Re: Some other odd buildfarm failures

Tom Lane wrote:
> Andres Freund <andres(at)2ndquadrant(dot)com> writes:
> > Can't we just move the test to run without parallelism? Its quite quick, so I don't it'd have noticeable consequences timewise.
>
> That just leaves the door open for somebody to add more tests parallel to
> it in future.

I've been long wanted to add declarative dependencies to tests: each
test file would declare what other tests it depends on, and we would
have a special clause to state "this one must not be run in concurrence
with anything else". Of course, this is just wishful thinking at this
point.

> TBH, I think we could have done without this test altogether; but if we're
> going to have it, a minimum expectation is that it not be hazardous to
> other tests around it.

The number of assertion failures in get_object_address without all the
sanity checks I added in pg_get_object_address was a bit surprising.
That's the whole reason I decided to add the test. I don't want to
blindly assume that all cases will work nicely in the future,
particularly as other object types are added.

--
Álvaro Herrera http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services