20:00:57 <david-lyle> #startmeeting Horizon
20:00:58 <openstack> Meeting started Wed Jun  3 20:00:57 2015 UTC and is due to finish in 60 minutes.  The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:59 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:01:02 <openstack> The meeting name has been set to 'horizon'
20:01:28 <woodm19791> Good afternoon!
20:01:31 <TravT> o/
20:01:32 <matt-borland> o/
20:01:34 <lhcheng> o/
20:01:34 <r1chardj0n3s> o/
20:01:35 <neillc> o/
20:01:35 <bpokorny> Hi all
20:01:37 <mrunge> hello there
20:01:39 <doug-fish> greetings
20:01:40 <clu_> hi!
20:01:41 <ducttape_> yo
20:01:43 <pauloewerton> hi there
20:02:46 <asahlin> hello
20:02:49 <rhagarty> howdy
20:03:11 <david-lyle> hello everyone
20:03:16 <hurgleburgler1> Hello
20:03:21 <tqtran> 0/
20:03:24 <david-lyle> as the horizon turns episode 2000
20:03:52 <lhcheng> https://wiki.openstack.org/wiki/Meetings/Horizon
20:03:53 <asahlin> lol, what season?
20:04:26 <david-lyle> so the agenda for today can be found on the link lhcheng provided
20:04:29 <david-lyle> thanks lhcheng
20:04:32 <lhcheng> wow, we have a lot in agenda today
20:04:45 <david-lyle> crazy, right?
20:04:46 <r1chardj0n3s> ok, who added all that stuff to the agenda? :)
20:05:15 <lhcheng> lol
20:05:26 <TravT> what's wrong with people?
20:05:39 <david-lyle> I have an extra topic that I'm going to start with
20:05:41 <david-lyle> just cause
20:05:53 <woodm19791> #drunkwithpower
20:05:54 <r1chardj0n3s> Surprise Topic!
20:06:31 <neillc> Nobody expects the surprise topic
20:06:45 <david-lyle> OpenStack is moving from a shake the magic 8-ball approach to release planning to reporting what we did
20:06:46 <tqtran> he's holding out the anticipation
20:06:52 <hurgleburgler1> darn, forgot to add my topic!!
20:07:07 <david-lyle> hurgleburgler1: go ahead
20:07:13 <david-lyle> on the wiki
20:07:24 <tqtran> david-lyle: what does that mean?
20:07:44 <r1chardj0n3s> (oh good, it wasn't just me)
20:07:47 <mrunge> we'll probably have fewer blueprints completed
20:07:48 <doug-fish> so ... no planning? just reporting results?
20:07:54 <david-lyle> so we used to slate all these blueprints for milestones and hope by some crazy chance it might actually land there
20:08:07 <doug-fish> lol - I thought that was the old process!
20:08:13 <doug-fish> oh wait
20:08:26 * doug-fish stops laughing and reads
20:08:33 <david-lyle> the success rate was quite low, not surprisingly
20:08:58 <david-lyle> At the summit we actually spent some time lining up what the priorities were
20:09:19 <david-lyle> they were collated here #link https://etherpad.openstack.org/p/YVR-horizon-liberty-priorities
20:09:41 <david-lyle> we also talked about having a page that showed review priority
20:09:58 <david-lyle> My suggestion would be just to use the etherpad to track the release
20:10:14 <david-lyle> rather than on a per milestone basis
20:10:22 <r1chardj0n3s> pile of poo here we come
20:10:33 <david-lyle> link the bps and bugs on that etherpad directly
20:10:43 <david-lyle> I've found two things over time
20:11:08 <david-lyle> 1) reviewers don't actually look at the milestone list to prioritize blueprint reviews
20:11:34 <david-lyle> 2) moving around random things in launchpad is not highly effective.
20:11:44 <david-lyle> Thoughts?
