2014/10/30/#juju-gui.txt

=== urulama___ is now known as urulama
[14:09] <hatch> kadams54: hey get my email this morning?
[14:09] <kadams54> Yeah, looking things over right now
[14:09] <hatch> great thanks
[14:10] <hatch> the fix ended up being rather extensive lol
[14:11] <hatch> kadams54: I'm going to take this branch and work on the services side now
[14:11] <kadams54> OK
[14:25] <kadams54> hatch: when you have a few minutes, I'd like to talk over the changes.
[14:31] <hatch> yeah I'm good
[14:31] <hatch> standup?
[14:32] <hatch> ^ kadams54
[14:32] <kadams54> yup
[14:50] <rick_h_> jujugui call in 10 please kanban
[14:50] <hatch> jujugui hi mom
[15:37] <kadams54> hatch: all changes are now pushed to the PR
=== alexpilotti_ is now known as alexpilotti
[15:37] <kadams54> Will get the tests working and then (hopefully) land this puppy.
[15:38] <hatch> thanks
[15:58] <hatch> kadams54: did you want to rebase your branch to get it down to less commits? 
[15:59] <kadams54> That'll be the last step before landing.
[15:59] <kadams54> Once I get the tests passing.
[15:59] <hatch> ok I'll make sure that my stuff can be added on after
[15:59] <hatch> just lemme know if you make any code changes :)
[16:08] <kadams54> hatch: just changes to test code
[16:09] <kadams54> Which should be all fixed now. *fingers crossed*
[16:12] <hatch>  haha nice
[16:12] <hatch> kadams54: without any changes the 'highlight' and 'fade' work for the source service but any related stuff doesn't work
[16:13] <hatch> darn heh
[16:13] <hatch> was hoping for better results by default
[16:13] <kadams54> I'm actually surprised anything works, so it's glass half full for me.
[16:16] <kadams54> Just pushed test fixes, so hopefully it'll have a green light in about 20 minutes.
[16:18] <hatch> great
[16:18] <hatch> kadams54: there is also a css bug in the tooltip when deploying a bundle, it's shifted to the left
[16:19] <hatch> this likely happened when I did the header changes to make the header tooltip notificatiosn line up
[16:19] <hatch> maybe that'll be a simple fix you can do as a driveby with the other css stuff
[16:23] <kadams54> hatch: Yeah, I can take a look as I'm updating the icon.
[16:23] <kadams54> guihelp: Need another QA on https://github.com/juju/juju-gui/pull/631
[16:24] <hatch> thanks
[16:24] <rick_h_> hatch: you did the first pass can you make sure the issues are all better?
[16:24] <hatch> yeah I already have locally and we did a shared qa
[16:24] <hatch> but we both have been in here a while :)
[16:25] <kadams54> This is complex enough that I want a fresh set of eyes on it.
[16:25] <rick_h_> kadams54: hatch rgr
[16:26] <rick_h_> Makyo: you were a test away from  your branch up for review, can you qa while you get reviews?
[16:26] <Makyo> rick_h_, yep
[16:26] <rick_h_> ty
[16:27] <rick_h_> fabrice: rather than the cookie one can you look at the clicking on deploy card next?
[16:28] <rick_h_> fabrice: updated the board to reflect the priority/new card there if that's ok?
[16:31] <hatch> kadams54:  I created the cards for your next tasks 
[16:31] <kadams54> Thanks
[16:31] <hatch> I THINK this is everythinig
[16:31] <hatch> (famous last words)
[16:31] <hatch> lol
[16:43] <kadams54> luca__: any updates on the highlight icon SVG -> PNG?
[16:43] <luca__> kadams54: not yet, sorry. Give me a sec.
[16:44] <hatch> luca__: YO
[16:44] <hatch> I have a problem with your service spec
[16:44] <luca__> :(
[16:44] <hatch> wait no I don't
[16:44] <hatch> ignore me
[16:44] <hatch> I just figured it out
[16:44] <hatch> too many names for things lol
[16:45] <hatch> you keep doin good things :)
[16:48] <rick_h_> luca__: kadams54 is asking about the highlight icon? ^
[16:48] <luca__> otp
[17:14] <hatch> :/
[17:15] <hatch> having multiple hidden/shown in the canvas is seriously messing things up :)
[17:21] <hatch> kadams54: got a few minutes to be a sounding board?
[17:24] <kadams54> hatch: yeah
[17:24] <hatch> ok joining standup
[17:28] <hatch> Makyo: you able to hop into standup?
[17:29] <hatch> rick_h_: maybe you will know, able to hop into standup?
[17:30] <hatch> nm Makyo joined
[18:05] <hatch> Makyo: were you able to qa 631?
[18:09] <Makyo> hatch, no, not yet, I'm still struggling through mine.
[18:14] <hatch> ahh ok np
[18:41] <kadams54> jujugui: need reviews and QA on https://github.com/juju/juju-gui/pull/633
[18:41] <hatch> on it
[18:41] <kadams54> hatch, rick_h_, Makyo: that's the toggling for highlight buttons, so that only one can be active at a time.
