Chewing international (unicode) filenames on windows?

[Mark Hammond]

Andrew Markebo wrote:
> | > For me on my windows 2000 machine, to get them into UTF-8 format I do
> | > something like:
> | > for filename in os.listdir():
> | >         utffilename=unicode(filename, "latin-1").encode("utf-8")
> | 
> |           utffilename=unicode(filename, "mbcs").encode("utf-8")
> 
> Hello!
> 
> So do I get it right now, reading it as mbcs ("Multibyte-character")
> will read all ntfs filenames correctly (as unicode defines it) into a
> unicode string?

This will work as long as the filename can be represented in your 
current code page.  If not, you will get a '?'.  Other comments in this 
thread point to a proposal to use Unicode objects making every file 
available regardless of your codepage.

> | > p.s.2 what is the status on win9x? Give it up or ;-)
> | 
> | I *think* it still works when accessing a Unicode aware file system
> | (ie, network)
> 
> What I want is to read a local file-name on a win9x fat/fat32 disk and
> be able to convert it into unicode.. And the other way around.
> 
> I will probably add a if windows9x do that, else if win2k do
> this.. ;-) Hmm how do I determine windows-version?

win32api.GetVersion(Ex)

Mark.