jodh | xnox: Hi - have you had a chance to look at that async issue (iterating through job_change_state() too many times?) | 09:01 |
---|---|---|
xnox | jodh: nope. (you mean the one exposed when dropping valid from process_data? or the one from the racy test-case?) | 09:01 |
xnox | jodh: the builds are all happy, but on powerpc. And alas, I cannot reproduce it on the porter box for powerpc. | 09:02 |
jodh | xnox: the former. That issue should be in lp:upstart too. | 09:03 |
xnox | powerpc builds hang in test_job_process and i'm not sure why. | 09:03 |
jodh | xnox: saw the ppc MP - I'm not sure that's going to be reliable. I need to pick over that test case more closely... | 09:04 |
jodh | xnox: I thought we removed the state_only param from job_process_terminated()... ? | 10:08 |
xnox | i did, it caused regressions, we didn't merge. | 10:09 |
xnox | it was the same type of story as with ->valid | 10:10 |
jodh | xnox: sure? seems to be in both lp:upstart and lp:upstart/async | 10:12 |
xnox | jodh: what do you mean? removing state_only was in a merge proposal, that didn't get merged. Thus lp:upstart still has state_only arg on job_process_terminated | 10:14 |
xnox | and lp:upstart/async is fully merged into lp:upstart now, so shouldn't be used any more. | 10:14 |
* xnox requested ppc64el and arm64 to be added on daily non-virt builds | 10:15 | |
jodh | xnox: sorry - misremembering the order of changes :) We need to resolve http://paste.ubuntu.com/7588618/ - the behaviour is incorrect atm fwics - we should not be allocating these objects twice. Do you want me to look at this? | 10:24 |
xnox | looks like arm64 is also hanging - https://code.launchpad.net/~canonical-foundations/+recipe/upstart-daily-nonvirt | 11:40 |
xnox | =( | 11:40 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!