnus still uses iso-8859-5.
> so, it seems that gnus does not extract that info from coding system
> properties (or elsewhere), to guess the preferred charset to use.

Hm.

What does

(mm-find-charset-region (point-min) (point-max))

and

(mm-find-mime-charset-region (point-min) (point-max))

evaluate to for you in a buffer with some cyrillic text?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: setting default MIME charset not works with emacs 20.4

[Vladimir Volovich]

"LMI" == Lars Magne Ingebrigtsen writes:

 >> Well, i set
 >> 
 >> (set-language-environment "Cyrillic-KOI8")
 >> 
 >> and i expect that in this language environment gnus should prefer
 >> koi8-r when sending cyrillic letters. but gnus still uses
 >> iso-8859-5.  so, it seems that gnus does not extract that info
 >> from coding system properties (or elsewhere), to guess the
 >> preferred charset to use.

 LMI> Hm.
 LMI> What does
 LMI> (mm-find-charset-region (point-min) (point-max))
 LMI> and
 LMI> (mm-find-mime-charset-region (point-min) (point-max))
 LMI> evaluate to for you in a buffer with some cyrillic text?

We have
(set-language-environment "Cyrillic-KOI8")

then, in the *mail* buffer with cyrillic letters,
(mm-find-charset-region (point-min) (point-max)) gives
(ascii cyrillic-iso8859-5)
(mm-find-mime-charset-region (point-min) (point-max)) gives
(iso-8859-5)

(note that current language environment is Cyrillic-KOI8, but mime
charset returned is iso-8859-5!)

After executing
(put-charset-property 'cyrillic-iso8859-5 'prefered-coding-system 'koi8-r)
we get the following results:

(mm-find-charset-region (point-min) (point-max)) gives
(ascii cyrillic-iso8859-5)
(mm-find-mime-charset-region (point-min) (point-max)) gives
(koi8-r)

So, i wonder, is gnus supposed to guess mime charset for cyrillic to
be koi8-r when the Cyrillic-KOI8 language environment is active, or i
should execute the above statement
(put-charset-property 'cyrillic-iso8859-5 'prefered-coding-system 'koi8-r)
in my .emacs?

	Best regards, -- Vladimir.

Re: setting default MIME charset not works with emacs 20.4

[Lars Magne Ingebrigtsen]

Vladimir Volovich <vvv@vvv.vsu.ru> writes:

> So, i wonder, is gnus supposed to guess mime charset for cyrillic to
> be koi8-r when the Cyrillic-KOI8 language environment is active, or i
> should execute the above statement
> (put-charset-property 'cyrillic-iso8859-5 'prefered-coding-system 'koi8-r)
> in my .emacs?

I thought the language-environment thingie was supposed to set the
preferred-coding-system property on the Mule charset?  Is this a bug,
I wonder?

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: setting default MIME charset not works with emacs 20.4

[Arcady Genkin]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> > So, i wonder, is gnus supposed to guess mime charset for cyrillic to
> > be koi8-r when the Cyrillic-KOI8 language environment is active, or i
> > should execute the above statement
> > (put-charset-property 'cyrillic-iso8859-5 'prefered-coding-system 'koi8-r)
> > in my .emacs?

FWIW, using Xemacs without Mule, I have to keep these in my .emacs:

(setq message-default-charset `koi8-r
      gnus-default-posting-charset `koi8-r)

-- 
Arcady Genkin
"It seems as though I had not drunk from the cup of wisdom, but had
fallen into it." (S. Kierkegaard)