Re: PXGS vs TAP tests

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Andrew Dunstan <andrew(at)dunslane(dot)net>
Cc: Noah Misch <noah(at)leadboat(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: PXGS vs TAP tests
Date: 2021-06-20 14:45:21
Message-ID: 1478865.1624200321@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> The recipe for running TAP tests in src/Makefile.global doesn't work for
> the PGXS case. If you try it you get something like this:
> ...
> Notice those bogus settings for top_builddir, PG_REGRESS and
> REGRESS_SHLIB. The attached patch fixes this bug.

OK, but does the 'prove_check' macro need similar adjustments?

> I removed the REGRESS_SHLIB setting altogether for the PGXS case - it's
> not clear to me why we need it in a TAP test recipe at all.

After some digging in the git history, it looks like it's there because
of Noah's c09850992, which makes me wonder whether 017_shm.pl requires
it. If so, it'd make more sense perhaps for that one test script
to set up the environment variable than to have it cluttering every TAP
run.

(In any case, please don't push this till after beta2 is tagged.
We don't need possible test instability right now.)

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Andrew Dunstan 2021-06-20 14:56:40 Re: PXGS vs TAP tests
Previous Message Andrew Dunstan 2021-06-20 14:15:55 Re: pgbench logging broken by time logic changes