[00:08] hi huwshimi [00:08] bac: Hey! [00:09] huwshimi: you coming to vegas? [00:09] bac: Yep! [00:09] cool. look forward to seeing you there. [00:09] * bac dinners [00:09] bye [00:09] seeya [00:15] huwshimi: are you cool or did you need a chat or another pair of eyes? [00:25] rick_h_: I'm not quite sure what's going on, but I'm sure I can figure it out. [00:25] rick_h_: Did you have specific things you wanted me to test, there's not really anything going on at the moment... [00:26] huwshimi: ok, well going offline for the night. [00:26] rick_h_: Yeah, head off, it sounds like tonight you shouldn't be working :) [00:26] huwshimi: just that the render works ok, maybe verify some of the height calc stuff if we need to. [00:26] rick_h_: OK sure [00:26] rick_h_: Night [00:26] huwshimi: mostly basics, nothing specific in my head yet. It'll get a lot more as we integrate with other things. [00:27] huwshimi: have a good day [00:27] rick_h_: OK sure. Will do. [10:37] morning [10:40] hi rick_h_ [10:45] we having fun yet frankban ? [10:45] frankban: you able to make the rescheduled interview tonight? [10:45] rick_h_: sure [10:46] frankban: cool thanks [11:48] benji, hey, have you used teambeamer in trusty? [11:49] dimitern: nope, I haven't (I just installed trusty recently) [11:50] benji, I'm wondering how difficult it is to set it up on trusty - perhaps building from source? [11:51] I don't think it'll be too hard. Termbemer itself is trivial to "build" because it is Python, but the dependencies might give you some trouble. [11:52] have you tried installing from the PPA? [11:52] * benji tries [12:22] rick_h_: actually I'm inclined to go with the default values for MongoClient, which is fsync of False. [12:22] bac: ok, we can try it. worst case we know where the code is [12:22] rick_h_: the defaults for MongoClient are "safe writes" as opposed to the unacknowledged unsafe writes of the defaults for Connection() [12:23] bac: gotcha, than that's great. I just remember debugging strange test issues too much from Connection() as the unsafe default writes made it a headache [12:23] rick_h_: Even without fsync=True you'll still get a fail message on conflicts. the reading i've done says avoid fsync "if you can help it". our use should allow it to be safe. [12:24] rick_h_: i'll push the latest changes for you to eyeball on RV [12:25] bac: ok thanks [12:26] rick_h_: it looks like defaults are sane. using w=0 is the only thing that gets you into trouble, breaks locks, etc [12:26] ok [12:26] yea, I worried we went too far safe with w=1 [12:26] tmi. :) [12:27] but knew we originally added the safe=true for hard to debug test failures [12:29] hwody kadams54, welcome to jujugui [12:29] err howdy [12:29] Morning all. [12:38] dimitern: the PPA wasn't set up to build packages for trusty, I have set that up now; we'll see if they build without error [12:38] benji, cheers! [12:47] benji, I see python-vte is being built right now, but can't see pending builds for termbeamer, here: https://code.launchpad.net/~benji/+archive/termbeamer/+packages [12:48] dimitern: build started [12:48] (and the VTE build worked, which is the one I was worried about, so that's a good sign) [12:49] benji, great! [13:04] welcome kadams54 [13:05] rogpeppe: morning, do you have time today for reviewing https://codereview.appspot.com/77420043 ? thanks a lot [13:08] kadams54: welcome! [13:17] thanks for the review frankban! [13:17] hatch: mp [13:34] * frankban lunches [13:36] hatch: when you have the else you don't need the return :P [13:36] hatch: having BOTH is the issue [13:36] I like it being explicit in case someone adds code outside of the conditional [13:37] right, but it's harder to find the return value of a func when it's returning all over the place [13:37] "oh, this function has a condition on line 10 that returns something else" [13:37] I generally have a dislike for those, but I admit it's a personal pref [13:37] especially because you were just using the return to avoid an else in this case. There's no explicit return value [13:38] heh and I prefer to use returns :) [13:39] s/return/break and maybe... [13:39] empty