Friday, 2018-02-16

*** kumarmn has joined #openstack-tc00:27
*** kumarmn has quit IRC00:28
*** kumarmn has joined #openstack-tc00:28
*** kumarmn has quit IRC00:29
*** kumarmn has joined #openstack-tc00:29
*** kumarmn has quit IRC00:40
*** kumarmn has joined #openstack-tc00:41
*** kumarmn has quit IRC00:41
*** kumarmn has joined #openstack-tc00:41
*** persia has joined #openstack-tc00:54
*** kumarmn has quit IRC01:25
*** kumarmn has joined #openstack-tc01:26
*** kumarmn has quit IRC01:31
*** mriedem1 has joined #openstack-tc02:29
*** mriedem has quit IRC02:29
*** mriedem1 is now known as mriedem02:29
*** kumarmn has joined #openstack-tc02:31
*** kumarmn has quit IRC02:35
*** kumarmn has joined #openstack-tc02:46
*** kumarmn has quit IRC02:55
*** mriedem has quit IRC03:07
*** harlowja_ has quit IRC03:23
*** ykarel has joined #openstack-tc04:44
*** harlowja has joined #openstack-tc04:46
*** kumarmn has joined #openstack-tc04:55
*** harlowja has quit IRC04:59
*** kumarmn has quit IRC04:59
*** harlowja has joined #openstack-tc05:19
*** harlowja has quit IRC05:27
*** rosmaita has quit IRC05:39
*** coolsvap has joined #openstack-tc06:03
*** openstackstatus has quit IRC06:27
*** openstack has joined #openstack-tc06:29
*** ChanServ sets mode: +o openstack06:29
*** openstackstatus has joined #openstack-tc06:30
*** ChanServ sets mode: +v openstackstatus06:30
*** dtantsur|afk is now known as dtantsur07:46
openstackgerritMerged openstack/governance master: Team diversity tags update, February 2018  https://review.openstack.org/54356608:17
openstackgerritMerged openstack/governance master: Update Rocky PTLs  https://review.openstack.org/54475308:22
openstackgerritDougal Matthews proposed openstack/governance master: Update IRC nick for Dougal Matthews (me)  https://review.openstack.org/54523608:29
*** david-lyle has quit IRC08:46
*** jpich has joined #openstack-tc08:54
openstackgerritDougal Matthews proposed openstack/governance master: Update IRC nick for Dougal Matthews (me)  https://review.openstack.org/54523609:14
*** bauzas is now known as bauwser09:30
openstackgerritMerged openstack/governance master: Update IRC nick for Dougal Matthews (me)  https://review.openstack.org/54523610:05
*** openstackgerrit has quit IRC10:18
*** cdent has joined #openstack-tc10:46
*** pbourke has quit IRC11:32
*** pbourke has joined #openstack-tc11:34
*** kumarmn has joined #openstack-tc11:56
*** kumarmn has quit IRC12:00
*** cdent has quit IRC12:08
*** cdent has joined #openstack-tc12:38
*** rosmaita has joined #openstack-tc12:42
*** coolsvap has quit IRC13:22
*** cdent has quit IRC14:00
*** mriedem has joined #openstack-tc14:03
*** ykarel has quit IRC14:13
*** kumarmn has joined #openstack-tc14:30
*** dansmith is now known as superdan14:43
*** ykarel has joined #openstack-tc15:01
*** cdent has joined #openstack-tc15:38
smcginnisJust mentioning here, but I'm seeing more from distros going to python 3.6.15:38
smcginnisWe currently only support 2.7 and 3.5, and there are some things that break with 3.6.15:39
smcginnisSo we might need a bigger community wide "goal" to get moving on 3.6 support since it is now going to be the default for several distros.15:39
pabelangeryah, fedora 26+ has been 3.615:40
mtreinishsmcginnis: what breaks with py3.6?15:40
smcginnismtreinish: I haven't looked too close, but there are a lot of little things.15:40
mtreinishsmcginnis: but we'll get 3.6 jobs in a couple months when 18.04 ubuntu is released15:40
mtreinishsmcginnis: not in my experience, I only have 3.6 on my systems. While I don't run 3.6 for my cloud, all my local testing on my desktop and laptops is with 3.6 and I haven't seen anything15:41
persiamtreinish: The one that caught me yesterday is that in 3.5 1_2_3 is a string, and in 3.6 1_2_3 is an integer.15:41
smcginnismtreinish: You can take a look at all the bugs I duped to https://bugs.launchpad.net/cinder/+bug/174995315:41
openstackLaunchpad bug 1749953 in Cinder "Add support for Python 3.6" [Wishlist,New]15:41
