[00:03] huwshimi doing alright, the mrs is trying to plan a winter trip to somewhere warm.....I just don't get it [00:03] winter is for snow [00:04] snow-boarding, snow-mobiling, snow-shoveling [00:04] I don't see snow-beaching anywhere [00:04] :) [00:06] my android watch is still sitting in Minneapolis with 'Exception' in red on the ups site.... [00:07] I wonder if these watches weren't cleared with some standards body or something [00:07] it's been there for 2 days [00:13] hatch: doh [00:18] yep [00:18] very doh indeed [00:18] hatch: what was the topic from today to talk to UX about? [00:19] hatch: nvm, the walk through from scale up UX to MV/etc [00:19] the transition from manually placed and auto placed, what should the user see when they switch to the machine view, and should unplaced units auto place on commit [00:19] yeah [00:19] :) [00:21] Is it just me or is it strange to ever see the inspector and the machine view at the same time? [00:21] The inspector has lost its context (the service) [00:21] huwshimi: yea, it'll be replaced by the deployed services view. [00:22] huwshimi: which we've decided we can release MV 1.0 without, but want to come back to as MV 1.5 later this cycle [00:22] rick_h__: Ah cool [00:22] huwshimi the issue is what to do when you click 'manually place' then scale up....you have to switch to the machine view....so that should probably happen right away [00:24] yea, lots of love needed there [00:24] I think I've got this unit thing figured out - tomorrow I'll start the real implementation [00:25] cool [00:25] this env stuff is like the shibboleth of the GUI [00:26] you're not REALLY a GUI dev until you can figure out how the env interactions work lol [00:26] heh [00:27] we should probably document the flow in a flowchart of some sort [00:27] it's definitely complicated [00:27] which part is this? [00:28] the flow from delta to db and back again, and the UI interactions to delta/db and back again [00:28] ah yea [01:29] huwshimi: you rock. Going through bugs and see it's already marked fix committed :) [01:30] rick_h__: I'd fixed it before it was reported :) [01:34] huwshimi: you're even better than I realized! === uru-afk is now known as urulama [10:02] checking Hacking.md file and there is a reference on /var/lib/juju/units ... however, units dir is no longer there, there are containers and agents, right? [10:35] jujugui how does lbox work with github? or how is proposing/merging process done there? [10:45] (probably through web interface, but, well, better to ask) === urulama is now known as uru_ === uru_ is now known as uru-urnotallama === uru-urnotallama is now known as urulama [11:22] urulama: i haven't used lbox with github [11:22] urulama: mornin' BTW [11:23] rogpeppe: g'day to you, sir [11:25] rogpeppe: what's the process of api proposal now that Francesco is out for a week? [11:26] urulama: since frankban is away, do you perhaps fancy continuing the store work along with me? [11:26] urulama: i guess we'll address the comments we can [11:27] rogpeppe: sure, we can try, but you'll spent more time on explaining the details, 'coz i'm not familiar with them. is that ok? [11:28] urulama: i think that will probably useful for both of us actually [11:29] rogpeppe: agree. let me finish something in the ftests, then i'll grab a lunch and come back [11:29] urulama: cool [11:29] urulama: i've got a few things to do first anyway [11:30] rogpeppe: my father some mushrooms in the woods yesterday, it would be a waste not to eat them :D [11:31] urulama: nice! [11:31] urulama: what kind of mushrooms? [11:31] rogpeppe: ah, translation ... just a sec [11:31] urulama: (i'm keen on hunting mushrooms too) [11:31] rogpeppe: these ... http://www.gobe.si/slike/Boletus_appendiculatus.jpg [11:32] urulama: not Boletus Edulis? [11:32] urulama: i like boletes [11:33] rogpeppe: you're right. it's edulis [11:33] urulama: perfect [11:33] urulama: we usually dry the good ones and eat the rest [11:34] urulama: enjoy! [11:34] rogpeppe: have you tried just freezing them? the flavour is kept. those are great for risotto [11:35] urulama: we do that too - usually after sauteeing for a little bit [11:35] urulama: which reminds me, we've got a few bags still in the freezer that need eating [11:35] :D [11:36] rogpeppe: we can start international juju-gui mushroom day :D [11:36] what an excellent idea [11:37] it's not yet the season for boletes here, but we found a load of st. georges mushrooms earlier (Calocybe gambosa) which were delicious [11:38] nice [11:41] urulama: what do you mean the