Win32 Non-ASCII Filename Access

Matz,

I have been working on stuff at work that involves non-ASCII filenames
on Windows, with differing character sets (such as "日本語" and 
"jalapeño"). Windows stores these filenames as UCS-2 entries on all
modern filesystems (FAT32 and NTFS).

The win32 directory and filename handling is using FindFirstFile
instead of FindFirstFileW; this means that it will never be possible
to handle certain filenames in Ruby.

I don't have a problem working on fixing this -- I have the knowledge
now to do so -- but the best fix that I can determine would produce
UTF-8 file names. Since Ruby strings are by default unencoded, how
would I handle this? Should I use $KCODE = 'u' and look to see what
the $KCODE is before handling this?

I'd like to fix this, as there are things that I could do with Ruby if
this were the case (it would also allow for filenames up to 32k wide
characters long, in chunks of 256 wide characters). Please advise on
the preferred way to handle this sort of thing.

I don't think it's as big an issue on Unix-based operating systems
except perhaps MacOS, because most other Unix OSes can't handle wide
character filenames and will therefore encode most international
filenames as UTF-8, if the environment setting is correct.

-austin
-- 
Austin Ziegler * [e-mail]
               * Alternate: [e-mail]