Re: two utf-8 questions

[B. T. Raven]

"Stefan Monnier" <monnier@iro.umontreal.ca> wrote in message
87slr0kp0b.fsf-monnier+gnu.emacs.help@gnu.org">news:87slr0kp0b.fsf-monnier+gnu.emacs.help@gnu.org...
> > 1)
> > Even though the following is in my .emacs:
>
> >  (setq default-buffer-file-coding-system 'utf-8)
>
> > when I type 'C-x ret f' I see the prompt:
>
> > Coding system for visited file (default, nil) instead of (default,
utf-8)
>
> I think it never said anything else than nil.

Then it appears that there is no way to set the default (at least on my
w32 build, GNU Emacs 21.3.1 (i386-mingw-windows98.2222) of 2004-03-10 on
NYAUMO)

>
> > There are contexts where I need to specify utf-8 here, or else the
file
> > won't be saved in the correct format. This seems to be true even if
the
> > file header ;; -*- coding: utf-8 -*- is present. Why is this?
>
> This varies a lot between different versions of Emacs (because it's
still
> being continuously improved).  Also it depends on many more details.
> So it's hard to give a useful answer without knowing the version of
Emacs
> you're using, the locale under which you're running, how much you've
munged
> the default Mule setup, and then how you've loaded the file, what
operations
> you've done in between and which command gave you the above problem.
>
> As for obeying the `coding' tag when saving: this has only been added as
> a new feature recently in Emacs-CVS.  In earlier code, the `coding' tag
is
> only used when opening a file (and it's checked for consistency when
> saving).
>
>
>         Stefan

As far as I know, these are the only settings in my .emacs that might have
munged the default setup:

(set-language-environment               'UTF-8)
        (set-default-coding-systems             'utf-8)
        (setq file-name-coding-system           'utf-8)
        (setq default-buffer-file-coding-system 'utf-8)
        (setq coding-system-for-write           'utf-8)
        (set-keyboard-coding-system             'utf-8)
        (set-terminal-coding-system          'utf-8)
        (set-clipboard-coding-system            'iso-8859-1)
        (set-selection-coding-system            'iso-8859-1)
        (prefer-coding-system                   'utf-8)
        (modify-coding-system-alist 'process
"[cC][mM][dD][pP][rR][oO][xX][yY]" 'utf-8-dos)

and via custom:

 '(current-language-environment "UTF-8") ;; may be synonymous with
set-language-environment above
 '(unify-8859-on-encoding-mode t nil (ucs-tables))


Apparently the clipboard and selection coding systems affect the
locale-coding-system value which shows as iso-8859-1. Still, a buffer that
correctly shows unicode glyphs and which has -u in the mode line should be
saved with the utf-8 encoding. When I issue C-x C-s, C-x k, and then C-x
C-f with the same file name, the utf-8 encoding has been lost. I would be
satisfied with a setup which could handle only the utf-8 encoding if that
were possible by updating some of my *.el files without installing the
newer CVS binaries. But I suppose that's not advisable. Thanks for your
input anyway.

Ed