On Thu, Jun 21, 2018 at 4:00 PM, Tom Lane wrote:
> We could probably fix it by using a significantly larger test case,
> but that's not very attractive to put into the regression tests.
> Anybody have a better idea about how to improve this? Or even a
> clear explanation for what's causing it? (
I've been poking at why coverage.postgresql.org shows that the second
stanza in int8_avg_combine isn't exercised, when it clearly should be.
I can reproduce the problem here, so it's fairly robust. Eventually
it occurred to me to try a straight EXPLAIN ANALYZE VERBOSE, and what
I find is that with