[18:58] <hatch> kadams54: review and qa done - just need one more for both brancheds
[19:13] <hatch> jujugui - I know everyone is busy but could we get another QA on 631 and a review on 633?
[19:14] <Makyo> On it
[19:14] <Makyo> I need out of test hell.
[19:14] <hatch> haha u rock
[19:19] <Makyo> hatch, kadams54 just to confirm, 631 doesn't affect service view?
[19:19] <hatch> orrect it's very broken
[19:19] <hatch> :)
[19:19] <Makyo> Assuming that's fixed later?
[19:19] <hatch> yup in my current local branch
[19:20] <hatch> just need to get 633 and 631 landed so I can finalize my branches code
=== kadams54_ is now known as kadams54
[19:41] <hatch> kadams54: so looks like we can ship 631 now
[19:41] <kadams54> Yay
[19:41] <hatch> odly enough I no longer get user icons in github 
[19:41] <hatch> ahh had to clear the cache
[19:42] <kadams54> Shipped
[19:43] <hatch> awesome
[19:45] <kadams54> Gotta switch locations, be back shortly.
[19:56] <hatch> jujugui one more review required - pretty small, qa already done https://github.com/juju/juju-gui/pull/633
[19:58] <Makyo> hatch, kadams54 +1
[19:58] <hatch> awesommmmmme
[19:58] <hatch> thanks for clearing these Makyo
[19:59] <hatch> can you +1 on gh? 
[19:59] <kadams54> Woot woot.
[19:59] <hatch> awesome so....
[20:01] <hatch> kadams54: so I think the only thing else for this (besides my cards) is the tooltip thing I mentioned
[20:02] <kadams54> Yeah, looking into that.
[20:36] <hazmat> why does quickstart try to bootstrap a running environment?
[20:36] <hatch> hmm I've never seen that
[20:37] <hatch> but it has been a while
[20:38] <hazmat> looks like it always trys to bootstrap which is inane
[20:39] <hatch> hmm odd maybe I've just never noticed :)
[20:39]  * hazmat files a bug https://bugs.launchpad.net/juju-quickstart/+bug/1387867
[20:40] <mup> Bug #1387867: always tries to bootstrap .. check the api endpoint instead <juju-quickstart:New> <https://launchpad.net/bugs/1387867>
[20:42] <hatch> thanks
[20:49] <hatch> kadams54:  I found another bug
[20:49] <kadams54> Gah
[20:49] <hatch> maybe you can work on this one
[20:50] <kadams54> K
[20:50] <hatch> well, can you try and reproduce it first heh
[20:50] <hatch> it is...
[20:50] <hatch> load up the gui so it lands on the service view and charmbrowser on initial load
[20:50] <hatch> deploy
[20:50] <hatch> the mongodb bungle
[20:50] <hatch> click to highlight mongosvr
[20:50] <hatch> configsvr*
[20:50] <hatch> switch to mb
[20:50] <hatch> mv*
[20:50] <hatch> no machines are hidden
[20:51] <hatch> you will need to add this code to models.js:1424
[20:51] <hatch>         if (!machine.units) {
[20:51] <hatch>           // If the machine view hasn't been opened yet then there won't be a
[20:51] <hatch>           // units collection on the machines.
[20:51] <hatch>           return;
[20:51] <hatch>         }
[20:51] <kadams54> K
[20:51] <hatch> to avoid it from falling over 
[20:54] <kadams54> I'll make a card and get on it.
[20:55] <hatch> thanks 
[20:55] <hatch> I really hope this isn't yet another oddity
[21:03] <hatch> Makyo: hey - we don't have a 'hide relation line' method?
[21:03] <hatch> we have show and fade but no hide?
[21:04] <hatch> I can just copy the other methods but I thought we had something to do this already
[21:05] <hazmat> is there a config option yet to display icons for non official charms?
[21:05] <Makyo> hatch, no, nothing in place
[21:07] <hatch> allllllllrighty
[21:19] <hatch> Makyo: just to confirm - in a relation topology d.source is the 'near' and d.target is the 'far' ? There is a bug where it's showing relation lines where the 'far' is hidden so I'm just wondering if this terminology is consistant 
[21:19] <hatch> it's around line 770 in topology/relation.js
[21:19] <hatch> it appears to be so
[21:20] <Makyo> It shooould be?
[21:20] <hatch> heh ok I'll asume so as well :)
[21:23] <hatch> d3 is pretty awesome even if it's api is horrible ;)
[21:33] <hatch> ahhhhh just rebased in develop and it broke it!
[21:33] <hatch> blargggg
[21:35] <hatch> kadams54: so the branch which did the toggle thing actually has these events propogate everywhere
[21:35] <hatch> we need to use method calls instead of events
[21:35] <kadams54> Why?
