On Thu, Mar 05, 2009 at 08:19:05PM +0100, Magnus Hagander wrote:
> Peter Eisentraut wrote:
> > On Thursday 05 March 2009 18:04:42 Joshua Tolley wrote:
> >> As an aside, is access() adequately portable, ok to use within the
> >> backend, etc.? I just sort of took a shot in the dark.
> >
> > Using access() is usually not a good idea. In this case it would be better to
> > check the return of the actual open() call for EPERM (or the equivalent for
> > fopen(), whatever is used).
>
> That's what we do in the proper fix in HEAD. It requires an API change
> to backport it...
>
> Given that I think this is the first time we've heard of this issue, I'm
> thinking we should probably just not bother to backpatch it.
I'm inclined to agree, FWIW.
- Josh / eggyknap