SP v4 alpha2 sessions ?

[Scott Cantor]

I've patched those bugs and I added a unit test that at least on a Mac so far does indicate open(CREAT|EXCL) leaves an empty file behind and that a subsequent attempt to open it that way fails. That was meant to be sufficient to prevent a race condition based on what the code actually does, not the theoretical pattern you're talking about.

The update logic, OTOH, is a mess. I doubt it's safe to use, and my fear is that the blasted happy eyeballs nonsense causing clients to just flip around between IP4 and IP6 is pretty well guaranteed to mess it up, but it at least only messes up a single session.

Thanks,
-- Scott