On Thu, Feb 9, 2023 at 5:17 PM Andres Freund <andres@anarazel.de> wrote:
> I've seen more failures than I'd like. Permission errors, conflicting names,
> and similar things. But mainly that was a reference to running initdb, which
> I've broken temporarily many a time.
We've all temporarily broken initdb literally thousands of times, I'm sure.
> > > Of course we could still name the suite tmp_install (or to be consistent with
> > > the confusing make naming, have a temp-install target, which creates the
> > > tmp_install directory :)). I guess that'd still be less confusing?
> >
> > Yes, that definitely seems like an improvement. I don't care about the
> > tiny inconsistency that this creates.
>
> Then lets do that - feel free to push something, or send something for
> review. Otherwise I'll try to get to it, but I owe a few people work before
> this...
I'll try to get to it soon. Note that I've been adding new stuff to
the meson Wiki page, in the hope of saving other people the trouble of
figuring some of these details out for themselves. You might want to
take a look at that at some point.
> > I wonder if this could be addressed by adding another custom test
> > setup, like --setup running, used whenever you want to just run one or
> > two tests against an ad-hoc temporary installation? Offhand it seems
> > as if add_test_setup() could support that requirement?
>
> What precisely do you mean with "ad-hoc temporary installation"?
I mean "what make check does".
> I was wondering about adding a different setup that'd use the "real"
> installation to run tests. But perhaps that's something different than what
> you have in mind?
I wasn't thinking about the installation. Actually, I was, but the
thought went no further than "I wish I didn't have to think about the
installation". I liked that autoconf had "make check" and "make
installcheck" variants that worked in a low context way.
It's great that "meson test" runs all of the tests very quickly --
that should be maintained, even at some cost elsewhere. And it would
be nice to do away with the tmp_install thing. But as long as we have
it, it would be nice to make the way that we run a subset of test
suites against a running server similar to the way that we run a
subset of test suites against a throwaway installation (ala "make
check").
> The only restriction I see wrt add_test_setup() is that it's not entirely
> trivial to use a "runtime-variable" path to an installation.
I personally have no problem with that, though of course I could have
easily overlooked something.
--
Peter Geoghegan