[00:02] hatch: Just wondering if you know how the ambiguous relation menu is supposed to look? [00:38] huwshimi: hey - it's supposed to look better [00:38] :) [00:39] hatch: OK, better :) [00:39] yeah there are no mockups so just better :) [00:41] alrighty then === kadams54-away is now known as kadams54 [02:46] hatch: I saw the PR of all PRs finally landed. === kadams54 is now known as kadams54-away === urulama_ is now known as urulama [14:32] jujugui lf a review and qa on removing the as flag https://github.com/juju/juju-gui/pull/643 [14:32] hatch: looking now. [14:32] thanks [14:41] jujugui need one more review for ^ [14:42] hatch: looking [14:42] thanks [14:42] rick_h_: are you ok with the flag removal landing? [14:43] hatch: is it ready? :) [14:43] hatch: yea, I need to find time to get over to the GUI side and QA and such. Will try to do that once we get the release going today [14:43] well there is one known bug - so i'd like to get the flag removed so we can catch any related bugs while doing the remaining work [14:44] hatch: rgr [14:44] there were so many mv related bugs after removing the flag that I want to avoid that [14:44] hatch: what's the one known bug? [14:44] * kadams54 really hopes it's the card he's working on. [14:45] kadams54: it's theorange card in this project [14:45] about the toggling of the highlight button [14:47] kadams54: is that the one you are working on? I was just about to grab it [14:47] No [14:47] ok taking [14:48] But I disagree with the card that unhighlight needs to be changed to method calls [14:48] Huw did a great job styling the ambiguous relation dialog last night [14:48] Or rather, I'm highly skeptical ;-) [14:48] because huw is awesome :) [14:48] haha he is [14:48] kadams54: yeah first I need to figure out why it only is an issue with ghost services and not deployed ones [14:48] then will figure out an appropriate fix [14:48] Good stuff - looking forward to seeing it in action === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [15:17] for manage.jujucharms.com, I only get 20 results back everytime I search which makes queries like type=approved&series=trusty worthless [15:18] marcoceppi: &limit=1000? [15:18] changes nothing [15:19] kadams54: fixed it [15:19] without changing it to a method === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [15:19] just going to write the test now and then I'll have it up [15:19] hatch: sweet. What needed fixing? [15:19] marcoceppi: looking [15:20] Sign me up for QA and review :-) [15:20] heh will do [15:20] you were on the right track with your fix but looking the wrong direction ;) [15:22] now the question is...does it work with relations [15:23] And does it work with machine view [15:23] *snif* it does *snif* it's...so....beautiful [15:23] And does it work when you turn highlight on for one service and hide 3 other services, two of which aren't related at all to the highlighted one. [15:24] And does it work when you go in the reverse direction [15:24] kadams54: snap out of it! *slap* [15:24] kadams54: Thanks, I needed that. [15:24] yep everything is good - but ghosted services don't hide properly in the machine views [15:24] I'll make a card for that for you [15:25] *sob* [15:25] It took us a week to implement 90% of added services. [15:25] hah - well you know the mv stuff the best [15:26] And then someone *ahem* said, "hey, could we tweak how this works?" And bam, three weeks later. [15:26] yeah but it's stable now [15:26] :) [15:26] marcoceppi: ugh, not sure man. We've not landed any new code there in forever but yes, all queries are maxing out at 20 results. [15:26] \o/ [15:27] cool, well that's fine for now I just won't mess with charm-tools for a bit [15:27] marcoceppi: https://api.jujucharms.com/v4/search?series=trusty&limit=600 [15:27] :OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOo [15:27] if you need some info right now the new api has it working [15:27] ITS SO FAST [15:27] :) only the ids returned [15:27] but yea, much faster [15:27] marcoceppi: but don't rely on this yet. It's still early and the url is going to change/etc [15:27] also, whats the way to get just approved [15:27] or promulgated or whatever [15:28] too late, already dependant on it [15:28] marcoceppi: but can help