[Haskell-cafe] Re: ANN: System.FilePath 0.9

[Simon Marlow]

Neil Mitchell wrote:

> I have also added a canonicalPath function, support for spotting
> file\con as invalid and fixing it, support for \\?\ paths (if you
> don't know what they are, don't look it up, they are quite painful!)
> and a few very obscure corner cases which broke some of the
> properties.

I haven't been following this discussion very closely, but this caught my eye. 
Has anyone pointed out yet that eliminating ".." in a FilePath isn't valid in 
the presence of symbolic links?  I vaguely recall that this is why Krasimir's 
System.FilePath library doesn't include normalisation.  Normalisation that just 
changes separators, removes redundant separators and removes "." components is 
probably ok.

I don't really see the point of fullPathWith.  Isn't it just combine?

I think shortPath is on dodgy ground, given the difficulty with normalising 
FilePaths, and the fact that paths are not unique in the presence of symlinks.

The temporary file stuff is wrong - see System.IO.openTemporaryFile.  The only 
way to reliably create a temporary file is to open it at the same time, 
otherwise there's a race condition.

I have some other issues with naming, and the fact that the library mixes IO and 
non-IO functions.  We should avoid referring to $PATH as the "path", since we 
already have FilePath.  eg. spiltPath and getPath are talking about two very 
different things.  Where you use the term Filename, I think it should probably 
be FilePath.  I think we should consistently use a single term, preferably FilePath.

Cheers,
	Simon