Re: Remove backend warnings from SSL tests - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Remove backend warnings from SSL tests
Date
Msg-id 227642.1689774279@sss.pgh.pa.us
Whole thread Raw
In response to Re: Remove backend warnings from SSL tests  (Aleksander Alekseev <aleksander@timescale.com>)
Responses Re: Remove backend warnings from SSL tests
List pgsql-hackers
Aleksander Alekseev <aleksander@timescale.com> writes:
>> When looking at a patch in the CFBot I realized that the SSL tests generate
>> backend warnings under ENFORCE_REGRESSION_TEST_NAME_RESTRICTIONS

> Good catch. I can confirm that the patch corrects the named WARNINGs
> appearing with:
> CPPFLAGS="-DENFORCE_REGRESSION_TEST_NAME_RESTRICTIONS"
> There are plenty of similar warnings left however.

Yeah.  We have not worried about making TAP tests clean under this
restriction, because the point of it is to limit the hazards of
running "make installcheck" against a cluster containing useful data.
TAP tests always use one-off test clusters, so there is no hazard
to guard against.

If we wanted to extend the rule to TAP tests as well, I think we'd
have to upgrade the WARNING to an ERROR, because otherwise we'll
never find all the violations.  Not clear to me that it's worth
the trouble, though.  And it's definitely not worth the trouble to
fix only one TAP suite.

> Alternatively we could get rid of
> ENFORCE_REGRESSION_TEST_NAME_RESTRICTIONS entirely since its practical
> value seems to be debatable.

Strong -1 on that, for the reason given above.

            regards, tom lane



pgsql-hackers by date:

Previous
From: Dmitry Koval
Date:
Subject: Re: Add SPLIT PARTITION/MERGE PARTITIONS commands
Next
From: "Hayato Kuroda (Fujitsu)"
Date:
Subject: RE: [PoC] pg_upgrade: allow to upgrade publisher node