=== kadams54 is now known as kadams54-away === urulama__ is now known as urulama === urulama is now known as urulama__|home| === urulama__|home| is now known as urulama [14:27] hatch: https://github.com/kadams54/juju-gui/commit/252f24ad38649027e64aff7e2252968a0b37fabc#diff-8d19084702f50f400641361a9cb78591 [14:27] hatch: tests for setMVVisibility [14:28] hmm is that enough? [14:29] there are 4 conditionals in the code so shouldn't there be a test for each path [14:30] it's definitly missing adding a key to the _highlightedServices prop [14:32] does one of the tests test the !machine.units story? (sorry I'm pre coffee but I don't see it) [14:33] speaking of which.....running real quick to grab a coffee :) [14:51] 1H 9M until I can buy linkin park tickets - I may miss our standup :P [14:53] the ticketmaster experience *refresh* *cuss* *refresh* *cuss* *get seat assignment* *website crash* *cuss* *refresh* *sold out* *wait 2 minutes* *refresh* *get same seat assignment* *cuss* [14:54] surely something better to do with your time :P [14:54] hatch: no link for the branch you've got in review [14:54] nah I almost booked a half day for this [14:54] :P jk [14:55] https://github.com/juju/juju-gui/pull/648 [14:55] that one? [14:55] hatch: yea, got it thanks [15:00] rick_h_: that thing you tweeted about would make for a cool charm for people who wanted to host it in their own env [15:00] jujugui need one more review https://github.com/juju/juju-gui/pull/648 [15:02] hatch: looking [15:02] hatch: right, like us or other properties in canonical [15:02] hatch: I want to look more but ran across it in my rss feed last night [15:02] it's a pretty cool idea if it works :) [15:03] yea [15:03] need actions so people can click 'update dictionary' :) [15:03] heh yea [15:04] was thinking they should go in the inspector [15:04] under another tab [15:04] unit and service details is the first thought, maybe running actions in the inspector for results? [15:04] because it's unit level as well the inspector is a bit harder [15:04] there is no service level? [15:04] there is, service + unit [15:05] well that's a pretty big oversight no? [15:05] if I wanted to update the internal dictionary from a remote location for all units I'd have to manually loop through them all? [15:06] hatch: missing key added to _highlightedServices prop is tested implicitly in both "highlights XXX properly" tests. Ditto for !machine.units, since none of the machines setup in beforeEach have units set. [15:06] huh? you can trigger actions at the service and unit level [15:07] ohh ok then that's good - sorry I missunderstood [15:07] hatch: so yes, you can do what you want, and with leader elections there's more power to them as well [15:07] perfect [15:07] where the service level trigger can get a single unit nominated to do the action/etc [15:08] yeah that will be nice - we've needed this functionality for a long time [15:08] yea [15:08] Alright, this is getting ridiculous. Just had my third merge build fail. Two failures due to NPM errors, one due to a failed notification test. [15:08] really would like an 'upgrade ghost' button which just goes and pulls the latest version [15:09] not sure how one would do a rolling update [15:09] but....yeah [15:09] :) [15:09] there was talk in brussels of apis to upgrade single units [15:10] oh? [15:10] yea, not sure when but the idea is out there [15:11] *notes that ticketmaster is down again* [15:11] well that's good to hear at least [15:11] kadams54: ok now I see how the tests work [15:11] could you maybe add some comments to that effect [15:11] it's not really clear that it's hitting those conditionals [15:13] Will do [15:17] kadams54: working on this latest branch I've noticed that there is a LOT of noise around the machine and unit events [15:18] easily a 5x multiplier on thenumber of machines/units per highlight [15:18] Yeah. [15:19] I noticed that we fire off change events for every machine in setMVVisibility [15:19] When in reality we should be checking for machines that actually change their hide attribute and only trigger events for those [15:20] I think re-doing things so that the service token modifies service flags directly, instead of firing an event, will also help quiet things down [15:21] I'll be creating some cards once I get this branch to land [15:23] kadams54: let's chat