[00:01] hatch: does jcsackett's branch work with yours then? e.g. can I qa them together and get a drag/drop with a ghost unit? [00:01] I think so, his branch doesn't auto update, you have to close then re-open the machine view for his to render the units [00:01] there is a card on the board to databind that list [00:01] k [00:17] hatch: is the other one in review in kanban up on github? [00:17] going to get ice cream, will look at more soon [00:19] rick_h_, my deployer summary branch is ready for another review [00:20] rick_h_, Kyle QA'ed but not merged yet [00:29] rick_h_ this branch does both of them [00:29] unless you've found otherwise [00:29] antdillon: I'm currently working on tests for the branch; I'll merge once those tests are finished. [00:31] kadams54, Ah thanks, wasnt should if I needed to put in another pull request or not [00:31] antdillon: Nope, I'll take it from here. Thanks! [00:32] kadams54, Thanks man [00:47] antdillon: cool, will look [00:47] oh kadams54 is the case [00:47] hatch: oh cool, yea didn't notice it tbh, but now that you mention it [00:47] :) [00:47] 2-birds [00:51] I may end up doing the token drag and the dropping [00:51] the tasks may end going hand in hand [00:51] hatch: yea, I wasn't sure how much parallel we could get out of it [00:52] and I figure once one is done, the other is almost a dupe to the same calls, just with a parent id [00:52] yeah [00:52] for the containre to know what machine it's in [00:52] will play it by ear [00:52] brb gota go bring the plants in [00:52] supposed to get to -4 tonight [00:52] ouch [00:52] 10 here [01:04] yeah it's a little odd it's getting so cold [01:04] my guess is that it doesn't get that cold [01:07] well we had a frost warning this week [01:07] but think it'll be the last [01:26] ok, emailed Mark S with a link to the latest on comingsoon. /me ducks [01:29] :) [01:30] how goes huwshimi? [01:30] rick_h_: Good thanks. Nice to have some family time :) [01:31] (laptop next to me as we play with Lego) [01:31] rick_h_: Sounds like your family time is interrupted by 4g :) [01:31] huwshimi: well, I think I'm going to bed on time tonight [01:32] but yea, told the wife I'd take her camping for mothers day so got the camper out here [01:32] didn't say I'd stay off the laptop :P [01:32] haha [01:32] but about beat, crazy day today [01:32] rick_h_: Make sure you assign anything you want to me for Monday. I'll see if I can pick up something over the weekend too. [01:33] huwshimi: sure thing, thanks. I honestly think we're in good shape. If matt and frankban get the change set stuff working and the rest of us get drag/drop working it's demo-able [01:33] huwshimi: so it'll probably be asking you to follow behind us and clean up any little things to help polish it up. [01:34] there's definitely some spacing things atm, I'll try to look at it with clear eyes tomorrow [01:35] rick_h_: Yep, no problems. I'll try and do a bit of general QA once those bits are in. [13:12] morning, phew crashed last night [13:12] love a good sleep outdoors [13:29] rick_h_: Morning [13:30] rick_h_: Saw your comments on the tests… FYI, I have the problem narrowed down to one line of code, but I don't understand why that one line is problematic or how to fix it. [14:16] rick_h_: Looking for my next card… as I understand, the drag/drop stuff is closely tied to Jeff's work. Maybe the card for databinding the unplaced units column? [14:29] kadams54: +1 [14:29] kk [14:29] kadams54: I'm not sure if that needs to hook into databinding or just YUI events watching the db now [14:29] he made the card to say databinding, but I think it's just db watching like other places in the machine panel view [14:38] kadams54: a couple of QA things on the summary, I noted them in the card under polish in the description [14:42] ok [15:35] ok, status email sent, couple cards added for polish/etc. Going to take a break an walk the dog [15:36] biab [17:30] hello fellow weekenders [17:46] hatch, Hey [17:46] hows it going antdillon ? pulling some weekend hours too? [17:46] hatch, Na it official like, just checking all's good [17:47] so far :) [17:47] rick commented that there were some qa issues with your branch [17:47] not sure what they were yet [17:48] I probably won't get a chance to look at it this wkend [17:53] hatch, I'll take a quick look [17:54] hatch, I fixed up Kyle QA last night. He's landed it all now [17:55] oh cool [17:55] jcsackett are you around? [18:00] hatch, I'll be around a bit tomorrow to help out if your guys have any styling stuff [18:00] ok cool [18:11] hatch, ping [18:11] hey [18:11] Running into a weird JS problem. [18:12] ok, shoot [18:13] this.changeSet has one property, addMachines-866, but this.changeSet['addMachines-866'] is undefined. In the JS