https://api.jujucharms.com/v4/search?series=trusty&owner=&limit=600 [15:28] marcoceppi: setting owner= (empty) will show approved [15:28] \o/ [15:28] kadams54: are there no tests for the onHighlightToggle methods? [15:28] marcoceppi: raising a bug on charmworld and will see. For now let me konw and we can help get you the info needed from the new api endpoint to unblock. [15:29] that unblocks me for now [15:29] hatch: There should be. One moment… [15:29] ahh they are tested as side effects [15:29] won't put it in the charmtools stuff [15:30] hatch: test_added_services_view.js, around line 304 [15:30] That tests directly. [15:32] well it doesnt' send any service names [15:32] looks like I'll have to modify the suite so I can test with a ghost service id too [15:36] rick_h_: thanks, here's the bug whenever you guys get time https://bugs.launchpad.net/charmworld/+bug/1390127 no real rush [15:37] Bug #1390127: All queries are limited to 20 results regardless of limit parameter [15:37] marcoceppi: ty [15:53] jujugui call in 8 please kanban [15:56] jujugui lf a review and qa https://github.com/juju/juju-gui/pull/645 [15:56] kadams54: ^ [15:57] Will look after standup [15:57] hows your branch coming along? [15:57] jcsackett: I'll do the other review on huws branch [16:19] kadams54: your card in the maintenance lane - is it landed? [16:21] hatch: ah, yeah, missed that [16:26] jcsackett: I'm not able to reproduce this bug any longer - could you give it a try on comingsoon to see if it's indeed fixed? https://bugs.launchpad.net/juju-gui/+bug/1339093 [16:26] Bug #1339093: Drag and drop failing with syntax error [16:27] hatch: looks fixed to me. [16:27] thanks, closing [16:30] jujugui anyone need any reviews? [16:35] hatch, Makyo: https://github.com/juju/juju-gui/pull/642 is ready for reviews and QA. [16:36] kadams54: you didn't reply to my comment [16:37] actually I'm pretty confused by the override thing too [16:38] hatch: I just replied to your comment. [16:38] And I addressed the override thing. [16:41] kadams54: ok so your reply should be in a comment in the code - github is being odd and not showing replies in the code for some reason :/ [16:41] I did put a comment in the code… the XXX bit. [16:41] I do't see any comment about the override [16:41] Oh, by "comment in the code", do you mean on the PR? [16:41] Or in the code in the code? [16:41] right but that doesn't say why you couldn't use the real property in the template [16:42] Ah, OK [16:42] and I still don't understand override [16:42] that api seems very....fragile [16:42] It is [16:42] And it's all temporary [16:43] There needs to be some way to distinguish between renders that happen after a button click and those that happen on initial view render. [16:43] The override flag does that job. [16:44] We need to be able to tell, because in one situation (initial render) we want to override the local attributes with the service attributes. In the other (after a button click) we don't want to override. [16:44] We don't want to override because the service attributes are being set to their new values asynchronously and may actually still be the old values at render time. [16:45] Which would result in the just-clicked button being rendered as inactive. [16:45] All of this is way too much for code that is intended to be very short-lived, so I'd rather just stick with the current XXX comment and get a card in for the long-term fix. [16:47] kadams54: ok review done [16:48] rick_h_: I'm going to work on https://bugs.launchpad.net/juju-gui/+bug/1375918 - it might take a while, I have no idea :) [16:48] Bug #1375918: units can be created without a service causing cascading failures [17:00] kadams54: maybe I'm doing something wrong but I can't get your branch to qa properly [17:01] hatch: how so? [17:01] it doesn't work heh [17:01] even after clearing cache and all that [17:01] it's like nothing has changed wrt the button statuses [17:01] maybe could you try rebasing your branch with develop and checking locally [17:02] possible something landed that broke it? [17:05] frankban: hey are you still in? [17:05] hatch: yes on call [17:05] np ping when you have a moment [17:05] hatch: Bah, no, I broke it myself in the last commit, the one called "review