16:01:10 <david-lyle> #startmeeting Horizon
16:01:10 <openstack> Meeting started Tue Feb 18 16:01:10 2014 UTC and is due to finish in 60 minutes.  The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:14 <openstack> The meeting name has been set to 'horizon'
16:01:23 <david-lyle> Hello everyone!
16:01:32 <absubram> hi!
16:01:32 <lsmola> hello
16:01:39 <akrivoka> hi o/
16:01:40 <julim> hi there
16:01:43 <tmazur> hello! o/
16:01:46 <MaxV> hi
16:02:12 <jrist> o/
16:02:17 <jrist> I was confused where we were meeting...
16:02:23 <jcoufal> \o o/ \o
16:02:37 <enykeev> hey!
16:03:32 <amotoki_> hi
16:03:55 <doug-fish> hello all
16:04:19 <david-lyle> The only general items I have for today, are the today is the feature proposal deadline for Icehouse, if you have a blueprint that you want in icehouse and it's not assigned to i-3 let's talk about it now
16:04:50 <absubram> https://blueprints.launchpad.net/horizon/+spec/neutron-subnet-mode-support
16:04:54 <david-lyle> If you have code all the better, as the deadline to merge is March 5
16:04:56 <absubram> I've mentioned it before..
16:04:58 <absubram> small change
16:05:04 <absubram> code almost done
16:05:09 <mrunge> o/
16:05:19 <absubram> uploading a version by this afternoon
16:06:09 <absubram> I'll mark it as a Workinprogress though since we need the neutron and neutronclient code to be merged first
16:06:20 <david-lyle> absubram, I added it to i-3
16:06:33 <absubram> saw it :) thanks!
16:06:38 <david-lyle> if the neutron side doesn't make it, we'll pull it back out
16:06:43 <absubram> soudns good
16:07:09 <tshirtman> hi o/
16:07:57 <david-lyle> any others, I saw a couple of new ones were added this past week
16:08:42 <amotoki_> absubram: how is the neutron side patch of IPv6 support? I haven't caught up the progress well, but IPv6 feature is one of the important topics in the neutron community.
16:09:05 <absubram> yes.. it is still in code review but it is looking in better shape now
16:09:07 <lsmola> david-lyle: Ceilometer is definitely J, but all the prerequisites should be done in I3, yaaay
16:09:15 <absubram> still hoping to be in i-3
16:09:34 <amotoki_> i think the blueprint above is tightly coupled with the neutron side, so we need to take care of the progress around I-3.
16:09:46 <amotoki_> i will take care of it as neutron bridge.
16:09:53 <david-lyle> lsmola, that's good news, we can tackle it better in J
16:10:28 <absubram> amotoki: thanks.. is there any way I can put my horizon code out for review as "being dependent" on the neutron merge?
16:10:38 <lsmola> david-lyle: yeah eglynn will finish it probably very close to deadline, so there wouldn't be a time
16:10:57 <amotoki_> absubram: we need to do it manually :-(
16:11:12 <david-lyle> absubram, yeah no cross repo dependent patches
16:11:34 <david-lyle> without managing them yourself
16:11:35 <jomara> oh boy, glad i checked this room, my time is off by an hour
16:11:37 <amotoki_> the review comment or commit message is a good place to share the information.
16:12:10 <absubram> amotoki: david-lyle: got it.. that's what I thought.. I'll reach out to you offline to find out the manual steps if it is somehting I can do on my end..
16:12:44 <absubram> amotoki: yep.. will add detaiuls in the commit message
16:13:06 <david-lyle> absubram, sounds good, but I think amotoki's addition is all I would add
16:13:15 <david-lyle> the bp is clear that it has a dependency
16:13:48 <absubram> ok got it.. I mistook what you meant by manually.. you meant I need to keep an eye on the neutron merge myself and then when done try to get the horizon side merged yes? :)
16:14:15 <absubram> no gerrit tool support
16:14:17 <david-lyle> yes, it's that glamorous
16:14:22 <absubram> haha
16:15:05 <david-lyle> i-3 is looking really good, we have 8 bps completed so far and 22 that need review
16:15:31 <absubram> so should I mark it as a workinprogress until such time or how do we prevent a merge? Or I'm guessing Jenkins will probably -1 it without the neutron code anyway?