console, it shows as (...) [18:13] Have you run into this before? [18:13] I remember I have, but I can't remember the workaround [18:14] try console.dir(this.changeSet['addMachines-866']) [18:14] console looks like https://gist.github.com/makyo/25153ef6db7b152dc8e5 [18:14] yeah, try using console.dir [18:14] Just says undefined. [18:14] hmm [18:14] Trying to figure out where it's getting set to that. [18:15] ohh, are these defined properties [18:15] Nope. [18:15] Makyo, I used the changeSet ok in the deployer-bar.js is that helps [18:15] has antdillon's branch landed in yours? The one with the Object.observe pollyfill? [18:16] I wonder if that's converting it to a defined property [18:16] Hmm, maybe. [18:16] the getter and setter is throwing me off [18:16] why would those be there if there wasn't a defined property [18:16] Ah, yep. [18:16] That's probably what's happening. [18:17] Alright, thanks. [18:17] np, not sure what the solution is though :) [18:18] Me either, but it gets me on the right path as to why it's still in the object. [18:19] Or not, hah. Well, I'll keep poking around for another few minutes. [18:25] hey guys [18:26] bah, missed ant [18:26] if ant comes back we could use some info on what design he was doing the deployment summary to [18:26] the qa issues were after it landed, so they're current issues [18:26] hatch: let me take jc's card then you and you run with drag/drop stuff [18:26] Makyo darn...well I'm about to land jc's branch [18:27] hatch: ok, well never mind [18:27] yeah the tokens are super broken though [18:27] but they show up [18:27] ok, well we can update them [18:27] yea, that's good enough, to have divs to start to drag/drop [18:27] I can look at them then and see if I can update them to less broken state [18:27] yep got that :) [18:27] just running local lint and tests [18:27] coolio [18:40] hatch: prioritize dropping on a container in an existing machine over new machine [18:41] hatch: sounds like that's the demo, have an existing bnudle running, add nagios server, place it on an existing machine, hit deploy [18:42] ok the demo wireframe has it dropping on a machine as well [18:42] If you download the pdf you can zoom into it more [18:42] hatch, it's Object.observe's fault. I have a workaround. [18:42] hatch: yea, forget that wireframe [18:42] hatch: new plan [18:42] Makyo ahh, well that was a pretty good guess then :) [18:42] rick_h_ lol oh boy [18:43] jujugui just copied everyone on an email where I'm working with othres to get into the official demo plan on things [18:43] hatch: well it's simpler tbh [18:43] yeah - I'm happy with simpler :) [18:43] hatch: no relations, no config, just place one service on an existing machine [18:43] hit deploy, get the summary, confirm, done [18:43] drag from unplaced to a container.....got it [18:43] hatch: rgr [18:44] Makyo: ^ I don't think it savesus anything on the ecs-machine end but fyi [18:48] css shapes are pretty cool http://alistapart.com/article/css-shapes-101 [19:03] hehe python interpreter written in JS is faster than Cpython :P https://rfk.id.au/blog/entry/pypy-js-faster-than-cpython [19:12] :P [19:30] rick_h_ I've figured out what locks us in the machine view panel [19:31] hatch: orly cool [19:31] if you try and add units to a service which doesn't exist, then click machine view, you can't click back [19:31] the addUnit code isn't resilient I guess [19:32] gotcha, ok. Well we can avoid that thankfully for now and will add a card [19:32] card added [19:34] hatch: ok, so with JC's branch and yours we don't see tokens from undeployed units because creating a service doesn't create the one ghost unit yet right? [19:34] correct [19:34] you actually cannot deploy a service with 0 units [19:34] it errors [19:35] but that doesn't mean that that unit couldn't be unplaced I suppose [19:35] I haven't looked into why it errors yet [19:35] hatch: ok, I wonder if kadams's branch touches this at all [19:40] umm [19:40] no it wouldn't [19:40] but I'd like to chat with him about that branch because he can use the same code that's used in the machine list and mass scale up service list [19:40] save him a bunch of time writing it himself [19:40] rick_h_ was he putting in time this weekend? [19:41] hatch: yea, he was around this morning [19:41] hatch: k, use email then for now to send that out [19:41] and I'll try to catch him and point him at it [19:41] ok cool, will do [19:46] email sent, and cc'd u [19:46] ok, so I'll start look at that getting a single unplaced service going. Then we can look at styling the token shown to be pretty [19:46] hatch: cool thanks [19:48] yeah that sounds good - have fun in environment land....there be dragons in there :) [19:48] ruh roh [19:48] ben wrote it, blame him [19:48] lol [19:48] lol, good to know [19:48] this demo is nuts [19:49] at least he made the bundle look pretty [19:49] ish [19:49] :) [19:51] it would be cool if you clicked on a service if it would hide all services not related to that one [19:51] in a complex env like that [19:52] yea, that's the plan :P [19:52] in the filter settings ui for the left panel [20:09] rick_h_ Once the user starts to drag, the headers turn into their drop target UIs, The container/machine tokens stay the same UI. And if the user drops on a blank space the unplaced token being dragged snaps back to the original place and everything returns to normal [20:09] that all sounds like the proper UX? [20:12] hatch: sounds good to me [20:13] :+1: [20:16] Woo, only 233 failures. [20:17] lol [20:23] party on! [20:23] :beer: [20:27] Oo boy I forgot about IE testing.... [20:27] heh [20:29] follow up card, we don't be using IE [20:29] now the question is....how do I do testing in IE with vagrant running... [20:29] hmm [20:31] fire up a sauce IE session and point it at your home network [20:31] or bring up an IE VM using anything else [20:31] can't you run vbox/fusion/etc locally? [20:31] and point it at the vagrant ip [20:31] vagrant uses virtual box, and my win 8 vm is on parallels [20:31] opening up both vm engines causes a kernel panic [20:32] parallels runs windows extremely well [20:32] oh, well that sucks [20:32] so create an IE VM in virtualbox? [20:32] the good news is that dragging and dropping a unit on a container token (third column) now works [20:36] rick_h_ when I drag and drop a charm then click deploy the deploy list shows a service to deploy then switches to 'no uncommitted changes' right away..... [20:36] confirm still works....but did you know about this bug? [20:37] no, I've not seen it [20:37] oh now I do :/ [20:37] hmm, it was just working [20:37] I added cards to fix broken icons and such on it [20:38] actually it works in my fresh checkout [20:38] hatch: try before your last landing of pr/298? [20:39] hatch: hmm, only does it the first time the button is pressed [20:39] yeah [20:39] close and go back and it's there [20:39] this was somehow caused by the changes from jc's branch you're saying? [20:40] woot got a token to show up [20:40] Down to 2 failures, but now tests crash., [20:41] hatch: well I'm asking. I hadn't seen this behavior until you just mentioned it and I qa'd it several times today [20:41] hatch: so pondering something changed [20:41] Makyo: heh, comment out the deployer bar tests and see if it still happens [20:41] Makyo: in the index.html, those were crashing for me though kadams thinks he got the one doing it [20:42] oh, I've got a few too many changes right now to bisect back but I can in a while [20:42] hatch: rgr, I'll add a card to look into it [20:42] Makyo tests always crash for me, I have to use test-server [20:42] hatch, this is test server [20:42] it's painfully slow, but.....such is life [20:42] oh.... [20:42] :'( [20:42] Something's inflooping [20:46] Huh, must just be a browser thing. All tests pass in test-debug; the two failures were about the cookies being set. [20:46] So...bonus! [20:46] heh woot [20:46] does it crash in both FF and chrome? [20:47] Haven't tried FF yet. Will now. [20:49] rick_h_, Everything's fine in FF. Just some weirdness in chrome, I guess. [20:50] Makyo: hmm, ok good to know [20:50] Oh, I had a another window of tabs open. Wonder if something's stealing resources. [20:51] nom nom resources [20:52] bah, this is such a twisted bit of code [20:53] was never meant to do this [20:57] Chrome works, but still has the two cookie banner failures. [20:57] Oh well. [20:57] Let me commit and push. guihelp if anyone has a moment, I might ask for a test-run on chrome [20:57] Makyo: rgr [20:58] I might need a pre-imp or help with this branch [20:58] I somehow got it to work once but now can't again and no idea how it worked the first time [20:59] guihelp: https://github.com/makyo/juju-gui/tree/more-ecs-calls [21:00] Makyo: pulling down [21:01] ty verizon for letting me know I've not hit 75% of my monthly bandwidth allotment [21:05] Makyo: ran here, one failure but can't find it yet [21:06] when I click on the 1 failure it opens "removes the drag handlers" test [21:06] no cookie issues here [21:08] Hmm. [21:08] rick_h_ how much data do you have on that hotspot? [21:09] hatch: 10GB/mo [21:09] oh haha that's tons [21:09] crossed 7.5 with about 8 days left in the cycle [21:09] lol netflix? [21:09] well, with the sprint and such [21:09] no, no movies [21:09] I've only ever once gone over and that's from downloading hd movies on itunes [21:09] did that once, downloaded a movie for the boy while camping, 3.3GB eats a ton of space [21:10] I did get a new hotspot now that should