15:00:11 <krotscheck> #startmeeting storyboard
15:00:12 <openstack> Meeting started Mon Apr 21 15:00:11 2014 UTC and is due to finish in 60 minutes.  The chair is krotscheck. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:15 <krotscheck> Who’s here?
15:00:15 <openstack> The meeting name has been set to 'storyboard'
15:00:18 <SergeyLukjanov> o/
15:00:35 <NikitaKonovalov> hi!
15:00:38 <krotscheck> Hey hey!
15:00:44 <krotscheck> Ok, so ttx is on vacation
15:00:55 <krotscheck> Probably at the bottom of a wine barrel ;)
15:01:01 * krotscheck envies the french sometimes.
15:01:31 <krotscheck> #topic ElasticSearch vs. Sphinx (krotscheck)
15:01:49 <krotscheck> No progress. Mostly because there hasn’t actually been any need for a fulltext style search yet.
15:01:58 <ruhe> o/
15:02:01 <NikitaKonovalov> looks like we are not considering Sphinx anymore
15:02:07 <NikitaKonovalov> right?
15:02:34 <krotscheck> NikitaKonovalov: Actually, last meeting that came back up - jeblair pointed out that it takes at least .5 FTE’s to maintain an ES cluster.
15:02:50 <krotscheck> So there’s the question of whether we really want to overkill on the search.
15:03:05 <ruhe> there also were debates in the infra channel about ES vs Lucene
15:03:33 <NikitaKonovalov> ruhe: isn't ES running on top of Lucene?
15:03:45 <krotscheck> So I suggest we just punt the discussion another week or until we actually have feature need.
15:03:59 <NikitaKonovalov> krotscheck: agree
15:04:06 <krotscheck> Anyone disagree?
15:04:17 <jeblair> i do not disagree
15:04:17 <ruhe> NikitaKonovalov: it is, but sdague had an idea to use Lucene directly. which i do not support since lucene is too low-level
15:04:32 <jeblair> solr might be the middle ground there
15:04:45 <krotscheck> Alright, next topic.
15:05:01 <krotscheck> #topic Migrating to Task Timeline (NikitaKonovalov )
15:05:22 <NikitaKonovalov> ok, there are 2 CRs right now
15:05:38 <jeblair> krotscheck: (agenda link?)
15:05:46 <krotscheck> jeblair: https://wiki.openstack.org/wiki/StoryBoard
15:05:51 <NikitaKonovalov> one for the server side which I would like to talk more about
15:06:02 <krotscheck> Go for it.
15:06:09 <NikitaKonovalov> And the client which is actually adding one directive
15:07:02 <NikitaKonovalov> It'll be a bit tricky to migrate the db without losing the existing comments
15:07:40 <NikitaKonovalov> Now the comments are attached to the events, which should be created somehow
15:08:18 <krotscheck> I’ve had trouble with running data migrations as well. It’s super easy for schemae, but trying to run both Alembic and SQLAlchemy led me down a path of annoying transaction conflicts.
15:08:22 <jeblair> NikitaKonovalov: alembic lets you write python migration methods, right?  so can we iterate over existing comments and create event records?
15:09:23 <jeblair> i think this is going to be the sort of thing we're going to have to get good at because we're running in CD
15:09:50 <krotscheck> Right.
15:09:57 <NikitaKonovalov> jeblair: ok, I thought alembic was about manipulating with schema
15:10:15 <NikitaKonovalov> but if it work with data, It's great
15:10:22 <mordred> yah - you can def. do data
15:10:54 <jeblair> i think people prefer to use the db when possible, for speed; but when not possible, you can still use python
15:10:55 <NikitaKonovalov> the I'll put server-side patch to WIP
15:11:11 <NikitaKonovalov> then*
15:11:16 <krotscheck> kk
15:11:32 <krotscheck> We should code review that.
