[00:00] oh....yeah that whole thing is a typo [00:00] lol [00:00] oops, that was a comment delete error [00:00] windowNode.simulate('keydown', { keyCode: 104, shiftKey: true, altKey: true }); [00:01] I can't say that that's the proper syntax (I'm pretty sure it is though) but it should work [00:01] are you saying that when you add an extra modifier it stops working? [00:02] Nevermind, I'm an idiot [00:03] well I wasn't gona say anything.... [00:03] but since you brought it up..... [00:03] :P [00:03] :) [00:08] ok now I want my new watch.....patience is not a virtue [00:09] hatch: heh, just closed my tab checking my xps build status [00:09] still in China is it? [00:15] I remember when I bought my last computer form Dell it took a month and a half heh [00:16] granted it wasn't one of those 'fast ship' ones [00:16] yea, still 'in production' [00:21] think it'll get here before london? [00:21] I give up on ever hoping my tests will pass our CI [00:22] lol [00:22] lemme take a look [00:23] huwshimi that looks like it's the intermittent failure [00:23] It's actually getting a bit ridiculous at this point. [00:23] hatch: Yeah, I realise that's the problem, it's just a hassle that that happens every time [00:24] yeah I'm not going to disagree there [00:46] yea, I mean it's an issue, but try watching juju-cores landings :) [00:47] I actually feel better about our CI lately. It really rarely fails 2 times in a row [00:47] and really I think it fails 1 in every 7 ish branches [00:47] which is still :( [00:51] we'll get there [00:51] in time...in time [05:33] huwshimi: mornin' [05:33] rogpeppe: Woah, a very early morning to you too :) [05:34] huwshimi: :-) [12:01] rick_h_: hey! Got some time to chat about juju-quickstart and status in Ubuntu? [12:01] * rbasak isn't sure who else would need to be in that conversation [12:01] rbasak: on a call atm but should be free shortly [12:01] rick_h_: OK, thanks! [12:12] rbasak: free as a bird! [12:13] rick_h_: ten minutes? I just grabbed some food! [12:13] rbasak: ok, sounds good [12:37] rick_h_: https://wiki.ubuntu.com/StableReleaseUpdates#Procedure [12:39] https://bugs.launchpad.net/ubuntu/+source/juju-quickstart/+bug/1311321 [12:39] <_mup_> Bug #1311321: ascii can't decode error in 14.04 server install

[12:39] https://bugs.launchpad.net/ubuntu/+source/juju-quickstart/+bug/1309678 [12:39] <_mup_> Bug #1309678: a value is required for the control bucket field

[12:39] https://bugs.launchpad.net/juju-core/+bug/1306537 [12:39] <_mup_> Bug #1306537: LXC local provider fails to provision precise instances from a trusty host

[12:41] rick_h_: https://bugs.launchpad.net/ubuntu/+source/juju-quickstart/+bug/1325013 [12:41] <_mup_> Bug #1325013: juju-quickstart has no dep8 tests so can break when juju-core changes === urulama is now known as uru-food [12:55] rick_h_: https://bugs.launchpad.net/juju/+bug/1081247 [12:55] <_mup_> Bug #1081247: maas provider releases all nodes it did not allocate [does not play well with others]

[13:06] * rick_h_ runs for coffee and breakfast. early meetings ftw === uru-food is now known as urulama === rogpeppe2 is now known as rogpeppe [14:06] hey rick_h_ [14:06] jcastro: otp what's up? [14:07] oh good, you can just listen then. [14:07] so you know how the cluttered search sidebar is my favorite? [14:07] jcastro: yep [14:07] I am now realizing that every time we accept a charm into trusty, we double the horribleness [14:07] shortcut to hide it has landed [14:07] heh yea [14:07] for example, try "elasticsearch" [14:09] thats strange. Trusty ES shows up under other, and not the recommended charms. [14:48] jcastro: ok, phone calls over. [14:48] jcastro: yea, we try to not confuse users and show one 'recommended' charm [14:49] jcastro: and the precise one is 'more popular' and wins [14:49] rick_h_, I really want the full screen search back [14:49] this isn't going to get any better as the number of charms goes up === makyo_ is now known as Makyo [14:50] jcastro: understood, wip to help [14:58] jujugui call in 2 [15:00] rogpeppe: ^ antdillon [15:15] Was expecting "to your scattered bodies go", which I think pegs me as a nerd. [15:19] rogpeppe: available for a hangout? [15:20] rick_h_: yup [15:20] rick_h_: got one to hand? [15:20] rogpeppe: https://plus.google.com/hangouts/_/gwumte2sj2qezub4v54iy37dvqa?authuser=1&hl=en [15:33] rick_h_ so I've been messing with this hide sidebar branch and I think we have to expand its functionality a bit. [15:33] hatch: ruh roh [15:33] if the user has hidden the