mtreinishpersia: wouldn't 1_2_3 be a syntax error in 3.5?15:42
mtreinishpersia: and where would we hit that in openstack?15:42
smcginnis1_2_3 is an integer?!15:43
* smcginnis goes home15:43
dhellmannsmcginnis : https://www.python.org/dev/peps/pep-0515/15:43
mtreinishsmcginnis: yeah, it's weird15:43
smcginnis"This is a common feature of other modern languages"15:44
smcginnisI guess I'm old and don't use those "modern" languages. :)15:44
*** david-lyle has joined #openstack-tc15:44
jrollruby has this too, it's nice for writing long numbers: 12_765_45615:45
*** zaneb has quit IRC15:52
persiamtreinish: Onlly a syntax error if you assume a non-castable type.  If you parse it as input, and treat castable-to-integer as integer and otherwise consider it a string (say, you have a YAML parser or some such), then it gets more interesting.15:53
mtreinishah, ok15:54
*** zaneb has joined #openstack-tc15:55
*** ykarel_ has joined #openstack-tc16:10
*** ykarel has quit IRC16:13
*** david-lyle has quit IRC16:15
*** hongbin has joined #openstack-tc16:16
*** zaneb has quit IRC16:19
*** zaneb has joined #openstack-tc16:21
*** david-lyle has joined #openstack-tc16:22
fungihuh, i've been using 3.6 and 3.7 for a while and didn't know about type(1_2_3)16:44
funginifty16:44
fungibut yes, when we get ubuntu 18.04 servers, whatever migration goes on to move testing for projects from xenial to bionic nodes will also mean requiring they switch from testing with 3.5 to 3.6 (probably 3.6.4 from the looks of it)16:46
*** ykarel_ is now known as ykarel|away16:47
fungior they can customize their jobs to add python3.5 packages from somewhere else (or build from source if they really want, though rebuilding python is slooow)16:47
smcginnisfungi: When is 3.6 testing expected then?16:50
fungismcginnis: not until at least april16:50
fungisince that's when ubuntu 18.04 is scheduled to release16:50
smcginnisfungi: And we expect to pick it up right after release?16:51
fungiodds are probably more may-ish depending on how quickly we get people interested in getting images building for it16:51
dhellmannthere was a thread from the RDO folks about setting up something on Fedora, but I don't know their timeline16:51
fungiyeah, i mean we already provide fedora nodes you can run jobs on if you want. those are more problematic from a stable testing perspective though since they go out of support before we stop supporting the stable branches which were released testing on them16:52
fungiwhereas if we switch to 3.6 testing for rocky or stein then we can continue testing them on the same platform and python version to the end of our support for those stable branches16:56
fungiswitch to 3.6 testing on ubuntu 18.04 that is16:57
pabelangerYah, I'm already working on getting ubuntu-beaver DIBs working. But looks like need a backport on xenial for debootstrap info16:57
fungisimilarly, we could do 3.6 or 3.7 testing on centos 8 to similar effect, but no clue when that's releasing16:58
fungitesting on non-lts distros is useful for short periods of time, but isn't something we can carry sanely into stable branch testing16:59
fungibecause we'd either have to drop the platforms providing those tests or start testing old stable branches on new platforms, neither of which is likely to bring joy to those trying to maintain them17:00
persiaThe third possibility is continue testing old stable branches on unmaintained platforms, but that should be dismissed as an unboundedly large security problem.17:12
dhellmannyeah, I was just wondering if the fedora support might become a thing before the ubuntu lts update17:13
dhellmannit was proposed as a way to get to some precursor to centos/rhel 8 before that is locked down, since fedora is the upstream17:13
*** jpich has quit IRC17:15
thingeecdent: there's a hack session with contributor guide happening at the ptg, if you want to get some fast review on anything for governance related content. https://etherpad.openstack.org/p/OUI-Rocky-PTG17:49
