03/27/2020, 7:47 PM
found a subtle ordering bug in
that explains some weird behavior Ive been frustrated with
BlockJUnit4ClassRunner expects instantiation to happen before testStarted, but the current impl of ^^ instantiates the test after that, which broke an assumption I was making
While I’m poking at this, would it be neat to have flaky tests that were retried report all their failures rather than just the 1st one?
03/27/2020, 9:08 PM
I'm bad to ask. I am ~philosophically opposed to automated flaky test re-runners.
03/28/2020, 1:28 AM
lol--maybe it should just go away?
or be turned off
fwiw, pants doesn’t set retries
03/28/2020, 1:54 AM
Aha! I'd love for it to be killed. Since it's dead code, regardless of philosophy about the functionality, it probably should be killed anyhow.
03/28/2020, 3:30 AM