[physfs] windows eccentricities...

Adam D. Moss adam at gimp.org
Tue May 3 20:34:17 EDT 2005

Miles Vignol wrote:
> is it within the scope of physfs to sort out the differences between 
> window's file system and more sane file systems?  for example, would a 
> standardized mapping from windows style file names 
> ("c:\folder\file.ext") to unix/osx ("/folder/file.ext") be worthy of 
> adding to physfs?  obviously, there would be some issues in how you'd 
> deal with drive letters, but by folding that into physfs, perhaps a 
> reasonable standard could be achieved.
> by incorporating such conversions, it would also make sense for 
> enumeration of "/" to return drive letters access points (whatever they 
> may convert to) to mirror unix's behavior of returning mounted file 
> systems.  this implies that converting from the windows filename "c:" 
> should probably be someting like "/driveC" or some other "top-level" 
> looking drive name.  or maybe "c:" isn't even really the top level... 
> windows is a bit convoluted in this regard.

Most of this has always been in PhysicsFS, as a side-effect of
PhysicsFS's API always using unix-like filenames and being able
to virtually mount directories as well as archives.

The remainder ('/driveC' sort of stuff) is also covered as of
recently, in the CVS version as a side-effect of mount point

Adam D. Moss   -   adam at gimp.org

More information about the physfs mailing list