/srv/irclogs.ubuntu.com/2013/02/22/#juju-gui.txt

=== Makyo is now known as Makyo|out
bcsallerlbox propose took 7 minutes :(05:39
jovan2Morning! Considering icons for our service providers - AWS etc - does anybody know if there is an icon for LXC?10:26
gary_posterjovan2, hi.  no, I do not know of one, and it looks pretty clear to me that LXC does not have an official icon.  However, hallyn or stgraber in #ubuntu-server will probably have authoritative answers, and I have asked them.  I'll report back.13:17
gary_posterI was going to sat that to jovan2 but he disappeared as I was writing that :-P13:18
gary_posterI'll email him13:18
hatchmorning14:01
hatchbrb gota reboot14:41
=== Makyo|out is now known as Makyo
hatchI've got my new webcam, are you guys pumped to see me in HD?15:06
hatchI know I am!15:06
hatchbcsaller: are you in yet?15:09
bcsallerhatch: I am15:09
hatchhttp://bazaar.launchpad.net/~hatch/juju-gui/1130790-create-subapp/view/head:/app/app.js#L467 and 453 have an assignment operator in the if() statements15:10
hatchis this intentional or a typo?15:10
bcsallerintentional15:11
hatchalright :)15:11
bcsallerbasically, if I was able to take another item from the sequence 15:12
hatchgotcha - I think I have figured out why this is erroring out but I'll wait until later so we can work out a sollution - it looks like it's evaluating all routes regardless of the 'namespace' property15:14
bcsallerhatch: have you merge the change with match()?15:15
bcsallerhatch: looks like you have an older version15:15
hatchhmm ok let me do another merge15:16
hatchso when I do these merges `bzr merge ../../trunk` it does the merge but then shows changed files which I have to re-commit, am I doing something wrong or is that just how bzr works?15:19
bcsallerhatch: yes, you have to commit the merge, if there are conflicts for example you'd resolve them now15:20
hatchgotcha15:21
hatchhttps://code.launchpad.net/~hatch/juju-gui/1130790-create-subapp here is the most recent codebase15:21
hatchonce it loads you'll see that it trys to call the callbacks that aren't on the proper namespace15:22
benjiman, after getting used to actually being able to quickly see where entities come from (Python) wandering around blind is irritating (Go)15:22
hatchwe can discuss after the meetings, just putting it there :)15:22
hatchhow do you like Go?15:22
hatchya know - assuming you knew the language :P15:23
benjiI actually like a lot of the ideas in Go; unfortunately it seems to have missed some important things we have learned over the last decade or so.15:26
benjione of those things being "I should be able to quickly figure out where a name comes from just by looking at the current source file."15:26
hatchaww that's unfortunate15:27
hatchI've heard Ruby also misses some of the life lessons15:27
hatchI don't have any first hand experience though15:27
bcsallerhatch: I see a null callback, its using callback = self[callback] rather than the routes subapp || self15:27
benjiyeah, me neither; I'd like some, if only to be able to complain about it afterward ;)15:27
hatchhaha15:27
hatchbcsaller: ahh I see that now15:28
gary_posterjujugui call in 215:28
hatchin copying my music over to my NAS it decided to duplicate about 25% of them15:29
* hatch is less than impressed15:29
gary_posterhatch goodspud call now15:31
hatchbcsaller: did you want to work on this router bug now?16:33
hatchI have 2H16:34
bcsallerhatch: did you change the code to resolve from the proper context object?16:34
hatchnope - I thought that was psudo code :) changing16:35
hatchbut but even if I do that, the callback shoudln't be being called anyways16:35
bcsallerhatch: around like 459, that should fix it16:35
bcsallerhatch: you're saying its matching improperly? I haven't seen that yet16:36
hatchthrow a console.log(callback) on ln479 and you'll see that it's calling the namespace routes even on root localhost requests16:37
hatchif you can't see that I can screenshare to show you what I mean16:38
bcsallerhatch: ok, seeing if g+ is open16:40
bcsallerhatch: it is16:40
hatchbcsaller: sorry needed to install plugin16:44
hatchalmost up16:44
bcsallerhatch: I think I found the source of the problem, its the way we are mutating the routes ATTR, we mix in options like namespace but this only happens when routes are appended, not inserted at offset so the property (namespace in this example is lost). I can generate a patch17:23
hatchahh - figures there would be some weird stuff going on17:24
hatchthanks!17:24
gary_posterbenji bac frankban teknico please join https://launchpad.net/~canonical-cloud-engineering17:27
benjigary_poster: page not found17:27
gary_posterbenji, works for me, but never mind:17:28
gary_posterbenji bac frankban teknico nm I am doing it for you17:28
