19:59:36 <david-lyle> #startmeeting Horizon
19:59:37 <openstack> Meeting started Wed Apr  8 19:59:36 2015 UTC and is due to finish in 60 minutes.  The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:59:38 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:59:40 <openstack> The meeting name has been set to 'horizon'
19:59:47 <david-lyle> hello horizon folks
19:59:52 <lhcheng> o/
19:59:52 <mrunge> o/
20:00:02 <bpokorny> Hi
20:04:34 <crobertsrh> Hello/
20:04:35 <david-lyle> so much excitement an hour ago, so few hands
20:04:35 <TravT_> ?
20:04:35 <robcresswell> Wrong time again? :p
20:04:35 <mrunge> TravT_, you missed us trying to start the meeting one hour earlier
20:04:35 <robcresswell> Evening all :)
20:04:35 <TravT_> ahh, glad i missed it!
20:04:35 <david-lyle> everyone was wanting a meeting an hour ago
20:04:35 <TravT_> let's pretend it was an hour now
20:04:36 <TravT_> pretend it was an hour ago, and just call it good
20:04:36 <david-lyle> Alright folks, we are almost done with RC-1
20:04:36 <david-lyle> #link https://launchpad.net/horizon/+milestone/kilo-rc1
20:04:36 <esp> nice!
20:04:36 <david-lyle> 5 FFEs merged and 1 was deferred to Kilo
20:04:36 <robcresswell> Nice work
20:04:36 <doug-fish> cool
20:04:36 <david-lyle> err libery
20:04:36 * david-lyle failing
20:04:36 <david-lyle> Liberty
20:04:36 <david-lyle> there I can type it
20:04:36 <david-lyle> a Few more bugs that would be nice to have fixed
20:04:46 <david-lyle> at least the fixes for said bugs
20:04:53 <david-lyle> we have enough bugs
20:05:10 <david-lyle> a couple more items for launch instance are not in that list for some reason
20:05:18 <david-lyle> I blame the PTL
20:05:29 <david-lyle> will look to remedy that
20:05:34 <mrunge> we can elect a new ptl
20:05:37 <doug-fish> :-)  I have more bugs on deck that I'm planning to spring on you over the next day or 2.  Related to hardcoded English strings.
20:06:04 <david-lyle> I don't think any of the remaining bugs will block the RC
20:06:11 <mrunge> doug-fish, I think we're already in string freeze?
20:06:13 <david-lyle> which I expect to be tomorrow or Friday
20:06:18 <doug-fish> that makes sense to me.
20:06:22 <david-lyle> Monday at the latest
20:06:28 <mrunge> if not, we're nearly there
20:06:32 <doug-fish> I think we are only in _mostly_ string freeze at this point
20:06:34 <doug-fish> right?
20:06:41 <doug-fish> rc1 is the real freeze
20:06:57 <david-lyle> Yes, we need to give the translators time to do the final translations
20:07:14 <TravT_> we have fixes up for the launch instance bugs.
20:07:18 <TravT_> i just tested all of them.
20:07:18 <david-lyle> that will hopefully be the only patch merged for RC2
20:07:24 <mrunge> david-lyle, when you cut rc1, could you release another d-o-a, too?
20:07:38 <TravT_> looks like a couple really minor styling things to fix up.
20:07:40 <mrunge> I mean django_openstack_auth
20:07:41 <lhcheng> there have been couple of bugs creeping in, but mostly bugs we think important enough to include and was already approved
20:08:05 <david-lyle> mrunge: I need to wait for the kilo release to finalize
20:08:21 <david-lyle> I will right after that
20:08:21 <doug-fish> david-lyle: will we hold up rc1 for https://bugs.launchpad.net/horizon/+bug/1436903 ?
20:08:22 <openstack> Launchpad bug 1436903 in OpenStack Dashboard (Horizon) "integration tests failing blocking gate" [Critical,Confirmed] - Assigned to David Lyle (david-lyle)
20:08:33 <david-lyle> doug-fish: I don't think so
20:08:52 <david-lyle> I certainly would like to have a fix, but I have not tracked down the root cause
20:09:01 <doug-fish> yeah that was going to be my next question
20:09:09 <lhcheng> david-lyle: when do we consider kilo release finalized? is that after rc1 or even later?