20:11:52 <doug-fish> agreed
20:12:00 <TravT> Well, I liked using etherpad for launch instance last release
20:12:10 <r1chardj0n3s> I think it's worth a shot, perhaps we just need to remind folks occasionally
20:12:21 <doug-fish> The etherpad is probably clearer and simpler - and no reason to think it will be less effective
20:12:24 <r1chardj0n3s> also, could people with testing patches please update https://etherpad.openstack.org/p/june_2015_testing_push :)
20:12:34 <matt-borland> sure thing
20:12:40 <r1chardj0n3s> (that's linked from the priorities page)
20:12:47 <tqtran> yeah, im fine with either approach, but having etherpad helps a bit. my only concern is when it starts to grow and becomes a tangle of mess
20:13:09 <r1chardj0n3s> at least - unlike blueprints - you can have an actual conversation in an etherpad
20:13:09 <david-lyle> I had the same issue with launchpad
20:13:12 <mrunge> I would expect to become a mess more sooner than later
20:13:28 <mrunge> there is no sync back to etherpad
20:13:31 <r1chardj0n3s> ... or are we still expecting to use blueprints?
20:13:36 <mrunge> (once patch is reviewed)
20:13:44 <r1chardj0n3s> mrunge: good point
20:13:47 <woodm19791> mrunge's point is the biggest
20:13:48 <r1chardj0n3s> it's very manual
20:13:51 <tqtran> i think a hybrid approach might make sense, have your bp in launchpad, and link it in etherpad
20:14:06 <mrunge> tqtran, that's double effort then
20:14:18 <bpokorny> Anyone can modify the etherpad, so we'd have to watch out for who is changing/deleting things.
20:14:24 <tqtran> this way, the details and stuff is out of the way and reviewers can just focus on line items they need to knock down
20:14:26 * david-lyle pulls back the curtain a bit
20:14:35 <david-lyle> I update the bp status manually in launchpad too
20:14:40 <david-lyle> it's not automated
20:14:48 <TravT> Well, one thing i like about the current etherpad linked here is that we can see themes.
20:14:57 <mrunge> aand, easiest way to sneak own patches to top prio ;-)
20:15:06 <bpokorny> Yes!
20:15:12 <TravT> true, no governance mrunge.
20:15:55 <TravT> etherpad does have a timeslider...
20:15:55 <TravT> https://etherpad.openstack.org/p/YVR-horizon-liberty-priorities/timeslider
20:16:03 <TravT> but, no way to enforce changes
20:16:28 <mrunge> TravT, true, so revisions are recorded.
20:16:29 <doug-fish> I like the etherpad idea - maybe we revisit if we find it's being abused/vandalized?
20:16:37 <david-lyle> if you all like launchpad, we can stick with that, I just waste a lot of time on managing those lists
20:16:44 <TravT> although, it doesn't seem to work right (at least playing with it).
20:16:56 <david-lyle> and get very little out of it
20:16:56 <TravT> wait, now it is working.
20:17:06 <r1chardj0n3s> woah, woah, who alleged I like launchpad?
20:17:06 <david-lyle> TravT: you have to save revisions as well
20:17:49 <david-lyle> I would also hazard to guess 75% of what's in the etherpad has neither a bp or bug in launchpad right now
20:18:02 * david-lyle points finger of guilt around the room
20:18:09 <r1chardj0n3s> there should be more bugs linked, yes. I will do one now
20:18:15 <lhcheng> Doesn't hurt to try it, I think we should just go for it.
20:18:24 <lhcheng> It is better than what we do now.
20:18:25 <mrunge> david-lyle, that'd be another reason not to use an etherpad
20:18:29 * ducttape_ points a different finger back at david-lyle  ;)
20:18:32 <tqtran> yeah, lets try it and see how well it works out
20:18:38 <lhcheng> We can revisit it after a few months if it working or not.
20:18:52 * TravT feeling guilty
20:19:21 <david-lyle> the thing is we have an idea what we want to accomplish, it's there in the etherpad, it's more about fleshing out the details
20:19:27 <TravT> so, process wise, last release, we'd just use the cross out thing when something got done.
20:19:28 <mrunge> I'd say: let's try and revisit next week
20:19:42 <david-lyle> TravT: +1
20:19:51 <mrunge> TravT, but you were under 10 people
20:19:57 <TravT> yep, agreed
20:20:06 <mrunge> and you met each day on that pad
20:20:16 <TravT> i am a little worried having all blueprints linked here will get messy.
