Re: Minimal logical decoding on standbys

On Thu, Apr 6, 2023 at 7:50 PM Drouvot, Bertrand
<bertranddrouvot(dot)pg(at)gmail(dot)com> wrote:
>
> Thanks! Will update 0005.
>

I noticed a few typos in the latest patches.

0004
1.
+ * Physical walsenders don't need to be wakon up during replay unless

Typo.

0005
2.
+# Check if all the slots on standby are dropped. These include the 'activeslot'
+# that was acquired by make_slot_active(), and the non-active 'inactiveslot'.
+sub check_slots_dropped
+{
+ my ($slot_user_handle) = @_;
+
+ is($node_standby->slot('inactiveslot')->{'slot_type'}, '',
'inactiveslot on standby dropped');
+ is($node_standby->slot('activeslot')->{'slot_type'}, '', 'activeslot
on standby dropped');
+
+ check_pg_recvlogical_stderr($slot_user_handle, "conflict with recovery");
+}
+
+# Check if all the slots on standby are dropped. These include the 'activeslot'
+# that was acquired by make_slot_active(), and the non-active 'inactiveslot'.
+sub change_hot_standby_feedback_and_wait_for_xmins
+{
+ my ($hsf, $invalidated) = @_;
+
+ $node_standby->append_conf('postgresql.conf',qq[
+ hot_standby_feedback = $hsf
+ ]);
+
+ $node_standby->reload;
+
+ if ($hsf && $invalidated)
+ {
...

The comment above change_hot_standby_feedback_and_wait_for_xmins seems
to be wrong. It seems to be copied from the previous test.

3.
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has
been updated
+# we now expect 3 conflicts reported as the counter persist across reloads
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 4) from
pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot updated') or die "Timed out waiting
confl_active_logicalslot to be updated";

The comment incorrectly mentions 3 conflicts whereas the query expects 4.

--
With Regards,
Amit Kapila.