On Tue, Mar 20, 2007 at 07:41:32AM +0100, Magnus Hagander wrote:
> > > Does it seem like I've overlooked anything obvious in this? I do get the
> > > feeling that this is too simple, but I don't know exactly where the
> > > problem is :-)
> >
> > I think you do still need the on_shmem_exit detach callback. Consider
> > the situation where the postmaster is trying to reinitialize after a
> > child crash. The Unix code is designed to detach and destroy the old
> > segment then create a new one. If that's not what you want to do then
> > this code still seems not right.
>
> Ok, will look into that. Haven't tested that scenario.
That was indeed so. Added in new version, attached.
> > There seem to be a lot of system calls not checked for failure here.
> > Do they really not have any failure possibilities?
I looked it over, and didn't find "a lot". I found one or two (which are
now fixed). Are you referring to anything in particular?
//Magnus