feedback" [17:05] :) ok lemme know when it's fixed [17:08] Fixed. [17:10] testing [17:12] kadams54: +1 [17:15] does demo.jujucharms.com use a different font than when deployed locally? [17:16] they appear to be the same [17:19] https://bugs.launchpad.net/juju-gui/+bug/1390161 [17:19] Bug #1390161: demo.jujucharms.com and locally deployed GUI has different weighted fonts [17:19] hatch: what's up? [17:20] frankban: well I'm going to be working on https://bugs.launchpad.net/juju-gui/+bug/1375918 [17:20] Bug #1375918: units can be created without a service causing cascading failures [17:20] so this is the one about creating units using 'new' instead of trying to guess the id [17:21] hatch: ok [17:21] I was just wondering if there were any issues you could forsee using this approach [17:21] I can't htink of anything - the machine view seems to work well with it [17:21] * frankban thinks [17:24] hatch: the more I think about it, the more I feel that units and machines are very similar on that perspective. so I don't see problems (excluding the required changes on the ecs) [17:24] yeah this change is going to be substantial for sure [17:24] what I don't want to do is implement it then think of a better approach [17:24] either for units or machins [17:24] machines [17:25] newX is certainly prettier than some random id string [17:25] hatch: we can never know what's the next id assigned to new entities in juju [17:26] hatch: that's why the new prefix succeeded in putting ghost machines under a different namespace, avoiding the clashes we have for units [17:26] hatch: so mysql/new1, wordpress/new2 sounds reasonable and also does not look bad [17:26] true true [17:27] this will all fall apart if juju gets ecs in core [17:27] so I suppose that's not a concern either [17:28] hatch: exactly, once we have more state from juju we can avoid pretending to have a state in the browser [17:29] frankban: ok thanks - I'm now more confident that I can emulate the same approach with units :) [17:29] hatch: yes np [17:33] rick_h_: I think this is a critical regression https://bugs.launchpad.net/juju-gui/+bug/1390165 [17:33] Bug #1390165: container tokens are not rendered [17:33] it exists in the released version [17:33] hatch: rgr, then please start that, I'm afk for a little bit [17:33] will do [17:54] kadams54: I see you have picked up the card I created - were you able to reproduce the issue or do you need more details [17:54] re the ghost service machine token hiding [17:54] Yeah, I reproduced it [17:54] ok great [17:54] I'm going to grab some lunch [18:48] hatch: I know why ghost machines aren't being highlighted properly. I just don't know how to fix it. [18:52] Well, maybe I do. [18:57] Seems that I do. Yay. [19:03] haha yay [19:04] jujugui need one more review on https://github.com/juju/juju-gui/pull/645 [19:05] jujugui trivial 1 review/qa for https://github.com/juju/juju-gui/pull/646 [19:06] sure [19:06] I'll look at both. [19:06] Or hatch and I will split. [19:06] Either way. [19:06] kadams54: he still needs 2 [19:06] and I need one that's not him :) [19:06] so yep you'll do both ;) [19:08] +1 Makyo [19:14] Makyo, hatch: For the trifecta: https://github.com/juju/juju-gui/pull/647 [19:14] On it [19:14] kadams54: interesting... [19:15] kadams54: needs tests [19:15] :) [19:15] might as well start on that now :P [19:15] hatch: got a merge conflict when trying to pull in your #645 [19:16] ahh probably in the tests? [19:16] ok fixing [19:16] hatch: IIRC, setMVVisibility was tested indirectly. I did verify that all tests passed. That said, having some direct tests would not be a bad idea, but I'd prefer to handle it in a separate card. [19:17] kadams54: right - but there was a bug [19:17] you fixed it [19:17] the tests didn't change [19:17] so the tests were insufficient [19:17] Not my fault the original author didn't write any direct tests :-) [19:17] well even the indirect tests are likely insufficient then [19:18] I agree, I just think that's outside the scope of this card. [19:20] how do you figure? [19:20] you didn't update the tests to test for the failure [19:20] so how can we be sure that the next branch that lands doesn't break it again? [19:21] It would be one thing if there had been tests originally