#general
- b
bored-art-40741
03/31/2016, 6:42 PMI've also needed to xfail two tests that I'm confident are broken because of jdk8 stuff - b
bored-art-40741
03/31/2016, 6:42 PMDue to pants bugs, not jenkins issues - b
bored-art-40741
03/31/2016, 6:42 PMI'm going to put together three RBs today. One deletes haskell, one xfails those tests with explanations (I'd love to fix them, but it's a bit out of scope imo--the tests weren't being run at all before, so it's not a big deal), and one with the new lock. - b
bored-art-40741
03/31/2016, 6:43 PMThen I'm going to start working on a runbook for jenkins and checking in our job configurations. - b
bored-art-40741
03/31/2016, 6:43 PMAny high level objections to this process? - b
bored-art-40741
03/31/2016, 6:45 PMAs an aside, @ancient-stone-50795 and friends, I'm pretty confident the existing ivy lock is broken - b
bored-art-40741
03/31/2016, 6:45 PM - b
bored-art-40741
03/31/2016, 6:45 PMInstead of waiting for the lock, it seems to fail fast - b
bored-art-40741
03/31/2016, 6:50 PMI see this lock leak all the time, both on jenkins and my local machine - w
witty-crayon-22786
03/31/2016, 6:50 PMthat sounds fine. thanks. - b
bored-art-40741
03/31/2016, 6:50 PM:\ - b
bored-art-40741
03/31/2016, 6:51 PMI'm amazed no one else is seeing this - w
witty-crayon-22786
03/31/2016, 6:51 PM@bored-art-40741: still willing to ship a revert. - b
bored-art-40741
03/31/2016, 6:52 PMAt this point no one is going to be releasing before I get this new lock out, so let's just wait I suppose - a
ancient-stone-50795
03/31/2016, 6:56 PM@bored-art-40741: maybe we should have set a timeout on the lock to get the blocking behavior - b
bored-art-40741
03/31/2016, 6:56 PMReading the code, I'm confused by why it's not working. Passing timeout=None should block indefinitely, as we want - b
bored-art-40741
03/31/2016, 7:27 PMAny ideas on this failure? http://jenkins.pantsbuild.org/job/pants_ci.integration_tests/149/SHARD=5/console - w
witty-crayon-22786
03/31/2016, 7:30 PM@limited-terabyte-34878: ^ - w
witty-crayon-22786
03/31/2016, 7:30 PMpretty darn well documented at least - w
witty-crayon-22786
03/31/2016, 7:31 PM@bored-art-40741: possible that the signal handler doesn't actually work on that host? - w
witty-crayon-22786
03/31/2016, 7:32 PMguess it depends whether it fails deterministically - h
happy-kitchen-89482
03/31/2016, 7:56 PMLooking at build break - l
limited-terabyte-34878
03/31/2016, 8:37 PMKris Wilson pointed that out to me earlier - l
limited-terabyte-34878
03/31/2016, 8:38 PMit is non-deterministic - l
limited-terabyte-34878
03/31/2016, 8:38 PMi have no idea what to do about that - l
limited-terabyte-34878
03/31/2016, 8:38 PMi can’t replicate at all locally - m
many-engine-77019
03/31/2016, 8:38 PMfwiw I've hit issues with pytest timeouts when running code that used sigalrm - m
many-engine-77019
03/31/2016, 8:39 PM(since signals and handlers are global state) - h
happy-kitchen-89482
03/31/2016, 8:41 PMSo https://travis-ci.org/pantsbuild/pants/jobs/119877734 is not due to my change? - b
bored-art-40741
03/31/2016, 8:42 PMhttps://rbcommons.com/s/twitter/r/3631/ is up cc @freezing-motorcycle-14222