Re: kevent latch paths don't handle postmaster death well

Thomas Munro <thomas(dot)munro(at)gmail(dot)com> writes:
> I couldn't resist digging further into the Apple sources to figure out
> what was going on there, and I realised that the code path I was
> looking at can only report EACCES if you asked for NOTE_EXITSTATUS,
> which appears to be an Apple extension to the original FreeBSD kqueue
> system designed to let you receive the exit status of the monitored
> process. That is indeed much more privileged information, and it's
> only allowed for your own children.

Ah.

> So it's possible that commit
> 70516a17 was a waste of electrons, but I don't think it can hurt;

Yeah, I'm not inclined to revert it. If we did get that errno,
it'd be hard to interpret it in any way that didn't involve the
postmaster being gone.

regards, tom lane