* Andrew Dunstan (andrew@dunslane.net) wrote:
> Are you really sure we can do this consistently? The regression
> tests have to run against all sorts of settings, including those we
> have no control over via installcheck.
Sure? No. However, there are quite a few existing regression tests
that do the same (with costs off, of course), and things on the
buildfarm look more-or-less clean again, so I'm at least hopeful.
I can simplify the tests some if they end up causing issues.
Thanks,
Stephen
