On Mar 17, 2011, at 2:30 PM, exar...@twistedmatrix.com wrote: > On 12:40 pm, ita...@itamarst.org wrote: >> On Wed, 2011-03-16 at 20:55 -0600, pa...@wolfwood.twistedmatrix.com >> wrote: >>> + >>> + skippedReactors = ["Glib2Reactor", "Gtk2Reactor"] >>> + reactorClassName = reactor.__class__.__name__ >>> + if reactorClassName in skippedReactors and >>> platform.isWindows(): >>> + raise SkipTest( >>> + "This test is broken on gtk/glib under Windows.") >>> + >> >> Shouldn't that be a todo, rather than a skip? Or maybe file a ticket. > > A ticket would be nice, indeed. We are collecting quite a nice set of > gtk2-does-not-work-on-Windows issues in our test suite.
Ticket plus skip would be my preference. The terminology of 'skip' vs. 'todo' is slightly misleading. 'skip' really means 'this is a test that I know fails, and I want to fix it later', so it doesn't bother to run it. Whereas 'todo' means 'this is a test that sometimes fails and sometimes succeeds, I don't know what's going on with it, let's keep running it all the time to gather some data about how it fails differently over time'. In this case, we basically know what the problem is, and running the test on the buildbot in a broken configuration is likely to cause more problems than it solves by a long shot. Frankly, I am on the fence about the functionality of 'todo' being useful, ever. Mostly it seems to just slow down test suites for no particular gain. _______________________________________________ Twisted-Python mailing list Twisted-Python@twistedmatrix.com http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python
