https://pantsbuild.org/ logo
Powered by
#general
  • l

    limited-terabyte-34878

    10/20/2015, 4:21 PM
    we won’t notice
  • w

    witty-crayon-22786

    10/20/2015, 4:21 PM
    what bug
  • w

    witty-crayon-22786

    10/20/2015, 4:21 PM
    the log message is not a way to detect any bug
  • l

    limited-terabyte-34878

    10/20/2015, 4:21 PM
    the bug was that a test that takes 2 seconds to run will trigger the timeout but still take 2 seconds
  • l

    limited-terabyte-34878

    10/20/2015, 4:21 PM
    (say if the timeout is 1 second)
  • w

    witty-crayon-22786

    10/20/2015, 4:22 PM
    have an actual test for that, imo.
  • w

    witty-crayon-22786

    10/20/2015, 4:22 PM
    but if you also want to leave the message, then that's fine.
  • l

    limited-terabyte-34878

    10/20/2015, 4:23 PM
    how would i test for that? In the integration test i could time pants_run but there’s all the setup and such
  • l

    limited-terabyte-34878

    10/20/2015, 4:23 PM
    in the unittests all the timeouts are mocked, so they don’t take any time to run
  • l

    limited-terabyte-34878

    10/20/2015, 4:24 PM
    i could ensure that kill() gets called (which I think I’m doing, would need to double check) but it doesn’t ensure that the test process is actually killed and that the testing will stop.
  • w

    witty-crayon-22786

    10/20/2015, 4:24 PM
    make the bad actor's runtime something ridiculous
  • w

    witty-crayon-22786

    10/20/2015, 4:24 PM
    and then time it out after 1 second
  • w

    witty-crayon-22786

    10/20/2015, 4:24 PM
    totally fine for a unit test... zundel would agree that it is better to have this tested then not tested.
  • w

    witty-crayon-22786

    10/20/2015, 4:26 PM
    @limited-terabyte-34878: and actually, i think you'll want the integration tests to use more ridiculous times as well
  • l

    limited-terabyte-34878

    10/20/2015, 4:26 PM
    ok that makes sense
  • w

    witty-crayon-22786

    10/20/2015, 4:26 PM
    hundreds of seconds, with a timeout of 1
  • w

    witty-crayon-22786

    10/20/2015, 4:26 PM
    it's under testprojects, which we assume is full of wonky cases, and don't actually try to run directly
  • d

    dry-analyst-73584

    10/20/2015, 4:34 PM
  • a

    ancient-stone-50795

    10/20/2015, 4:47 PM
    @dry-analyst-73584: done.
  • d

    dry-analyst-73584

    10/20/2015, 4:47 PM
    Thanks!
  • w

    witty-crayon-22786

    10/20/2015, 4:52 PM
    @happy-kitchen-89482: sorry for the first bad diff on https://rbcommons.com/s/twitter/r/2991/ ... should be reviewable now though. last change adds a self-contained test of the incrementalism
  • h

    happy-kitchen-89482

    10/20/2015, 4:53 PM
    will TAL
  • w

    witty-crayon-22786

    10/20/2015, 4:54 PM
    thank you
  • h

    happy-kitchen-89482

    10/20/2015, 5:05 PM
    Who here has permissions for our org on bintray?
  • e

    enough-analyst-54434

    10/20/2015, 5:26 PM
    Eric and I - I can add you if you give me your bintray account name
  • c

    cuddly-architect-27013

    10/20/2015, 6:04 PM
    anyone with rights mind killing this: https://travis-ci.org/pantsbuild/pants/builds/86454808
  • c

    cuddly-architect-27013

    10/20/2015, 6:05 PM
    it’s doc and ws only changes
  • c

    cuddly-architect-27013

    10/20/2015, 6:05 PM
    seems not necessary to run the ci
  • c

    cuddly-architect-27013

    10/20/2015, 6:05 PM
    uuh, wait it’s failing
  • c

    cuddly-architect-27013

    10/20/2015, 6:05 PM
    lol, sorry need to fix something apparently
1...568569570...2420Latest