16:00:44 <jpich> #startmeeting horizon
16:00:46 <openstack> Meeting started Tue Oct 14 16:00:44 2014 UTC and is due to finish in 60 minutes.  The chair is jpich. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:47 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:49 <ericpeterson> hello all :D
16:00:49 <openstack> The meeting name has been set to 'horizon'
16:00:50 <jpich> Hi everyone
16:00:59 <david-lyle> Hello
16:01:01 <tzumainn> hiya
16:01:02 <bradjones> \o
16:01:04 <jgravel_> Hi
16:01:14 * david-lyle on phone irc
16:01:24 <jpich> #link https://wiki.openstack.org/wiki/Meetings/Horizon
16:01:28 <akrivoka> hello!
16:01:32 <rbertram> hi
16:01:35 <jpich> david-lyle: I expect interesting auto-correct typos :)
16:01:37 <tsufiev> hi!
16:01:41 <sambetts> o/
16:01:51 <amotoki> hi
16:02:07 <TravT> o/
16:02:24 <jpich> #topic juno-rc-potential bugs
16:02:30 <jpich> #link https://bugs.launchpad.net/horizon/+bugs?field.tag=juno-rc-potential
16:02:32 <jtomasek> hey
16:02:47 <jpich> So we have a couple of bugs, including a critical one that must absolutely be fixed before the release
16:02:51 <jpich> Release which is planned in two days
16:02:55 <deshipu> hi
16:03:02 <jpich> Additional help and eyes on the problem would be extremely welcome :)
16:03:03 <crobertsrh> hello/
16:03:13 <jpich> #link https://bugs.launchpad.net/horizon/+bug/1379761
16:03:57 <david-lyle> The hope would be to have a fix in the pipeline by tonight in the US
16:04:30 <david-lyle> As we don't have a proposed fix to review that may be difficult
16:04:35 <jtomasek> jpich: I was looking at that issue and managed to reproduce it. I think the main problem is that compilation won't happen without debug set to true
16:04:59 <gugl> hi
16:05:01 <jtomasek> jpich: and I think it has to be done manually at some point
16:05:05 <bpokorny> Hi
16:05:21 <david-lyle> At this point I'm even considering a monkey patch of the problematic method to consider debug to be always true
16:05:55 <jpich> jtomasek: Ok... Please, consider addding any additional information to the bug report too, that is helpful
16:06:09 <jtomasek> jpich: ok, I'll do
16:06:16 <jpich> Thank you
16:06:20 <david-lyle> Hoping a cleaner method is proposed before that
16:06:28 <ericpeterson> I ran into this a few weeks ago, but on master it worked yesterday (I'm behind the times)  (for the compress issue)
16:07:13 <david-lyle> It worked yesterday with debug=False?
16:07:36 <ericpeterson> yup, pretty sure.  can follow up later
16:07:40 <jpich> ericpeterson: Ok... If you have any idea/half-formed thoughts on what might causing this *sometimes* please add to the bug :-)
16:07:46 <jpich> Thank you!
16:07:57 <jtomasek> ericpeterson: are you sure that some previously compiled stylesheet was not in horizon/static folder?
16:07:59 <deshipu> david-lyle: the django-pyscss people werre ve  sponsive so far and always merged pattchs within a day
16:08:09 <ericpeterson> it was related to static assets and the xstatic stuff changing around for sure.  the bug is on the correct track
16:08:27 <david-lyle> Final release for openstack is thursday
16:08:33 <ericpeterson> ack
16:08:43 <david-lyle> I don't think we have time for a proper fix
16:08:48 <david-lyle> Unfortunately
16:09:10 <david-lyle> Obviously that is the desired path
16:09:12 <jpich> deshipu: That's very good to know, I'm not sure if that one has been reported back upstream though looking at the comments
16:09:16 <ericpeterson> so you're looking for duct tape.
16:09:34 <deshipu> I also have the impression that it used to work
16:09:35 <david-lyle> Luckily you're here
16:09:59 <jpich> Did we get a version upgrade recently?