sidebar and then clicks a service the sidebar should open again [15:34] else it just looks broken [15:34] hatch: but but but, that's the point [15:34] because the url changes but nothing shows up [15:34] it's just a manual hack for demo/etc [15:34] it's not a usable experience [15:34] at least for the moment, file a bug but it's going to be down the list tbh [15:34] right, but I'm not sure what demo is possible without a sidebar when they click a service [15:34] ok that's fine [15:34] thx [15:34] hatch: right, for the demo, they need to manually show/hide the sidebar [15:34] like manually picking the slides or the terminal during a presentation [15:35] you have to swap between then on your own [15:35] hatch: but agree, it's a definite nice to have to have it not suck [15:36] rick_h_, heya [15:36] jcastro: party [15:36] what are the chances you have a sec to talk bundles with carla/ant? [15:36] jcastro: changes are good for the next 2hr [15:43] rick_h_ ok cool, as long as people know that they have hidden the sidebar so anything which requires that will also be hidden heh [15:44] other than that his branch is gtg so I'll land it [16:05] so, my ISP saw we had a lot of traffic going to port 6667 so they concluded we had a TROJAN and helpfully blocked the port for us. [16:05] rick_h_ we are now getting the intermittent ci failures on the merge job too [16:05] freenode == 6667 [16:05] looks like we need to put some time into fixing it now [16:06] bac lol - lots of traffic? freenode? [16:07] hatch: k, otp will lok [16:07] look [16:15] https://saucelabs.com/jobs/3eeb9e9d9dba479cbf5621b069399007 [16:15] hatch: that's the common issue on that crap stuff [16:16] right - but now it's also happening on the merge ci [16:16] it never used to happen on that one [16:16] it appears the frequency of the failures has gone up [16:16] and it's spreading.....like a virus....a test virus [16:16] :) [16:17] yeah it's happened twice in a row now on two different branches [16:17] ok, but that's bad luck [16:17] it's happened before [16:18] hatch: ok, off the phone [16:19] hatch: yea, that sucks. We've had that kind of mass sauce tests hate us before and it usually makes a day suck and passes [16:19] hatch: but yea, we definitely want to fix that as part of our code/clenaup work post machine view [16:19] right, it's just, how much time is wasted re-running ci builds to see if they pass [16:20] hatch: right, if today it takes 4 tests runs, that's 2hrs. Fixing notifications is several days of work noe 2hr [16:20] so it takes having this happen for a week in a row to be a blocker like that imo [16:20] and it's not like this solves all CI failures, this notification one in IE is the most regular of them for sure. [16:21] and it'll be first on the list of fixing up test stuff post machine view [16:21] this is in ff now [16:21] too [16:21] ok [16:21] it's only this test [16:21] but I don't see anything wrong with it, or the previous ones [16:22] right, and I spent a day trying to update that test and found it's a fundamental issue in how notifications are written and the fix means rewriting that code [16:22] blarg [16:22] notifications are one of the last original code things [16:22] heh [16:22] I took a day to try to fix it and the hack is more than a day of work, and the workarounds are uglies than this failures [16:22] the only right thing to do is to change it to be less timeout heavy and work properly in an event world [16:23] which is a base level rewrite of how it works [16:23] hatch: sorry it's sucking today. We will make it better, but not today. [16:23] I'm just curious why it happened all of a sudden [16:23] something must have changed [16:23] it's a timing thing [16:24] the sauce machine is slower than usual, the network, and it fails [16:24] it's depenent on the stack and it's a big stack of pita and that stack is now 10 inches tall every single day [16:24] it's 9.8" one day, and 10.2" another day [16:25] haha [16:25] and if it gets to 10.3" the tests tell us to go screw ourselves and die [16:26] I wonder if we went on a 'private' sauce instance if it would be faster [16:26] they don't say there is any difference in performance [16:26] hatch: right, it's something I want to bring up. [16:26] hatch: when we get theblues going and hopefully are doing more sauce-like tests I want to try to get us a real account [16:27] welcome BradCrittenden! [16:27] yeah cool [16:27] hatch: at that point it'll be easier to justify and make a case it's a cross project thing [16:27] right [16:27] hatch: but yea, not looked into it yet. Just talked with them at pycon [16:27] haha that louis suarez got banned for 9 games [16:27] ahh [16:27] ok, I have to refill my coffee, afk for a min [16:32] lazypower are you collecting the steam summer adventure cards? I need two of the ones you have ;) [16:36] rick_h_: can I suggest that quickstart prompts for ssh key creation before installing packages? Then the user can get over the interactive part without having to wait first. [16:37] +1 [16:37] (otherwise I come back to a test that I started earlier to find that it's waiting on a prompt) [16:37] rbasak I'd recommend filing a bug [16:38] I'd be happy to if it sounds like it's doable [16:39] I have no idea, but it would improve the experience so I'm all for at least investigating it [16:40] rbasak: yea please file a bug and we'll investigate [16:40] they are running the fiber through my neighbourhood right now [16:41] Filed https://bugs.launchpad.net/juju-quickstart/+bug/1334757 - thanks. I don't have permission to set Importance: Wishlist. [16:41] <_mup_> Bug #1334757: ssh key generation prompt happens after a long wait, mismatching the user's expectation that it will all be ready when he comes back [16:41] rogpeppe: had a talked with design and eco on bundles and there's a couple of notes for the bundle spec. I'll update the spec after lunch and then file a couple of bugs [16:42] rbasak thanks a lot [16:42] rogpeppe: but we wanted to support categories and description as optional items to the spec [16:42] rick_h_: how is "description" different from README ? [16:42] rogpeppe: and want to talk about pulling vcs info during a publish, so it's part of the publish process, but we'll have some metadata that the store will want to keep [16:42] rick_h_: (personally i'd be happy ditching README) [16:42] rogpeppe: the description is a paragraph while the readme is a full blown document going through optios, what it does, how it works, minute details [16:43] rick_h_: ok [16:43] Makyo when you have a second I'd like to have a chat about implementing the 'removal' story in the ecs when there are hierarchical changes to be made [16:43] rogpeppe: e.g. https://jujucharms.com/bundle/mongodb/5/cluster/#readme is the readme, but the desciption might be "Deploys a 3-shard mongodb cluster" [16:44] rogpeppe: anyway, will follow up after lunch in the right places but wanted to let you know. [16:44] * rick_h_ goes for food [16:44] rick_h_: i'm trying to launch against azure. i'm getting 'affinity group conflict' so i assume there is a machine associated with 'bactest' that is still alive, though juju doesn't know about it. on their dashboard i cannot figure out which machines are which. would you happen to know how? [16:44] bac: I just match the names in the dashboard with the juju status outpuyt [16:44] output, using the dns names [16:44] rick_h_: right but i now have no juju output [16:45] hatch, sure, now's good. [16:45] bac: :/ ok looking [16:45] Makyo ok going to the standup room [16:46] bac: go to the monitor tab and move the timeline to 24hrs and view which machine has data for things like disk going back the right timeframe? [16:47] oi [16:48] bac: can you join the other irc channel from this morning please? [16:48] rick_h_: i got a list of the real ones [16:50] bac: awesome, let me know if you need a hand [16:50] * rick_h_ goes for lunch [17:24] * rogpeppe is done for the day. [17:24] g'night all [17:31] see ya rogpeppe [17:56] juju-quickstart 1.4.0 uploaded to Utopic. It should land in the next hour or so. [17:57] rbasak: <3 [17:57] Thank you for making it easy :) [17:57] I didn't need to change the packaging at all. [18:00] bac and frankban rock on that [18:01] jujugui need add'l reviews/qa on https://github.com/juju/juju-gui/pull/405 - switching to new upgrade UX. I suspect I'll have conflicts on my next branch, so I'd like this to land ASAP. [18:02] on it [18:02] hatch, thanks. [18:08] jcsackett what's round 16? [18:14] hatch: soccer [18:14] oh.... [18:15] I saw 5 minutes of a game yesterday and saw a guy flop [18:16] he flopped, sat down holding his knee, when no one cared, he got up and started playing again... [18:16] stupid game [18:26] Makyo code is good [18:26] do you need another qa as well? [18:27] hatch, I should be good, the next few branches will be additional QAs. [18:27] hatch: got a gui question for ya - http://askubuntu.com/questions/488575/juju-configuring-a-service-by-clicking-on-charm-with-stand-alone-implementatio [18:28] Makyo