returns are so semantically just broken [13:39] a break could potentially introduce the issue I was trying to avoid [13:39] lol no they aren't [13:40] "yea, here's you're return value...nothing, and never will be...booya" [13:40] "but if you add a return value in one case, better go grepping through the whole function in case we return in othre places you're not looking in" [13:40] ugh :P [13:40] well it's returning undefined :) but in this case it's acting as a short on the function [13:41] right, like an if else [13:41] lol the function is only 10 lines long [13:41] if x do a else do b [13:41] // don't add extra code outside of this if block [13:41] vs [13:41] return; [13:41] it's the principal of the thing! /me gets his soap box [13:41] :) [13:41] * hatch gets bigger soap box [13:41] (and the soap box arms race begins) [13:41] lol [13:41] lol, canadaians, bigger? never! [13:42] bah canadians [13:42] hah fine - I'll remove the return.....under duress! [13:43] lol [13:48] hatch: Huw's branch isn't quite ready, I turned it into a pull request and it's failing CI, but if you get time could you pull it down and review/qa please? [13:48] hatch: and we can provide feedback for him to try to get that landed tonight [13:58] rick_h_ yeah sure [13:58] hatch: thanks [14:01] rick_h_: vacation request submitted for 1/2 week before sprint [14:01] bac: thanks === marcoceppi is now known as marcoceppi-mobil [14:06] kadams54: you should have an invite for kanban, let me know when you accept so I can give your account permissions [14:07] rick_h_: Accepted. [14:07] kadams54: awesome thanks [14:16] rick_h_, hatch: I agree it's a personal preference. I also like return to exit the function faster and then the rest of code being not nested, i.e. from now on I can read the function without thinking about conditionals. it's just that return + nested else seems a bit unclear to me [14:16] hmm [14:16] :) [14:23] kadams54: ok, you should be able to load the kanban board now [14:23] kadams54: please verify before the standup in 40 [14:23] rick_h_: Yup, I'm in and looking it over. [14:26] kadams54: oh, you need to get travel setup as well [14:28] frankban wow I don't think I've ever typed so much in a code review :P Mind taking a look and seeing if I answered all of your questions [14:28] and btw - you found a serious bug :) thanks [14:29] welcome kadams54 [14:29] hatch: I'll take a look asap [14:30] rick_h_ il? "Yo man this feature is so ILL" [14:31] hatch: :P [14:31] lol [14:31] good way to remember it [14:31] haha [14:31] my next feature flag will be steezin [14:37] rick_h_: did you see ppetraki's comments in #juju-dev? [14:37] the charm store is exporting null bundles [14:37] has anyone tried to download a bundle lately from the charm store? I've tried to export several different ones and all I'm getting is [14:37] envExport: [14:37] services: {} [14:37] relations: [] [14:37] series: precise [14:37] that one ^ [14:37] lazyPower: no, what's up? [14:38] ppetraki: oh looking [14:39] ppetraki: hmm, how are you downloading? Is this when you build an environment on jujucharms.com and hit export? [14:39] * ppetraki digs up example [14:40] ppetraki: I just did a qiuck test and it worked. I wonder if you're trying to do something differently or with something that's not working? [14:40] ppetraki: screenshot of your environment would be useful to replicate on jujucharms.com to test out [14:40] ack [14:41] ppetraki: thanks, we'll help look into it for sure. Thanks for the report of the issue. [14:41] * rick_h_ goes afk for a few to migrate home from coffee shop [14:42] rick_h_, http://picpaste.com/upload-guibad-aN53Tvlj.png [14:42] rick_h_, I just visit the bundle, click the '^', and open the file it downloaded [14:45] ppetraki, deploy the bundle to the GUI first. The ^ button exports the canvas, not the charm or bundle you are viewing. [14:45] Will take a note, though, that's worth clarifying. [14:46] * hatch is also relocating [14:47] ppetraki, if you wish to grab a bundle's code from the store without deploying, you can go to the code tab, eg https://jujucharms.com/bundle/~charmers/mediawiki/6/single/#code