16:10:04 <david-lyle> deshipu looking at the code not sure how
16:10:20 <david-lyle> Blame says that code is from feb
16:10:54 <deshipu> you have o run collectstaic wih debug disabled
16:11:19 <ericpeterson> and was compress_offline also turned on?  that's needed too I think
16:11:23 <deshipu> butt you always had"
16:11:33 <deshipu> no idea
16:11:56 <jtomasek> deshipu: running that just moves the stylesheets but won't do the compresion
16:12:04 <david-lyle> ericpeterson couldn't get offline compression to work either with debug false
16:12:34 <david-lyle> The compress step is looking for storage files not the collected assets
16:12:48 <deshipu> I'm afk for the rest of he day, can't help :(
16:13:25 <david-lyle> It may be possible to set STATICFILES_STORAGE to something that will work by default
16:13:57 <ericpeterson> looking into it, will post back when I have something to share / rebuilding venv etc to start fresh
16:14:09 <david-lyle> Thanks all!
16:14:11 <deshipu> ericpeterson: tthnk you
16:14:20 <jpich> Thank you. I think more people will be hanging out in #openstack-horizon to continue working on this after the meeting
16:14:23 <jtomasek> ericpeterson: great
16:14:26 <ericpeterson> don't thank me until I have something working ;)
16:14:36 <jpich> ericpeterson: Counting on you :P
16:14:47 <jpich> amotoki: You had another RC3 issue to bring up?
16:14:56 <amotoki> yes. I would like to raise https://bugs.launchpad.net/neutron/+bug/1378525 (due to neutron bug)
16:15:08 <amotoki> I  think It is not release blocking. It is because neutron disabled this feature in RC3.
16:15:42 <amotoki> I just proposed a patch for it. If we cut RC3, I hope this in.
16:15:47 <jpich> If the patch is ready we can get it in at the same time, I think
16:15:48 <jpich> yes
16:16:22 <amotoki> I think we can also mention this as a known issue.
16:17:40 <jpich> and put in the fix in the first stable release then if needed then... Sounds fair enough, thank you for putting up the patch too
16:17:44 <jpich> Maybe we can get it in on time
16:18:27 <ericpeterson> I have offline compression working, fwiw
16:18:32 <jpich> ericpeterson: Eep??
16:18:50 <jpich> Could you post steps on the bug report so we can see if everyone is following the same process?
16:18:58 <ericpeterson> yep
16:19:10 <jpich> Also if it's devstack or elsewhere, wonder if there is a common pattern...
16:19:54 <jpich> Ok, anything else of relevance to the release on Thursday?
16:20:21 <jpich> david-lyle: Are there release notes the community can help review somewhere or is that tbd?
16:21:02 <david-lyle> I have them mostly done. Will post soon as patch and to the wiki. Hack up the wiki to fix my errors
16:21:33 <jpich> Awesome, thanks a lot
16:21:35 <david-lyle> Looks like my chapter on l3 ha is going to have to be redacted :)
16:21:56 <jpich> Hopefully just a lil' correction :-)
16:22:09 <david-lyle> Well yes
16:22:19 <jpich> Ok, let's move on to the next topic then
16:22:30 <jpich> #topic Adding Sinon.JS (http://sinonjs.org/) dependency to horizon. -- tsufiev
16:22:43 <david-lyle> This was a hold over
16:23:01 <jpich> Ah, ok
16:23:06 <jpich> An update was posted to the list yesterday: http://lists.openstack.org/pipermail/openstack-dev/2014-October/048301.html
16:23:10 <jpich> so maybe we can follow up there for now
16:23:22 <jpich> and continue in a meeting if no consensus is reached by email
16:23:22 <david-lyle> Sure
16:23:33 <tsufiev> jpich, yep, post something in ML :)
16:23:42 <jpich> Thanks tsufiev :)
16:23:54 <jpich> #topic Summit Session Selection (david-lyle)
16:24:02 <jpich> #link https://etherpad.openstack.org/p/horizon-kilo-summit
16:24:14 <jpich> Is that a holdover too? Vote, people, vote?