work internationally so can only have one wheee [21:11] oh rock on [21:11] what's that plan cost? must be pricey [21:11] ok, well the 10G is shared between hotspot, my phone, wife's phone [21:11] for all three it's aroud $200/mo [21:12] my phone usually does 1-2GB and she does around 500MB and rest is for hotspot [21:12] where the $#@$#@ does the ghost service get its unique id from? [21:14] ouch your cellular service is such a ripoff [21:14] woot! [21:15] wow and I do get a machine with the service [21:15] crazy [21:15] app/models/models.js:494 [21:17] Makyo: thanks, I realized I could just use the ghostService there. So I can drag/drop and get an unplaced unit token to show up [21:17] total hack but wheee [21:20] hmm, wonder if you can deploy a 0 unit service to lxc [21:22] ok tokens can now be dragged and dropped [21:23] jujugui should we destroy/re-create units when the user places an unplaced unit in a machine....or modify it the original [21:23] ^ this will have to be ecs'able [21:24] good morning huwshimi [21:26] hatch: Morning [21:26] hatch: How's it all going? [21:27] well I think [21:28] Good : [21:28] *:) [21:28] I landed jc's branch so it can be styled now [21:28] are you working today? [21:28] it's like 730am there right? [21:29] hatch: https://github.com/juju/juju-gui/pull/299 is what I've got working [21:29] looking [21:29] hatch: so I think the only thing we'd have to modify is the machine [21:29] though I'm still working through how this is going to work. [21:29] right, but the act of dragging an unplaced unit to a container still needs to be added to the ecs right? [21:30] so we need a 'modify unit' in the ecs as well [21:30] cool hack.....not sure that feels right though as an official implementation :) [21:31] or maybe it does..... [21:31] hatch: yea, no kidding. But if addUnits is ecs-ified... [21:31] a service without a unit.....is kind of pointless..... [21:31] so they need at least one [21:31] hatch: though, what about our local charm story of adding a charm to the env without deploying it? [21:31] well that's a charm, not a service [21:32] and in this case you still end up with a unit, just not as part of env.deploy call [21:32] true [21:32] I'm wondering if a deploy call should, no matter what, create a unit, but be unplaced [21:32] so revert your changes, but change what it does in the back end to mirror your changes [21:33] I was looking in there but not all the data is there [21:33] hmmm [21:33] I'll look there, just working on how the parts fit atm [21:34] I've tracked what's been going on but not actually been diving in here a ton :/ [21:34] hatch: anyway, I'd expect the drag/drop to just modify the machine info on the unplaced unit data in the db [21:35] ok, but via the ecs [21:35] hatch: 100% [21:35] so ecs.modifyUnit('unitId', 'field', 'newVal') [21:35] ish [21:35] hatch: I'll try and get some stuff done a bit later. It's mother's day here so we're going out for breakfast and then I'll see what I can get done. [21:36] hatch: (yeah, it's 7:30am) [21:36] or maybe ecs.placeUnit() which would sync up the pending machine/container and the pending unit or what not [21:36] huwshimi cool np, yeah mothers day here tomorrow too [21:36] huwshimi: tell mom we say hello! [21:36] ahh placeUnit much better [21:36] rick_h_: hehe, will do :) [21:36] new card! [21:36] hatch: lol [21:39] ok back to writing tests *grumble grumble grumble* [21:47] hmm, ok so we're getting crazy version numbers from somewhere and that's what caused it to not work in lxc [21:51] crazy version numbers? [21:51] it's part of the deploy stuff modifying the serviceName [21:52] ohh, yeah I added a random number generated to the end of the serviceName [21:52] without the inspector there is no way to modify it, so I didn't want it to blow up if people used two of the same charms [21:52] right, but that looks like a service-version so it tries to deploy rev XXX of mysql charm [21:52] gotcha [21:53] yea, we'll have to figure that out [21:53] ohhh heh, that can be fixed by removing the - then [21:53] mysql1234 [21:53] for example [21:53] oops, thats my bad [21:53] I didn't realize that it would have just split on the - [21:53] all good [21:53] $500,000 for 1300sq/ft house....this city is going bonkers [21:54] woo! my house will make me rich! [21:54] except here it's 1/4 of that [21:54] so you should move down here [21:55] ok, taking the boy to the park for a bit, biab [21:55] haha, cya [23:13] hmm... how do I get an unplaced unit? [23:15] oh, I see we now have numbers as part of the service name [23:22] huwshimi: check out the QA instrcutions for the branch [23:22] huwshimi: https://github.com/juju/juju-gui/pull/296 [23:22] for instance [23:28] rick_h_: Thanks === alexpilotti_ is now known as alexpilotti