[PATCH 1 of 4 STABLE] i18n: use UTF-8-ed string to check use of characters reserved for Windows

[FUJIWARA Katsunori]

At Sun, 25 Dec 2011 15:56:58 +0900,
Shun-ichi Goto wrote:
> 
> 2011/12/25 FUJIWARA Katsunori <foozy at lares.dti.ne.jp>:
> >> Is there any problem with win32mbcs? > Katsunori
> >
> > This patch is for non-Windows platform, like cygwin.
> 
> win32mbcs is for platform using problematic encodings,
> not only for windows nor win32 native python.
> So cp932 issue of cygwin on Japanese WIndows could be
> fixed by win32mbcs,  I think.
> 
> > And sorry, this 1st patch causes problem with win32mbcs extension in
> > "encoding.fromlocal()".
> >
> > I'll post patches that (1) fixes win32mbcs for cygwin, and (2) folds
> > test into exising ones, so pleaese discard this series.
> 
> Installed python2.6 on my cygwin returns
> os.path.supports_unicode_filenames as False.
> We need to remove the check code in the beginning of
> extsetup() of win32mbcs.py to use in cygwin.
> 
> And it seems cygwin can handle unicode path as if the unicode path
> could be encoded into cp932.
> So I guess, cygwin can use win32mbcs almost as-is.
> Please try and test.

Thank you for your suggestion.

Current win32mbcs can not be enabled on cygwin as-is, because it wraps
Windows specific function "os.path.splitunc".

So, rewritten patch will separate this from others and wrap it only on
Windows platform.

----------------------------------------------------------------------
[FUJIWARA Katsunori]                             foozy at lares.dti.ne.jp