16:24:17 <tsufiev> jpich, we've discussed this topic with Radomir on #openstack-horizon yesterday
16:24:32 <david-lyle> Not a holdover
16:24:39 <jpich> tsufiev: Eep, maybe I moved too fast then. Or maybe the summary of the conversation can be posted on list? :-)
16:25:30 <david-lyle> The selection process is open to everyone. Please vote and help decide what are the most important items you would like to have covered
16:25:32 <tsufiev> jpich, the main conclusion I made myself after discussion with Radomir is that we both don't know what are the plans for Horizon client-side code
16:26:06 <tsufiev> jpich, thus, it's not clear whether rewrite every JS unit-test in Jasmine or don't bother with it
16:26:09 <david-lyle> My personal preference would be to spend the half day on clientside
16:26:26 <jpich> david-lyle: Seems relevant to the current cross-topic happening :-)
16:26:43 <jtomasek> david-lyle: +1
16:26:49 <david-lyle> Yes indeed
16:26:50 <tqtran> +1 to half day
16:27:12 <tsufiev> david-lyle, half a day each day of summit :)?
16:27:35 <tqtran> hahaha
16:27:43 <david-lyle> We will have a pod as well, we should use it
16:28:13 <jpich> Should we requests PoCs to be created in advance for sessions where it makes sense? To avoid "wouldn't it be nice" topics that can't really reach a conclusion or move anything forward?
16:28:21 <jpich> Oh cool I thought there wasn't enough space for pods this time
16:28:47 <david-lyle> My understanding is there will be pod space
16:29:12 <david-lyle> We may not have a fully dedicated one, but we'll have access to space
16:29:29 <jpich> Gotcha
16:29:31 <tqtran> jpich: i think POCs is a good idea
16:29:53 <david-lyle> PoCs are encouraged
16:29:54 <tqtran> jpich: its generally better to have something to show rather than theorycraft
16:30:01 <david-lyle> Indeed
16:30:19 <jpich> Ok I updated the etherpad
16:30:31 <david-lyle> A big concern regarding client-side is extensibility
16:30:31 <ericpeterson> I'd like to see theorycraft, never heard that term (but I get the idea / point) ;)
16:30:53 <jpich> people with topics, PoC away! ideally early enough that people have time to poke at it if possible :-) No pressure...
16:31:06 <david-lyle> I'd love to see some proposals around extensible angular
16:31:26 <david-lyle> Or is that a pipedreamm
16:31:43 <tqtran> david-lyle: what exactly is extensible angular? not sure what that means
16:32:19 <david-lyle> One of the current strengths of horizon and one of its guiding design goals is extensibility
16:32:36 <david-lyle> With django you can extend at almost any point
16:32:53 <david-lyle> From templates to blocks, etc
16:33:11 <deshipu> I tried to make an extension for horizon using angular
16:33:17 <ericpeterson> i think the templates for display etc are in that area w angular
16:33:19 <david-lyle> What can we expect from angular?
16:33:21 <deshipu> I failed, but I'm very new to that
16:33:47 <deshipu> but we do have an extension to tuskar-ui written in angular
16:33:50 <david-lyle> Say on the project overview panel
16:34:21 <david-lyle> Can I remove the usage summary because its a nova extension underneath?
16:34:46 <david-lyle> Or do I have to rewrite the entire page
16:35:04 <david-lyle> Of course a modular implementation helps
16:35:19 <david-lyle> But it's something we need to keep in mind
16:35:30 <david-lyle> Very few use horizon just as shipped
16:35:42 <david-lyle> Because that's not realistic
16:36:11 <ericpeterson> support from version to version would also be interesting for angular
16:36:11 <tqtran> david-lyle: so i think on the extensibility part, it will most likely be reliant on django still
16:36:35 <tqtran> david-lyle: i think its more of, how can we leverage django to extend and maintain angular
16:37:19 <david-lyle> Hmm, if the template just loads angular, that's very coarse granularity
16:37:28 <deshipu> tqtran: the angular components need to be reusbale somehow too
16:37:48 <deshipu> tqtran: I don't want to have to rewrite everything
16:37:54 <david-lyle> Let's talk about it together someplace, maybe paris?