process of the api proposal? [11:42] rick_h__: morning [11:43] urulama: morning [11:43] morning [11:43] rick_h__: they were sometimes on hangouts, working together, and franban is out for some time now [11:43] morning, jrwren [11:44] jrwren: i've made a small change to ftest so that you can run on local [11:44] rogpeppe: let me know when you've got time to chat, urulama can chat as well. I've finished going through the docs and such and would like to hash out some questions and such [11:44] jrwren: will push proposal later when i test it with other env that everything still works [11:44] rick_h__: will do. 15 minutes? [11:44] urulama: the goal would be to get the work discussed broken into small single branches of work and added to the kanban board, then started to work on. [11:44] rogpeppe: sounds good [11:45] urulama: oh cool! [11:45] urulama: lbox is specific to launchpad (LP) and reitveld. [11:45] rick_h__: yes, went to see the code and found out :D [11:45] urulama: the normal process for github related things is based on the GUI process https://github.com/juju/juju-gui/blob/develop/HACKING.rst#typical-github-workflow with small tweaks [11:46] http://ci.jujugui.org:8080/ has the charmstore CI for instance and does the tests/landing [11:46] we're working with core on it for other parts of the code [11:46] rick_h__: tnx ... so, if i push proposal on LP, does this reflect the github repo? [11:46] urulama: what project are you pushing to? [11:47] urulama: LP and github are seperate. we use one or the other [11:47] rick_h__: will push for juju-gui, ftests support for local (really trivial change, but works for local now as well if one requires) [11:47] urulama: the charm? [11:48] urulama: ok, so that's only in LP and yes, uses lbox. jrwren submitted some docs to it yesterday. The hacking doc should help with that process [11:49] rick_h__: np with the lbox, just curious about the LP vs Github [11:49] urulama: we used to have all code on LP and we're slowing moving new things (and the big projects like the GUI) to github [11:49] urulama: so we live in a dual tool world. [11:50] urulama: but anything new should be in LP [11:55] jrwren: https://code.launchpad.net/~uros-jovanovic/charms/trusty/juju-gui/hackingreview/+merge/226270 [11:56] that's the simple change [11:56] and it doesn't take much time to run on local [11:57] i tried something similar - removing the forced_machine paramter to the juju_deploy call - and it failed for me [11:57] as long as this works, I'm happy :) [11:58] jrwren: yes, the deploy.go checks the value and add --to if it is not None ... [11:58] urulama: but what does this do to the test time on non-local? [11:58] urulama: the majority of our users are using the gui on various clouds. That's the more important functional test run to some opinions. [11:58] rick_h__: it is pushed to 0 if it is anything than local [11:59] rick_h__: it's the same as before [11:59] ah gotcha [11:59] * urulama needs to make lunch now ... starving :D [12:00] morning bac, any better interwebs today? [12:03] rick_h__: btw, whoever came up with debug-hooks is a genius :D [12:05] urulama: +1. I love it. [12:05] hey rbasak how goes? [12:08] jrwren: so can you QA the changes from urulama on both local/ec2 and review? [12:08] urulama: can you add a card to the kanban board in review/maint for the work please? [12:10] rick_h__: sure. [12:10] rick_h__: good thanks. Hoping to get juju 1.20 in Utopic soon :) === urulama is now known as uru-afk [12:10] rbasak: cool, any word on the quickstart update into trusty? [12:11] rbasak: oh, nvm see it's fixed release in the bug now [12:11] rick_h__: it's uploaded and waiting on the SRU team now. I should chase them really - it's been a week now. [12:11] Fix Released for Utopic - waiting on Trusty SRU approval. [12:11] rbasak: ah gotcha [12:14] rick_h__: sorry, call went on a little longer than the 10 minutes i thought it was going to [12:14] rogpeppe: all good [12:14] rick_h__: we could do it now, or we could wait for uros to get back from lunch [12:15] well let's chat and if he gets back in time to join we can bring him in. [12:15] * rogpeppe has grilled halloumi salad, mmm. [12:15] rogpeppe: hangout room? [12:15] rick_h__: sure === uru-afk is now known as urulama [12:16] rick_h__: want me to join? [12:16] pulling bzr branch lp:~uros-jovanovic/charms/trusty/juju-gui/hackingreview to test it [12:17] urulama: sure thing [13:09] morning all. [13:10] good morning jcsackett [13:41] hi rick_h__ [13:54] bac: howdy [13:54] * urulama runs for that lunch now [13:54] rick_h__: hi. unping. [13:54] unping? [13:54] rick_h__: nm [13:54] hah [13:55] meaning i resolved my issue and don't need to bug you any longer [13:55] woot [14:00] jrwren: so, i'll change the ftest some more, coz it's not the name that counts, it should be the type of the env and I'll add another parameter like JUJU_ENV_TYPE and set a check if it's local. [14:05] urulama: oh, ok. I didn't mind the magic name. [14:05] I did get an error, although I am not convinced it is related. [14:05] http://pastebin.ubuntu.com/7775331/ [14:09] while you are making changes, could you add charm-tools and firefox to SYSDEPS? :) [14:14] apparently the guys who cut my power/phone lines are coming in a bit to schedule a fix so I might pop away for a few minutes if and when that happens :) [14:15] http://tholman.com/giflinks/ [14:16] lol [14:16] definitely run that past the UX team [14:21] omg kadams54 that is great. [14:23] hatch: we should open the machine view ux sync meeting with that. [14:23] lol [14:32] btw, found two more core bugs in the morning, bug 1340077 and bug 1340133 [14:32] <_mup_> Bug #1340077: if "default" is named to undefined env, the parsing fails

[14:32] <_mup_> Bug #1340133: debug-hooks don't work in manual env

[14:33] urulama: I think debug hooks not working is a known issue/missing feature. [14:38] urulama: indeed, that failed test was some other issue. It worked second time I ran. [14:38] and runs in 56% the time as on ec2. yay local. [14:38] jrwren: ;) [14:42] rick_h__ #1340147 bug is really the same as card 1326193 in project 1 [14:42] <_mup_> Bug #1340147: Ghost services should be available to scale out in the mass scaling UI [14:42] at least I think they are one and the same [14:42] hatch: yea, I was going to chat with you about it [14:42] hatch: because it seemed like that to me :) [14:42] ohhh wait that's the MV mass scaling UI? [14:43] hatch: but wanted to get them down on record [14:43] then that's what Im doing right now? [14:43] heh [14:43] hatch: right [14:43] hatch: it's your WIP [14:43] so I wanted to see if this fits into your current work or not [14:43] but been on calls all morning, sorry [14:44] no problemo, I think it'll be solved but I'll keep an eye on it [14:44] hatch: ty much [14:45] jrwren are you a C# guy? twitter seems to think you're similar to a lot of C# people :) [14:47] I am. [14:47] I wrote C# for 7 or so years. [14:47] cool, I've dabbled in it for Unity (3D) could never get used to the different uppcercased methods for working with arrays [14:47] ha! [14:48] if that is the only gripe, you'd adapt very quickly. [14:48] psh, you mean it wasn't javascript [14:48] UnityScript is pretty close to javascript. [14:48] ah [14:48] yeah other than that it was pretty nice - learning it was somewhat difficult because it was always hard to find stuff that -wasnt- about some .net framework or something [14:48] Unity still runs on Mono 2.6.x IIRC, so some features are lagging [14:49] yeah I didn't like unity script [14:49] felt that c# was a better lang for that stuff [14:49] I would of course prefer golang now :) [14:49] hatch: why not? too different from javascript or too similar? [14:49] jrwren the tooling wasn't there [14:49] ah. that makes sense [14:49] c# provided a better dev/debug experience [14:50] taking a 5 min break before standup [14:51] jrwren if I remember correctly there is a c# framework that abstracts away those utility method oddities? [14:51] jujugui call in 10 [14:51] hatch: which utility method oddities? [14:52] like requiring different utilities to loop through arrays vs multi-dimensional arrays [14:53] oh. there might be. But that should be straightforward. [14:53] everything I learned about such actions I mastered thanks to the perllol man page :) [14:54] yeah I just remember it was very odd to have to choose different array utility methods based on it's dimensions [14:55] oh! yes, is... odd. [14:55] I guess i contrast to C where there are no such methods and you are on your own and I was happy to have those methods. [14:58] jujugui call in 2 kanban please [14:58] jujugui I might be a few min late, wrapping up cross team call now [14:58] yeah good point [14:58] rick_h__ np, we'll sit and wait for you like good children [14:58] * hatch grabs a knife and sticks it in the outlet [14:58] lol [15:09] jcsackett yes that was some epic lag [15:09] :) [15:09] jrwren: no need to do the review, i'll provide another pathc [15:09] patch [15:09] hatch: stand up over? [15:10] hatch: i'm at our local hackerspace, and someone is using *all* of the