I've seen test runners that report the number of assertions checked, not just the number of tests. Is this useful?
I understand if you have non-fatal assertions (like EXPECT_TRUE in googletest), but I think this is a rare feature in testing libraries.
miniblog.
Related Posts
7
Igalia's post on self-hosting CI runners is a masterclass in build engineering. It shows dealing with weird API constraints, but still finding good designs.
I also learnt that some CI systems use OCR on Apple's installers to automate!
Some test runners, such as MiniTest, randomise the order of tests to ensure tests are isolated. Neat idea, but I worry about repeatability.