Re: Non-ASCII characters in @include search path

[Gavin Smith]

On Sun, Feb 20, 2022 at 01:45:19PM +0100, Patrice Dumas wrote:
> On Sun, Feb 20, 2022 at 01:10:16PM +0100, Patrice Dumas wrote:
> > 
> > I think that the correct way to do that is to use
> > Encode::encode($text, 'utf-8');
> > Also I think that it should be done as late as possible, so it would be
> > better on $possible_file.

My thought was that the argument to -I could have been any sequence of bytes,
not necessarily correct UTF-8.  It would be wrong then to attempt any
encoding or decoding to a string formed from such an argument.

> 
> It is Encode::encode('utf-8', $text), but in my tests it does not work.
> Only utf8::downgrade() work, although it should probably not be the
> function used as it converts to some 8-bit encoding but we are not sure
> that it converts to utf-8.  I am puzzled.

I'd guess that you would get a kind of "double UTF-8" where the individual
bytes of the directory name each become their own UTF-8 codepoint.