Danek Duvall wrote: > On Tue, Jul 22, 2008 at 03:57:48PM -0700, Brock Pytlik wrote: > > >> http://cr.opensolaris.org/~bpytlik/ips-1753/ >> > > My main comment here is on the names. I'd rename baseline.py to be > pkg5testcase.py. "Baseline" is too tightly coupled to the meaning of the > known state of the tests. Likewise, BaselineTestCase should be called > Pkg5TestCase. And cli/testutils.py should rename Pkg5TestCase to > CLITestCase or something. > Ok > testutils.py: > > - line 523: should we warn that we failed to tear down the depots? > Trying once and silently giving up doesn't seem like the right way > forward here. > > I'm open to suggestions on what to do here. The problem is you might get exceptions from trying to remove non-existant files for example. The tearDown could be doing almost anything really and throwing any exceptions. Any errors we report there might make the output of the real test more confusing. If we try and fail (in the sense of leaving depots running) is there any expectation that trying again would succeed? Also, it won't fail silently unless it happens only as the last test of a set since any subsequent test will fail and complain a pkg.depotd instance is already running.
Maybe the right thing is to ignore the existence of tearDown, and simply kill the depots directly and not worry about the rest of tearDown's mechanisms since (I think) they won't affect subsequent tests. > t_setUp.py: > > - lines 46, 51: docstring needs changing > > Sorry about that. > Danek > _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