gary_posterthough you might want to join mailing list17:28
benjigary_poster: it is private, doesn't that prevent anyone from seeing it until they are a member?17:28
gary_posterbenji, ah maybe so17:28
teknicogary_poster, it looks like you did the mailing list subscription too :-)17:30
gary_posterteknico, heh, maybe.  unintentional if so :-)17:30
benjiteknico: there is a setting somewhere that lets you choose to automatically subscribe to team mailing lists when you become a member of the team17:31
frankbangary_poster: done17:31
gary_posterthank you17:31
teknicobenji, apparently I have that set already :-)17:31
gary_posterbcsaller, landscape-aggro (which I keep on thinking of as landscape-aggression) is clear for take off :-)17:36
bcsallergary_poster: thanks for the review, I'll submit it after I get a patch for hatch ready17:37
gary_postercool17:37
bcsallerhatch: http://paste.ubuntu.com/5555767/17:43
hatchinnnnnteresting17:44
hatchthanks a bunch!17:44
hatchare you going to be putting this in your branch?17:44
bcsallermy branch is already landed, you should take it 17:46
bcsallerhatch: ^17:46
hatchoh I didn't get the email17:46
hatchthanks I'll merge it in17:46
bcsallerhatch: I forgot that the insertion would invalidate an assumption 17:46
hatchI don't see the branch in LP, did you push?17:47
bcsallerhatch: you merged the match method from trunk, no? Thats all I was talking about. I don't have a new outstanding branch for this so take the patch and apply it17:48
hatchoh yeah sorry that's what I was asking :)17:49
hatchok taking patch......now17:49
hatchand as a gift for fixing this17:49
hatchhere is a tai fighter17:49
hatch|-o-|17:49
hatchpew pew pew17:49
bcsallerqueue the robot chicken17:50
gary_posterheh17:50
hatchJust incase you guys ever need to insert an array of values into another array I had to figure this out the hard way yesterday so I figured I should document it somewhere http://fromanegg.com/post/43733624689/insert-an-array-of-values-into-an-array-in-javascript18:05
gary_posterI figured it was an apply trick :-)18:06
hatchhaha yep, using call/apply is pretty powerfull18:07
benjidoes anyone else who is running quantal been able to deploy anything using go-juju?18:15
hatchmy laptop has 12.10 on it, I can try if you like18:15
hatchalthough I'd need some sort of guide :)18:16
gary_posterI thought frankban had, benji18:16
gary_posterhe is past EoD though18:16
benjihatch: I suspect the setup entailed would be too involved to ask that of you18:16
benjiI was trying to use the OS the core developers were using and assumed (bad me) that it was quantal.18:17
gary_posterbenji, they are on precise?18:17
gary_posterI know jam is18:18
* hatch finds it easier to remember version numbers than names :) 12.04, 12.1018:18
benjigary_poster: I don't know, but when I look at the charm store, the number of quantal charms is paultry.18:18
gary_posterbenji, oh that's different18:18
benjiyeah, I hate the code names; unfortunately they infect everything18:19
gary_posterbenji, that's what is run on the vm side18:19
gary_posterwhich you can change to whatever you want, whatever version your own machine is, at least with pyjuju18:20
benjiis it?  Given that the dev platform is so intamately tied with the deployment platform, I would expect the OS version to be critical.18:20
gary_posterno, what I mean is that the charm runs on the remote machine benji.  Usually we use LTSs for those18:20
benjigary_poster: right, but this is go-juju18:20
hatchbcsaller: awesome - with your fix in route and then chaging the callback in _dispatch to18:21
hatch              callback = self[callback] || self.get('subApps')[namespace][callback];18:21
hatchit all looks like it's working18:21
hatch:D18:21
hatchthat's one ugly line though eh? :)18:21
gary_posterbenji understood.  (1) go juju should do this too--that's one of the goals, and (2) I think frankban did it18:21
benjiI'll try changing the preferred series.18:21
gary_posterdefault series, I think18:21
benjiyep18:22
hatchI'll be cleaning that line up because it doesn't need to fetch the sub apps in a loop18:22
bcsallerhatch: the order is backwards or callback resolution I think, if the sub has it use it there first18:22
hatchbcsaller: ok I'll do some testing to determine the proper order18:22
hatchI'll probably have to flesh out the sub apps a little more to get some more use cases18:23
frankbanbenji, gary_poster: that's the problem with the gui charm too, i.e. bug 113160818:23
_mup_Bug #1131608: deployed series is arbitrary <juju-core:New for fwereade> < https://launchpad.net/bugs/1131608 >18:23
gary_posterfrankban, I thought that only was true for the upload tools option?18:23
gary_posterI must be confusing the two issues18:24