20:09:11 <david-lyle> the only patch I saw attempting to fix it other than mine was to lengthen the timeout
20:09:16 <david-lyle> which I'm loathe to do
20:09:56 <david-lyle> lhcheng: I think once most/all projects RC-1s are cut
20:10:24 <lhcheng> david-lyle: okay
20:10:28 <david-lyle> we just don't want to cause gating issues as a side effect
20:10:48 <david-lyle> so releasing now is frowned upon
20:10:55 <lhcheng> sounds reasonable
20:11:25 <lhcheng> ++
20:11:26 <david-lyle> Ok, other general items
20:11:51 <david-lyle> The PTL nomination season is upon us if you didn't see the email on the dev ML
20:12:20 <david-lyle> I nominated myself again, it is always open to self-nomination
20:12:39 <david-lyle> I think the deadline is 6:00 UTC tomorrow
20:12:39 <tqtran> how does one vote for it?
20:13:20 <david-lyle> So if someone declares their candidacy, there would be an election
20:13:33 <mrunge> tqtran, you'll just send a mail to openstack-dev and announce, you're running for PTL. Bonus points, for reasons to vote for you
20:13:55 <david-lyle> and you would receive email about voting, otherwise if I am the only one running, no election
20:13:56 <mrunge> at least, we'd then have an elected PTL
20:13:57 <tqtran> no i meant, once the candidates are announced, how do we vote?
20:13:58 <lhcheng> mrunge: only folks that contributed to that project can vote?
20:14:02 <tqtran> ah ok
20:14:05 <mrunge> yes
20:14:15 <david-lyle> yes ATCs would get the email
20:14:30 <mrunge> lhcheng, that's right, and only contributors can declare their candidacy
20:15:09 <david-lyle> A week after the PTL elections close is the TC self-nomination period
20:15:10 <mrunge> so: folks, think twice, nominate yourself
20:15:33 <mrunge> just to have an elected PTL
20:15:37 <david-lyle> if you're interested in that
20:15:43 * mrunge hides
20:15:58 <david-lyle> I meant the TC
20:16:37 <TravT_> Is TC only self-nominate?  or are all of these also things that others can nominate?
20:17:02 <david-lyle> TravT_: I think only self nominate, but unsure
20:17:16 <david-lyle> I think there is a willingness thing that comes into play :)
20:17:35 <david-lyle> the board seats are true nominations
20:17:43 <david-lyle> and you can decline the nomination
20:17:54 <david-lyle> the PTL and TC elections are not that sophisticated
20:17:55 <TravT_> Can one explain exactly what all the TC does?
20:18:08 <david-lyle> TravT_: :/
20:18:16 <mattfarina> lol
20:18:40 <mrunge> TravT_, can we please move such merely unrelated questions to somewhere else?
20:18:45 <david-lyle> no, the TC (Technical Committee) is the technical governing body of the developers in openstack
20:19:24 <mrunge> https://www.openstack.org/foundation/tech-committee/
20:19:40 <david-lyle> but the prose mrunge linked will be better and more effective
20:19:44 <TravT_> mrunge: thx
20:19:55 <ttx> also governance.oenstack.org for current data on TC
20:20:05 <ttx> err.. governance.openstack.org
20:20:55 <TravT_> ttx: thx
20:22:10 <david-lyle> ok, release stuff in another room, back
20:23:23 <mrunge> we have a single item on today's agenda
20:23:28 <mrunge> https://wiki.openstack.org/wiki/Meetings/Horizon#Agenda_for_April_8_2000_UTC
20:23:40 <david-lyle> #chair mrunge
20:23:41 <openstack> Current chairs: david-lyle mrunge
20:23:45 <david-lyle> back in a minute
20:24:15 <mrunge> lhcheng, you put django-1.7 and d-o-a and python-2.6 there
20:24:41 <lhcheng> mrunge: yeah
20:25:03 <lhcheng> so we have some work to add support for Django1.7 for horizon
20:25:21 <david-lyle> I think we'll be ok to drop 2.6
20:25:22 <lhcheng> however, the DOA patch to update the django version to 1.7 have been failing: https://review.openstack.org/#/c/167942/
20:25:32 <david-lyle> just trying to determine the timeline right now
20:25:36 <lhcheng> this is due to some django tests classes code that doesn't work with py26
20:26:00 <mrunge> yeah. correct. django-1.7 dropped py26 support
20:26:18 <mrunge> so, django-1.6 is the last version supporting py26
20:27:21 <doug-fish> so what happens with d_o_a for icehouse or juno? no updates are available if you have py26?