16:37:57 <tqtran> david-lyle: its still a huge advantage because each component is broken down into smaller parts
16:38:25 <tqtran> david-lyle: so we can display our web app in a more ajaxy way
16:38:30 <david-lyle> Agreed, but the angular needs to be modular and reusable
16:38:43 <ericpeterson> offline it for paris methinks
16:38:48 <tqtran> yea haha
16:38:50 <jpich> Also on the ML, describing the current challenges and what is possible or not (like a summary for people to read before Paris and start from the same understanding)
16:40:28 <tqtran> jpich: so do we wait for the topic to get selected before writing out the summary?
16:40:55 <jpich> tqtran: I think it's a conversation we want to have whether or not it happens officially at summit or not
16:41:05 <jpich> probably I should have picked another word than "summary" :)
16:41:10 <david-lyle> Client-side will be a topic
16:41:21 <tqtran> more like a battle lol
16:41:59 <tqtran> jpich: ok make sense
16:41:59 <woodm1979> tqtran: Always the optomist.
16:43:26 <jpich> Great way to make people be relaxed and mind open for collaboration and new ideas before summit :)
16:43:33 <jpich> Anything else around the summit?
16:44:02 <tqtran> last year, we had a pre-summit meeting, would be good to have that again
16:44:04 <rbertram> How do we "vote" on topics? just adding comments to the etherpad?
16:44:05 <deshipu> I will need to hunt down the oslo people to talk about websockets
16:44:12 <deshipu> rbertram: yes
16:44:22 <jpich> tqtran: Yes.
16:44:42 <jpich> Totally :)
16:45:02 <jpich> Let's plan that closer to summit
16:45:16 <david-lyle> ++
16:45:34 <jpich> Ok, moving on to the last topic on the agenda so
16:45:38 <david-lyle> I think it helped a lot
16:46:07 <jpich> yep
16:46:26 <jpich> #topic Design / Blueprint approval process (akrivoka)
16:46:33 <akrivoka> yes
16:46:49 <akrivoka> I wanted to bring up this topic, as I've felt we are lacking a proper design phase in our development process
16:47:03 <akrivoka> we've had this problem many times now
16:47:33 <jpich> Somewhat related, has the new "spec" template been agreed on? (Or like, not strongly objected to?). Should we bring it up on the list/update whatever documentation needs updating?
16:47:36 <tsufiev> akrivoka, +1
16:47:38 <akrivoka> a typical situation is that only after someone submits a patch, is it revealed that the design of the solution is fundamentally wrong, or at the very least controversial
16:48:05 <tsufiev> akrivoka, several times reviewing a commit addressing some BP I wasn't sure what the Author is trying to do
16:48:16 <jpich> Hopefully the more detailed spec will help with that, two liner blueprint descriptions certainly get old...
16:48:24 * jpich probably guilty of this too, though
16:48:24 <akrivoka> that then leads to design discussion in the Gerrit review - which I don't think is where they belong
16:48:39 <jpich> #link https://blueprints.launchpad.net/horizon/+spec/template
16:48:40 <mrunge> I'd propose to clean up blueprints
16:48:43 <amotoki> doesn't the way David proposes work?
16:48:58 * mrunge thinks of removing/clean *all* blueprints
16:49:08 <jpich> I don't remember how we decided to do approvals after the spec is decided though
16:49:10 <akrivoka> jpich: oh interesting, I didn't know about this - will take a look
16:49:32 <jpich> it'd be cool if everyone felt free to participate in commenting at the design level too
16:49:48 <jpich> akrivoka: Cool! Feedback welcome obviously :)
16:49:58 <akrivoka> jpich: sure, thanks for the heads up
16:50:02 <jpich> People should remember to indicate their name on the whiteboard though, it's not obvious who's saying what otherwise
16:50:16 <amotoki> the proposal suggests us to describe more detail and requirements of blueprints. I believe it helps us.
16:51:08 <david-lyle> Ux  is a step in the process that needs more consideration
16:51:20 <mrunge> that template is awesome david-lyle !
16:51:36 <mrunge> can we make it mandatory for new blueprints?