gary_posterSorry18:24
frankbangary_poster: not sure, it's worth to try18:24
benjibtw, I need --upload-tools *and* to deploy a charm18:24
gary_posterah, benji, I think that's the broken part.  Could be wrong.18:24
frankbanbenji: you can try recompiling and bootstrapping juju-core from a precise lxc. it's not ideal, but, if it works, it can be a (uncomfortable) workaround.18:28
benjifrankban: that's a good idea; I may have to do that18:28
frankbancool, eod, have a nice weekend18:29
benjihave a good weekend, frankban 18:29
benjigary_poster: cool, Danilo is comming back into the fold18:30
gary_posterbenji, yeah, I heard!  I don't know where though--do you know that?18:30
benjiBlue18:31
benjiSquad18:31
benji(they added a newline in the middle of their name... hipsters)18:31
gary_posterheh18:31
Makyo99% there with win8, missing only network.  Anyone have to change settings for that away from NAT?18:52
Makyoguihelp ^^^18:56
bcsallerMakyo: I still don't have it working, I was planning on trying again today18:57
hatchMakyo: are you using parallels?19:03
gary_posterMakyo, I had to use bridged19:03
Makyohatch, virtualbox.19:03
Makyogary_poster, ah, trying that now.19:03
hatchahh I actually had to do the same with parallels19:03
hatch:)19:03
MakyoThat did it, thanks gary_poster, hatch 19:05
bcsallerkinda of have ie working in the VM now :-/20:56
bcsallerkind of.. blah20:56
hatchwhat issue are you having?21:10
hatchugh itunes is such a junk program21:13
hatchbcsaller: I've been trying to track down a little bug maybe you have some insight...._dispatch() is being called 4 times on load22:00
hatchthis is almost certainly a bug no?22:00
benjihatch: I think it is past his end-of-day, but yes, that sounds like a bug 22:03
hatchahh - we need a map where everyone is located so we can see when EOD is ;)22:04
hatchok found it....looks like it was being dispatched 4 times lol22:10
hatchbut it didn't matter until now!22:10
hatchohhhh boy - how to solve this22:10
hatchgary_poster: still around?22:13
hatchlooks like everyone might have checked out for the weekend22:14
bcsallerhatch: its a known issue, the app calls dispatch aggressively and it gets called on delta updates23:09
bcsallerhatch: some pages it works out to more and some less, plus if there is a delta depending on the page it can appear as quite a lot23:10
bcsallerdispatch alone being called doesn't have to be a big deal though, we mostly treat it as an event saying to any active views, "do you have to update yourself now?"23:10
hatchohh ok so we'll need to change that23:11
bcsallerbut some views currently do a full rerender to that question23:11
hatchcalling dispatch causes the sub apps to rerender23:11
hatchso they render 4x which is obviously no good23:11
hatchthat might be a large refactor though23:11
bcsallerit is23:11
bcsallernotifications and env view both do it better23:12
hatchyeah see dispatch is designed to initiate the routing callbacks based on url changes23:12
hatchif the url doesn't change, it shouldn't dispatch23:12
bcsallernotifications is model bound and env view has an update() phase and then is event rather than render driven23:12
bcsallerbut callback doesn't have to equal full render23:13
hatchno but in my opinion the views should be pretty dumb...the singletons should be outside of the views23:14
bcsallerto be clear I'm not trying to defend the model, but its part of a much larger issue which to my mind steams from a lack of model bound views 23:14
bcsallernon-persistent views can be dumb and we let them be23:14
bcsallerpersistent views need to be smarter, "most" of the app is the persistent env view23:15
hatchyeah I'm going to need to understand more of how the app operates to see how we can fix this23:15
bcsallerthat component framework helps handle incremental update (which is what we have been calling it)23:16
bcsallerby drawing lines between initial render and event driven update cycles 23:16
hatchyeah maybe this weekend I'll find some time to really step through the execution order of this app23:17
hatchdispatching 4x on load has to kill the startup time heh23:17
bcsallerhatch: it really doesn't, thats not the issue, its often 1/2 that (still an issue) followed by the 1st delta window23:18
bcsallerthen look at app.on_database_changed23:18
hatchahh ok23:18
bcsallerthats deals with an async change to the data from the websocket w/o the app triggering it23:18
bcsallerto keep the views dumb they can just dispatch on that23:19
hatchok well as long as it's a known thing because I was sure it was a bug :D23:20
bcsallerits a compromise 23:20
bcsallerok, signing off23:20
hatchalright at least now I understand its purpose :)23:21
hatchhave a great weekend!23:21

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!