Re: Race condition in InvalidateObsoleteReplicationSlots()

Hi,

On 2021-04-08 17:03:41 +0530, Amit Kapila wrote:
> I haven't tested the patch but I couldn't spot any problems while
> reading it. A minor point, don't we need to use
> ConditionVariableCancelSleep() at some point after doing
> ConditionVariableTimedSleep?

It's not really necessary - unless the CV could get deallocated as part
of dynamic shared memory or such.

Greetings,

Andres Freund