20:20:37 <david-lyle> it's all messy
20:20:40 <TravT> maybe we should copy this etherpad to somewhere for reference.
20:20:48 <asahlin> +1
20:20:53 <mrunge> we have a little over 240 blueprints currently
20:21:02 <TravT> since it is more thematic at the moment without all the details.
20:21:02 <r1chardj0n3s> wow
20:21:08 <david-lyle> mrunge: and how many of those are legit?
20:21:10 <mrunge> not to speak about bugs or patchsets
20:21:23 <mrunge> david-lyle, right. maybe 50?
20:21:39 <woodm19791> We should clean that out then.
20:21:51 <hurgleburgler> Legit as in Accepted to work on?
20:21:56 <woodm19791> Switching to etherpad won't help that.  :-)
20:22:23 <mrunge> hurgleburgler, in the past, a not accepted blueprint was not a blocker for review
20:22:33 <mrunge> we might think again about that
20:23:04 <david-lyle> lets take the week to consider and move on to the scheduled items
20:23:05 <mrunge> we could simply drop unaccepted blueprints after... 1 month?
20:23:43 <david-lyle> honestly I find none of the options overly appealing
20:23:44 <TravT> if you are going to do that, there probably should be a weekly blueprint review meeting among drivers or something like that
20:24:25 <hurgleburgler> not a bad idea, but more overhead for the bp reviewers
20:24:28 <mrunge> +1 for discussing blueprints
20:24:35 <david-lyle> +1 for more overhead
20:24:39 <david-lyle> oh wait
20:24:58 <david-lyle> I think it makes sense though
20:25:11 <david-lyle> let's take this offline for now
20:25:20 <david-lyle> #topic Angular translation, continue using tweaked Django or switch to Babel? (tqtran/r1chardj0n3s)
20:25:23 <hurgleburgler> I would like it, cause I honestly never know who I need to bribe for a bp
20:25:38 <david-lyle> hurgleburgler: when in doubt, all
20:25:42 <r1chardj0n3s> ooh, translation cage match
20:25:51 <david-lyle> and in this corner
20:25:52 <TravT> we all take bribes around here.
20:26:05 <r1chardj0n3s> oh man, I gotta get in on this bribery action
20:26:36 * TravT digresses - let tqtran have the floor with r1chardj0n3s
20:27:00 <tqtran> ok next topic
20:27:15 <tqtran> so basically, we have settled on 2 different ways to translate angular static html
20:27:30 <r1chardj0n3s> tqtran: got the bp link? I've lost it because launchpad
20:27:51 <tqtran> https://blueprints.launchpad.net/horizon/+spec/angular-translate-makemessages
20:28:02 <tqtran> its quite a long read, but it outlines my thought process
20:28:28 <tqtran> and list alternatives with pros/cons
20:28:53 <tqtran> right now, i think its agreed that we will go with approach 3 since its something angular-gettext supports
20:29:05 <r1chardj0n3s> (for markup)
20:29:15 <tqtran> yes for markup and future compatibility
20:29:35 <tqtran> the part where we need a discussion around is the translation extraction
20:29:41 <tqtran> which is listed as options in the bp
20:30:14 <tqtran> richard is advocating for option 2 and using babel for extraction
20:30:19 <tqtran> his reasons are as follows:
20:30:29 <tqtran> 1) use angular-gettext now, not later
20:30:34 <r1chardj0n3s> (except that option 2 I'm advocating is less sucky than the BP implies ;)
20:30:37 <tqtran> 2) use babel to extract messages from django templates and angular templates, exrtacting into the current messages catalog
20:30:44 <tqtran> 3) use django's translation system as we do today over the top of that messages catalog
20:30:51 <tqtran> 4) need to use babeldjango plugin
20:31:15 <tqtran> and today, i found out that we already have babel in the requirements and are using it to some degree
20:31:22 <r1chardj0n3s> oh!
