Open wodow opened 10 years ago
Nice idea!
We should make this behaviour configurable, though. Least surprise and all that.
Any idea if any other tools do that? I bet they do.
Nose probably does. People keep telling me nose does everything that junorunner does, already. (Not 100% true, IIRC).
FWIW, nose wouldn't, I imagine, inherit/extend other DiscoverRunner implementations the way junorunner can.
In a large codebase, a breaking change can cause many tests to fail.
To help find a root cause, it might be useful to sort test failures heuristically:
(a) length of tests (number of non-comment lines?) - the simpler the test the more likely it covers the root cause specifically, or (b) depths/heights of traceback - shorter the call depth the closer to the root cause.