bug#23595: 25.1.50; file with chinese/japanse chars, vc-diff fails (HG, Git, RCS)

[Eli Zaretskii]

> Cc: oub@mat.ucm.es, 23595@debbugs.gnu.org
> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Tue, 24 May 2016 23:19:05 -0700
> 
>     - Like Eli pointed out, (coding-system-get coding-system-for-read
>     :ascii-compatible-p) should work about as well. Why doesn't it?
> 
> It doesn't work for EBCDIC.

It's easy enough to exempt EBCDIC (and any other similar encodings).

>     - Shouldn't that change be in vc-coding-system-for-diff?
>     - It seems to try to fix a separate issue (whether all files use the same 
> coding
>     system).
> 
> Yes. For emacs-25 that's probably too much, as you suggest. So we can fix the 
> problem in vc-coding-system-for-diff. Revised (more-conservative) patch 
> attached. 

That patch relies on subtleties of comparing unibyte and multibyte
strings, so I don't like it, even if we ignore the issue of computing
something whose result is known in advance.

I still think we can come up with a simple (and safe for emacs-25)
method of identifying the problematic encodings, and leave the general
issue of having a new attribute for master.  Can we have a patch along
those lines, please?