bandwidth. [15:10] rick_h__ yup we r efficient [15:10] woot! [15:10] jcsackett: crack the router and lock them out :) [15:10] jrwren: got a sec? [15:11] jrwren: :p [15:11] jrwren: I did get a list of api endpoints from rogpeppe today. http://paste.ubuntu.com/7774899/ so on your card, I think getting familiar with stuff is valuable to continue [15:11] jrwren: but just fyi I did get some data and maybe you can compare against it for a wrap up of that card of work [15:14] rick_h__: i got secs. [15:14] rick_h__: thanks. [15:14] man my watch is STILL sitting in what I can only assume is customs in Minneapolis [15:16] hatch: canadians get no love [15:16] lucky they shipped it out to you at all [15:16] lol truth [15:16] UPS is always a disaster when it comes to crossing the border [15:17] $50 'customs' fees and all that bs [15:17] rick_h__: tests/20-functional.test:from selenium.webdriver import Firefox [15:17] the tests are firefox tests. [15:18] jrwren: ah gotcha cool [15:18] jrwren: ignore me then [15:18] Its a good question. [15:19] on my machine I use firefox nightly and link firefox to firefox-trunk [15:19] so the idea of apt-get install'ing firefox was :( to me [15:20] hrm... good point. [15:21] maybe a different way to meet that dependency? [15:21] naw it's all good [15:21] I'm a freaky case [15:21] if I want to I can skip sysdeps and mange them manually [15:21] i think we all are freaky cases [15:21] :) [15:30] rick_h__: so, i'm getting familiar. Is there anything you would like to see from that pastebin that is not already in https://github.com/juju/charmstore/blob/master/README.md ? Those 4 endpoints are documented there with examples [15:30] jrwren: lol, no. I guess just confirmation that's up to date [15:30] maybe details like "the request can have unspecified serious" [15:31] jrwren: naw, that's enough for my needs [15:31] rick_h__: I’m getting an error when I join the hangout [15:31] rick_h__: I’m going to restart [15:32] luca: rgr [15:32] rick_h__: using COLO or even JUJU_ENV_TYPE is not that great ... the juju-test does not pass another parameter to the test and setting global attribute is not a good idea, so, i propose that local test requires an environment that is named "local" ... [15:33] urulama: :( ok, then I'd suggest wording the docs around it to be about the local specific and less around colocation [15:33] what struck me was that the sentences leading up to the command all mentioned colocation, but the command didn't have anything about colocation in it [15:34] rick_h__: i'm not happy with the solution either, the solution is to change juju-test command to pass parameters to the .test file [15:35] urulama: that's fine. I'd rather we move forward with the imperfect solution but make the docs match up/explain [15:35] rick_h__: but for now, let's just to the "local" thing. it does speed up the tests [15:36] jrwren, rick_h__: i've also removed the firefox dependency [15:36] urulama: I think it's ok [15:36] could have the sysdeps make target depend on /usr/bin/firefox :) [15:37] and a /usr/bin/firefox target that does apt-get install firefox :) [15:39] jrwren: but then there are /usr/local or /opt or ... [15:39] let's keep it out for now [15:40] ok [15:40] we've talked about it so much, i'll never forget it now === mhall119_ is now known as mhall119 [16:05] jcsackett any luck on increasing te # of tries? [16:27] * Makyo goes to lay down for a bit [16:30] * rick_h__ runs away for a while, finally done with phone calls for a bit === urulama is now known as uru-qfk === uru-qfk is now known as uru-afk [16:45] hatch: no, that didn't work, but i've figured out the issue, i think. almost done with an experiment to verify. [16:45] hatch: bad news is we may need a better solution to this whole quandry. [16:45] cool i'll be interested to know what the issue was [16:45] :( [17:26] my ec2 env is taking a very long time to finish loading new source for my experiment. [17:38] it does [17:39] it hates you [17:44] rick_h__ when you have -only- a ghost service the mass-scale-up UI in the MV is still there, it should be hidden because we don't allow mass-scale-up on ghost services (or we should allow it) [17:45] only ghost service(s) [17:45] that is [17:45] hatch: so the bug we talked about htis morning is allowing mass scale up of ghost services [17:45] hatch: so the idea is that a ghost counts as a scalable service [17:46] ahh ok right [17:46] looking into that [17:47] oh we just simply skip them [17:48] heh [17:48] dddddd [17:48] fixed [17:48] :) [17:52] hah, yay then [17:52] hatch: make