15:12:05 <krotscheck> I’ll take a look today after my morning meetings, anyone else want to give an eye-over onto what NikitaKonovalov is working on?
15:12:39 <NikitaKonovalov> here are the changes https://review.openstack.org/#/q/storyboard+AND+topic:timeline,n,z
15:12:42 <mordred> NikitaKonovalov: neutron also uses alembic already - so looking in neutron/db/migration/alembic_migrations/versions/ in the neutron tree might be a decent place for inspiration
15:12:55 <krotscheck> #topic Search/Filter Stories
15:12:58 <NikitaKonovalov> mordred: thanks
15:13:00 <krotscheck> Patches are up for that.
15:13:28 <krotscheck> I’ll leave this on the agenda until we get it reviewed.
15:13:53 <krotscheck> #topic Task assignment
15:14:20 <krotscheck> The UI for task assignment is about 75% done, I hit a wall when I realized that I have to provide a not-shitty way of searching for users to assign things to.
15:14:27 <sdague> (sorry for late comment) however you can do interesting changes in db with data, we've done them a bunch in nova to change data formats. I'd be happy to help review those
15:14:43 <krotscheck> So that’s my big task this week.
15:15:30 <mordred> krotscheck: oy. searching for users is apparently hard, given how much the UI for it kinda blows in gerrit
15:15:55 <sdague> how big is the "all user" list?
15:16:14 <krotscheck> sdague: I’ve heard “6000” bantied around
15:16:20 <krotscheck> For the UI changes: https://review.openstack.org/#/c/88707/
15:16:32 <sdague> if it's not prohibitively big we could ghetto through it with chosen being applied to that field
15:16:32 <mordred> sdague: between 2k and 10k ish - there are 2000 devs who have written code, there are 10k foundation members
15:17:03 <sdague> mordred: ok, but that's not a now thing right? that's an in 6 months thing.
15:17:35 <NikitaKonovalov> right now there are 20
15:17:36 <jeblair> sdague: can you translate "ghetto through it with chosen" for those who don't speak jivequery? :)
15:17:41 <sdague> :)
15:18:00 * krotscheck shrugs.
15:18:02 <mordred> right now there are 43 entries in the user table
15:18:05 <sdague> chosen is this nice jquery plugin for select fields - http://harvesthq.github.io/chosen/
15:18:07 <krotscheck> I’m just going to do a LIKE filter.
15:18:16 <krotscheck> on the API
15:18:38 <krotscheck> And then do a typeahead dropdown very similar to what sdague is suggesting.
15:18:48 <krotscheck> But “
15:19:05 <mordred> do you match on username, fullname, email or all of the above?
15:19:16 <krotscheck> mordred: That’s where the ghetto comes in.
15:19:24 <mordred> yay ghetto!
15:19:28 <sdague> so with only 43 users, you could just put them all in the list and let chosen select client side with type ahead
15:20:18 <sdague> which wouldn't scale to 6000, but could be a fast path to be able to punt that problem till later. Though if krotscheck has a better option that's quick, that's cool too
15:20:26 <jeblair> i don't want this to fall over as soon as we actually use it for openstack though, so thinking ahead for up to 10k users is probably a good idea
15:20:49 <jeblair> if we do something that only scales to 100 (to get something on the page), we will need to swing back around and do it again fairly soon
15:20:53 <mordred> btw - we have one entry in the users table with no username. I'm not sure how that's possible...
15:20:57 <NikitaKonovalov> \me just found that an unauthorized request to /users returns non-public fields which it sholdn't
15:21:04 <mordred> NikitaKonovalov: woot!
15:21:15 <mordred> NikitaKonovalov wins for the day
15:21:20 <jeblair> our first cve? ;)
15:21:26 <jeblair> (just kidding)
15:21:36 <krotscheck> Goody goody, who wants it!
15:22:56 <krotscheck> Alright,
15:23:18 <krotscheck> #action NikitaKonovalov File story for security issue on users endpoint.