20:31:30 <tqtran> thanks to doug-fish for pointing that out
20:31:37 <doug-fish> sure, np
20:31:50 <tqtran> i'm advocating for option 3
20:32:28 <tqtran> 1) use angular-gettext as well
20:32:37 <tqtran> 2) extend a django makemessages command
20:33:12 <tqtran> 3) create a temp file that we dump all static translatable html strings into
20:33:12 <tqtran> 4) let django do its work
20:33:12 <r1chardj0n3s> (with a hack that leaves a horrid legacy for future horizon devs)
20:33:12 <tqtran> 5) delete file
20:33:13 <tqtran> its not a hack lol
20:33:16 <r1chardj0n3s> :)
20:33:31 <tqtran> https://review.openstack.org/#/c/187321/
20:33:45 <tqtran> here is the patch for the single translation version, with interpolation and plural added in later patches
20:33:47 <r1chardj0n3s> it messes with the workings of makemessage in a way that is potentially fragile and not clear to future devs
20:34:26 <tqtran> it doesn't mess with makemessages, it adds a temporary file that contains all extracted strings so that django can do its thing
20:34:30 <r1chardj0n3s> that's my reasoning to use a tool that's designed to be extended and already handles django templates
20:34:45 <r1chardj0n3s> it replaces makemessages command by overridding the built in class
20:34:58 <r1chardj0n3s> monkeypatching, yeah :)
20:35:29 <tqtran> so for me, i dont understand babel enough to know how the plugin would fit into horizon
20:36:01 <tqtran> if you can put up a patch and show how it would all fit together, i will be more than happy to create a plugin for it
20:36:25 * TravT gets some popcorn
20:36:31 <tqtran> thoughts?
20:36:54 <david-lyle> I'd like to see an investigation for using babel, the temp file seems "interesting"
20:37:18 <neillc> I think a look at babel before writing our own tool would be a good idea
20:37:24 <r1chardj0n3s> babel plugins are just a python file that knows how to extract messages from text
20:37:31 <tqtran> yeah, its a work around since we don't have access to override the makemessages class in django
20:37:39 <neillc> I don't understand either option well enough yet to offer a definitive opinion
20:37:47 <TravT> i'd agree at this point that seeing an alternate patch with babel would help to clarify.
20:37:48 <r1chardj0n3s> eg here's the existing django one http://babel.edgewall.org/browser/contrib/django/babeldjango/extract.py
20:37:49 <david-lyle> but if that goes poorly we have a fallback
20:37:59 <r1chardj0n3s> which uses django's built-in parser to pull messages from its templates
20:38:32 <r1chardj0n3s> all we need to do is use tqtran's regex he's already written for his tempfile approach and put it in an extract_angular function and register it with babel
20:38:33 <tqtran> but how would we use that in horizon today?
20:38:49 <r1chardj0n3s> and then instead of invoking manage.py makemessages we invoke babel
20:38:55 <r1chardj0n3s> that's it, afaict
20:39:15 <r1chardj0n3s> the message catalog produced is the same as today
20:39:18 <tqtran> so we invoke bable for the angular stuff and leave the js/py/templates to django?
20:39:21 <r1chardj0n3s> so all the downstream is the same as today
20:39:39 <r1chardj0n3s> tqtran: no, babel extracts the django stuff as well using its plugin. one command
20:39:56 <tqtran> so we need 2 plugins, a django one and an angular one
20:40:04 <TravT> in either case, would the HTML markup look similar (attempt to be forward compatible with angular gettext)?
20:40:08 <david-lyle> ok, r1chardj0n3s can you work on a prototype for that? or someone else? we have a lot to get through today
20:40:10 <r1chardj0n3s> yes, the django one that already exists
20:40:16 <tqtran> TravT:  yes it would be
20:40:25 <r1chardj0n3s> david-lyle: I can work on it, but I won't get to it until later next week
20:40:54 <tqtran> ok cool, lets wait on the prototype, we can work on it together if you like r1chardj0n3s
20:41:01 <r1chardj0n3s> tqtran: ok, thanks
20:41:03 <david-lyle> that would be ideal
20:41:09 <david-lyle> thanks!
