mm-util.el should know as many charsets as iconv

mm-util.el should know as many charsets as iconv

[Dan Jacobson]

In the buffer *Messages* I have painstakingly traced the non "$0"
prefixed message "Unknown charset: windows-874" to the file
/usr/share/emacs/site-lisp/gnus/lisp/mm-util.el

And how unfair it is that windows-874 is unknown. Compare
$ iconv -l |grep 874|xargs
874// CP874// CSISO153GOST1976874// GOST_1976874// IBM874// WINDOWS-874//

So maybe if mm-util.el doesn't know a charset, perhaps it should
upcase the name, and pass the work over to iconv to give it a try.

OK, I did
(codepage-setup 874)(define-coding-system-alias 'windows-874 'cp874)
like the Info page said. The Info page should say why we must do this
by hand as it is not done automatically for us for all "windows-" sets.

$ man Encode::Supported says for example,
   All cp* are also available as ibm-*, ms-*, and windows-* .

(By the way, the characters shown were squares unless I used emacs -nw
or used echo in a shell buffer. Perhaps I don't have a bold Thai font
installed.)

Re: mm-util.el should know as many charsets as iconv

[Reiner Steib]

On Mon, Jun 19 2006, Dan Jacobson wrote:

> In the buffer *Messages* I have painstakingly traced the non "$0"
> prefixed message "Unknown charset: windows-874" to the file
> /usr/share/emacs/site-lisp/gnus/lisp/mm-util.el
>
> And how unfair it is that windows-874 is unknown. Compare
> $ iconv -l |grep 874|xargs
> 874// CP874// CSISO153GOST1976874// GOST_1976874// IBM874// WINDOWS-874//
>
> So maybe if mm-util.el doesn't know a charset, perhaps it should
> upcase the name, and pass the work over to iconv to give it a try.

mm-util.el doesn't "know" any charset.  The available charsets are
provided by (X)Emacs.

> OK, I did (codepage-setup 874)(define-coding-system-alias
> 'windows-874 'cp874) like the Info page said.

See also `mm-charset-eval-alist' (I didn't have time to document this
properly yet) and `mm-charset-synonym-alist'.

> The Info page should say why we must do this by hand as it is not
> done automatically for us for all "windows-" sets.

I recall some discussing on emacs-devel or emacs-pretest-bugs the not
all `cp...' charsets should be aliased to `windows-...'.  AFAICS,
`windows-874' is not IANA registered:
<http://www.iana.org/assignments/character-sets>.

> (By the way, the characters shown were squares unless I used emacs -nw
> or used echo in a shell buffer. Perhaps I don't have a bold Thai font
> installed.)

Probably.

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