From: Jeff Garland (jeff_at_[hidden])
Date: 2004-02-16 19:40:59
On Mon, 16 Feb 2004 12:38:47 -0500, Beman Dawes wrote
> At 09:49 AM 2/16/2004, Jeff Garland wrote:
> >Also, I believe filesystem::path could and should be refactored to
> >wide strings on platforms that provide this capability. Obviously
> using >this interface would limit portability...
> My intent for filesystem::wpath (note the "w") is that it should
> work whenever possible on operating systems which use char as the
> external representation type. There would certainly be some
> limitations; the external encoding would have to be one that did not
> generate invalid names.
> For example, wpath on Linux would presumably use UTF-8 as the
> default external encoding, so that should work just fine. On Windows,
> the wpath internal and O/S external encoding are the same, so the
> same code would work on both Windows and Linux, and probably quite a
> few other O/S's too.
Sounds like a solid plan to me :-)
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk