[11:22] is there a way in the LoCo directory to see all the events you are registered to attend? [11:22] perhaps a personalised ical feed? [11:23] Fairly sure there isn't a personalized ical feed. [11:26] so am I. Would be cool though [11:27] AlanBell: file a bug [11:28] write a patch if you can [11:29] RSS would be complicated [11:29] But having a profile page would be awesome. [11:29] yeah, will file a bug [11:30] are there any recent patches with test scripts to look at? [11:31] I couldn't work out how to do a meaningful test for https://code.launchpad.net/~alanbell/loco-directory/backbutton/+merge/73526 [11:54] who should bug 773243 be assigned to? [11:54] Launchpad bug 773243 in ubuntu-website "Ubuntu website advertises "Fully compatible with Microsoft Office" (affects: 4) (heat: 18)" [Undecided,Confirmed] https://launchpad.net/bugs/773243 [12:52] AlanBell: summit should have good tests [15:26] mhall119, https://code.launchpad.net/~james-w/awstrial/remove-maverick-text/+merge/75572 [15:28] james_w: I was actually joking about you doing awstrial work, that's in ISD's hands now [15:28] yeah [15:28] I was hoping it would still work though :-) [15:28] I'll take it, don't get me wrong, it's less work for me [17:00] james_w: I got an error running your multi-sprint tests [17:01] oh [17:01] pastebin and I'll fix it [17:01] it's in the MP [17:01] looks like maybe you're using a newer version of Django? [17:01] it can't find assertIs [17:02] ah [17:02] assertEqual will suffice [17:04] mhall119, fixed [17:06] also argle, non-deterministic behaviour to try and write a test for [17:07] well, I'll just fix the code and keep the test I think [17:07] this is some private meeting fixes I'm working on, not the multi sprint branch [17:07] ok [17:10] james_w: my only concern about the multi-sprint is that once you add a Sprint record, you lose the default which uses the summit name [17:11] this may cause confusion [17:11] hmm [17:11] we'll just have to make sure people know about it [17:11] ok [17:11] so right now we'll have uds-p spring [17:11] sprint [17:11] I would have put some text on the admin page, but I wasn't sure how [17:11] and if Marianna adds the Linaro sprint, suddenly it won't check the uds-p sprint anymore [17:12] I think just making them aware will suffice for this cycle, but we might want to discuss how to keep it consistent going forward [17:13] either dropping the use of a default, or adding the default as a sprint on initial save of a Summit object [17:13] ah, that's a good idea [17:14] I'll set this one up, so it won't be a problem unless anyone changes them [17:14] ok [17:16] awesome test case for it though [17:26] james_w: both your branches have landed in trunk, I also updated http://ec2-50-16-76-22.compute-1.amazonaws.com/ to the latest trunk and migrated the DB [17:26] james_w: you can test importing multiple sprints on there if you have one for linaro ready to go [17:26] woop [17:26] thanks [17:27] I don't, but now that I know this is approved I will create one for testing [17:27] I just ran lpupdate on there with the default sprint [17:27] hmm, it will happily import from staging, I'll try that [17:28] man, having this ec2 running trunk sure is nice [17:30] yep [17:37] cjohnston: nigelb: james_w: summit now has 43 unit tests, and they're all passing \o/ [17:38] yay [17:38] You guys are awesome :) [17:38] I propose a hacking session every UDS. [17:38] *after* or *before* every UDS [17:38] *at* [17:38] ok, but no deployments [17:38] Yeah [17:39] This puts all the stakeholders + people who can fix into one place. [17:39] well we've had sesssions at the last 2 UDSs [17:39] this seems to be working ok [17:40] Yeah :) [17:40] it fails to import blueprints from staging, as it's hardcoded to production [17:40] but it imported users fine [17:40] so I think we're good to deploy this [17:40] Hrm, we should take it out into a config [17:40] any objection to a deployment on trunk to summit.ubuntu.com? [17:41] not from me, but it's going to require some extra work this time because we re-set the South history [17:41] ah [17:41] so we'll need to ./manage.py reset south [17:41] I can do it, but some clues about that would be appreciated [17:41] the ./manage.py migrate schedule 0001 --fake [17:41] and ./manage.py migrate sponsor 0001 --fake [17:42] then ./manage.py migrate [17:42] ok [17:42] should I merge trunk to stable and production branches first? [17:43] not to stable, once trunk merges to production we're going to be done with stable [17:43] ah¸ok [17:43] I'll get some lunch and tackle this [17:43] ok, when do you want to try for a production deployment? [17:44] we need to let jcastro know to expect it [17:44] mhall119: why do you want to get rid of stable? :( [17:44] nigelb: because once trunk merges to production, trunk will be our new stable [17:45] But what if want to make unstable fixes and stable fixes? [17:45] so we'll go back to only 2 branches, trunk and production [17:45] Like james_w did have to do a bit back. [17:45] you make unstable fixes in feature branches from now on, and propose them back into trunk [17:45] ah, ok. [17:45] Nothing undeployable goes into trunk. [17:45] Makes sense. [17:45] let's say 1:15 from now [17:45] james_w: wow, that soon? ok [17:46] james_w: do you know how to make a proper backup of postgres? [17:46] is there a reason to wait? [17:46] not that I know of [17:46] except letting people know it's going to happen [17:46] and backing up the db prior [17:48] yeah, I'll see about the db backup [17:48] mhall119: pg_dump [17:48] downtime should be short, and I don't think anyone is using it much right now [17:48] nigelb: yeah, but what parameters do we want? [17:48] james_w: I'm checking with jorge on that [17:48] database name? [17:48] and I'd guess username and password. [17:48] https://code.launchpad.net/~james-w/summit/fix-autoscheduling-conflict-with-private/+merge/75594 was the problem I was fighting with earlier [17:50] mhall119: "pg_dump summit" should work. [17:50] unless it fails for authentication [17:50] you can get credentials from settings.py anyway. [17:51] james_w: can't we compare user records by their primary key? [17:52] also, why did you remove local_settings.py.sample? === Ronnie2 is now known as Ronnie [18:12] mhall119, that wasn't intentional, I must have used mv rather than cp [18:12] mhall119, compare by primary key, you mean implement __eq__ for Attendees? [18:21] james_w: no, I mean using attendee.pk instead of attendee.name [18:21] ah [18:22] that should work [18:23] james_w: I've got to pick up the kids from school in a few minutes, so just ping me before you deploy to make sure i'm around [18:23] ok [18:23] you can go ahead and merge trunk->production if you want though [18:23] I'll hang back to start anything in production [18:23] ok [18:25] james_w: once this goes out, let's talk about how to fix the caching (or at least providing a means of force-clearing it) [18:25] yeah [18:25] I was thinking about that in the shower this morning [18:26] I've made the .pk change and re-instated local_settings in the autoscheduler branch [18:26] and pushed up a small js branch for a bug I found while testing private meetings [18:27] cool, I'll review later, it can go out after this deployment [18:27] yeah [18:27] they are important fixes, but don't block deployment [18:27] and once we've done this one we should be able to do frequent deployments of well-tested changes :-) [18:27] I'm thinking, about the cache, that we make a signal that gets fired whenever some code changes the schedule [18:28] and invalidate everything? [18:28] that way we can start with a specific URL that you can hit to fire the signal, but also start adding it into the places where it needs to be and eventually not need the URL [18:28] james_w: that's what we should discuss [18:28] ok [18:28] ok, leaving now, bbl [18:28] let's do so later [18:28] bye [18:53] james_w: okay, I'm back [18:56] cool [18:56] just did the merge to production [18:58] james_w: I <3'd your RT. [18:58] "summit is such a fuster cluck" [19:01] pg_dump done [19:01] new code in place [19:01] did you touch the .wsgi files? [19:01] not yet [19:02] I'm thinking that should happen after the migrate commands [19:02] ah, right [19:02] starting those now based on your instructions [19:02] You have requested a database reset. [19:02] did you just do the --fake migrations? [19:02] This will IRREVERSIBLY DESTROY any data for [19:02] the "south" application in the database "summit". [19:02] Are you sure you want to do this? [19:02] Type 'yes' to continue, or 'no' to cancel: [19:03] reset first right? [19:03] yes [19:03] we have a backup anyway, right? [19:03] yeah [19:04] migrations done [19:04] touching wsgi files [19:04] crossing fingers [19:05] looks good to me [19:05] looks good to me too [19:05] anyone see any problems? [19:05] we have the new code as well :-) [19:05] Hrm, today link is falling off the top menu :( [19:06] not for me [19:06] what browser? [19:06] nigelb: no it's not [19:06] firefox [19:06] nigelb: I get "There are no sessions scheduled for today." [19:07] what do you get? [19:07] mhall119: No, I meant its not staying in the top menu [19:07] also, are you going to /today ot /uds-p/today ? [19:07] but dropped down [19:07] hmmm, not for me [19:07] nigelb: what URL are you seeing this on? [19:08] summit.ubuntu.com [19:08] nigelb: try ctrl+refresh? maybe you have old theme dadta [19:08] mhall119: http://people.ubuntu.com/~nigelbabu/today.png [19:09] or maybe your font settings are causing it? [19:09] maybe because I don't have Ubuntu font. [19:09] Right. [19:14] whats up with microblogging? [19:15] are all of those curacao rooms for hacking? [19:15] cjohnston: the Hackfest is from a track being assigned to the room [19:15] I'm not sure what's up with twidenash [19:16] Right.. I know that... I assume that those rooms are only for the hackfest tho? [19:17] cjohnston, they are Linaro rooms, assigned to teams to use for hacking [19:17] as Linaro isn't planning to schedule a full roster of sessions, but spend some time hacking as well [19:18] mornings for discussion, afternoons for hacking [19:18] mhall119, let me know when you want to discuss caching [19:18] gotcha [19:20] james_w: should those be private rooms then, so that public meetings aren't auto-scheduled into them? [19:20] hmm [19:21] double hmm [19:21] probably shouldn't be openly scheduled [19:21] We don't have a way of displaying the private rooms yet [19:21] maybe they should be private, I'm not sure [19:22] I'll have a look [19:23] What does marking a room closed do? [19:23] james_w: can you either give g+w ./summit/media/js or run: bzr branch bzr+ssh://bazaar.launchpad.net/~django-foundations-dev/twidenash/2.0/ ./twidenash [19:23] Does that allow manually scheduling? [19:23] cjohnston: only through the django admin [19:23] hmm [19:23] So we need a way to mark a room as avoid auto scheduler [19:24] Like a 'Manual Scheduling' [19:25] because, unless Linaro has a reason to have them private, there isn't really a reason that they shouldn't be displayed to anyone. [19:26] cjohnston: maybe "Reserved" would be a good type description [19:26] That would work.. [19:30] mhall119, done both [19:31] james_w: what do you think about a reserved status.. That way the room is still "Public" but the autoscheduler doesn't schedule stuff in there.. similar to private, just being visible to everyone [19:31] on the edit page stuff can be manually scheduled [19:33] fixed microblogging [19:33] cool [19:36] cjohnston, maybe [19:36] I need to look in to whether we want the schedule for these rooms to be displayed at all [19:37] ok [19:37] we are using the tracks feature though [19:37] so it would only autoschedule sessions in the linaro-hacking track [19:37] in to those rooms [19:37] right... [19:37] so if we just schedule it how we want the autoscheduler will leave it along [19:37] alone [19:38] wont the conflict resolver mess with it if conflicts arise? [19:39] nope [19:39] not if we manually schedule everything [19:39] the conflict resolver doesn't change hand-scheduled stuff [19:39] gotcha [20:21] mhall119: oops [20:21] james_w: the rescheduler does. [20:22] cjohnston: yeah, sorry about that [20:22] nigelb, does what? [20:22] statik has requested an official build and release of the latest django-openid-auth though [20:22] james_w: move things around [20:22] nigelb, where? [20:22] whats the eta on that mhall119 [20:22] cjohnston: dunno, let me ask [20:22] james_w: I think its commented out now. So we may be okay. [20:22] be even more awesomer if the memory issues get fixed instead [20:23] yeah, that's the big concern [20:23] thats why i asked about getting them switched [20:23] maybe you can say thats a good idea or something [20:23] Worst case, we could just spawn an ec2 for summit. [20:24] Really, I'm not that scared about the memory issues. [20:24] do you actually think they would point summit to an ec2? [20:24] Yes, given enough pressure. [20:25] nigelb, the reschedule code only acts on autoscheduled events [20:25] james_w: AH! [20:25] the "auto" flag, right. [20:25] and you don't think that when we actually add load to summit cranberry will get worse? [20:25] ywp [20:25] cjohnston: I can't say that for sure. [20:25] No one can. [20:25] Because we don't know what's causing the problems [20:26] Its intermittent. [20:26] thats my point... that we need to figure it out [20:26] Personally, I think its something to do with apache config. [20:26] But someone needs to spend time figuring that out. [20:26] yup [20:26] I don't have access, neither do you. [20:26] That leaves IS or mhall119 / Daviey / james_w. [20:27] but do mhall119 Daviey and james_w have the correct access to figre it out? [20:27] And we probably need root at some point. [20:27] So, they don't have all the access. [20:27] I only have write access to the summit directory [20:27] We're screwed with a ghost of a problem. [20:27] throw harvest back on cranberry and give us guanabanana [20:27] And, to make things even more complicated, its one of those things you have to be watching happen. [20:28] or wtf ever it is [20:28] Its guanabana. [20:28] whatever [20:28] :-P [20:38] nigelb: mhall119 are there any of the emails that would be better than another to attach to the RT? [20:38] pick any [20:41] cjohnston: for the memory issue? [20:41] none of the django ones will be especially helpful, because the error gets thrown at random places, whatever happened to be executing when it ran out [20:42] I