15:23:37 <krotscheck> #topic Stupid Task Priority (via enums)
15:24:02 <krotscheck> Does anyone feel like they have the bandwidth to tackle this this week?
15:24:34 <krotscheck> I _may_ be able to.
15:24:56 <krotscheck> But I’m not comfortable committing to it because of the size of the patch chain I have to maintain on Gerrit.
15:25:31 <NikitaKonovalov> It looks like it's a simple migration to server side, so I can take that
15:26:05 <krotscheck> NikitaKonovalov: Awesome. Adding a simple enum dropdown is also not a problem from my perspective (unlike assigning users turned out to be)
15:26:21 <NikitaKonovalov> The only thing is what are the enum values?
15:26:24 <krotscheck> Do we know what the priorities are going to start up being?
15:26:26 <krotscheck> Oh.
15:26:28 <krotscheck> Right, brainwave.
15:26:51 <krotscheck> Anyone have a disagreement with “Low, Medium, High, RTFN”?
15:27:11 <jeblair> i think we can get by with low,med,high
15:27:30 * mordred kinda likes the RTFN status - hasn't seen that in other bug trackers, its a differentiator
15:27:47 <mordred> s/status/priority/
15:27:50 <krotscheck> jeblair: ….because from a practical standpoint, anything that’s critical at the RTFN level people are going to be working on immediately anyway?
15:28:34 <jeblair> krotscheck: yeah, and i don't want to over-think it or add too many status fields since we know this has a limited life
15:28:41 <krotscheck> Point
15:28:45 <krotscheck> Any disagreements?
15:28:50 * krotscheck agrees
15:29:13 * NikitaKonovalov also
15:29:35 <krotscheck> #agreed Initial task priorities will be Low/Medium/High because we don’t expect this implementation to live for very long.
15:29:52 <krotscheck> Wait, is this a story or a task level field?
15:31:20 <jeblair> i think task?
15:31:22 <krotscheck> Given that we don’t really have a way of searching tasks right now, and this is a short lived implementation, shall we put it on status?
15:31:25 <krotscheck> Sorry
15:31:26 <krotscheck> story?
15:32:07 <krotscheck> I feel that task priority won’t be useful unless we can sort by it.
15:32:35 <mordred> yeah - I feel like we're more likely to end up with more smaller stories with fewer tasks, rather than large complex stories with tons of tasks - so I'd be fine with it on story
15:32:36 <jeblair> that's true; but we do need that anyway, right?
15:32:37 <krotscheck> So either we’ll have to put it on task and roll it up to stories much like we’re doing task status, or we put it directly on the story.
15:33:21 <jeblair> i think it's worth putting it on task
15:34:12 * krotscheck is ambivalent.
15:34:19 <NikitaKonovalov> So as I understand the task is a thing that actually says what should be done to comlete a story
15:34:20 <jeblair> whatever we end up with is likely to be task-oriented, so we may as well start thinking of things in that way
15:34:33 <NikitaKonovalov> that's why it should have a priority
15:35:03 <jeblair> yep
15:35:32 <krotscheck> NikitaKonovalov: Sounds like you’re more in favor of Task, yes?
15:35:38 <NikitaKonovalov> yes
15:36:11 <krotscheck> I’m ok with that.
15:36:24 <krotscheck> Ok, so three statuses on tasks. Done.
15:36:58 <krotscheck> #action NikitaKonovalov Work on task priority on the API layer.
15:37:08 <krotscheck> #action krotscheck Work on task priority on the client layer.
15:37:26 <krotscheck> #topic Stories with all landed tasks should not be in primary UI filter.
15:37:40 <krotscheck> Patch up. Feels like a repeat of Search/Filter
15:37:47 <krotscheck> Tasks can’t be edited: Done.
15:37:58 <krotscheck> Support fro Project Groups, outstanding.
