Michael Paquier <michael.paquier@gmail.com> writes:
> And I just wrote a test case breaking my previous patch, please see
> attached. The idea is just to use ADD CONSTRAINT foo { PRIMARY KEY |
> UNIQUE } and check if default_tablespace works as expected. Perhaps
> that's worth having in the test suite? There is currently no coverage
> regarding for index additions done with ALTER TABLE and tablespaces.
Looks like a good idea, except that I reordered the tests so that
the stanza leaves default_tablespace back at its usual value, so
it won't affect later tests.
regards, tom lane