and walk through it before hand. I'm nervous of tokens making model changes. [15:24] Sure. I can chat whenever. tldr version is that changing the attributes directly would simplify a great deal. [15:38] kadams54: did you review 648? [15:39] hatch: No, no I did not. I forgot about it and made a breakfast sandwich instead. Lo siento. [15:39] Looking now. [15:40] do you? do you feeeeeeel it? [15:41] I think lo siento literally translated means "feel it" [15:41] :) [15:44] "Siento" is also the I form of "sentar", which means "to sit" [15:52] hatch: I'm worried about 648. [15:52] why? [15:52] hatch: the lines that you removed were put in by Makyo as part of the added services work [15:52] yeah, but they don't fit the spec [15:52] I tested highlighting and showing and they still seem to work, but I'm not entirely sure why. [15:53] because the code I removed was only for hiding container tokens [15:53] Specifically I'm worried about highlighting [15:53] right, what lines are we talking about in 648? there's no lines there really, just MV tokens? [15:54] Deleting lines 206-213 in container-token.js [15:55] What's supposed to happen if you have multiple containers on a machine and only one has the service that is highlighted? [15:55] Should the other containers be hidden (as we do with other machines)? [15:55] Or should they be shown, just without their units? [15:55] I don't think the specs we have cover that. [15:56] kadams54: according to the spec it doesn't say - but it says it should represent the environment [15:56] so keeping them visible makes sense [15:56] Why? [15:56] jujugui call in 6 kanban please [15:56] We hide other machines on highlight. [15:58] this is true [15:58] but I also think that's weird :) [15:59] no sense highlighting if the one you're highlighting is 3 scrolls down the page [15:59] I think showing a bunch of machines with nothing on them is pretty useless in highlighting. [15:59] I would sort them to the top [15:59] Yeah, what rick_h_ said :-) [15:59] so we either remove machines not highlighted or sort and redorder but now you're doing another evil thing and a pita to do :) [15:59] because hiding them when trying to be a representation of the hardware doesn't make sense [16:02] hatch: it's tough because there's not an easy 'grey it out' like in services [16:02] FWIW: both develop and this branch behave the same way when you have multiple containers and only one has a highlighted service: [16:02] hatch: let's try it one way and get user feedback and update [16:02] I'm with kadams54 on containers acting like machines for the first pass if it's not in the spec [16:03] They both display all containers and all services on those containers, even the ones that aren't highlighted. [16:03] As long as one service on the machine is the highlighted one. [16:06] ok it'll be considerably more difficult to do it the other way than I did [16:07] kadams54: did you want to start on that then while I continue on what I'm on? [16:08] hatch: No, not really. I'd like to talk with rick_h_ about having the tokens update the service flags directly and then work on that :-) [16:08] well this one is blocking release, that isn't [16:09] Given that this branch doesn't actually change behavior currently on develop wrt containers being hidden on highlight, I'm fine with landing it and will +1 after standup. [16:09] We can then bike shed about correct container behavior separately :-) [16:09] well it does change it - it makes it so it's not broken lol [16:10] atm containers with no units are hidden [16:10] Which has nothing to do with highlight. [16:10] So it doesn't change highlight behavior and it DOES fix brokeness elsewhere [16:10] Which is why I'm in favor of landing it. [16:11] yeah ok cool [16:11] you make a good point about it matching the machines column [16:12] see the issue is that there is no way (from within the token) to know if a container should be hidden based on some other service being hidden [16:12] it just knows that it has no units [16:18] kadams54: so can you +1 with your comment on the PR so I can land it? [16:18] Yeah. [16:31] uiteam [16:31] #first [17:06] kadams54: when on Monday did you want to start pairing on that stuff? [17:06] Right away? [17:07] so.....your 9:30 ? [17:07] * rick_h_ goes and takes a deep breath after thinking hatch was going to blame us for