[00:10] alexpilotti: howdy, aroud? [00:10] around that is [00:44] rick_h_: hi [00:44] alexpilotti: hey, how goes? [00:46] well, confirmed you can deploy a service with 0 units. [00:55] Back in the saddle for 2nd shift [00:55] * rick_h_ starts playing 'raw hide!" [00:55] Code wrangling [00:56] * kadams54 suddenly feels the need for a 10 gallon hat and some spurs. [00:57] rick_h_: you available to chat some about this unplaced units card? [00:57] hey, been on my AMZ wishlist for years :P [00:57] kadams54: sure thing [00:57] LOL [00:57] let me get the mac out [00:58] kadams54: link me whenever your want === alexpilotti_ is now known as alexpilotti [01:02] kadams54: https://github.com/juju/juju-gui/pull/296 [01:02] http://cl.ly/image/3t1x3L2k1y2e === alexpilotti_ is now known as alexpilotti [01:56] kadams54_: Are you still around? [01:58] Yup [01:58] kadams54_: Actually, I need to duck out. If you happen to return, I was just looking in the serviceunit-token and there is a container '.icons' but I can't see from the mockups where there will be more the the token-move icon. Will the be issues if I rename that block? [01:59] :-) [01:59] Not that I know [02:04] evening [02:05] party [02:05] yeah, I could use a good :beer" [02:05] : [02:05] bah [02:06] rick_h_ out of data yet? [02:07] hatch: almost, crossed 9gb [02:07] haha how did you use up 2 gigs this afternoon? [02:07] I'll be getting a bill probably [02:08] hatch: lots of clean-all I guess [02:08] and I was at 7.5 [02:08] ohh 1.5GB :) [02:08] :) [02:08] that's a lot of clean-all haha [02:08] you sure you don't have some background process eating it up? [02:08] something, I don't know. [02:08] I did do a hangout with kadams54_ [02:08] but yea, something is eating it up [02:09] hatch: welcome to where all the cool people hang out [02:11] if all the cool ppl are here should I expect a wedgie and have all my lunch money stolen soon? [02:26] ok, EOD status email sent out to everyone [02:26] give it a read and if anything is unclear let me know please [02:27] http://constitutionconnectedyouth.files.wordpress.com/2011/04/moe.gif [02:27] I'll check in in the morning, but will be afk until we get the camper back home and I can get unhooked/online again [02:27] Have a good night [02:27] kadams54_: you too [02:41] This is the best test error Error: Script error. (:0) [02:46] Brilliant. [02:52] github is 503'ing for me [02:53] https://status.github.com/ [02:53] heh [02:53] that availability graph sure tanked hard [02:58] hatch: It's the weekend, there's not like anyone is doing any work anyway. [02:59] haha - someone must have kicked a powercord or something :P [02:59] although its been down for almost 30mins [02:59] that's quite a long time for an entire site to be down [04:20] Anyone still around? [04:26] kadams54_: I am [04:27] Working on having the list of unplaced service units update as the DB updates… [04:27] Ah yes [04:27] Do you know if that needs to update intelligently, similar to the machines, or can we just re-render the entire list? [04:29] kadams54_: I would think intelligently, as when you've clicked the 'move' icon it will be displaying a form for that token, and you'll lose that if they all re-render. [04:29] If that makes sense... [04:29] Ah yes, I think you're right === _mup__ is now known as _mup_ === _mup__ is now known as _mup_ [13:06] jujugui: is there anybody out there? [13:07] sunday morning and I'm falling... [14:02] frankban: morning [14:02] hi rick [14:02] hi rick_h_ [14:02] frankban: saw the pull request, will check it out when I get back. Thakns for picking up on that [14:02] np [14:02] will just have to coodinate that with hatch's branch of drag/drop and think we're about good for the demo. Will do some QA and such tonight [14:03] and make sure we have any last notes out for you and huw monday morning [14:03] sounds great [14:03] we should be able to do a little walk through and then cut a charm release monday [14:03] and then we'll just have the demo team update the charm they've got [14:04] rick_h_: I'll propose another branch for some small fixes on how the units in machine and sub-containers are retrieved [14:04] frankban: awesome, thanks. [14:06] rick_h_: is Makyo working today? === alexpilotti_ is now known as alexpilotti [14:18] frankban: I'm not sure. He was in here yesterday with some test failures/etc [14:18] frankban: but didn't hear anything after that [14:18] rick_h_: ok thanks [14:18] * rick_h_ goes offline to pack up camp. Back in a few hours. [15:15] rick_h_, any gotchas to be aware of re gui machine view/ [15:21] hazmat` yes quite a few still [15:21] hatch, got a moment to review them re hangout? [15:21] can you give me an idea of what you're trying to do and I can let you know