=== matsubara is now known as matsubara-lunch [12:43] rick_h_: jcsackett sinzui are you guys awake yet? :) [13:01] I am luca_ [13:01] you are not luca. you are sinzui! [13:02] sinzui: great, I have a meeting with someone from the Webapps team to discuss how to make Juju responsive and would like someone involved with the front end to join the talk. [13:02] Sorry, more coffee is needed to find myself [13:02] gary_poster: the new ppa, i'd like to put it in ~juju-gui-charmers. ok by you? [13:02] gary_poster: haha [13:02] :-) [13:02] bac +1 [13:02] coolio [13:02] sinzui: its in 1 hour and would be about half an hour long [13:02] gary_poster: we need to talk also about transitioning away from python-charmhelpers [13:03] okay. [13:03] That works for me. luca_ do you want more than one person? [13:03] gary_poster: it turns out it never got promoted to an ubuntu repo as i thought. it has only every lived as part of ppa:juju/pkgs [13:04] bac, it was python-shelltoolbox that was promoted [13:04] sinzui: anyone can come, he started asking me techy questions yesterday that I couldn't answer. [13:04] gary_poster: i had thought both were. [13:04] ah, no. [13:04] luca_, fab. [13:04] bac, anyway, we want to migrate to wedgwood's thing, right? [13:04] sinzui: I've sent you an invite, feel free to give anyone interested the hangout link [13:04] I mean, that's what you want to talk about [13:05] orangesquad, staging is again a victim of zookeeper/juju-agent problems. I am going to reboot the unit [13:05] gary_poster: i think so. he is advocating for not putting in "client-side" helpers, those that we use in our tests [13:05] by advocating, i mean marked them as deprecated and not promoting them out of contrib [13:05] bac, ah right, I saw that. what's his argument, and what is his suggested approach for us? [13:06] his argument is simply client side stuff doesn't belong. he suggested putting them in charm-tools, which is a non-starter for me. [13:06] what's an example of said "client-side" helper? what does "client-side" mean? [13:06] in context [13:07] gary_poster: http://bazaar.launchpad.net/~charm-helpers/charm-helpers/devel/view/head:/charmhelpers/contrib/charmhelpers/__init__.py [13:08] client-side means things that you don't use in hooks [13:09] ah I see. have we already argued against this, casting these as test helpers, and suggesting a test-helper directory? [13:10] testutils [13:10] something like that [13:12] orangesquad looks like charmworld/3 's agent is truly dead. The site is up. I will let it run for now [13:14] I tend to call the module holding things like that "testing" [13:17] sinzui: roger. Should we reboot so that landings don't work? I'm not expecting to land anything soon. [13:17] s/landings don't work/landings work/ === matsubara-lunch is now known as matsubara [13:18] gary_poster: no, i raised the issue in review but didn't make that specific suggestion. i will do so now and propose a branch. [13:18] abentley, I did reboot it, the agent did not start [13:18] great thanks bac [13:19] sinzui: Oh. We could remove-unit charmworld/3 and add-unit charmworld maybe? [13:20] abentley, after my rollback of elastic search last night, I think I want to be very cautious about next steps, but I think add-unit is right. that is what I did a few weeks ago. [13:21] sinzui: It's certainly more convenient than destroy-service followed by deploy because no reconfiguration is needed. [13:32] benji, btw, you can plan your coffeescript revolution with bcsaller. he is sold too. :-) [13:32] bcsaller, to the coffeecave! [13:38] good morning [13:39] jujugui, heads up: npm appears to be having some pretty serious issues today. https://github.com/isaacs/npm-www/issues/325 is making it hard to very install npm dependencies. If anyone wants details, lemme know. [13:39] morning hatch. safe trip? [13:39] very hard to install, that is :-P [13:39] this will likely affect CI :-( [13:40] it was thanks, long trip though - the planes were pretty old so none had wifi and only one had tv's :) [13:40] heh [13:40] solar flares affect CI [13:40] -/ [13:40] :-/ [13:40] bac: lol [13:50] hatch, how did angular go? working through examples slowly but still optimistic from what I see [13:51] ugh, this npm thing is really annoying [13:52] gary_poster: right now I'm trying to figure out how to 'pre compile' templates so that we can send them over the wire like we do our hbs templates [13:53] cool hatch. other things working though? [13:54] technically yes, the passing data between angular and yui is pretty trivial [13:55] you also need to do special techniques if you want to minify it because of how the dependency injection works [13:55] and I'm a little concerned about performance on our apps because our DOM is huge [13:56] gary_poster, hi there, a question regarding the tarmac setup for juju-gui: Currently you use JUJU_BRANCH to define which branch to run tests on. if we start using tarmac, we want to run the tests from the trunk branch + the merged in changes of the approved merge proposal, right? [13:56] trivial: cool hatch, good. looked like it. special techniques to minify: tricky/problematic, or you have resolved to your satisfaction, at least mostly? performance: what do you plan to show at mtg? [13:57] the angular guys have a workaround for the minification issue [13:57] matsubara, right. the JUJU_BRANCH thing is for the post-commit CI approach [13:57] ok [13:58] my performance issues are based around the idea that it queries the whole dom for these directives [13:58] so the time to do that will increase with the amount of elements in the dom [13:59] hatch, that's one-time on load? [14:00] on load and then any time you initiate a module [14:00] I `think` because it's undocumented :) [14:03] hatch, any time you initialize a module for the first time, or any time you instantiate a model for a given view? I'd hope for the first. Also, I expect we could test/time this pretty trivially, but hooking angular in to the app and finding the start of the DOM walk and the end. just because we don't actually have any directives isn't relative to your concern AIUI. :-) [14:08] just called my ISP and she said my radio was broadcasting at too high a signal. she's going to turn it down from 11 to 9. [14:10] gary_poster: yeah it's probably not a valid concern [14:10] it's quite a bit less code and easy to modulize [14:10] I will likely need to spend the weekend reading more on it [14:10] hatch, do we have something for the team to look at today, or is that premature? [14:12] honestly I think it's a little premature - i'd like to spend the weekend on it but I have learnt a lot about how it works and still think that it is probably the right approach for such a DOM centric portion of the app [14:13] I have found the best documentation is blogs, youtube videos, and stackoverflow [14:15] hatch, ack. I have 15 min before a call--can we hang out in 5 for about 10 minutes so I can get a bit more detail? [14:15] yeah sure [14:18] the build fails with window undefined if I try and use window in 3rd party files [14:21] thought that was odd [14:24] hatch, heading to guichat, if you are available [14:25] yeah one sec I just gota grab my headset out of my bag [14:27] gary_poster, can you add https://launchpad.net/~jujugui-lander as part of the team allowed to change lp:juju-gui ? [14:30] matsubara, done [14:30] thanks gary_poster [14:30] gary_poster, my meeting is running over. I will be a few minutes late [14:31] ok thanks sinzui, ping me when you are ready [14:40] gary_poster, I am coming to the meeting now [14:44] gary_poster: npm is out to get me? ack. With all the deps the dir is 27M so I am reluctant to push that up but I can if you want [14:44] bcsaller, :-) on call. maybe you will need to screen share [14:45] gary_poster: that can work [14:49] the node crowd is saying it's fixed now once dns updates [14:50] that would be good [14:50] bcsaller, could you fix bug 1185982 [14:50] <_mup_> Bug #1185982: "=" and "-" keybingings are too agressive [14:50] gary_poster: I can look at it after the meeting [14:51] thx [14:51] jujugui call in 9; kanban now [14:55] wow, glipper got up to 1GB of resident memory [14:57] npm is working for me again [14:58] using googles dns at least [14:59] does google show ads on dns too? ;-) [15:02] jujgui call now, I'm coming, sorry :-) [15:25] teknico, I am sorry I was miskta [15:25] mistaken, marco is traveling today. [15:25] arosales: ok, thanks [15:26] teknico, if you are blocked m_3 may be able to assist until marco gets back. [15:26] arosales: great, thanks again [15:26] teknico, there are a couple of alternatives.. opendns is one [15:27] hazmat: that's what I used to use before moving to google dns :-) could not stand the 404 intercept [15:29] teknico, I caught marcoceppi_ for a bit between traveling [15:29] hey teknico [15:29] if you have any questions may be good to get them in now :-) [15:31] marcoceppi_: hi, sorry to interrupt your flight :-) still having problem with selenium and display [15:32] marcoceppi_, does this look ok, if you can see it? http://pastebin.ubuntu.com/5719876/ [15:32] teknico: that shoudl be good enough, as long as the host as it set properly [15:33] marcoceppi_: the host? where? [15:34] teknico: the host being your maching [15:34] I considered just mirroring everything in os.environ to the env in Python but figured it'd be better to piecemeal them over. I may have been incorrect in that assumption [15:34] marcoceppi_: well, forwarding the whole env would definitely be simpler, if maybe a little cavalier :-) [15:35] teknico: I wanted to try to make sure the env wouldnt' mess anything up between people's machines by having a set number of things, least common denominator or variables, used [15:36] marcoceppi_: yes, it sounds like a reasonable approach [15:36] Though, it might just be in vain. If it turns out to not really be a problem I'll just have the whole thing mirrored and amend the few env variables like JUJU_ENV, etc [15:37] marcoceppi_: it would be worth trying it to see if it improves matters, even temporarily [15:38] teknico: if you want to patch your local install to use the whole of the os.environ I can send you a patch. Not sure if I want to put it in trunk just yet, I