and select the bundle.yaml file. [14:47] Makyo, it says it'll be deployed to my local provider immediately, does that mean my current juju-env? [14:48] ppetraki, if you are viewing it at jujucharms.com, the provider is the demonstration provider, which is just a sandbox within the browser. [14:48] Makyo, yeah, as a regular user I have no idea what the difference is [14:48] and assumed the worse [14:48] * ppetraki well, im not exactly a regular user [14:49] ppetraki, that's totally fair. We'll see if there's a way to make that more clear. [14:49] Makyo, thanks [14:50] jujugui call in 10 [14:50] ah thats something I didnt think about, that charms are transient until you actually "deploy" them on the gui [14:50] see? this is why you guys make the big bucks. [14:50] * lazyPower thumbs up [14:51] https://bugs.launchpad.net/juju-gui/+bug/1294694 cc ppetraki [14:51] <_mup_> Bug #1294694: Bundle deployment in sandbox is confusing [14:53] * Makyo coffees self thoroughly before call. [14:54] Makyo, sub'd, thanks === marcoceppi-mobil is now known as marcoceppi [14:57] jujugui call in 3 [14:59] hatch: thanks. also your comments look good. what do you think about my last proposal (re: functional approach)? [15:00] yeah I think that will be a good change [15:00] sorry I meant to comment on that :) [15:00] kadams54: https://plus.google.com/hangouts/_/canonical.com/daily-standup?authuser=1 in case you can't see the link in the calendar [15:00] hatch: oh, and it would be nice if the changeSet hierarchical structure is documented in that branch [15:00] thanks for that branch hatch [15:00] benji: standup pingu [15:00] yeah I can do that - I kind of didn't on purpose because I figured the structure might change a bit === marcoceppi is now known as marco-traveling [15:13] luca is there an overview image of the new zoom assets? [15:14] hatch: like a mock up? [15:15] luca yeah, you mention getting rid of the slider, so are there just going to be two buttons floating there now? [15:16] yeah, sec [15:16] can you add it to the bug plz and thanks :) [15:16] hatch: same style as the bundle visualisation zoom slider [15:16] https://docs.google.com/a/canonical.com/file/d/0B_l975nRB6BuSEVfQXhxU2g2SXc/edit [15:16] ahh ok cool [15:17] I'll update the ticket [15:17] ok, thanks [15:17] hatch: try to upload the orig png please [15:17] hatch: as the gdocs link isn't public bug the bugs are [15:17] hatch: so it keeps more info in the open [15:17] ok cool [15:17] ah [15:17] right [15:17] sorry [15:18] luca: np, just want to keep us all nice and open for community/etc [15:18] thanks for the update and bug report! [15:19] addeed [15:21] thanks hatch [15:28] I really need to figure out how to setup znc [15:36] someone just parked a van across the street on the sidewalk and snowbank......intentionally.... [15:37] that's got to be a first [15:42] hatch: its the nsa! [15:42] hatch: also - have you looked at quassel? the ZNC component is dead simple to setup/use [15:42] lol - or a, now-stuck, idiot [15:42] + it has a mobile component for those of us on the go that like to keep irc in our pockets. [15:43] lazyPower well I'd like to set up a ZNC bouncer so that I could log into it with textual in osx and something like quassel in Ubuntu [15:43] quassel has a osx client to... [15:43] oh? [15:43] interesting [15:44] yeah it runs on every flavor of OS, i'm pretty sure there's a chrome extension being dev'd too [15:45] I really like how textual uses safari so you can theme using css [15:45] safari webview that is [15:47] hatch: well, DO has you covered - https://www.digitalocean.com/community/articles/how-to-install-znc-an-irc-bouncer-on-an-ubuntu-vps [15:47] and that smells of an opportunity to be charmed [15:47] * lazyPower nudges gently [15:47] oh awesome [15:48] lol - I need way more hours in the day [15:48] you and me both brother [15:48] I wish DO had a juju provider [15:49] uhm [15:49] hazmat wrote one [15:49] hazmat has written an alpha provider plugin [15:49] :) [15:49] yeah [15:49] https://github.com/kapilt/juju-digitalocean [15:49] hatch, jump on in.. its fresh out of the oven ;-) [15:50] that's when