where we should be by Monday night [15:22] hatch, i'm gonna be dry running demos, and doing booth setup later today [15:22] in a few I could, just running around getting ready for mothers day [15:23] hatch, i was figuring we'd have machine view at the booth, but we can not enable the feature flag if its going to be problematic for self-directed exploration [15:23] hatch, ack [15:23] ahh ok, well gimme a few and I'll ping when I can [15:23] hatch, thanks [15:30] i'm reading through rick_h_'s email from yesterday re current status, i guess that covers a good portion [15:32] ohh cool, yeah I didn't think you were on peeps [15:32] yeah that basically covers it [15:32] a few of us have been working over the weekend to get to that point, so it's definitely not ready for self-driven exploration [15:32] but the good news is that everything is under flags [15:33] hazmat` is there anything that's not in the email that you would like clarification on? [15:34] hatch, i was skipping inbox on peeps.. re-reading.. saw everyone hacking yesterday.. just curious as to state.. machine view is still behind feature flag i assume? [15:34] hatch, what's the url for enabling feature flag on comingsoon? [15:34] yes, very much so :) [15:34] /:flags:/ [15:35] just append that to the url with the flags [15:35] so for example [15:35] comingsoon.jujucharms.com/:flags:/il/mv/ [15:36] hatch, thanks [15:43] * hazmat` realizes half the people around him are talking about openstack at the airport [15:43] lol [15:43] nerd airport [15:44] frankban I found a QA issue with your branch with the unplaced units [15:44] * hatch is wearing his juju shirt for mothers day [15:44] hatch: it should be related to the fact we add a unit but the ecs add_unit call is not yet there [15:45] so you're suggesting we land it as-is with a follow-up? [15:47] hatch: I confirm the error is due to the missing add_unit. Matt is working on that, and I guess he has a branch almost ready. [15:48] ok I'll update the PR and then you can shipit [15:48] hatch: so basically the unit in the db is not updated with the real name of the service [15:49] Pr updated, thx [15:49] hatch: as I described in the XXX comment, this can only be done when the ECS call is ready and we call env.add_unit with a callback [15:49] yeah ok np [15:50] hatch: I guess the callback could just delete the service unit and wait for it to be re-added by the mega-watcher, but this is just a supposition [15:51] yeah there has been a bit of discussion around that - but I think the proper way to do it is the same we have always done - fake it, then update it when the deltas come in [15:54] hatch: could you please take a look at https://github.com/juju/juju-gui/pull/302 too? [15:54] on it [15:55] thanks [15:57] frankban test failure looks like the spurious one [15:58] hatch: good to know [15:58] you can re-run it manually [15:58] just QA"ing now [15:59] hatch: to re-run manually should I just delete the test failed comment? [15:59] oh no you need to go into jenkins and flip some switches [15:59] I can do it for you [15:59] one sec [15:59] hatch: thanks [16:00] np, it's re-running, remind rick or I sometime and we can show you how to do this :) [16:01] cool [16:02] frankban here is the current job - you'll have to watch it to know when it's done, it won't update the PR http://ci.jujugui.org:8080/job/juju-gui/959/ [16:44] jujugui anyone available for a review? no qa https://github.com/juju/juju-gui/pull/303 [16:44] Makyo hey are you around? [16:44] Yep, just now. [16:45] are you adding a 'placeUnit()' method to the ecs as well? [16:45] hazmat`: you all set? [16:45] hatch, no, just add units and add machines [16:45] hazmat`: can look in a bit, home and need a shower and lunch [16:46] bah hatch ^ [16:46] alright - I have to run some errands then off to mothers day stuff - I'll pick up whatever is next in the queue when I get back [16:46] hatch: where's the drag/drop branch? [16:46] yep no rush, I'll be offline for a few hours [16:46] hatch: is that the one there or close? [16:46] rick_h_ that's #303 [16:46] hatch: gotcha [16:46] took me forever to track down a stupid test failing with no information [16:47] "Script Error" OH THX! lol [16:47] :) [16:47] yea, hit that. It was an old fashioned "comment out tests in index.html and start to narrow it down" [16:47] rick_h_ basically once we have a 'placeUnit' method it'll be a single line to hook this drag and drop up to the ecs [16:47] hatch: ok cool [16:48] hatch: so we need placeUnit which no one is currently working on right? [16:48] correct [16:48] hatch: ok cool. Will finish catching up and see where we can get today [16:49] hopefully we should be good for huw to clean up some thing his monday tonight [16:49] yeah - I'll be gone for at least a few hours, but when I'm back I'll pick up whatever is next [16:49] and we'll finish and get a charm release out tomorrow [16:49] hatch: rgr, have a good time with mum :) [16:49] yeah I think we are in a pretty reasonable spot [16:49] go us [16:49] :) th [16:49] x [16:49] <3 [16:49] lata [16:50] hazmat`: if you've got any questions or need anything shoot me a call 248-956-1024 and I'll be back online in a bit. [18:30] jujugui could use a second set of eyes on this https://github.com/juju/juju-gui/pull/304/files I think I'm about done, but I want to be sure. === alexpilotti_ is now known as alexpilotti [20:34] Makyo: looking [21:01] Makyo: going to :shipit: is there anything you know of why I shouldn't? [21:03] rick_h_, I think everything that needs an XXX comment got one. [21:03] Only other note was switching away from object.observe [21:03] Makyo: yea, we've got a slack card to update that. I wonder if there's known issues we need to move that up to a todo [21:03] ok, going to ship this to get it in front of hatch before his branch lands [21:04] and try to convince him to integrate with the calls here [21:08] alexpilotti_: around? [21:09] guess not lol [21:34] Makyo: ok landed [22:11] hey hazmat` [22:11] bah hatch [22:11] tab complete failing me so hard today [22:11] hey, I am just replying to your comments [22:12] hatch: cool, matt's branch landed and landing huw's now [22:12] oh nice, did you fix up the line I commented about? [22:13] hatch: no, just made a card for it [22:13] you said it qa'd so added a card and moved it along [22:13] oh yeah - I suppose I just try to not land tech debt :) [22:13] usually haha [22:13] you and me both [22:14] but will declare next week tech debt week after the demo [22:14] just want to keep the list documented [22:15] ok so as per your comments about add machine you'd like me to update this branch with develop then make sure the add machine stuff works? [22:15] I can also do that as a follow-up [22:15] hatch: well just thinking while it's here waiting on landing we could do it and then the branch can be qa'd to an extext [22:15] hatch: I get nervous of 'un-qa-able' branches [22:16] well if you want to edit the code you can add a debugger in the drop handler to inspector the arguments [22:16] hatch: but it's one less moving part to get in a follow up. The follow up at this point needs to be placeUnit [22:16] I just figured most don't want to edit the code to do a qa [22:16] hatch: rgr [22:16] hatch: I'm fine either way I guess. Just tossing it out [22:16] hatch: I do think the UI updates should be added [22:16] yeah? Ok because that was left out for the demo [22:17] I can add them in, just another thing to potentially break haha [22:17] heh, but it should work [22:17] I mean this whole thing is a house of cards and we'll do walk throughs [22:17] alrighty [22:17] at least it should give us some UI updates that draw audience attention away from other things that don't update [22:18] hey kadams54, how goes? [22:19] Hi rick_h_ [22:19] I'll wait until huws branch lands so that we can QA with the styled stuff [22:19] Ready to move along more code doggies [22:19] should be merged fairly soon [22:19] kadams54: woot [22:19] hatch: yea, it hit a strange error in the lander on the last run :/ [22:20] oh? [22:21] hatch: yea, the lander itself died in a strange way. But it just landed matt's branch so not sure [22:21] maybe an api issue with github [22:23] ahh odd [22:23] brb got to run and grab some propane [22:37] You know, normally around this time most people head out to get food. [22:37] But not in Canada. Oh no. They get propane. [22:37] propane leads to food? [22:41] look at that, just in time [22:42] :) [22:42] uh oh, rebasing blew up :/ [22:45] Morning [22:46] yikes that was a rough rebase [22:46] morning huwshimi [22:47] hey huwshimi [22:48] uh oh trunk is broken [22:48] now my branch is broken [22:48] lol [22:48] Oh dear [22:48] :( [22:49] rick_h_ can you try and reproduce...... [22:49] on comingsoon with li, mv [22:49] D&D mysql, click deploy [22:49] click confirm [22:49] switch to machine view [22:49] food on the stove atm, can kadams54 or huwshimi help please? [22:49] mass scale up UI is gone, and now you can't switch back to the service view [22:49] rick_h_, hatch I'll take a look [22:49] Yeah, I'll check… [22:49] thanks [22:51] hatch: Yep, broken. [22:51] :/ [22:51] This isn't working: unit.icon = db.services.getById(unit.service).get('icon'); [22:52] Yeah [22:52] I actually fixed that in my branch [22:52] yeah, looks like we can skip that - but the real fix is we need add_unit [22:52] could it be that the css on comingsoon didn't kick over like happens sometimes [22:52] Which may be ready for QA [22:52] it's because there are no services yet [22:52] in ecs to be hooked up to deploying services [22:52] er units [22:52] oh wait, there is a service [22:52] yeah this is because we need matt's fixes for frankbans branch [22:53] My fix [22:53] http://pastie.org/private/ew3cx7lr4tsjlsdwasog [22:53] that landed hatch? [22:53] yeah that's not the 'proper' fix I don't think [22:53] rick_h_: The last set of CSS changes I made are on comingoon [22:53] rick_h_ this is not a css issue [22:53] *comingsoon [22:54] Oh wait, that's in a different place [22:54] I actually had to fix similar errors in two spots [22:54] https://github.com/juju/juju-gui/pull/301/files#diff-f3e78957af1c877e1b8c3d58b0dfa03aR93 [22:54] comment in frankban's branch [22:54] http://pastie.org/private/ako9by4d7irehbidopc9qg [22:55] kadams54 yeah that's a bandaid [22:55] Bandaid nothing! [22:55] there should never be units added without a service [22:55] Defensive coding is never a bandaid [22:55] lol, well then it should throw a console error :) [22:55] Bullocks on anyone who says "should never be" :-) [22:56] haha, that just doesn't translate wekk [22:56] well [22:56] "Bollocks" for you British Empire people. [22:57] doing that 'defensive' coding just makes it fail further down [22:57] haha [22:57] Anyhow, we should probably figure out a coordinated fix so that we don't have 4 different solutions landing in different branches. [22:58] yeah - need to hook up the add unit stuff frankban mentioned [22:58] that's the real fix [22:58] ok, well that's next on the list [22:58] Au contraire… properly done, the system should fail gracefully, even when faced with "impossible" scenarios. [22:58] Missing icon < uncaught error that interrupts all code execution [22:59] no there is much more wrong that just a missing icon [22:59] *sigh* [23:00] now it just fails silently [23:00] All my tests pass [23:00] Works for me [23:00] ;-) [23:00] hah [23:00] rick_h_ ok so I can't land my branch until the add_unit stuff is hooked up [23:00] hatch: :( ok, and you could before because? [23:00] I thought it was landed? [23:01] because now it REALLY can't be QA'd [23:01] it could before because you actually could create units and drag them to containers [23:01] hatch: ok, sorry trying to keep up while making omlettes for the fam [23:01] now you can't get containers or machines to render [23:01] hatch: they drag ot headers? Why do we care about rendering machinse/containers atm? [23:02] no, they drag to the container [23:02] oh, because you can't get a machine to get a container header [23:02] that, and you wanted this to drag to the container token first [23:02] :) [23:02] Man, whoever coded up the unplaced service units token was a dunderhead. [23:02] https://www.dropbox.com/sh/lvdydgiu7jeuuso/7MlRX5-IPu#lh:null-06-machine.png [23:02] kadams54: hey, let's keep it cool [23:02] right, but you said you wanted it to land on the container token first heh [23:02] * kadams54 heads back to fix his own mistakes. [23:02] kadams54: everyone's worked hard on this stuff and it's all been done by our team. So calling out like that isn't cool [23:03] rick_h_ it's ok it can sit for a bit, it's not going to go bad or anything - I'll work on hooking up the ecs add unit stuff [23:03] kadams54: oh, well if it's you then ok cool :) [23:03] hatch: ok [23:03] I'm just kidding of course :-) [23:03] hatch: thanks, will try to catch up and try things out after I feed the fam [23:04] yeah we have company coming over in an hour too, so I'll be back on after a few hours again [23:04] heh, what a gong show today has been [23:04] lol [23:04] I put {{ container }} in the handlebar template, which was intended to be… well, I'm not really sure? Since they're unplaced, they don't have containers yet. But regardless, the container attribute is actually set to the DOM container. [23:04] So that renders nicely. [23:05] hatch, rick_h_: Anything I can do to help get things back on track? [23:05] huwshimi are you familiar with the ecs stuff? [23:05] :) [23:05] hatch: I haven't touched it yet, but I can take a look... [23:06] it's ok, I'm guessing there are going to be some dragons with this add unit stuff [23:06] huwshimi: I've got a few cards your way for polish I was hoping you could get to today [23:06] hatch: OK sure [23:06] rick_h_: Yep, I can take a look at those [23:06] hatch: yea, the add unit/place unit stuff takes some background/etc [23:06] I'm not entirely sure I even know haha [23:06] I haven't done any of this stuff in a few weeks [23:06] * hatch pokes Makyo [23:07] poke poke poke [23:07] I had to run it on a live lxc env to watch wtf happened [23:07] haha - yeah it's complicated [23:07] wish core had it already [23:07] } // if [23:07] }, this); // foreach [23:07] } // for [23:07] } // if [23:07] lol someone was lost [23:08] yea, noticed that as well :) [23:08] hatch: I haven't entirely followed the conversation about develop being broken, but is there a fix we can land to allow us to work a bit easier? [23:08] haha, doesn't vim draw vertical lines on braces? [23:08] huwshimi we can make it fail silently but then it just breaks further down the line [23:09] huwshimi: not really, we're making the GUI do something it wasn't designed to, add a unit of a service before the service exists [23:09] hatch: So how are we going to move forward? [23:09] huwshimi: and before the machine/container does. So we're bending a lot of rules/laws internally to the GUI environment [23:09] There is no spoon. [23:10] huwshimi: so we've got to finish the last step in adjusting the environment to accept these new laws of juju [23:10] huwshimi atm I'm getting caught up on the ecs stuff [23:10] huwshimi: there's no 'quick fix' to make it work [23:10] hmm... just wondering how to test stuff with this broken then... [23:10] rick_h_: I've got unplaced units updating automatically. Not sure I can land this branch though since I had to shim in temp fixes for the broken stuff. [23:11] Now that they're showing up, they need a lot of CSS love, so I could work on make them pretty next. [23:11] Or I could move on to something else… [23:11] they look good on trunk [23:12] kadams54: rebase on trunk and huw's work to clean up the CSS on them should help [23:12] kadams54: if your branch works and is ready and can land go for it [23:12] k [23:12] we know where we're headed next and have to do it regardless so let's keep the ball rolling forward [23:13] kadams54: the next thing if you could look at is the bug card about the deployment summary switching to "no uncommitted changes" [23:13] kk [23:13] kadams54: there's that, and adding support for "new machine/container" now that it's landed in that summary [23:14] kadams54: check out the QA notes from Matt's branch that recently landed for how to manually add a machine into the ecs/db so that you can have it listed in the summary [23:16] Yeah, that came in handy for testing this branch [23:16] kadams54: ok cool, thanks [23:16] * rick_h_ goes off to dinner with for 20ish [23:20] LOL [23:20] https://twitter.com/iamdevloper/status/450905958139834368/photo/1/large?utm_source=fb&utm_medium=fb&utm_campaign=phessler&utm_content=465613662468997121 [23:23] so if someone tries to deploy a service without any placed units, it should.....? [23:26] hatch: All break horribly [23:26] so we need something to check that they placed at least a single unit before letting them hit 'confirm' in the deployer bar [23:26] :) [23:26] so many edge cases [23:27] Was at dinner. Sorry for the leftover comments, hatch. Trust me, it used to be worse :P [23:28] lol it's ok I was just poking fun [23:28] Makyo do you have time to hook up the add_unit now? [23:28] I'm working on learning it now, but company is coming in 30 [23:28] if not I can get on it after everyone gets outa my house!!!!! [23:28] :P [23:29] hatch, maybe later? just finished cooking, was about to sit down. [23:29] with a laptop???? ;) [23:30] lol jk it's ok [23:32] odly when I follow your qa steps I only get one machine rendered but there are two in the db [23:38] hatch: error on the summary page and make you go back and fix it [23:38] hatch: but yea, all these edge cases are stuff that luca wants to sort out as he tries it [23:38] hatch: behind the FF [23:39] hatch: because the second one is a container on the first (reason why you only get one machine) [23:39] it should show up in the container column then no? [23:39] hatch: when you click on the machine [23:39] hatch: the container column should be driven by the selection in the machine column [23:39] yeah, that's what I thought too, lemme try again [23:40] ohh, that's why, it's only adding a single machine [23:40] right, one machine with one lxc [23:41] now I see where I screwed up [23:41] I 'might' have the add_unit thing figured out [23:44] I lied [23:44] lol [23:46] a lazy added unit should probably still be created in the db right away [23:46] then updated [23:47] hatch: right, it needs to be updated when placed on the machine in the db/ecs [23:48] well I have something that I think will allow people to get unblocked at least [23:50] blah [23:50] ok people are arriving, consider me having accomplished nothing if anyone else wants to start on this add_unit stuff [23:51] ^ rick_h_ [23:51] hatch: rgr thanks [23:51] hatch: I'll try to send an email summary and maybe frankban can look at it his morning [23:51] thanks! have fun! [23:51] haha, I'll try [23:51] have a good night [23:51] cya [23:52] hatch: Night