sure to update the bug/card to go along with this branch then please [18:07] kadams54: call? [18:09] * hatch grabbing a bite [18:30] rick_h__: Bah, sorry, I'm back. Went out for lunch and forgot my plan was just to stay at Panera and use their wifi for the call. [18:33] jcsackett: call? [18:34] kadams54: k, let me know if you want to reschedule/chat or not then. [18:34] Things are good from my end. [18:34] * rogpeppe is stepping out for 30 mins to grab some supper [18:34] * rick_h__ is feeling unloved, back to back no shows [18:43] I am a jerk. === uru-afk is now known as urulama [19:02] rogpeppe: back for the resources call? [19:07] rick_h__: am now [19:49] hmm....hmmmmmmmm [19:52] so we have a slight problem with how the ecs was set up to handle units....each unit is its own record [19:52] so I propose that we just 'deal' with it sending multiple 'add_unit' calls and then go back and refactor this later [19:52] ^ rick_h__ [19:53] it's a pretty integral part to how the add unit ecs stuff was written, place unit, etc etc [20:20] hatch: sec otp [20:21] sure no rush [20:23] hatch: ok, so who da what? [20:24] i'm outtie. rick_h__ suggestion for next desired. [20:24] jrwren: have a good day, let's catch up in the morning then [20:24] k [20:24] call? [20:24] hatch: sure [20:25] but my ears are falling off :P [20:29] ok, running away until the AU call later [20:36] * rogpeppe is done for the day [20:36] g'night all [21:15] Had the branch working until I tried deleting a unit that hadn't been committed. Damn. [21:17] :( [21:17] Will do that in this branch, as it's a little broken otherwise. [21:25] hatch: can you look at this WIP pr? i've got a changeState that isn't getting picked up. [21:25] https://github.com/juju/juju-gui/pull/428 [21:26] sure [21:27] jcsackett what's `this` when it's fired? [21:27] is it the View? [21:30] jcsackett and this isn't handled by the retry? [21:30] I'm not sure that bailing just because the charm hasn't loaded yet is a good experience [21:30] they entered that url for a reason right? [21:57] odd I can't create a card for #1340147 it says it's already on the board...but a ctrl+f doesn't find it [21:57] <_mup_> Bug #1340147: Ghost services should be available to scale out in the mass scaling UI [21:57] and searching says it doesn't exist [21:58] hatch: make sure to show the backlog [21:58] ahah I found it [21:58] odd I had to show the backlog for the search to work... [21:58] hatch: and I told JC it's time to bail on these cards. They were nice to haves but not something blocking machine view and we need to move this work out of the way [21:58] hatch: ajax baby [21:59] hatch: so for now we'll bail, file a bug, and try to get back to it, but MV or bust [21:59] lol I mean "their search" also didn't find it [21:59] ohh ok [22:01] hatch: this is the view. [22:01] hatch: and no, this actually occurs regardless of the other retry. [22:01] and was actually occuring *before* my other branches. they just exposed it as an issue. [22:02] ohh ok [22:02] hmm [22:02] hatch: yeah, i'm perplexed. [22:02] lemme look deeper [22:02] hatch: ok. i'm running out to dinner--if you leave notes on the PR i can poke at it more tonight. want this on its way to done before i get started tomorrow so i can get cranking on MV. [22:03] sounds good [22:28] Morning [22:30] morning huwshimi [22:30] rick_h__: Hey [22:30] morning huwshimi [22:30] hatch: Hey [22:32] hatch: Makyo either of you joining the AU call? [22:32] omw [22:49] hatch: So, my question is how to update the unit count in the new scaleup UI. [22:49] hatch: In the old UI it has this data-bind https://github.com/juju/juju-gui/blob/develop/app/templates/serviceOverview.handlebars#L9 [22:49] huwshimi 2 seconds I've already written that I'll get you the diff :) [22:49] hatch: But I can't see how that's updated, to re-use it for our new UI [22:49] ah [22:51] huwshimi so I put this branch on hold while I do the ones I'm currently on but here is the diff that I had https://gist.github.com/hatched/dbdf85c23eeafb9daa44 [22:51] or at least...whatever was left [22:53] hatch: Oh great, I'll see if I can make that work. [22:53] you should be able to....as it does work ;) [22:53] but we don't want to put the databinding in the input as I understand it [22:53] the input I think is only for units we want to add [22:53] not the current number [22:53] at least that's how it's playing in my head and the mockups seem to go along with that [22:55] hatch: Yeah, that's right [22:57] I'm stepping away for a bit [22:57] bbl......likely [22:58] hatch: Thanks