Glenn Fowler writes:
> On Tue, 16 Jun 2009 15:59:42 -0400 James Carlson wrote:
> > Glenn Fowler writes:
> > The Solaris one does a modest check to see if the name is unused. But
> > since it doesn't hold onto that newly-created object, anyone else
> > could slip in a millisecond after mktemp(1) exit and create that same
> > object with nary a complaint observed.
>
> > I don't see now it's an architectural distinction.
>
> solaris does two extra calls { creat(2) or whatever and unlink(2) }
> either of those two calls could fail, causing mktemp(1) to fail
Arguably, the existence of a failure mode of any sort for -u is itself
just a bug.
> I don't like the -u option in either incarnation
> but the incarnations are distinct
I'm having a hard time seeing how the distinctions are actually
important, rather than being mere implementation artifacts.
--
James Carlson, Solaris Networking <james.d.carlson at sun.com>
Sun Microsystems / 35 Network Drive 71.232W Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
