Yeah that’s a pretty crappy situation for a new contributor.

You basically have to make some educated guesses. Do more tests fail on
average after your patch than before? Are the fails in tests that you added
or in tests that look related (test with backup in the name for example)?

If the fails doesn’t seem related to your changes, you pretty much leave it
on whoever might commit your code to use their deeper experience or luckier
dev environment to get your changes in regards to unrelated tests.

That pretty much dictates you want to run the tests a fair amount to help
you gauge. To do that reasonably, you need to crank up the parallel. Almost
all of the longer tests do very little actual work - so you can run a lot
of them together as long as you have the RAM for it. On any decent modern
hardware, you ought to be able to run them in at most 40 some minutes in
most cases. With good hardware and the right parralel test settings, you
should be able to do more like 10-20 min.

Reply via email to