would hate to set one expectation only to remove it again at a later date [15:39] marcoceppi_: no need to commit it, just send it to me when you can, thanks [15:41] teknico: ack, I'll send it over in a few [15:42] marcoceppi_: thank you. don't miss flights over this, though :-) [15:42] teknico: sadly I'm in a car. I'd be far happier flying :P [15:43] marcoceppi_: maybe try a flying car then ;-) you're not driving, hopefully :-) [16:02] teknico: I've not created just a patch in bzr before, so I have a bzr send, which google told me I should use [16:03] orangesquad, the new ES is queued to build. The issue was my merge script to repack the build. The deb contain the new and old versions of lucence jars. ES 0.90.0 is not compatible with the 4.1 jar that were from the previous repack. [16:03] marcoceppi_: the output of bzr diff would be enough [16:03] teknico: http://paste.ubuntu.com/5720274/ [16:05] marcoceppi_: perfect, thanks again [16:05] teknico: np, I'll be afk for a few but feel free to ping me in here, just expect a delayed response! [16:06] marcoceppi_: I'll be EOD in one hour anyway, have a safe trip! [16:06] teknico: thanks! have a good weekend! [17:27] hatch: api1 helping with your browser issues in dev? [17:28] haven't tried yet :) [17:30] somehow I managed to screw up my node install but I think it's fixed now so I'll be trying shortly [17:56] orangesquad: does anyone need staging? I have had three failed efforts to add-units. I want to deploy a replacement service. [17:57] sinzui: Not I. Perhaps this is a good time to try out Kapil's deployer thing. [17:57] abentley, capture the stack and redeploy it? [17:58] sinzui: If we can. I kind of assumed we'd have to write it by hand. [17:58] The deployer config, I mean. [17:59] I don't want to write one by hand this week. I want to stop get this distraction out of the way so that I can work on the charm model branch. [18:00] sinzui: If using deployer is not a quick fix, then I have a script you can use. [18:02] I was just going to deploy a new charmworld service...and remember to set the charm_import_limit [18:03] sinzui: Oh. Have you tried doing remove-unit first, terminating the machine, and adding a unit? [18:05] no, actually I haven't since I wanted to be certain I got a new machine. I don't really need a new machine. I just want a working one [18:06] sinzui: If the agent on the machine is borken, you may need to get a new machine in order to get a working agent. [18:06] That is what I was thinking [18:08] abentley, I just used juju-jitsu export. It gave me something to deploy, but I cannot use it until the replacement ES is built. I will to remove the service for now [18:09] sinzui: Here is my script anyhow, in case it's useful in the future. https://pastebin.canonical.com/91952/ [18:09] thank you! [19:07] http://gizmodo.com/hell-yes-razer-made-the-most-powerful-small-windows-la-510358373 [19:07] finally 'possibly' an alternative to a MBP [19:11] hatch: Heh, I was thinking about it too. [19:12] :) [19:12] of course I couldn't buy one until the new MBP comes out anways so that I could compare lol [19:14] abentley: this razer would probably be a lot easier to install Ubuntu on too haha [19:15] hatch: Especially if we can get other Canonifolk to buy it :-). My Laptop Refresh Benefit should kick in just after 13.10 is released. [19:16] haha - I think I have to wait 2.5years lol [19:59] "TypeError: '[object BlobConstructor]' is not a constructor"... thanks [19:59] never seen that error before :) [19:59] sounds like an old browser w/export [20:00] yep; unfortunately it is the one that runs hour hedless tests (phantomjs) [20:05] this makes me wonder about the test coverage of the other code that builds blobs [20:08] are we running an older version of phantomjs? [20:09] current version is 1.9 and I have 1.8.1 installed [20:09] might want to see if you can update to 1.9 [20:10] npm has 1.9.0-6 [20:10] not sure what the -'s mean [20:20] hatch: thanks for the hint. The "-" is probably a packaging change (but not a change to the code). I'll try to update and see if my tests start to pass. [20:21] if that does fix let me know so I don't run into the same :) [20:23] k [20:26] well, now I can't get through a test run because phantomjs segfaults [20:27] no dice; the updated phantomjs didn't help [20:29] skip that test based on UA? [20:46] finally set up dual google accounts - wow I can't believe I didn't know about this sooner [20:46] so much easier hah [20:49] having two may be good, but I can't recommend having a half-dozen [20:49] I can't really see needing more than two :) [20:49] personal & work [20:50] and I love how it automatically picks the proper account based on the url [21:05] hatch: it hasn't always been so smart [21:11] gary still around? [21:11] gary_poster: ^ [21:12] hatch yes [21:12] shoudl go now though :-) [21:12] I have added a few things to the evaluation notes [21:12] just wanted to fill you in on a couple things I found out today [21:13] cool hatch. quick call? [21:13] sure [21:13] in guichat [21:13] yeah [21:20] \o everybody [21:22] bye bac [21:48] rick_h_: hey still around?