Please have a look at Ivy master now. There are still a couple of cases where fail() is used: one, where a clean post-mortem is made impossible by mock listeners that include assertions which fail and pelt the test with assertion errors; second, where an assertion message is too expensive to construct beforehand.
Gintas 2017-12-04 18:00 GMT+01:00 Jan Matèrne (jhm) <apa...@materne.de>: > IMO catching an exception only makes sense if you check that exception > against certain assumptions or you expect that. > Otherwise the test runner will catch that (not excepted) and mark the test > as failed. > > If the current codebase differs from that, I would agree with simplifying > that. > > > Jan > > > -----Ursprüngliche Nachricht----- > > Von: Gintautas Grigelionis [mailto:g.grigelio...@gmail.com] > > Gesendet: Montag, 4. Dezember 2017 17:23 > > An: Ant Developers List > > Betreff: Unit tests in Ivy and Ant > > > > Hello, > > > > I decided to spend some time on unit tests in Ivy and Ant; I did some > > work on Ivy tests this summer and I feel that work is unfinished. > > > > In particular, there was a conversation about the use of fail(); my > > feeling is that it is somewhat superfluous: there are assumptions for > > (de)activating tests, assertions and expected exceptions, and that > > should suffice for defining a test. > > > > The way fail() is used in Ivy now mostly looks like a rethrow of > > unexpected exceptions to AssertionErrors (why?). > > > > So, in a nutshell: I would like to minimize try/catch and if/else in > > unit tests and make them work without fail(). What do you think? > > > > Gintas > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org > For additional commands, e-mail: dev-h...@ant.apache.org > >
