2 macos rust tests are failing for me in CI only. ...
# developmente
early-needle-54791
09/03/2019, 6:02 PM2 macos rust tests are failing for me in CI only. I can’t repro then locally, and I have checked that the rust / cargo versions are the same. Anyone have any ideas kindof at a loss
a
aloof-angle-91616
09/03/2019, 6:02 PMwhich tests?
e
early-needle-54791
09/03/2019, 6:03 PMsome speculation tests that do stuff with the delay futures.
a
aloof-angle-91616
09/03/2019, 6:03 PMlink?
e
early-needle-54791
09/03/2019, 6:03 PMearly-needle-54791
09/03/2019, 6:04 PMtraceback isn’t very helpful its just an assertion failure
a
aloof-angle-91616
09/03/2019, 6:05 PMoh
aloof-angle-91616
09/03/2019, 6:05 PMi suspect that's flakiness
aloof-angle-91616
09/03/2019, 6:05 PMlet me check the actual code
e
early-needle-54791
09/03/2019, 6:06 PMhm okay. I haven’t seen it flake ever locally,
a
aloof-angle-91616
09/03/2019, 6:06 PMok, but that's what happens in CI sometimes, especially with osx testing where flakiness seems to occur more often
e
early-needle-54791
09/03/2019, 6:07 PMalright I’ll leave it for now
a
aloof-angle-91616
09/03/2019, 6:07 PMit looks like
run_speculation_tests()Duration::from_millis(speculation_delay_ms)aloof-angle-91616
09/03/2019, 6:07 PMcould you increase the value of
speculation_delay_msaloof-angle-91616
09/03/2019, 6:08 PMi usually try something like making it 100 seconds to make absolutely sure
e
early-needle-54791
09/03/2019, 9:20 PMyeah that is a good idea. I tried to keep the delays and timeouts short so we don’t make the tests too slow, but maybe on the ci infrastructure the runtime does weird things?
a
aloof-angle-91616
09/10/2019, 10:30 PMi just meant temporarily for the purposes of checking flakiness! but yes, the ci infrastructure is drastically different from laptops.