Re: nnmail-pathname-coding-system breaks my XEmacs.

[Katsumi Yamaoka <yamaoka@jpl.org>]

>>>>> Aidan Kehoe <kehoea@parhasard.net> wrote:
> I made this change yesterday to do that:

> http://hg.debian.org/hg/xemacs/xemacs?cs=774e5c7522bf
--8<---------------cut here---------------start------------->8---
--- a/lisp/mule/mule-cmds.el    Sun Jan 11 13:18:42 2009 +0000
+++ b/lisp/mule/mule-cmds.el    Tue Jan 13 12:07:27 2009 +0000
[...]
-      (define-coding-system-alias 'file-name 
[...]
+      ;; These variables have magic handlers to make setting them equivalent
+      ;; to setting the file-name, terminal and keyboard coding system
+      ;; aliases. See coding.el. 
+      (setq file-name-coding-system 
+            (or 
+             (let ((fncs (assq system-type system-type-file-name-coding)))
+               (and fncs (cdr fncs)))
+             native)
--8<---------------cut here---------------end--------------->8---

> This version of your macro works with versions of 21.5 both with and without
> that patch, and with 21.4:

> (defmacro nnmail-with-pathname-coding-system (&rest forms)
>   "Bind `file-name-coding-system' while running FORMS.
> A non-nil value of `nnmail-pathname-coding-system' will be used for
> decoding and encoding file names."
>   (if (featurep 'xemacs)
>       `(progn
>         (when (featurep 'file-coding)
>           ;; Work around a bug in many 21.5 betas:
>           (setq file-name-coding-system (coding-system-aliasee
>                                          'file-name)))
>         (let ((file-name-coding-system
>                nnmail-pathname-coding-system))
>           ,@forms))
>     `(let ((file-name-coding-system nnmail-pathname-coding-system))
>        ,@forms)))

IIUC `(setq file-name-coding-system ...)' here is for XEmacs 21.5
to which your patch has not been applied yet.  Is it necessary to
do that in advance whenever one binds `file-name-coding-system' to
something?  In other words, supposing `file-name' will not vary
after having been set to a certain value according to the system,
isn't it enough to do it once when XEmacs starts?  That is,

>>>>> In <b4mtz82duj7.fsf@jpl.org> Katsumi Yamaoka wrote:
> BTW, does this workaround do the trick?

> (if (featurep 'xemacs)
>     (if (featurep 'file-coding)
> 	(setq file-name-coding-system
> 	      (or file-name-coding-system
> 		  (coding-system-name 'file-name)))))

I planned to introduce such a wrapper, however there are many
places that need it in Gnus:

$ grep '(file-name-coding-system' *.el | wc -l
88

So, I'd like to use a much simpler way to fix the problem if
possible.  Note that the Gnus release (i.e. the Emacs 23 release)
is around the corner.

Regards,

P.S. I wonder why this is brought up now although binding of
pathname-coding-system (the predecessor of file-name-coding-system)
appeared first in Quassia Gnus v0.1 (1997). ;-)