16:15:58 <david-lyle> there a couple of openstack_auth patches I need reviews on so we can release a new version of django_openstack_auth for icehouse
16:16:54 <david-lyle> https://review.openstack.org/#/c/70479/ and the second depend on my finishing the django 1.6 compatibility done in Horizon
16:17:14 <david-lyle> which I believe is close, a couple more test failures to go
16:18:10 <david-lyle> #topic Open Discussion
16:19:06 <MaxV> I have some patches for the improvement of client side
16:19:24 <MaxV> if you want to take a look and see how to develop and test with angular
16:19:52 <lblanchard> o/
16:19:59 <MaxV> last of the serie https://review.openstack.org/#/c/59580/
16:20:22 <MaxV> and I have a question about mock and mox
16:20:27 <lsmola> MaxV: cool
16:20:52 <MaxV> this one is the last https://review.openstack.org/#/c/73795/
16:21:25 <tshirtman> david-lyle: did you have some time to review https://blueprints.launchpad.net/horizon/+spec/instance-boot-source-enhancement ? we are doing work in this direction for our own needs, and hope to push code to review for it soon (recent changes with client side are both useful and making things move fast), so, hopefully things can be settled for Ice House
16:21:58 <MaxV> for mox and mock I talked with some people of other project and it seems that mock will be the default for testing in openstack
16:21:59 <tshirtman> so if people have feedback about the proposal, they are welcome :)
16:22:09 <MaxV> do we have any status on this?
16:22:36 <lsmola> MaxV: we should use mock for the new tests
16:22:48 <MaxV> agreed
16:22:55 <tshirtman> +1
16:22:56 <david-lyle> I thought the issue with mox vs mock had something to do with nose integration, I could be wrong though
16:23:26 <david-lyle> that could be testr though
16:23:34 <lsmola> david-lyle: and probably python3 ?
16:23:39 * david-lyle can't follow all pending transitions
16:23:40 <MaxV> mox has an issue with python3
16:24:08 <david-lyle> I may be mixing issues
16:24:37 <lsmola> MaxV: yeah I saw somebody tried to add mox3
16:24:47 <MaxV> I think we should integrate mock as soon as possible
16:24:59 <MaxV> and stop developing mox tests
16:25:21 <lsmola> MaxV: we are already writing all tests in mock, in tuskar-UI :-)
16:25:28 <david-lyle> MaxV, I haven't dug into this, but I assume there's a fair amount of overhead to transition to mock?
16:25:55 <MaxV> maybe but mox is just a pain
16:26:01 <amotoki_> the ways to write tests are quite different. It is not straight forward.
16:26:07 <lsmola> david-lyle: yeah it will be pretty dull to rewrite it all :-(
16:26:26 <david-lyle> not saying we shouldn't transition, just not an undertaking for the next two weeks
16:26:38 <lsmola> :-)
16:26:42 <tshirtman> i feel mox tests prevent improvements just as they can prevent failures, the way they work just prevent code changes
16:26:47 <tshirtman> but i may be missing the point
16:27:02 <amotoki_> it seems we should first look tuskar-ui way to know how to write tests.
16:27:20 <tshirtman> but yeah, redoing a test suit is some work :]
16:27:46 <lsmola> amotoki: hehe, we will be rewriting them, cause some of them are not very nice :-)
16:28:02 <lsmola> amotoki: but yeah, it is good for inspiration
16:28:28 <david-lyle> tshirtman, we may want to transition as suggested writing new tests using mock and possibly maintaining the mox based tests for some time
16:28:31 <amotoki_> yeah. personally i am writing both :-(   neutron and nova  use mock, neutronclient and horizon use mox.
