Re: Re: [COMMITTERS] pgsql: Change the way that LWLocks for extensions are allocated.

Andres Freund <andres(at)anarazel(dot)de> writes:
> Committed after simplifying the Makefile.

While I have no particular objection to adding these tests, the
commit log's claim that this will improve buildfarm testing is
quite wrong. The buildfarm runs "make installcheck" not
"make check" in contrib. What I'm actually seeing in the
buildfarm reports is

make -C pg_stat_statements installcheck
make -C ../../src/test/regress pg_regress
make -C ../../../src/port all
make -C ../backend submake-errcodes
make[4]: Nothing to be done for `submake-errcodes'.
make -C ../../../src/common all
make -C ../backend submake-errcodes
make[4]: Nothing to be done for `submake-errcodes'.
../../src/test/regress/pg_regress --inputdir=. --bindir='/Users/buildfarm/bf-data/HEAD/inst/bin' --port=5678 --temp-config ../../contrib/pg_stat_statements/pg_stat_statements.conf --dbname=contrib_regression_pg_stat_statements
(using postmaster on Unix socket, port 5678)
============== dropping database "contrib_regression_pg_stat_statements" ==============
NOTICE: database "contrib_regression_pg_stat_statements" does not exist, skipping
DROP DATABASE
============== creating database "contrib_regression_pg_stat_statements" ==============
CREATE DATABASE
ALTER DATABASE
============== running regression test queries ==============

=====================
All 0 tests passed.
=====================

which is a rather blatant waste of cycles. I would suggest an explicit
do-nothing installcheck rule rather than the hack you came up with here.

regards, tom lane