and I just needed to update them. But since they need to be written from scratch, I feel like it's a big enough chunk of work to justify a separate card. [19:21] Which I will tackle next. [19:23] I'm pretty sure it's just an extra couple assertions in the browser events test file [19:24] in an existing test [19:24] but if you would like to create a new test suite for the unit tests I suppose that would be more thorough [19:24] I'd like to get something pretty thorough wrapped around it. [19:24] It's a pretty important bit of logic. [19:25] Especially with that crazy nested looping going on. [19:26] conflict resolved on #645 [19:30] hatch: commented on #645. I'll get to work on a test suite for setMVVisibility. [19:32] kadams54: so I had thought of that initially but was concerned about people modifying the ghost display name then playing with the added services stuff [19:33] tbh I'm not sure it's an issue [19:34] actually it shouldn't matter because it's in a blocking loop [19:34] will make the change [19:35] afp is a frustrating protocol, if it's not being used it closes the connection [19:35] then doesn't reopen when you try to access [19:36] and unfortunately there is no --keep-alive [19:43] Why are you dealing with AFP?!? [19:43] (Unless you mean something other than Apple Filing Protocol.) [19:43] lol [19:44] kadams54: my host os is OSX so in order to properly mount my NAS into OSX I need to use AFP [19:44] You can't use SMB? [19:44] OSX has huge issues with mounting anything using 'normal' prototcols [19:44] Apple has had AFP deprecated for years now, though they only started migrating away from it in Mavericks. [19:45] But still, you're on Yosemite, so you ought to be able to use SMB. [19:45] are you sure you'r enot mixing it up? [19:45] Pretty sure. [19:45] I moved to AFP in mavericks because other techniques woudln't work [19:45] And you're sure the problem wasn't PIBKAC? [19:46] I was originally using SSHFS I believe [19:46] which was poorly supported [19:46] NFS etc [19:47] maybe that's fixed and I can stop using this horrible bs [19:47] hah [19:47] http://appleinsider.com/articles/13/06/11/apple-shifts-from-afp-file-sharing-to-smb2-in-os-x-109-mavericks [19:48] lemme check if my synology supports smb [19:48] What file system are you using on your NAS? [19:50] ahh just updated it and there is a smb2 [19:50] option [19:50] it's under Windows though [19:50] very odd haha [19:51] well I know what I'm going to be doing later [20:04] kadams54: ok updated [20:07] +1'd [20:07] hatch: ^ [20:07] word [20:07] Linkin Park is coming to Stoon [20:07] w00t [20:07] too bad they likely won't be playing their first two albums wholesale :) [20:11] kadams54: looks like you can land #647 [20:12] then hopefully rick_h_ can do a real deep qa on it all :) [20:12] I've got this test suite half written, so I'll get that included, just to make you happy hatch :-) [20:12] lol [20:13] I think we have all the functionality hammered out now [20:13] I haven't noticed any bugs since these most recent branches [20:17] wheee [20:27] * rick_h_ crush qa boom [20:37] haha [20:38] rick_h_: it's going to be probably an hour before all the pending branches land then i'll be ready for qa [20:38] or you could merge them into your local branch if you feel like you want to do it now :) [20:38] hatch: all good, I've got broken kid at home so won't be able to go through it well until after he goes to bed [20:38] hah ok np [20:39] Heading out to grab an early dinner with the family. Will be back in about an hour to land #647. [20:39] coolio === kadams54 is now known as kadams54-away [20:50] hmm for some reason the container tokens have a hidden class applid [20:50] applied even [20:55] crap I found another bug [21:18] ok - container bug fixed and pushed to PR [21:18] and new bug created [21:18] https://bugs.launchpad.net/juju-gui/+bug/1390228 [21:18] Bug #1390228: Hiding services should update container column machine === urulama is now known as urulama__ [21:58] Morning [22:26] hatch: we're going to skip the AU call tonight if that's cool [22:45] yeah np [22:45] sorry i missed the ding [22:47] hatch: You didn't miss it, we just started very early :) [22:56] go luca go [22:57] :) === kadams54 is now known as kadams54-away