the cookies are best [15:50] lol [15:50] it's true! [15:50] hazmat does it support deploying using --to ? :) [15:50] bootstrap + gui + znc + whatever other garbage I want to put on their smallest node :P [15:51] hatch, ofc.. its a little differrent than a normal provider.. you $ juju docean bootstrap .. $ juju docean add-machine -n 5 [15:51] heh 512mb isn't much of a node [15:51] before deploying.. and then you can do deploy to .. etc [15:51] hatch, you can even do envs that span the globe ;-).. i did one env split between ams and nyc. [15:52] nice [15:52] manual provisioning solves everything :-) [15:52] hazmat is that a decision you made or a technical limitation of DO not being able to deploy and auto add machines? [15:52] hatch, its a limitation of juju's ability to do external provisioning plugins [15:52] * rick_h_ goes to nuke lunchables [15:52] hatch, i'd have to api intercept to do more interesting things [15:52] ahh [15:53] hatch, but really its a trivial burden imo.. to type one extra line to allocate machines before deploying or add-unit. [15:53] yeah for sure [15:54] I'd have to think of something else to put on that node.....$60/yr for a znc bouncer is a little nuts haha [15:54] hatch: if you go the route of quassel i'd have no problem putting you on my bouncer [15:55] but you have this need for textual - sooooo [15:55] can I trust you not to read my messages????? [15:55] CAN I!!! [15:55] no [15:55] i'm the nsa dude [15:55] :P [15:55] I knew it [15:55] haha, yeah man idc [15:55] i never mess with the db [15:55] if you trust me i trust you kinda trust [15:57] interesting that there are no 'znc as a service' startups [15:57] lol [15:57] such a niche market [15:57] I bet people would pay $5-10 a year [15:57] oh right..... znc and food picture sharing as a service [15:57] ^ startup idea [15:57] *patented* [15:57] hatch: i'll sync with antonio i'll ping you with the quassel core details [15:58] s/i'll/after my [15:58] cool thanks [15:58] np [16:49] frankban https://gist.github.com/3dd5ddad7b0e765af3e5 a litle more functional [16:49] thoughts? [16:49] hatch: looking [16:51] thanks [16:52] hatch: looks good [16:52] * Makyo runs to grab prescription, will have other laptop at coffeeshop. [16:52] cool - now to fix all the tests [17:02] bac https://twitter.com/dam/status/446323090612420608 hehehe [17:06] guihelp: I need one review/QA for https://github.com/juju/juju-gui/pull/186 Anyone available? thanks! [17:06] sure I can [17:07] thanks hatch [17:10] frankban: I'll take a look at your review and perhaps you'll have time to look at https://codereview.appspot.com/76860047 while I do that [17:10] frankban: oh, you already have a reviewer, well, I'll let you review my branch anyway [17:11] benji: sure, I'll take a look [17:11] frankban: note that the branch has undergone extensive QA, so you can feel free to skip that if you feel so inclined [17:12] benji: cool, ok, I'll just look at the code [17:12] frankban: did you determine if we should file a -core bug on the issue? [17:13] rick_h_: I planned to investigate that in core and file a bug if required, with instructions to dupe [17:13] frankban: ok cool thanks [17:15] DrabMakyo always has me looking for ShinyMakyo === DrabMakyo is now known as ShinyMakyo [17:15] Given that this is the Air, I might as well. [17:15] lol [17:15] lol [17:16] frankban +1 but I can't QA right now unfortunately [17:16] Damn, missed opportunity for FabMakyo [17:17] hah! [17:17] hatch: thanks. benji: could you please QA my branch? (only if you have local envs up and running) [17:18] frankban: I don't have local set up yet. I am willing to get it set up if you don't think it'll be too painful. [17:20] ooo move state out of browser.... I like the sound of that card ShinyMakyo :) [17:21] Haha. [17:21] Part of the other card with my face on it. [17:21] benji: well, I am not on trusty, but maybe it's worth trying the following: sudo add-apt-repository ppa:juju/stable && sudo apt-get update && sudo apt-get install juju-quickstart && juju-quickstart [17:21] * benji tries [17:22] ShinyMakyo what I have been doing is, when I create a sub card, add a count