Rethink blocking annotations in detach-partition-concurrently-[34].
authorTom Lane <tgl@sss.pgh.pa.us>
Mon, 5 Jul 2021 18:34:47 +0000 (14:34 -0400)
committerTom Lane <tgl@sss.pgh.pa.us>
Mon, 5 Jul 2021 18:34:47 +0000 (14:34 -0400)
commitc04c767059b8460b99f6aa4aae5450ab3ee257a3
tree26417b42a867023361f52247571ede8e6600593f
parentf025f2390e13d7da69da595086fb982bbaf7f329
Rethink blocking annotations in detach-partition-concurrently-[34].

In 741d7f104, I tried to make the reports from canceled steps come out
after the pg_cancel_backend() steps, since that was the most common
ordering before.  However, that doesn't ensure that a canceled step
doesn't report even later, as shown in a recent failure on buildfarm
member idiacanthus.  Rather than complicating things even more with
additional annotations, let's just force the cancel's effect to be
reported first.  It's not *that* unnatural-looking.

Back-patch to v14 where these test cases appeared.

Report: https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=idiacanthus&dt=2021-07-02%2001%3A40%3A04
src/test/isolation/expected/detach-partition-concurrently-3.out
src/test/isolation/expected/detach-partition-concurrently-4.out
src/test/isolation/specs/detach-partition-concurrently-3.spec
src/test/isolation/specs/detach-partition-concurrently-4.spec