bug#21302: 25.0.50; buffer coding

[Eli Zaretskii]

> From: Daniel Wu <daniel.wu@cenx.com>
> Date: Wed, 19 Aug 2015 15:55:22 -0400
> 
> a long story on `buffer coding' when non-ascii characters are involved in
> a path.
> 
> it happens with cider-mode, and i don't know how to repeat this issue
> without using cider-ode, so please have a look at:
> http://github.com/clojure-emacs/cider/issues/1224

I'm sorry, but I see no coherent story there, at least not one that
can be understood by someone ignorant about cider-mode and clojure,
such as myself.  Is it possible to have the Emacs-related details
reported here?  For starters, how did that file name get into Emacs?
Was it typed by the user, reported by some subprocess, read by Emacs
from a file/buffer, something else?

Also, it is not clear to me on what OS the original problem happens.
Someone mentioned Windows there, someone else mentioned Linux, you are
reporting on Darwin...

> i've tried to report this issue to the maintainers of cider-mode, and
> they decided it's an emacs bug and recommended me to report to you.

I'm not quite sure I understand the details, but from what little I
could glean and guess, it looks like the offending file name is a
unibyte string, i.e. it wasn't decoded before being used.  If the file
name is received from some subprocess, then you need to decode it.

Also, people there are talking about "buffer coding" and
buffer-file-coding-system, but since this is a file name,
file-name-coding-system might be a more relevant variable.

But all this is based on guessing.  Without a clear description of
Emacs details related to this, I don't see how this could be solved,
or even attributed to Emacs core.