16:28:52 <lsmola> tshirtman: I believe you can do the same things with mock, it is always about how good it your test :-)
16:29:02 <tshirtman> david-lyle: ack
16:29:15 <lsmola> david-lyle: +1
16:29:17 <tshirtman> lsmola: probably, i have yet to see a lib you can't misuse ^^
16:29:44 <david-lyle> it's coming up with new and creative ways to misuse the libraries that's interesting
16:30:20 <lsmola> hehe
16:30:50 <david-lyle> tshirtman, I need to review the bp you mentioned above, I will do that today, looks promising at first glance
16:31:04 <tshirtman> david-lyle: thanks :)
16:32:32 <david-lyle> I have a plug, HP is going to run a general usability test on the stock OpenStack Dashboard.  They are looking for takers with OpenStack Horizon experience and without, the survey to screen participants is https://www.surveymonkey.com/s/R8QZ2RB
16:33:02 <tshirtman> good call, i saw that and wanted to have a look
16:33:30 <lblanchard> +1
16:33:36 <lsmola> +1
16:33:45 <david-lyle> The idea is to provide the results back to the community to emphasize areas of work, and provide a baseline to judge future improvements against
16:33:57 <lsmola> cool
16:33:58 <absubram__> cool.. will share it with the folks here at Cisco too ;)
16:34:25 <david-lyle> the screener (survey) benefited from community input as well
16:34:47 <david-lyle> I think the actual testing will start on Feb 24
16:37:35 <david-lyle> anything else?
16:37:52 <absubram__> amotoki: can I get the +2 that was promised for https://review.openstack.org/#/c/48400/
16:37:54 <absubram__> :p
16:38:16 <amotoki_> absubram__: i can look it tomorrow.
16:38:18 <absubram__> I have a couple more bugfix reviews pending that are dependent on this getting merges
16:38:23 <absubram__> great.. thanks! :)
16:38:43 <amotoki_> not directly related to horizon, related to review 70479 above (keystone v3), what is the good way to use keystone v3?
16:38:58 <amotoki_> I know horizon support keystone v3, but most CLI do not. what is the recommended way to test?
16:39:41 <david-lyle> with that patch, keystone v3 will be used and the identity panel group in the admin panel will now show groups and roles
16:39:52 <david-lyle> panels in addition to users and projects
16:40:45 <david-lyle> if you want to test multi-domain keystone v3, then you need to add OPENSTACK_KEYSTONE_MULTIDOMAIN_SUPPORT = True to you local_settings.py
16:41:01 <david-lyle> that part is optional
16:41:30 <david-lyle> if you do turn in on, on the login form, the domain field should be "default"
16:41:44 <david-lyle> once you create other domains, you could log into them as well
16:42:07 <david-lyle> plus a domains panel will appear in the identity panel group
16:42:58 <amotoki_> david-lyle: it is really helpful, thanks.
16:43:00 <tshirtman> sounds cool!
16:43:10 <tshirtman> will certainly test this
16:44:00 <david-lyle> there is a bug without that patch, where if you set the identity API version to 3, you can't login
16:44:01 <enykeev> may i ask someone to review that one https://review.openstack.org/#/c/71851/ and also shed the light on how should we deal with that kind of circular imports?
16:44:35 <david-lyle> keystoneclient changes illustrated us abusing some auth mechanisms
16:46:26 <david-lyle> enkeev, what was the circular part?
16:47:14 <david-lyle> enykeev, oops
16:48:09 <tshirtman> david-lyle: https://bugs.launchpad.net/horizon/+bug/1276057 says login on non-default is not working, i understand it's now fixed?
16:49:03 <david-lyle> tshirtman, only with the patch mentioned before to openstack_auth
16:49:13 <tshirtman> ok
16:49:28 <enykeev> david-lyle: it imports policy.py on openstack_dashboard.settings and then goes all the up to django_openstack_auth where it tries to import django.auth which requires openstack_dashboard.settings to be loaded
16:50:22 <tshirtman> i can't seem to see where the patch number was mentionned
16:50:41 <david-lyle> should not be loading openstack_dashboard.settings from openstack_auth, that's where the fix needs to happen
16:50:48 <david-lyle> enykeev ^^^
16:51:13 <david-lyle> tshirtman: https://review.openstack.org/#/c/70479/
16:51:30 <tshirtman> thanks :)
16:53:13 <enykeev> david-lyle: openstack_auth didn't load it directly, django.auth imports it
16:53:38 <david-lyle> how would django.auth have a reference to openstack_dashboard?
16:53:59 <david-lyle> oh, pulling in the settings
16:54:37 <enykeev> david-lyle: through DJANGO_SETTINGS_MODULE env variable
16:55:33 <david-lyle> well, we shouldn't have a circular dependency, but we should take this offline
16:55:44 <enykeev> kk
17:00:08 <david-lyle> Thanks everyone.  Icehouse 3 is looking really good, and we have a lot of things waiting for review.  The gate is in great shape, so let's get the patches in shape and merged.  Have a great week!
17:00:12 <david-lyle> #endmeeting