attached the most recent I have [20:44] james_w: keeps spamming me :-P [20:44] I can stop working if you want :-) [20:44] nope [20:48] yay [20:49] mhall119: nigelb james_w are we good with marking [summit-hackers] figure out a way to import data locally for testing to postpone? [20:49] Yes, go ahead [20:49] same with [summit-hackers] JSON export of data to be used for local testing: TODO [20:49] [summit-hackers] Add more content/direct links on the front page: TODO [20:49] the last one is DONE [20:49] is it it? [20:49] [summit-hackers] Making user roles for different users providing different levels of access: TODO [20:49] oh my [20:49] more content direct links on front page is different than main nav matching [20:50] ah [20:50] I reeeeeeeeeeeeeeaaaaaaaaaaaaaaaaaaaaaaaaaaalllllllllly hate model_mommy [20:50] You shouldn't forget a single model [20:50] Or mummy doesn't work ;) [20:50] The launchpad way is more awesome. [20:50] [summit-hackers] Create a mutable item to where meetings in the past are muted: TODO <--- this one is where the rescheduler doesnt reschedule stuff in the past? [20:51] cjohnston: That was also for rooms being closed etc. [20:51] Like sessions happen in Room X on day 1. [20:51] We close that room. [20:51] To facilitate that, we delete the room. All hell breaks loose. [20:51] So we need to fix that prior to this uds [20:52] Actually, we need to make room availability better. [20:52] Like mark time periods where the room is unavailable. [20:52] we have that [20:53] Then we should just tell jcastro to use that next time. [20:53] There is busy times... [20:53] the ability to add three different times [20:54] i dont know if you can add more if you use all three [20:54] like say monday wednesday thursday afternoon and friday arent available [20:55] I am wondering with all the additional rooms if the schedule is going to be too wide for the screen [20:55] we should get the schedule to be better. [20:55] that still bothers me [20:56] i wonder if the screens will be available a few days prior to uds [20:56] ask elmo [20:58] mhall119: nigelb james_w https://blueprints.launchpad.net/ubuntu/+spec/community-o-summit what else do you think we can postpone? [20:59] the render.py [20:59] nigelb: did you ever finish the api bug you were working on? [21:00] No [21:00] I don't think I will get to it either. [21:00] I need more time. [21:00] Fix scheduling conflict resolution and notification [21:00] what's that one? [21:01] Making user roles for different users providing different levels of access <- I'm not sure we'll get to that one [21:01] I'm not hugely familiar with the API (read don't know how to do it on my own) but I think thats an important bug to fix to make summit hugely improved [21:01] hmm, we should likely turn the reschedule script off on the last day [21:01] oh no, it won't fight with the admins, so it's ok [21:01] james_w: it usually is turned off. [21:02] we forgot last time :-) [21:02] well it shouldn't be [21:02] perhaps add a rule to not reschedule anything that starts in the next N hours? [21:03] Allow people to define their own Busy times, check them for scheduling conflicts <- busy times are implemented, there's just no non-admin interface [21:03] can we hackup a form for that? [21:03] I think it should be "easy" [21:03] we could [21:04] yup [21:04] I'm not inclined to [21:04] well.. i dont know about easy [21:04] I wish we had more hands for ubuntu webdev. [21:04] teach czajkowski to hack summit [21:24] cjohnston: we can still get a JSON read-only API in before UDS [21:25] does anyone know if we are getting Guidebook again this time? [21:25] cjohnston: django-openid-auth_0.4 is building in our PPA now, I should be able to re-open that RT tomorrow [21:25] james_w: I haven't heard either way, I assume we are [21:38] james_w: we are afaik [21:38] mhall119: sweet [22:23] james_w: i dont know that we should mark the private rooms bug as released.. [22:23] we still dont have a way to display that someone is in a private meeting in a private room === Ronnie1 is now known as Ronnie [23:05] cjohnston, true, but we can live without it if we have to [23:05] I say file another bug [23:22] 5 files changed, 7 insertions(+), 50 deletions(-) === Ronnie2 is now known as Ronnie [23:29] james_w: have a moment to help me debug code? [23:44] sure [23:45] james_w: https://code.launchpad.net/~chrisjohnston/summit/trackleads [23:46] I'm getting an error with inner() which is in decorators.py [23:46] inner() takes at least 3 arguments (2 given) [23:46] I'm sure my lead isn't right.. I'm just lost on getting it right [23:47] why does the lead need to go in there? [23:49] inner() gets the arguments from the url [23:49] so it's taking the summit-name from the start [23:50] and expecting to get the lead next [23:50] but urls.py isn't set up to pass that through [23:51] ok