to it and then subtract that from the 'parent' card....I'm not sure if it works yet but it at least keeps the estimated work somewhat accurate [17:22] Ah, yeah, that makes sense. [17:22] hatch: ShinyMakyo yes please [17:22] Done and done [17:30] lately I've been getting test failures where the traceback points to another test entirely.....very odd [17:35] benji: I am not sure about how that nrpe hooks work. did you create relation-broken/departed sym links in your branch? [17:36] frankban: I did [17:36] benji: oh, ok, Rietveld does not show that :-/ [17:36] frankban: quickstart is bootstrapping ec2, is that expected? Should I set up a local environment config and make that the default? [17:37] benji: I tend to [17:37] benji: as it's a lot faster and nicer to work with and doesn't cost $$ [17:38] benji: yes, so you already had an environment. so now you can stop quickstart, destroy the ec2 env, and then create a local one with "juju quickstart -i" [17:39] benji: also, since you already have juju configured, before running quickstart -i, it's better to apt-get install mongodb-server lxc [17:40] frankban I pushed the latest changes up, mind taking a look for a +1 ? [17:40] frankban: too late, but quickstart told me to install it, which is nice [17:41] cool [17:42] benji: review done [17:43] thanks [17:48] Okay, prescription's ready. Will be back home in a few. [17:49] frankban: local environment setup seems to be working (but not done yet) [17:49] once it's done and I'm sure it's working I'll QA your branch [17:50] benji: cool, setting up a local env with the GUI is also the first step of my QA instructions [17:51] I'm glad you told me because I would have destroyed it before looking at the QA instructions :) [17:51] frankban: uh oh: juju-quickstart: error: juju-gui/0 is in an error state: error: hook failed: "install" [17:52] benji: interesting, what's in ~/.juju/local/log/unit-juju-gui-0.log? [17:56] frankban: oops, I have already destroyed the environment and am trying again [18:02] frankban: http://paste.ubuntu.com/7121143/ [18:02] Makyo did you ever get a chance to look at my branch? [18:04] benji: so apt failed. juju status? [18:04] frankban: http://paste.ubuntu.com/7121162/ [18:06] benji: ok, so the lxc is precise and that's correct. next step is trying the same command from inside the lxc: juju ssh juju-gui/0; sudo -i; apt-get -y install libapt-pkg-dev python-apt python-launchpadlib python-tempita python-yaml [18:06] hatch, Yes, but not with changes. Let me prowl through again. [18:06] let's see what the real error is [18:08] frankban: there's the problem: No space left on device [18:09] oh... [18:09] who needs a root partition larger than 20 gig I said, that would be crazy I said [18:09] :-) [18:09] I think you'll have to look somewhere else for your QA [18:12] ok. rick_h_, Makyo, bac: ^^^ anyone available for QAing https://github.com/juju/juju-gui/pull/186 ? [18:12] doubtful [18:12] frankban, I can. [18:12] Makyo: thanks [18:13] benji: it needs 20G free? [18:14] bac: I only have 20G total... oh, I only have 10G total on that partiion [18:15] benji: yeah, i had to go through the resize dance on my vm. was doable but unpleasant [18:30] Makyo ok cool thanks, I just want to get it landed [18:31] frankban: let me know I can after calls today if you still need someone [18:31] hatch, ack. [18:32] rick_h_: thanks, Makyo is doing the QA [18:32] frankban yeah sorry about the QA, updating juju borked my machine so I need to upgrade it and haven't had the time :/ [18:32] well juju and my machine are borked together :) [18:33] hatch: np, thanks for the review [18:38] * hatch lunching [18:43] benji: very good plot twist :-) [18:43] heh [19:00] frankban, I can't reproduce the first error [19:00] Let me get more info. [19:00] juju 1.17.4 [19:01] frankban, (ping re interview, though) [19:01] Makyo: maybe it's intermittent. uhm. I was using stable [19:01] frankban, ack, will keep poking. [19:01] Makyo: thanks [19:36] rick_h_ around? [19:37] hatch: otp, what's up? [19:37] ok I'll type :) was gona request a hangout [19:37] sec [19:38] when a user is calling setConfig we allow them