Hello, Tom.
> Since we're running out of time in the current commitfest,
> I went ahead and changed that, and made the cosmetic fixes
> I wanted, and pushed.
Great, thanks!
The small thing I was thinking to add in KnownAssignedXidsCompress is
the assertion like
Assert(MyBackendType == B_STARTUP);
Just to make it more clear that locking is not the only thing required
for the call.
> I'd be willing to
> make the memory barrier change anyway, because that seems like
> a simple change that can't hurt.
I'm going to create a separate commit fest entry for it, ok?
Best regards,
Michail.