Re: [bug-gnulib] ISSLASH on Woe32

[Paul Eggert]

Bruno Haible <address@hidden> writes:

> In approach (2) LIBDIR will be an UTF-8 encoded pathname. The ISSLASH
> operation will therefore work correctly. However, fopen() expects a
> string in locale encoding, not in UTF-8 encoding. Therefore we have
> to replace the last line with

OK, I'm starting to see the problem.  Yes, obviously that is
going to be a pain.

I agree with you that file names should just use one encoding.  If the
user wants an UTF-8 world, the user should specify all the file names
components in UTF-8, and then everything will work.  If the user wants
an EUC-JP world (not doable in Windows apparently, but the tradition
for Japanese Solaris) then all the file names should be specified in
EUC-JP.  The application shouldn't have to convert back and forth
internally: it should stick to just one encoding, and let the wrapper
functions (if any) deal with it.

So if I understand you correctly, yes, I'd favor (4) for gnulib code;
gnulib code shouldn't have to convert file-name pieces into a common
encoding.