to pass in a queued service or a real service. If the service is queued and they use the `immediate` flag, should we A) disregard or B) deploy the service and set the config? I vote A with a console warning [19:39] kadams54 hey just fyi you're still running through sourceforge's vpn for irc :) [19:39] hatch: hmm, throw an error for now? [19:39] hatch: lol, thanks man :-) [19:39] hatch: I mean we don't have a use case for this atm [19:40] rick_h_ throw as in throw()? [19:40] right, I just want to cover the bases [19:40] hatch: so it seems like a case of problems we don't have atm [19:40] haha I'm avoiding them before they are :) [19:40] yea, +1 on thinking about it, but don't want to get hung up on it [19:40] I think that the 'immedaite' is a real usecase and it's really not much more work to do it at the same time [19:41] the only use case I see is testing? [19:41] or is there more I'm missing? [19:41] well just if we want to give the user the ability to bypass the deployer bar for any request [19:42] right but that's not in any design for now [19:42] and testing [19:42] there's not real use case for the user to select it [19:42] and testing I think we can throw the error that shows "you forgit an immedite on the service, fix that in your setup" [19:42] forgot [19:42] ok I'll throw for now [19:42] does that make sense? [19:43] yep it does - I am pretty sure that it's an interaction we will want to allow though [19:44] hatch: k, yea that's why I'm fine thinking on it, but until we have a use case to walk through I'm not sure how to handle it [19:44] right ok cool [19:51] * hatch grumbles about using if/else vs multiple returns [19:51] :P [19:51] hatch: I give, if you and frankban like it I'll give. :P [19:51] VICTORY! [19:51] lol [19:52] how about short variable names? [19:52] too far? [19:52] NO! [19:52] lol [19:52] hahaha [19:52] or something like that [19:56] Not to start any holy wars, but… what are some of the different dev setups you guys use? I've been playing around with several different ones today (vagrant, regular Ubuntu VM, dual boot) and trying to figure out where I'm going to be most productive. [19:58] kadams54 atm I'm using OSX with the gui's vagrant config [19:59] I also have a box in the basement running Ubuntu to do real env testing [19:59] I also have a Ubuntu VM for various ubuntu stuff locally [20:00] I am trying to run Ubuntu on metal but the current kernel doesn't work on the new Apple haswell cpu's [20:01] but you will find that working IN Ubuntu is a lot less work when working with charms, juju-core hacking, etc [20:02] so...depending on your laptop the answer may be made for you :) [20:04] rick_h_, lost hangouts, sorry [20:04] Or maybe I lost Internet... [20:05] I see u [20:05] messages that is :) [20:06] jujugui I need a pre-imp call with someone about the set_config interaction with a ghost service [20:06] anyone available? [20:09] Another question: I'm not going through SF VPN, but I do have a SF cloak. Does Canonical have cloaks? [20:09] Yeah, I may setup a spare desktop I have lying around to run Ubuntu on the bare metal. Right now I'm leaning towards vagrant on my MacBook. [20:09] It's an older MacBook, but I use some stuff that's OS X-only… like coreStorage to tie together an SSD and a traditional HD into a single logical volume. [20:10] that's a good question - I have seen it....I just assumed they were running through the vpn [20:10] Hah, no. [20:10] And there was also the hitch this morning with getting my webcam to work in Ubuntu… [20:11] some may give you a hard time about it.....I have a ubuntu logo on my Apple symbol lol [20:13] dogfooding and all that :) [20:15] Yeah, I'm beginning to realize that :-) My intention is to spend as much time in Ubuntu as possible. [20:24] hmm there are quite a few ghost service interactions that need to be thought out for this state manager [20:24] hatch: :) [20:24] rick_h_ when you have a second lets have a hangout [20:24] hatch: k, shoot me a link [20:24] https://plus.google.com/hangouts/_/76cpiohup8sdmhrmvej1p946sg?hl=en [20:51] Makyo poke [20:52] hatch, sorry, got wrapped up :( 5 minutes? [20:52]