ok cool [18:28] +1 [18:28] lazypower looking [18:29] lazypower ok I'll reply thx [18:32] its easy points, thought you'd like that one === BradCrittenden is now known as bac [18:38] really wish people would ask one question at a time [18:38] :) [18:38] lazypower and answered [18:44] woot I broke 900 points on AU [18:44] look at me go [18:47] upvoted [19:41] man it's so loud here [19:41] they are directional drilling or something.... [19:48] * rick_h_ steps a way for a bit. [19:48] see you all at the AU call tonight [20:01] rick_h_ when you return - have you seen this error before? http://ci.jujugui.org:8080/job/juju-gui-merge/433/console usually it happens once then works the next time but this keeps happening on this branch and there isn't any information on the failure other than that it failed [20:05] trying again, 4th time is the charm! [20:19] hatch: that's github's api not responding nice [20:19] oh ok [20:19] hatch: we should add some error checking in the result of the api response to the merge command [20:19] if github has a timeout/etc you get stuff like that [20:19] if broke try again [20:19] yea, pretty much [20:19] nothing to do but ask github again [20:19] pretty please do this merge for me [20:20] * rick_h_ sneaks back into the shadows [20:36] "hey cool, it works" [20:38] Ooooo [20:39] after all that work I ended up solving it in 7 LOC's [20:39] *sigh* [20:40] rick_h_: those bugs for the SRU have been updated. i'm not sure i convinced myself on the 'regression potential' for the last one. [22:05] jujugui quick PR for review/QA https://github.com/juju/juju-gui/pull/406 [22:05] sure [22:05] Fixes #1328981 [22:05] <_mup_> Bug #1328981: Reloading the inspector on upgrade of a charm creates blank sidebar [22:12] Makyo so is clicking this change version button supposed to block out the entire inspector with no way to close it? :) [22:13] hatch, Did in the previous version, this should fix that. [22:13] Can you check your cache? This has been an issue in the past. [22:13] Also, report the URLs that you see in the process. Should clear sectionA, then repopulate with inspector/ [22:14] Makyo I mean, I click Change version then a list shows up [22:14] which I can't close [22:14] hatch, yes, you didn't QA the previous branch, but that was step 1, step 2 is styling/etc. [22:15] Which includes setting active tab to null, etc. [22:15] ok when I click to upgrade it changes the url to http://192.168.33.10:8888/precise/mysql-31/#/precise/mysql-31 [22:15] so it shows the charm details [22:15] but it doesn't always do this....hmm [22:18] Are you clicking the word 'upgrade'? there's two links there, just like there was in the previous impl, it just hasn't been styled yet. [22:18] the 'upgrade' link does not have an href and should not change your URL. [22:18] oh lol oops [22:19] Sorry, I promise styling is coming next. [22:19] 7 days was just too long for that branch. [22:19] lol np [22:19] ok I see I can't test this on sandbox [22:19] I can. Not sure what's up on your end. [22:20] well it doesn't say to reload the inspector [22:20] it just reloads [22:20] without issue [22:20] hatch, hmm, I noticed this with mysql too. Can you try wordpress while I see what's up? I have the feeling it's a databinding issue. [22:21] sure trying wp [22:21] (That button for reloading the inspector is databound to an attr on the service. [22:21] ) [22:21] jujugui .au weekly call in 9, for those going. [22:21] "error interacting with charmworld api" [22:21] Cannot set charm on a service with units in error without the force flag. [22:21] interesting [22:21] ^ that last one was because simulator [22:22] interesting non the less [22:22] Makyo wordpress also 'just changes' [22:22] hatch, yeah, that's a core thing. SetCharm api endpoint doesn't work when the service has units in error, reproduced in sandbox. [22:22] Butts. [22:22] Let me prowl some more. [22:22] Databinding is a little too magical. [22:23] ok lemme know [22:27] Looks like it has something to do with the delta stream. service.charmChanged doesn't get set to true unless things come back in the delta in the proper order. [22:27] Which is a little dumb. [22:27] I'll poke more after the call. [22:31] mm, no huw [22:32] there he is [22:32] Morning [22:33] hatch: coming along? [22:33] huwshimi: call? [22:33] hey huwshimi yep joining call [22:33] Joining [23:03] brb changing location [23:09] 13s relocation? :) [23:10] oh :) [23:22] jujugui lf a review and qa for https://github.com/juju/juju-gui/pull/407 plz and thx