Re: [boost] Filesystem rooted paths make_absolute proposal

From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-09-20 12:59:15

At 12:35 PM 9/20/2002, Peter Dimov wrote:

 >From: "Beman Dawes" <bdawes_at_[hidden]>
 >> At 08:49 PM 9/19/2002, Glen Knowles wrote:
 >>
 >>  >I think of the windows c:, d: as being analogous to the host name of 
a
 >> URL.
 >>  >From my viewpoint these are all absolute paths:
 >>  >
 >>  >http://host.com/foo/bar
 >>  >http:/foo/bar
 >>  >c:/foo/bar
 >>  >/foo/bar
 >>  >
 >>  >The first two are both defined as absolute URLs, with and without an
 >>  >authority. I believe "absolute" and "drive" are orthogonal.
 >>
 >> If you go down that road, then you need another term to describe a path
 >> that has both.  Peter Dimov's [make_]absolute needs another name, 
because
 >> it has a postcondition that the returned path both "has_drive" and
 >> "has_root_directory".
 >
 >To be more precise, it has a postcondition "the returned path does not
 >depend on the current directory, drive, or other implementation defined
 >global state".

Yes, that's even better.

 >/foo/bar (under Windows) has different meaning depending on which drive 
is
 >current, and is therefore not "absolute" by the above definition. It is
 >absolute under POSIX. Of course if we adopt the other meaning of 
"absolute"
 >we need another term for self-contained paths.

"self_contained"?

--Beman