On Fri, Jul 22, 2022 at 01:18:34PM +1200, Thomas Munro wrote:
> I wonder if we should move the noise suppression check closer to
> pg_regress, so that it works also for the "main" pg_regress run, not
> only the one in this new TAP test. As discussed in this thread,
> inconclusively:
Yes, perhaps. We could reduce the amount of junk generated in
regression.diffs in a more centralized way with this approach, and
I agree that this should use restart_after_crash=off as well as
something to prevent more tests to run if we are not able to connect,
as you mentioned there. At least this would reduce the spam down to
tests running in parallel to the session that crashed.
> https://www.postgresql.org/message-id/flat/CA%2BhUKGL7hxqbadkto7e1FCOLQhuHg%3DwVn_PDZd6fDMbQrrZisA%40mail.gmail.com
Ah, I forgot about this recent thread. Let's just move the discussion
there. Thanks!
I am planning to remove this open item from the list and mark it as
"won't fix", as this is a much older issue.
--
Michael