Re: ssl tests aren't concurrency safe due to get_free_port()

On 25.04.23 12:27, Peter Eisentraut wrote:
> These patches have affected pgxs-using extensions that have their own 
> TAP tests.
> 
> The portlock directory is created at
> 
>      my $build_dir = $ENV{top_builddir}
>        || $PostgreSQL::Test::Utils::tmp_check ;
>      $portdir ||= "$build_dir/portlock";
> 
> but for a pgxs user, top_builddir points into the installation tree, 
> specifically at $prefix/lib/pgxs/.
> 
> So when running "make installcheck" for an extension, we either won't 
> have write access to that directory, or if we do, then it's still not 
> good to write into the installation tree during a test suite.
> 
> A possible fix is
> 
> diff --git a/src/Makefile.global.in b/src/Makefile.global.in
> index 5dacc4d838..c493d1a60c 100644
> --- a/src/Makefile.global.in
> +++ b/src/Makefile.global.in
> @@ -464,7 +464,7 @@ rm -rf '$(CURDIR)'/tmp_check && \
>   $(MKDIR_P) '$(CURDIR)'/tmp_check && \
>   cd $(srcdir) && \
>      TESTDIR='$(CURDIR)' PATH="$(bindir):$(CURDIR):$$PATH" \
> -   PGPORT='6$(DEF_PGPORT)' top_builddir='$(top_builddir)' \
> +   PGPORT='6$(DEF_PGPORT)' top_builddir='$(CURDIR)' \
>      PG_REGRESS='$(top_builddir)/src/test/regress/pg_regress' \
>      $(PROVE) $(PG_PROVE_FLAGS) $(PROVE_FLAGS) $(if 
> $(PROVE_TESTS),$(PROVE_TESTS),t/*.pl)
>   endef

Any thoughts on this?  I would like to get this into the upcoming minor 
releases.

Note that the piece of code shown here is only applicable to PGXS, so 
given that that is currently broken, this "can't make it worse".