On Sat, Jul 31, 2010 at 1:56 PM, Alexandre Bergel <alexandre@bergel.eu>wrote:
Hi Laurent!
One thing with autotest. When I compile a method, even though it is not covered by the tests, the tests are run again. This does not always make sense I feel since it slow down the way I code...
See AutotestFinder>>methodsFor: . When no senders/test/setup/teardown found, it searches for tests which reference the class of the modifed method. That means it will run all tests related to the class, which may not cover the changed method. I've done this because you can have testMethod calls methodA calls methodB. If you modify methodB, there's no senders in tests. But I want to know if it's covered by tests actually. Or I don't want senders in tests because it's a private method. So when I modify methodB, Autotest run all tests which reference its class. In this case the "Hit count" info is really useful because it tells if the method is covered by tests or not. Is there another way to know if methodB is covered by tests or not ?
For example, when I am defining a meta-model, then running the test each time I add some business method it makes sense. However, when I define a mondrian visualization, all #visualize have to be excluded by autotest.
Is #visualize a slow method or there's too much senders of #visualize ? Is excluding long tests as we talked earlier can resolve it ? Can you send me your image and tell me which methods you change bother you ? So I can "feel" the problem too. Laurent
Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project