20:27:22 <mrunge> I think "we" (as OpenStack) decided in Paris(?) to drop support for Python-2.6 during Kilo dev cycle, right?
20:28:03 <mrunge> doug-fish, there was ttx email about stable branches for libs etc. today
20:28:34 <doug-fish> k, thx (looking)
20:29:00 <david-lyle> next release of doa will be 1.2.0 and not have python 2.6 support
20:29:17 <lhcheng> mrunge:  what if there are security patches needed for icehouse or juno, how do we distribute the fix?
20:29:35 <mrunge> lhcheng, I would expect via stable branch
20:29:48 <mrunge> like for horizon
20:30:05 <doug-fish> d_o_a doesn't have such a thing, does it?
20:30:06 <lhcheng> mrunge: ah so we backport to the last version and re-release it to pypi?
20:30:30 <mrunge> I would expect e.g 1.1.x to be pre-kilo
20:30:30 <david-lyle> lhcheng: we don't now
20:30:36 <david-lyle> those versions are capped
20:30:36 <mrunge> 1.2.x kilo
20:30:47 <mrunge> and maybe 1.3.x for liberty (or so)
20:30:48 <david-lyle> starting with kilo, we'll create a stable branch
20:31:24 <doug-fish> that's a good plan.  I can understand that!
20:31:31 <mrunge> I already have those for packages
20:31:38 <mrunge> stable branches
20:31:54 <lhcheng> mrunge: ah 1.1.x sounds good, thanks!
20:32:07 <david-lyle> ok looks like 1.2.0 is a possibility for Kilo
20:32:22 <mrunge> lhcheng, that's not set in stone yet
20:32:50 <david-lyle> mrunge: doa will have stable branch in its tree for kilo
20:32:56 <david-lyle> just have to get there
20:33:08 <mrunge> I would like to have some clearer versioning scheme
20:33:09 <david-lyle> then kilo will have a capped version of d-o-a
20:33:21 <david-lyle> mrunge: than?
20:33:39 <mrunge> david-lyle, maybe like 2015.1.x for kilo
20:33:47 <mrunge> 2014.2.x for juno
20:33:55 <mrunge> to have the clear relation
20:34:24 <david-lyle> not sure how that would work for requirements.txt
20:34:33 <mrunge> but if we'd move d-o-a back in horizon, as mentioned sometime ago...
20:35:05 <david-lyle> mrunge: that's a conversation for the summit, but certainly a possibility
20:35:25 <mrunge> 1.1.x < 2014.2 < 2015.1
20:35:33 <mrunge> so capping should be possible
20:36:18 <mrunge> back to d-o-a and py26: are we fine with dropping py26 support?
20:36:31 <mrunge> and disabling py26 based tests?
20:36:38 <david-lyle> mrunge: yes, checked and we are ok to doing that
20:36:45 <david-lyle> I have the patch almost ready
20:36:52 <mrunge> ok, great!
20:36:57 <david-lyle> will push for review after the meeting
20:37:15 <mrunge> thanks, could you please add me, I will review it asap
20:37:31 <david-lyle> going to send an email to the dev ML to explain the requirements mismatch and the high possibility of a late release and g-r bump
20:37:44 <lhcheng> ++ thanks david-lyle
20:37:52 <mrunge> yes, thank you!
