Re: Non-ASCII characters in @include search path

[Eli Zaretskii]

> Date: Sun, 20 Feb 2022 13:02:23 +0100
> From: Patrice Dumas <pertusus@free.fr>
> 
> I just read some code in the XS parser, and it seems that the XS parser
> works with utf8 encoded byte strings.  Therefore the stat within the
> XS parser will work if the file names are actually encoded using utf8 in
> the file system.  This should be right for Linux/Unix, but for Windows,
> I think that it is not ok.

Right, on Windows passing a UTF-8 encoded file name to 'stat' will
fail (unless the file name is pure-ASCII).

> So I think that the best would be to encode to the encoding expected for
> file names, could be utf8 in the default case, but maybe something
> platform specific, if needed.

Yes, we can deduce that from the locale.