Re: PG_TEST_EXTRA and meson - Mailing list pgsql-hackers

From Jacob Champion
Subject Re: PG_TEST_EXTRA and meson
Date
Msg-id CAOYmi+=jx=MWP2KjhjvM4viw2p3yPmsHMCQAXa6YwD29SBhRiQ@mail.gmail.com
Whole thread Raw
In response to PG_TEST_EXTRA and meson  (Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>)
List pgsql-hackers
On Fri, Aug 23, 2024 at 5:59 AM Ashutosh Bapat
<ashutosh.bapat.oss@gmail.com> wrote:
> If I run
> export PG_TEST_EXTRA=xid_wraparound; ./configure --prefix=$BuildDir
> --enable-tap-tests && make -j4 && make -j4 install; unset
> PG_TEST_EXTRA
> followed by
> make -C $XID_MODULE_DIR check where
> XID_MODULE_DIR=src/test/modules/xid_wraparound - it skips the tests.

Right.

> I thought this was working before.

No, that goes back to my note in [1] -- as of c3382a3c3cc, the
variable was exported at only the src/test level, and I wanted to get
input on that so we could decide on the next approach if needed.

> Anyway, now I have written a script to test all the scenarios. You may
> want to test your patch using the script. It just needs PGDir to set
> to root directory of code.

Thanks! I see failures in 110, 120, and 130, as expected. Note that
constructions like

    PG_TEST_EXTRA="" cd $XID_MODULE_DIR && make check && cd $PGDir

will not override the environment variable for the make invocation,
just for the `cd`. Also, rather than

    export PG_TEST_EXTRA; ./configure ...; unset PG_TEST_EXTRA

it's probably easier to just pass PG_TEST_EXTRA=<setting> as a command
line option to configure.

> If there's some other way to setting PG_TEST_EXTRA when running
> configure, I think it needs to be documented.

./configure --help shows the new variable, with the same wording as
Meson. Or do you mean that it's significant enough to deserve a spot
in installation.sgml?

--Jacob

[1] https://www.postgresql.org/message-id/CAOYmi%2B%3D8HVgxANzFT_BZrAeDPxAgA5_kbHy-4VowdbGr0chHvQ%40mail.gmail.com



pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: Eager aggregation, take 3
Next
From: Alexander Korotkov
Date:
Subject: Re: POC, WIP: OR-clause support for indexes