Re: Replacing TAP test planning with done_testing() - Mailing list pgsql-hackers

From Daniel Gustafsson
Subject Re: Replacing TAP test planning with done_testing()
Date
Msg-id 0097EBBB-6178-432C-AA8D-214778BEAAF1@yesql.se
Whole thread Raw
In response to Re: Replacing TAP test planning with done_testing()  (Michael Paquier <michael@paquier.xyz>)
Responses Re: Replacing TAP test planning with done_testing()  (Andres Freund <andres@anarazel.de>)
List pgsql-hackers
> On 10 Feb 2022, at 01:58, Michael Paquier <michael@paquier.xyz> wrote:
>>> Daniel Gustafsson <daniel@yesql.se> writes:

>>>> The attached patch removes all Test::More planning and instead ensures that all
>>>> tests conclude with a done_testing() call.

Pushed to master now with a few more additional hunks fixing test changes that
happened between posting this and now.

> Could it be possible to backpatch that even if
> this could be qualified as only cosmetic?  Each time a test is
> backpatched we need to tweak the number of tests planned, and that may
> change slightly depending on the branch dealt with.

I opted out of backpatching for now, to solicit more comments on that.  It's
not a bugfix, but it's also not affecting the compiled bits that we ship, so I
think there's a case to be made both for and against a backpatch.  Looking at
the oldest branch we support, it seems we've done roughly 25 changes to the
test plans in REL_10_STABLE over the years, so it's neither insignificant nor
an everyday activity.  Personally I don't have strong opinions, what do others
think?

--
Daniel Gustafsson        https://vmware.com/




pgsql-hackers by date:

Previous
From: Tomas Vondra
Date:
Subject: Re: postgres_fdw: using TABLESAMPLE to collect remote sample
Next
From: Andres Freund
Date:
Subject: Re: Replacing TAP test planning with done_testing()