Re: [COMMITTERS] pgsql: Improve isolation tests infrastructure.

Andres Freund <andres@anarazel.de>

From: Andres Freund <andres@anarazel.de>
To: Peter Eisentraut <peter_e@gmx.net>
Cc: pgsql-hackers@postgresql.org, Andrew Dunstan <andrew@dunslane.net>
Date: 2017-03-15T17:40:03Z
Lists: pgsql-hackers
Hi Peter, All,

On 2017-03-14 23:10:19 +0000, Andres Freund wrote:
> Improve isolation tests infrastructure.

There's:
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=frogmouth&dt=2017-03-15%2012%3A32%3A45
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=narwhal&dt=2017-03-15%2005%3A00%3A01

I suspect that's because previously src/test/regress/GNUMakefile's
installcheck provided --bindir='$(bindir)', but
$pg_isolation_regress_installcheck) doesn't (and didn't before the
changes).

ISTM that the only reason that this previously didn't cause issues on
these machines is that neither actually *runs* any of the isolation
checks in other directories.  I.e. they'd have failed before this, too,
if they ran tests using pg_isolation_regress_installcheck.

To me it seems like an oversight in the previous definition of
pg_isolation_regress_installcheck.  Looks like that was introduced in
https://git.postgresql.org/gitweb/?p=postgresql.git;a=commit;h=dcae5faccab64776376d354decda0017c648bb53

Peter, I guess it wasn't intentional that you added --bindir to
pg_regress_installcheck but not pg_isolation_regress_installcheck?

Regards,

Andres


Commits

  1. Specify bindir in pg_isolation_regress_installcheck.

  2. Improve isolation tests infrastructure.

  3. Improve speed of make check-world