pgsql: Add PGXS options to control TAP and isolation tests

On Sun, Nov 25, 2018 at 11:53:02PM +0000, Michael Paquier wrote:

Add PGXS options to control TAP and isolation tests

The following options are added for extensions:
- TAP_TESTS, to allow an extention to run TAP tests which are the ones
present in t/*.pl. A subset of tests can always be run with the
existing PROVE_TESTS for developers.
- ISOLATION, to define a list of isolation tests.
- ISOLATION_OPTS, to pass custom options to isolation_tester.

A couple of custom Makefile targets have been accumulated across the
tree to cover the lack of facility in PGXS for a couple of releases when
using those test suites, which are all now replaced with the new flags,
without reducing the test coverage. This also fixes an issue with
contrib/bloom/, which had a custom target to trigger its TAP tests of
its own not part of the main check runs.

Several buildfarm members complain about this commit, including dory and
longfin:
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=dory&dt=2018-11-26%2000%3A00%3A26
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=longfin&dt=2018-11-25%2023%3A59%3A03

The error comes from the TAP tests of contrib/bloom/, which we have
never run across the buildfarm until now:
# Failed test 'initial: query result matches'
# at t/001_wal.pl line 38.
# got: '0|d
# 0|9
# 0|3

I am not sure what to think of that yet, so as a short-term fix and to
keep the buildfarm green, I am going to disable the TAP tests for bloom,
but something is busted there, either in the code or in the tests.
--
Michael

On Mon, Nov 26, 2018 at 09:33:51AM +0900, Michael Paquier wrote:

Several buildfarm members complain about this commit, including dory and
longfin:
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=dory&dt=2018-11-26%2000%3A00%3A26
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=longfin&dt=2018-11-25%2023%3A59%3A03

The error comes from the TAP tests of contrib/bloom/, which we have
never run across the buildfarm until now:
# Failed test 'initial: query result matches'
# at t/001_wal.pl line 38.
# got: '0|d
# 0|9
# 0|3

I am not sure what to think of that yet, so as a short-term fix and to
keep the buildfarm green, I am going to disable the TAP tests for bloom,
but something is busted there, either in the code or in the tests.

I have been fighting with the buildfarm a bit this morning, but this is
proving to have a larger subset of issues than I anticipated first.
Here is the collection of my notes:
- MSVC scripts assume that REGRESS_OPTS can only use top_builddir. Some
test suites actually finish by using top_srcdir, like pg_stat_statements
which cause the regression tests to never run on Windows!
- Trying to enforce top_builddir does not work either when using VPATH
as this is not recognized properly.
- TAP tests of bloom are unstable on various platforms, causing various
failures.

As a result, I have reverted all the recent changes which introduced
those new PGXS options. I am going to spawn a set of thread to address
all the underlying issues first as those deserve more discussion. There
is a collection of tests not actually running on Windows.
--
Michael