[TIP] Comparing coverage between two runs of unit tests

francis francismb at email.de
Thu Dec 12 01:45:17 PST 2013


>> We have hundreds of program files in our system and sorting them by
>> coverage percentage is not going to help much.

IMHO is a good idea but the team has to have a process to look what's
happening with the top most ones.

>> I used to have daily builds that warn me about coverage regressions in
>> one particular work project.

>> People seemed to ignore them (myself included).  :-(

IMHO the same as above. While doing refactoring it could happen that some
coverage regressions happen (specially if you're making small steps) but
the team have to decide when the warning it's a failure (by maybe agreeing on
a max delta coverage percentage) to be forced to look at it or to write tests.


More information about the testing-in-python mailing list