#general
- e
enough-analyst-54434
04/26/2016, 12:41 AMOr upload a custom diff - w
witty-crayon-22786
04/26/2016, 12:41 AM./rbt --help post - m
mammoth-pharmacist-54215
04/26/2016, 12:45 AMupdated - e
enough-analyst-54434
04/26/2016, 1:28 AM@mammoth-pharmacist-54215, noted on RB, but the CHANGELOG should have 2 new items for the 2 new cherrypicks if I understand the consensus correctly. - e
enough-analyst-54434
04/26/2016, 1:28 AMIE: this process should be == to the master process mod the different branch. - m
mammoth-pharmacist-54215
04/26/2016, 1:34 AMmakes sense, will update later tonight - h
happy-kitchen-89482
04/26/2016, 2:22 PMI will try and get a review for all my docsite changes out tonight or tomorrow - h
happy-kitchen-89482
04/26/2016, 2:22 PMIt’s a big one - lots of style and content changes - e
enough-analyst-54434
04/26/2016, 3:00 PMLuckily this had not scrolled off our 10k slack limit: https://pantsbuild.slack.com/archives/general/p1460652027003446 - e
enough-analyst-54434
04/26/2016, 3:00 PMThat lockfile issue is unfortunately a real time bomb for pants users. - e
enough-analyst-54434
04/26/2016, 3:01 PMHit an Aurora user trying to build binaries for Aurora 0.12.0 (prior release / pants 0.0.70): http://markmail.org/message/3yllc5r2r2yg4gou - e
enough-analyst-54434
04/26/2016, 3:01 PMI'll send a note to pants-devel that explains the break and the workaround. - c
cuddly-flower-9506
04/26/2016, 3:35 PMapologies for not pointing this our earlier last week - e
enough-analyst-54434
04/26/2016, 3:35 PMNo worries, all kinds of ball drops on this one. - e
enough-analyst-54434
04/26/2016, 3:36 PMFor context - Medium hit this 4/20 as well, though they found a different workaround / not root cause. - c
cuddly-flower-9506
04/26/2016, 3:36 PMFWIW -- we just rebuilt the lockfile egg by hand and tossed it into our .cache - c
cuddly-flower-9506
04/26/2016, 3:37 PMwith the notion we should focus on moving forward as we are on a pretty outdated version of pants and knew upgrading would resolve it. - e
enough-analyst-54434
04/26/2016, 3:37 PMYup - I think most folks just used that route - e
enough-analyst-54434
04/26/2016, 3:38 PMThe Aurora case is interesting in that its OSS software and lots of folks build it from tarball releases - e
enough-analyst-54434
04/26/2016, 3:38 PMSo the issue basically will exist for years for them for older releases since lots of users are conservative in upgrading clusters - f
fast-megabyte-33579
04/26/2016, 3:40 PMwho owns the lockfile dependency? is it a pants thing? - f
fast-megabyte-33579
04/26/2016, 3:40 PM(and thanks for the write-up John) - e
enough-analyst-54434
04/26/2016, 3:44 PMpants used to depend on lockfile (==0.10.2, then later ==0.12.2). So we owned that pinned dep for pants <= 0.0.79. The lockfile dep on pbr, the problem child, is owned by ... openstack? - e
enough-analyst-54434
04/26/2016, 3:46 PMIts actually hard to say. I found this https://github.com/openstack/pylockfile but the pypi page only links to https://launchpad.net/pylockfile which I have a hard time reading, but its unclear who owns. - e
enough-analyst-54434
04/26/2016, 4:05 PMAnd twitter/commons of course has this issue. Turns out Antoine Tollenaere is trying to build commons to make py serverset edits for Twitter-internal aurora customisations and is hitting the same thing! This is a strange day to have this confluence of events - b
bored-art-40741
04/26/2016, 5:23 PMThis is super interesting: https://jenkins.io/doc/book/pipeline-as-code/ - b
bored-art-40741
04/26/2016, 5:24 PMI love the idea of defining the CI configuration in the repo and having that generate the full CI matrix per PR - a
adorable-glass-89507
04/26/2016, 5:29 PM@bored-art-40741: that's what we do - b
bored-art-40741
04/26/2016, 5:29 PMOh cool - b
bored-art-40741
04/26/2016, 5:29 PMDo you have like a multi-axis build per branch?