Re: Continuing instability in insert-conflict-specconflict test

The test material added in commit 43e0841 continues to yield buildfarm
failures.  Failures new since the rest of this thread:

 damselfly    │ 2021-02-02 10:19:15 │ HEAD          │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=damselfly&dt=2021-02-02%2010%3A19%3A15
 drongo       │ 2021-02-05 01:13:10 │ REL_13_STABLE │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=drongo&dt=2021-02-05%2001%3A13%3A10
 lorikeet     │ 2021-03-05 21:30:13 │ HEAD          │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=lorikeet&dt=2021-03-05%2021%3A30%3A13
 lorikeet     │ 2021-03-16 08:28:36 │ HEAD          │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=lorikeet&dt=2021-03-16%2008%3A28%3A36
 macaque      │ 2021-03-21 10:14:52 │ REL_13_STABLE │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=macaque&dt=2021-03-21%2010%3A14%3A52
 walleye      │ 2021-03-25 05:00:44 │ REL_13_STABLE │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=walleye&dt=2021-03-25%2005%3A00%3A44
 sungazer     │ 2021-04-23 21:52:31 │ REL_13_STABLE │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=sungazer&dt=2021-04-23%2021%3A52%3A31
 gharial      │ 2021-04-30 06:08:36 │ REL_13_STABLE │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=gharial&dt=2021-04-30%2006%3A08%3A36
 walleye      │ 2021-05-05 17:00:41 │ REL_13_STABLE │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=walleye&dt=2021-05-05%2017%3A00%3A41
 gharial      │ 2021-05-05 22:35:33 │ REL_13_STABLE │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=gharial&dt=2021-05-05%2022%3A35%3A33

On Tue, Aug 25, 2020 at 12:04:41PM -0400, Tom Lane wrote:
> I think what we have to do to salvage this test is to get rid of the
> use of NOTICE outputs, and instead have the test functions insert
> log records into some table, which we can inspect after the fact
> to verify that things happened as we expect.

That sounds promising.  Are those messages important for observing server
bugs, or are they for debugging/modifying the test itself?  If the latter, one
could just change the messages to LOG.  Any of the above won't solve things
completely, because 3 of the 21 failures have diffs in the pg_locks output:

 dory         │ 2020-03-14 19:35:31 │ HEAD          │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=dory&dt=2020-03-14%2019%3A35%3A31
 walleye      │ 2021-03-25 05:00:44 │ REL_13_STABLE │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=walleye&dt=2021-03-25%2005%3A00%3A44
 walleye      │ 2021-05-05 17:00:41 │ REL_13_STABLE │
http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=walleye&dt=2021-05-05%2017%3A00%3A41

Perhaps the pg_locks query should poll until pg_locks has the expected rows.
Or else poll until all test sessions are waiting or idle.