On 26 August 2011 17:19, Camillo Bruni <camillo.bruni@inria.fr> wrote:
well the problem with a non deterministic test is that the UI might not correspond to the opened debugger :)
I don't understand. #debug: simply runs the test without wrapping them into exception handlers, so it will just open a debugger on the spot where the problem happens. So if you test fails once in a while you should be able to reproduce it with 1000 timesRepeat: [ MyTest debug: #testMe ] and be exactly in the execution where the problem happend.
another application I see here is for instance when running tests in CI and you see only a couple of tests failing, rerun them again, save the image. when you open the tests locally you can directly see where it failed on the server (e.g. very annoying with platform dependent stuff) And I personally would prefer this as default behavior, even though it might not fully work on certain tests (so I will add it as an option). Ah, and long running tests which fail, are debuggable instantaneously ;)
Yeah maybe, but then again you can just run the tests with #debug: and leave the debuggers open in the image, no? Lukas -- Lukas Renggli www.lukas-renggli.ch