Thread: pg_regress.c also sensitive to various PG* environment variables

pg_regress.c also sensitive to various PG* environment variables

From
Michael Paquier
Date:
Hi all,

Following up with the recent thread that dealt with the same $subject
for the TAP tests, I have gone through pg_regress.c:
https://www.postgresql.org/message-id/YLbjjRpucIeZ78VQ@paquier.xyz

The list of environment variables that had better be reset when using
a temporary instance is very close to TestLib.pm, leading to the
attached.  Please note that that the list of unsetted parameters has
been reorganized to be consistent with the TAP tests, and that I have
added comments referring one and the other.

Thoughts?
--
Michael

Attachment

Re: pg_regress.c also sensitive to various PG* environment variables

From
Alvaro Herrera
Date:
On 2021-Jun-11, Michael Paquier wrote:

> Following up with the recent thread that dealt with the same $subject
> for the TAP tests, I have gone through pg_regress.c:
> https://www.postgresql.org/message-id/YLbjjRpucIeZ78VQ@paquier.xyz

Good idea.

> The list of environment variables that had better be reset when using
> a temporary instance is very close to TestLib.pm, leading to the
> attached.  Please note that that the list of unsetted parameters has
> been reorganized to be consistent with the TAP tests, and that I have
> added comments referring one and the other.
> 
> Thoughts?

I think if they're to be kept in sync, then the exceptions should be
noted.  I mean, where PGCLIENTENCODING would otherwise be, I'd add
/* PGCLIENTENCODING set above */
/* See below for PGHOSTADDR */
and so on (PGHOST and PGPORT probably don't need this because they're
immediately below; not sure; but I would put them in alphabetical order
in both lists for sure and then that wouldn't apply).  Otherwise I would
think that it's an omission and would set to fix it.

-- 
Álvaro Herrera                            39°49'30"S 73°17'W



Re: pg_regress.c also sensitive to various PG* environment variables

From
Michael Paquier
Date:
On Fri, Jun 11, 2021 at 10:08:20AM -0400, Alvaro Herrera wrote:
> I think if they're to be kept in sync, then the exceptions should be
> noted.  I mean, where PGCLIENTENCODING would otherwise be, I'd add
> /* PGCLIENTENCODING set above */
> /* See below for PGHOSTADDR */
> and so on (PGHOST and PGPORT probably don't need this because they're
> immediately below; not sure; but I would put them in alphabetical order
> in both lists for sure and then that wouldn't apply).  Otherwise I would
> think that it's an omission and would set to fix it.

Good idea, thanks.  I'll add comments for each one that cannot be
unsetted.
--
Michael

Attachment

Re: pg_regress.c also sensitive to various PG* environment variables

From
Michael Paquier
Date:
On Sat, Jun 12, 2021 at 09:10:06AM +0900, Michael Paquier wrote:
> Good idea, thanks.  I'll add comments for each one that cannot be
> unsetted.

And done, finally.
--
Michael

Attachment