[21:37] <hatch> kadams54: actually I was able to fix it reordering the events
[21:37] <hatch> it was highlighting then unhighlighting
[21:37] <hatch> hah
[21:37] <hatch> by switching them around it works
[21:37] <kadams54> I can make it less spammy
[21:37] <kadams54> Great
[21:38] <hatch> phew
[21:44] <hatch> Makyo: FYI target is not always 'far' 
[21:44] <hatch> :'(
[21:45] <Makyo> :/
[21:55] <hatch> sweet I got all the functionalities down
[21:55] <hatch> kadams54: any luck tracking down that bug?
[21:56] <kadams54> Not yet. My work time's been sporadic with dinner time, but the family is off to dance class now so I now have some quality bug squashing time.
[21:56] <hatch> ohh right it'a past your EOD
[21:56] <hatch> well you could just leave it for me
[21:57] <hatch> Now that I have this functionality done I need to write tests
[21:58] <hatch> rick_h_: are you around tonight?
=== kadams54_ is now known as kadams54
[22:00] <kadams54> hatch: It's no problem.
[22:02] <hatch> jujugui if anyone wants to do some preemptive qa work on this added services stuff it would be appreciated https://github.com/hatched/juju-gui/tree/canvas-fade
[22:03] <Makyo> On it
[22:03] <hatch> Makyo: thanks - the only currently known bug is the one I mentioned to kadams54 at 14:50 in the scrollback
[22:04] <hatch> and now to write tests....OOOOBOY!!!
[22:05] <hatch> Makyo: also I seem to have re-introduced the error when trying to hide/fade ghost services :/
[22:05] <hatch> will look at that now
[22:10] <hatch> ok that's being done as a follow-up
[22:10] <hatch> moving the card back 
[22:16] <kadams54> hatch: OK, I think I figured out what the bug is.
[22:16] <hatch> ok shoot
[22:16] <kadams54> It's actually the line you had me add at 1424 in models.js :-)
[22:16] <kadams54> if (!machine.units)
[22:16] <kadams54> machine.units isn't set in the DB
[22:16] <kadams54> We set it by making a call to db.units as part of render()
[22:18] <hatch> sooooooooo howdowefixit?
[22:18] <kadams54> _updateMachineWithUnitData, machine-view-panel.js, line 1409
[22:19] <kadams54> I think setMVVisibility needs to be moved out to the DB or a util method
[22:19] <hatch> so on render it should populate that unit data with the visibility statuses
[22:19] <kadams54> It will need to make a similar query to db.units first and then do the unit check
[22:19] <hatch> well setmv is in the db
[22:19] <kadams54> It's in MachineList
[22:19] <kadams54> I think?
[22:20] <kadams54> Yeah.
[22:20] <hatch> oh yeah
[22:20] <hatch> but that's ok
[22:20] <hatch> if we have the db we can access the machone list
[22:20] <hatch> but you're suggesting updating the machines with the unit data before we open the mv?
[22:20] <kadams54> No
[22:21] <kadams54> I'm saying setMVVisibility just needs to do this:
[22:21] <kadams54> var units = db.units.filterByMachine(machine.id, true);
[22:21] <kadams54> if (!units) {
[22:21] <kadams54> ...
[22:21] <kadams54> Instead of just:
[22:21] <kadams54> if (!machine.units) {
[22:22] <kadams54> That make sense?
[22:22] <hatch> nope
[22:22] <hatch> but I can try it
[22:22] <hatch> lol
[22:22] <hatch> sec
[22:23] <kadams54> We can't count on the internal list (machine.units) being initialized when setMVVisibility is called
[22:23] <kadams54> We need to go back to the DB (db.units.filterByMachine()) instead
[22:23] <hatch> oh wait I can't, now I see why you want it on the db
[22:23] <hatch> heh
[22:23] <hatch> ohhh
[22:24] <hatch> so you're saying we need to update the 'real' units
[22:24] <kadams54> Yes
[22:24] <kadams54> Those "real" units will get populated into machine.units at a later point
[22:25] <hatch> so does it make sense to do that OR do we run the update on mv load?
[22:25] <hatch> I think your idea is better
[22:25] <kadams54> Also: lines 1427-1432 in models.js is some crazy shit
[22:25] <hatch> yeah don't go there
[22:25] <kadams54> nested some() calls
[22:25] <hatch> it's dangerous
[22:25] <hatch> therebedragons
[22:26] <kadams54> #
[22:26] <kadams54> #therebedragons
[22:26] <hatch> truth
[22:26] <kadams54> Can you throw a comment on there?
[22:26] <hatch> no I don't remember what it does
[22:26] <kadams54> lol
[22:26] <hatch> lol!!
[22:26] <hatch> jk yeah I can
[22:27] <kadams54> kk, I gotta run again, but will check in again in about an hour
[22:27] <hatch> thanks! 
[22:27] <hatch> cya
=== kadams54 is now known as kadams54-away
=== Guest89108 is now known as bodie_
[23:48] <rick_h_> hatch: around what's up?