15:38:05 <krotscheck> #topic Support for project groups
15:38:25 <krotscheck> NikitaKonovalov, since you and I are really the only coders on this, do you have time this week to work on this?
15:38:45 <krotscheck> (My guess is timeline will keep you busy)
15:39:43 <NikitaKonovalov> krotscheck: I think so, but the missing thing is a Controller so if there is time by the end of the week, I'll bring it up
15:40:22 <krotscheck> Ok, so let’s not commit to it. I don’t want you to rush timeline, but let’s try to get events/story history solid and baked this week.
15:40:46 <krotscheck> Which means I need to step up code reviewing.
15:41:01 <krotscheck> #topic Story Activity
15:41:12 <krotscheck> In progress, feels like a repeat of Task Timeline
15:41:17 <NikitaKonovalov> it is
15:41:34 <krotscheck> #topic Story and Task Authors
15:41:41 <krotscheck> Patch for story authors is up.
15:42:06 <krotscheck> Patch for task assignees is in progress.
15:42:18 * mordred has some things up in his code review window currently
15:42:27 <mordred> will try to get some reviewing done this morning
15:42:38 <krotscheck> We’re starting to run out of screen realestate on the task list, but I can see if I can expand on that.
15:43:27 <krotscheck> Honestly, I want to break this up into three different issues.
15:43:37 <krotscheck> How about I do that.
15:43:51 <krotscheck> #action krotscheck Split author and assignee into three different stories.
15:44:15 <krotscheck> #topic Intelligent Priorities.
15:44:43 <krotscheck> Ok, design discussion: What problems do you imagine Intelligent Priorities will solve for you?
15:44:50 <krotscheck> (no implementation talk yet, please)
15:45:02 <jeblair> maybe we should defer this until ttx is back?
15:45:17 <krotscheck> I’m ok with that.
15:46:02 <krotscheck> Alright, does anyone else have any issues skipping over the rest of the agenda until ttx is back?
15:46:06 <krotscheck> It’s one more item.
15:46:38 <mordred> yeah- might also be a ttx topic - although I for one am in favor of subscription of things
15:46:46 <krotscheck> So am I.
15:47:06 <krotscheck> We can start working on some of the pieces that’ll be required, say, sending summary emails.
15:48:11 <krotscheck> mordred: Do you want to talk about subscription?
15:49:22 <krotscheck> #topic Open Discussion
15:49:27 <krotscheck> Anything else?
15:49:33 <krotscheck> I’m getting a lot of silence here.
15:50:04 * mordred is good for the day
15:50:16 <krotscheck> We’re way behind on code reviews. I was promised that some of Infra Core would help contribute just over a week ago. and that I shouldn’t be +2ing my own code anymore. So, what’s up with that?
15:50:31 <NikitaKonovalov> back to subscriptions, the thing I want is to see all the stuff that I'm subscribed too
15:51:02 <krotscheck> NikitaKonovalov: So emails aren’t necessarily a thing that’s super high priority?
15:53:10 <NikitaKonovalov> I'm not sure if it's time to send emails so early
15:53:26 <krotscheck> That’s fair.
15:53:48 <krotscheck> Well, honestly, I don’t think we’ve got enough time to address subscriptions yet anyway.
15:54:02 <jeblair> krotscheck: infra has a huge review backlog; the oldest storyboard review is about 20 reviews down on my list.  you can help by reviewing other infra patches if you're interested.
15:54:53 <jeblair> actually, it's at position 30
15:56:13 <krotscheck> jeblair: As soon as I’m done with all the additional features asked for on storyboard, sure!
15:56:53 <krotscheck> Also, I just got some UI review resources from HP which I may be able to hijack for storyboard UI reviews at large.
15:57:37 <krotscheck> jeblair: Given that I’m not really qualified to review most of the infra stack though I don’t know how helpful that would be.
15:57:59 <krotscheck> Anyway, sounds like we’re done.
15:58:03 <krotscheck> #endmeeting storyboard