cdentthingee: thanks for the reminder. I've been very bad about making that go, for the usual nova-related reasons. Eeking out some time at the PTG could help a lot and also be a nice break from the usual nova-related things.17:52
thingeecdent: no worries I understand :)17:52
*** harlowja has joined #openstack-tc17:55
*** mriedem is now known as mriedem_lunch17:58
*** bmace has quit IRC17:58
*** david-lyle has quit IRC18:12
fungidhellmann: yep, it seems like a reasonable stepping-stone as long as someone's volunteering to put in the effort to get it working and keep it maintained until we have ubuntu 18.04 or centos 8 providing the same, but i wouldn't want to claim, e.g., that stable/rocky will have 3.6 support if we can't switch from fedora to one of those lts releases for those jobs before release day18:29
*** dtantsur is now known as dtantsur|afk18:29
*** bmace has joined #openstack-tc18:29
fungiotherwise we're stuck having to swap them out for jobs running against the stable branch. not forcing all projects to transition to xenial in the same cycle as one another was a significant enough pain when we had to try to play catch up redoing jobs in stable, and i would rather learn a lesson from that experience18:31
fungiyes it's painful to have a testing platform change thrust on you, but either everyone transitions together or no one does18:32
*** harlowja has quit IRC18:44
*** harlowja has joined #openstack-tc18:48
*** cdent has quit IRC18:58
*** ykarel|away has quit IRC18:58
*** harlowja has quit IRC19:10
*** david-lyle has joined #openstack-tc19:13
*** mriedem_lunch is now known as mriedem19:44
*** harlowja has joined #openstack-tc19:50
*** bmace_ has joined #openstack-tc20:21
dmsimardfungi: I hope that RHEL 8 / CentOS 8 will be available by the time Rocky releases, it would make our lives much easier. Like dhellmann said, the current plan is to get started on Fedora with py3 and hope that gets us close to what RHEL8 will look like.20:37
dmsimardRDO has already been building packages on Fedora -- but they're not tested or recommended for public consumption. The idea is to at least get some testing coverage going and make that available.20:38
fungiyeah, but if we start with py36 testing on fedora during rocky and can't get it all switched over to centos 8 before rocky releases, i don't think we should claim python 3.6 is supported for that release nor continue testing python 3.6 on stable/rocky... just try again for the stein cycle and hope it goes better20:39
dmsimardDevstack is another story, I think ianw has had it working on Fedora20:39
dhellmannfungi : yeah, that's fair20:39
smcginnisMaybe start with experimental jobs?20:40
dmsimardI don't even know what version of py3 would land in rhel8, was there going to be a LTS release of py3 ? Is that 3.6 ?20:40
smcginnisThat would simplify our life if there was.20:40
fungii don't see any problem with check jobs if they're working and relatively stable, just wouldn't want to continue them on a stable branch unless we were doing them on an lts distro prior to the end of that cycle20:41
*** kumarmn has quit IRC20:56
pabelangersmcginnis: yah, legacy-tempest-dsvm-neutron-full-fedora-26 exists today in experimental for devstack21:09
pabelangerI believe that is python3, as much as possible21:10
smcginnispabelanger: Oh good, I should take a look at that. Thanks.21:11
pabelangersmcginnis: yah, I know ianw is spending time / effort working on it and python3. I should loop back with him next week and see how I can help21:12
*** bmace__ has joined #openstack-tc22:03
*** bmace has quit IRC22:03
*** bmace_ has quit IRC22:03
*** kumarmn has joined #openstack-tc22:23
*** kumarmn has quit IRC22:23
*** kumarmn has joined #openstack-tc22:23
*** openstackgerrit has joined #openstack-tc23:06
openstackgerritBrian Rosmaita proposed openstack/governance master: Update status of 'API deployed in uwsgi' goal  https://review.openstack.org/54548423:06
*** hongbin has quit IRC23:51

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!