20:38:04 <david-lyle> I'm worried we'll have install problems with 1.6 and 1.7 competing requirements for django cap
20:38:45 <david-lyle> #topic Open Discussion
20:39:11 <mrunge> as far as I understand that, we should be fine with django-1.7
20:39:34 <david-lyle> mrunge: I was worried about dependency resolution
20:40:16 <david-lyle> depending how the two are packaged
20:40:24 <mrunge> david-lyle, you mean, because of pip installing both?
20:40:44 <mrunge> pip overwrites installed files happily
20:41:19 <david-lyle> mrunge: I was more worried about distro packaging
20:41:36 <mrunge> david-lyle, yes, me too
20:41:46 <david-lyle> so even though horizon supports 1.7 you would be limited to 1.6
20:42:02 <david-lyle> and hopefully the various packaging systems could handle that
20:42:08 <mrunge> david-lyle, I'm usually ignoring requirements.txt
20:42:09 <david-lyle> which I'm not confident of
20:42:39 <david-lyle> mrunge: ok, the library already works with 1.7 in 1.0.9
20:42:47 <mrunge> I just take it as recommendation, and run the test suite during package build again
20:43:18 <david-lyle> just for the target python version?
20:43:24 <david-lyle> or py26 as well
20:43:25 <mrunge> yes
20:44:06 <mrunge> david-lyle, you can run run_tests.sh -N, which will run the test suite with installed versions
20:44:26 <david-lyle> mrunge: there is no run_test in django_openstack_auth
20:44:30 <mrunge> quite handy, and it gives me a more confidence
20:44:38 <david-lyle> but you just run it for horizon
20:44:45 <mrunge> yes, correct
20:44:48 <david-lyle> ok
20:45:07 <mrunge> d-o-a is quite simple here.
20:45:14 <mrunge> either works or not
20:45:42 <mrunge> but this reminds me, we should integrate that more tightly in integration tests
20:46:28 <mrunge> I mean, a d-o-a git checkout rather than released versions
20:46:57 <david-lyle> mrunge: that should be happening already as devstack clones d-o-a
20:47:18 <david-lyle> err, maybe not anymore
20:47:20 <david-lyle> :(
20:47:32 <mrunge> david-lyle, we had the situation, that d-o-a broke horizon not that long ago
20:47:49 <mrunge> imho lhcheng fixed that
20:48:08 <mrunge> and he might have a better memory than me about that
20:49:46 <david-lyle> looks like it's a configuration option in devstack now
20:50:02 <david-lyle> will look into more how to turn it on
20:50:21 <david-lyle> the overall stability of the integration tests is not good
20:50:40 <david-lyle> we need to rethink the tool sets a bit
20:50:43 <david-lyle> I think
20:50:50 <mrunge> yes!
20:50:58 <lhcheng> mrunge: I haven't done anything to fix the d-o-a checkout on devstack test
20:51:48 <mrunge> hmm, I thought so
20:51:59 <lhcheng> mrunge: just did a fix where horizon tests were failing for 1.8
20:52:12 <david-lyle> the behavior changed in devstack
20:52:27 <david-lyle> I think as part of the whole library stabilization efforts
20:53:44 <david-lyle> as a reminder: #link https://etherpad.openstack.org/p/horizon-liberty-summit
20:54:07 <david-lyle> a few minutes left today, anyone else?
20:54:19 <david-lyle> s/anyone/anything/
20:55:05 <mattfarina> david-lyle when are the proposed sessions for liberty summit due?
20:55:35 <david-lyle> mattfarina: the schedule is usually set a couple weeks before the summit
20:55:48 <david-lyle> to make sure people have time to prep
20:56:01 <david-lyle> but also want to make sure we get relevant/fresh topics
20:56:39 <david-lyle> The PTL for the cycle selects the sessions, so it can't happen until after election season
20:59:05 <david-lyle> ok 2.6 removal patch up on DOA
20:59:13 <david-lyle> running tests in the gate
20:59:22 <david-lyle> and we're out of time
21:00:08 <david-lyle> Thanks for all the hard work on RC-1, we'll be working on Liberty by the next meeting. That is once RC-1 is but, master is open for Liberty.
21:00:13 <david-lyle> #endmeeting