20:41:29 <david-lyle> #topic Horizon Keystone to Keystone Federation - current status and future directions (pauloewerton)
20:41:48 <stevemar> david-lyle, of course you bring this up as i'm getting ready to leave
20:41:58 <david-lyle> as I planned it
20:42:12 <david-lyle> should we punt a week on this topic stevemar
20:42:13 <david-lyle> ?
20:43:06 <doug-fish> (maybe he left?)
20:43:27 <david-lyle> ok, reshuffling deck, slipping K2K to the bottom
20:43:35 <pauloewerton> so I know that this work is probably blocked because of that keystoneclient auth plugin dependency
20:44:17 <doug-fish> it's true - it looks like they have some nice, not-yet-implemented ideas https://review.openstack.org/#/c/172155/
20:44:24 <stevemar> i trust doug-fish to know everything
20:44:38 <doug-fish> stevemar: let me know how that works out for you!
20:44:53 <stevemar> doug-fish, it's been working well
20:45:02 <doug-fish> "Oh yeah, the keystone guys will do that"
20:45:21 <david-lyle> I think this work can't make much progress until we have something concrete to build on, I know doug-fish has a patch up
20:45:36 <david-lyle> but we can't do much without the dependencies merged
20:45:37 <pauloewerton> I see
20:45:53 <doug-fish> true.  Here's the patch:  https://review.openstack.org/#/c/159910/
20:45:59 <stevemar> we're trying to figure out our auth plugins story on the keystone side
20:46:01 <david-lyle> We can discuss in the horizon room as well
20:46:15 <stevemar> we're trying to put all auth plugins into a new repo (keystoneauth)
20:46:19 <stevemar> and have that ready for L1
20:46:23 <pauloewerton> doug-fish, david-lyle ok, thanks
20:46:32 <david-lyle> I think there is visibility, we just need to hash out the details
20:46:32 <pauloewerton> just a question
20:46:46 <stevemar> (the k2k plugin should go into that repo)
20:47:12 <pauloewerton> once this work in django-openstack-auth is finished
20:47:46 <pauloewerton> would we have something like a service provider selector in horizon nav-bar, for instance?
20:48:15 <david-lyle> I think this is a discussion that should happen on the blueprint not in the team meeting
20:48:33 <doug-fish> https://blueprints.launchpad.net/horizon/+spec/k2k-federation
20:48:52 <pauloewerton> david-lyle, doug-fish, sure.
20:49:01 <david-lyle> I really need to get to the next two topics and time is dwindling
20:49:12 <david-lyle> pauloewerton: we'll work it out in the bp
20:49:21 <david-lyle> #topic A Quick searchlight update (TravT)
20:49:23 <TravT> so, at the summit, we showed project Searchlight in a fishbowl for both Glance and Horizon.  since then, have been doing the work to get it set up as a new project separated from Glance.
20:49:34 <TravT> today we put up for a vote to be included in the official Big Tent project list #link https://review.openstack.org/#/c/188014/
20:49:47 <TravT> the wiki is here: #link https://wiki.openstack.org/wiki/Searchlight
20:50:07 <TravT> horizon is intended to be a (the?) primary consumer of the project initially, so wanted to give an update and also set some context for discussion on the mid-cycle meet-up.
20:50:31 <david-lyle> #topic Midcycle meetup (david-lyle, TravT)
20:51:01 <david-lyle> so a mid-cycle meetup was strongly requested at the summit
20:51:24 <david-lyle> we've had several offers for hosts, thank you all
20:52:09 <david-lyle> One potential set of dates is July 8-10 in Boston, MA. This would coincide with the glance and searchlight mid-cycles as well
20:52:27 <TravT> a number of Glance cores are also going to be working on searchlight, so co-locating with them helps get approval.
20:52:31 <david-lyle> as there is an interest overlap that could potentially very productive
20:52:59 <david-lyle> I know of at least one person these dates don't work for
20:53:11 <hurgleburgler> Ah doh!  That's during comiccon
20:53:21 <woodm19791> :-)
20:53:28 <hurgleburgler> sorry, my priority will always be batman
20:53:37 <r1chardj0n3s> :)
20:53:44 <david-lyle> So there are a couple of other options
20:54:15 <david-lyle> July 20-22 or July 21-23 in Fort Collins, CO
20:54:17 <hurgleburgler> if you did it in san diego, that'd be an interesting mid-cycle meet up :-D
20:54:39 <hurgleburgler> the whole city becomes mardi gras for nerds
20:54:57 <david-lyle> or another undetermined data in San Jose, CA or maybe Portland, OR
20:55:11 <david-lyle> thoughts on people's availability?