16:52:01 <tqtran> david-lyle: agree, i think for big visual changes, we should recommend a mock-up of some sort so people can get an idea of what it ensues
16:52:05 <akrivoka> following a template like this would definitely be an improvement, imo
16:52:12 <akrivoka> huge +1 from me
16:52:14 <david-lyle> I would propose blocking pending ux input but I don't feel ux team is big enough to shoulder that burden
16:52:17 <akrivoka> (will review in detail)
16:52:27 <jpich> Yeah we definitely should publicise it a bit more now, I don't think it's been mentioned on the ML yet
16:52:37 <tqtran> david-lyle: agree, i think it should just  be recommended for now
16:52:48 <david-lyle> I think ux team input is desirable
16:53:01 <TravT> I have a question on mockups... where can we host them?  For example, if we have a pdf, I can't attach to wiki
16:53:06 <david-lyle> But horizon team ux feedback is also needed
16:53:35 <tqtran> TravT: thats a problem in code reviews also
16:53:47 <TravT> wiki only seems to accept picture files... unless I'm missing something.
16:54:17 <jpich> For things that aren't huge overhauls, simplish UX design by developers can also help understand what will happen, like what was done with the pdf at https://blueprints.launchpad.net/horizon/+spec/horizon-shelving-command
16:54:44 <tqtran> jpich: +1
16:54:59 <TravT> there's an example of a pdf hosted at some github location.
16:55:10 <david-lyle> TravT I think any free file sharing service is fine
16:55:10 <TravT> should we submit a patch with the pdf to horizon?
16:55:12 <jpich> People seem to use their own private accounts in other places for pdf/images so far. The wiki could be a potential location as well perhaps? (Not sure if attachments are possible though or if it can only link to external images)
16:55:23 <jpich> Not to Horizon no
16:55:46 <david-lyle> Binary files and git is not good
16:56:26 <tqtran> only issue is that sometime the host scrap those images
16:56:32 <tqtran> and we are left to wonder what was there
16:56:48 <tqtran> would be good to have a central place we encourage to host
16:56:51 <TravT> tqtran: yep, my concerns as well.
16:57:05 <deshipu> .oO( bitbucket has binary files uploads )
16:57:07 <TravT> maybe a question to send to openstack infra?
16:57:08 <david-lyle> I share your concern
16:57:16 <david-lyle> I will consult with infra
16:57:39 <deshipu> david-lyle: there is very nice pastebin for any file software written in python, bepasted
16:57:52 <deshipu> david-lyle: maybe we could get them to host it somewhere
16:58:07 <jpich> #topic Open discussion
16:58:12 <jpich> 2 minutes for any additional concerns
16:58:25 <clu_> how long can we expect the bp approval process to take?
16:58:25 <TravT> david-lyle: thanks.  let me know... maybe even if there were a horizon-specs repo, it could host mockups.
16:58:47 <clu_> there's no sort-by-date...
16:59:05 <bradjones> I'd like to draw attention to review.openstack.org/128295/ it fixes an upstream bug in UT with a KeyError
16:59:11 <jtomasek> jpich: david-lyle: any leads to when will master be open for Kilo patches?
16:59:14 <clu_> so some may just sit there...
16:59:14 <jpich> Yeah, and there's no notifications for new blueprints, it's awkward...
16:59:20 <jpich> jtomasek: It has been since the first RC :)
16:59:57 <akrivoka> bps are really awkward in many ways :\
17:00:01 <jpich> bradjones: Do you know how come the initial patch was able to merge?
17:00:04 <jtomasek> jpich: ok, I'd like to point out a few patches that fix most of the layout problems in Horizon and are in review proces for quite some time
17:00:15 <jtomasek> Fixed display issues on Details Page caused by bootstrap 3 update https://review.openstack.org/#/c/115693/
17:00:15 <jtomasek> Modal fixes - 2 columns, padding, header, membership https://review.openstack.org/#/c/113519/
17:00:15 <jtomasek> Restyled topbar to resemble UX guidelines https://review.openstack.org/#/c/117118/
17:00:15 <jtomasek> Restyled sidebar to resemble UX guidelines https://review.openstack.org/#/c/126289/
17:00:22 <jpich> Ok, let's continue in openstack-horizon or on the ML folks, that's our time for this week
17:00:28 <jpich> Thanks
17:00:30 <jpich> #endmeeting