using views as widgets :P [17:08] lol! [17:08] hatch: whenever you get in. 7:30 seems awfully early to me especially on a Monday. [17:09] events sent [17:09] we are only 1H apart now remember :) [17:09] TZ party! [17:09] Ah yet [17:09] yes [17:10] rick_h_: I invited you as well but it'll probably be a bunch of fumbling around for the first little while :) [17:10] uiteam make sure you put the pr urls in the kanban cards please? I've trained myself now but been a few missing today/yesterday [17:10] hatch: yea, all good. I can let you guys get a plan together without me being nosy :) [17:10] hatch: I think you know where I'm coming from and have been deeply involved in the lessons of the past so :) [17:11] yeah we've had some.....discussions.....before [17:11] :P [17:11] I think as much as we go back/forth together we've put some good stuff in place. [17:11] rick_h_: kadams54 so once I finish this current branch I THINK that's all that's necessary for the AS release [17:11] yup I agree [17:12] hatch: rgr, will do QA today. Going to deploy to guimaas later on and put it to good use with real env [17:12] someone say uiteam [17:12] uiteam [17:12] uiteam [17:12] * rick_h_ wonders if /reload works or if I have to close/reopen irssi to pick up the higlight [17:12] ooh yay ty [17:12] /reload ftw [17:12] say ui ... uuuuiiii ... say team ... teeeeeaaaam [17:13] lol [17:13] haha [17:13] I can't find where I change the highlights [17:13] * rick_h_ will like one highlight to rule them all [17:13] yeah that was a good idea [17:14] rick_h_, hatch: I'd like to look into getting bundle icons to show up in machine view when you first deploy. https://launchpad.net/bugs/1364493 [17:14] ok there we go [17:14] Bug #1364493: Machine token charm icons do not show when you add a bundle [17:14] kadams54: heh that's a deep one [17:14] pre-imp? [17:14] Yeah, but it annoys the crap out of me. [17:15] kadams54: hmm, I'd hold off on that. I've got a chunk of work this cycle to get first page loads to work without data and I think that falls into this? [17:15] kadams54: if there's a temp quick fix (1-2days) ok, but we knowwe've got a bigger problem and it's on the schedule for later this cycle [17:15] Yeah, sure [17:15] k [17:15] I'm pretty sure there isn't [17:15] at least not a clean one [17:16] someone say ui team :) [17:16] What about this one? https://bugs.launchpad.net/juju-gui/+bug/1379653 [17:16] Bug #1379653: Health bar in the inspector is in the wrong position [17:16] hatch: yea, that's my fear and why I've got a multi week chunk of work on the schedule for the cycle [17:16] uiteam [17:16] woot [17:16] thanks [17:16] kadams54: sounds good to me [17:16] and yeahgood idea with the time block [17:16] +1 [17:16] kadams54: the notification count one would be good pre-release [17:16] kadams54: as that's a bit of a regression [17:17] On it [17:17] kadams54: ty [17:17] * frankban removes juju-gui, jujugui and guihelp from the alerts [17:18] frankban: heh I think you were the last remaining guihelp user [17:18] heh still have that one [17:18] I still have guihelp [17:18] as well as jujugui [17:18] right, but who used it :) [17:18] I used it all the time! [17:19] Maybe that explains why I could never get anyone to do code reviews… ;-) [17:19] hatch: I suspected that [17:19] yea I never got two highlights that hit the same people [17:20] haha [17:20] it seemed like giving me two different cell phone numbers to call you on the same phone [17:21] did you guys ever have 'identicall' where you had multiple numbers that rang differently on the same phone? [17:21] this would have been back in the late 90s early 2k? [17:21] yea, when we got our second line when I was a teenager I had that a little bit [17:22] heh and the caller id separate box [17:22] ahh people have it too easy today [17:23] "back when I was your age" === kadams54 is now known as kadams54-away === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [18:45] afk - going to take the dog for a short walk and enjoy some rare sunshine. === kadams54 is now known as kadams54-away [19:08] hatch: if you think yosemite looks bad on non-retina display, imagine how yosemite looks on a hdtm tv! [19:09] hatch: that thing is horrible! [19:17] urulama: lol! [19:22] sigh got to reboot again, vm is being wako === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away