20:55:14 <TravT> we've (including nikhil_k - Glance PTL) put together an initial spreadsheet to help bring out ideas and for voting.
20:55:22 <TravT> #link https://docs.google.com/spreadsheets/d/1w0eI6SPCA2IrOyHiEYC2uDO3fbYGzahZRUQSva0UD3Y/edit#gid=0
20:55:32 <TravT> room idea are on top, but if you scroll down
20:56:13 <TravT> you can fill in your availability in the "Voting" section
20:56:59 <tqtran> cool will fill that in
20:57:08 <david-lyle> so if you can fill that in, we can close quickly
20:57:40 <david-lyle> if we went with the early dates, people will need to start booking travel in the next few days
20:57:59 <david-lyle> so please fill it in
20:58:01 * ducttape_ enjoys watching the names fill in.  mesmorizing
20:58:07 <david-lyle> magic
20:58:26 <david-lyle> last two topics real quick
20:58:33 <david-lyle> #topic Adding events information to instance audit tab https://blueprints.launchpad.net/horizon/+spec/instance-events (peristeri)
20:58:44 <david-lyle> peristeri: wasn't able to attend
20:58:52 <david-lyle> but he want to point out the bp
20:58:58 <david-lyle> there's a demo video as well
20:59:08 <TravT> i just watched it
20:59:22 <TravT> pretty interesting.
20:59:31 <david-lyle> I think it makes perfect sense, may need some tweaking, but overall I have no problem with the concept
20:59:32 <TravT> and some discussion should happen in context of searchlight as well
20:59:37 <david-lyle> TravT: sure
20:59:54 <TravT> i don't like that it is new functionality in legacy code, but that is a detail.
21:00:21 <david-lyle> and finally
21:00:24 <tqtran> lol yeah, that means more work for us when we do decide to convert
21:00:38 <david-lyle> #topic Themeing Additions (hurgleburgler)
21:00:44 <david-lyle> tqtran: WHEN
21:00:48 <hurgleburgler> yay!
21:00:59 <tqtran> when the fat lady sings
21:01:11 <hurgleburgler> I've been doing some work using the new themeing functionality and hit a few bugs as well as some things that can be improved quite a bit
21:01:15 <r1chardj0n3s> I think she's sung, it's time ;)
21:01:18 <hurgleburgler> if I can get some eyes on the following
21:01:18 <hurgleburgler> https://blueprints.launchpad.net/horizon/+spec/horizon-theme-templates
21:01:18 <hurgleburgler> https://review.openstack.org/#/c/188162/
21:01:18 <hurgleburgler> https://blueprints.launchpad.net/horizon/+spec/bootstrap-theme-preview
21:01:19 <hurgleburgler> https://review.openstack.org/#/c/187818/
21:01:21 <hurgleburgler> https://bugs.launchpad.net/horizon/+bug/1457188
21:01:21 <hurgleburgler> https://review.openstack.org/#/c/187797/
21:01:22 <openstack> Launchpad bug 1457188 in OpenStack Dashboard (Horizon) "Bootstrap variables overrides should only be default values" [Medium,In progress] - Assigned to Diana Whitten (hurgleburgler)
21:01:28 <r1chardj0n3s> woah, link!
21:01:30 <tqtran> dang calm down on the links
21:02:04 <david-lyle> these are patches to make the theming we based of hurgleburgler earlier work, far more useful
21:02:05 <mrunge> there's no need to click them all...
21:02:09 <david-lyle> please check it out
21:02:15 <david-lyle> s/it/them
21:02:28 <hurgleburgler> Thanks!
21:02:37 <david-lyle> ok, we're over. followup can be in the horizon room
21:02:44 <david-lyle> Thanks everyone!
21:02:46 <david-lyle> #endmeeting