[00:45] Is anyone about? [00:45] huwshimi: what's up? [00:46] rick_h__: I'm just a bit stuck with this branch... [00:46] rick_h__: I probably just need to talk it over with someone [00:46] huwshimi: sure, let me look at what you're up to [00:47] huwshimi: setup a call and I'll jump in [00:47] ok [00:48] rick_h__: https://plus.google.com/hangouts/_/gw3fj4cjdgws5hjs5dngj7og6ma [07:16] huwshimi, urulama: morning! [07:17] rogpeppe: Morning [08:59] rogpeppe: hi (got lost in the document land) :D [11:02] morning [11:19] Morning rick_h__ we are planning to do some MV testing next Tuesday, what state do you think comingsoon will be in? [11:58] luca: I'll see, I'm out tomorrow through tues so will leave instructions awnd try to keep an eye on it while away [11:58] luca: but I'll have a sever lack of interwebs while away in the woods. This place has no cell coverage, but a clubhouse with slow wifi I can try to visit [11:58] rogpeppe: got a sec to chat? [11:59] rick_h__: sure [11:59] rogpeppe: https://plus.google.com/hangouts/_/g27ldilysl2kzyynhvqc3ahbiea?authuser=1&hl=en [12:21] rogpeppe: google search ftw, found the resources draft spec, might be a bit out of date but something to look at. [12:21] rogpeppe: shared it with you so you should get an email/see it now [12:22] rick_h__: thanks [12:22] rogpeppe: also forwarding you another email [12:26] hi rogpeppe, i'm having some issues with godeps. do you have a moment? [12:26] bac: sure [12:27] rogpeppe: this is on jenkins, trying to build and test juju/charmstore. i have this script: http://paste.ubuntu.com/7736543/ [12:27] the godeps step is producing odd output: http://paste.ubuntu.com/7736539/ [12:48] bac: sorry, didn't see your second line there. looking. [12:48] rogpeppe: in this output http://paste.ubuntu.com/7736609/ i see two problems: 1) some specified bzr version revisions don't appear to exist. is it then trying again too quickly and hitting locking issues? [12:49] rogpeppe: ignore that second paste and look at the most recent [12:49] rick_h__: that’s no problem, we are just testing users reaction to how its designed and to see how they would use the drag and drop feature [12:51] luca: sounds good [12:51] bac: interesting. i'll try to repro [12:51] rogpeppe: thanks [12:53] rogpeppe: the script changed a bit to produce that output. it was: http://paste.ubuntu.com/7736632/ [12:59] bac: hmm, i tried (almost) exactly those steps and it works for me [12:59] rogpeppe: with a fresh GO directory? [12:59] bac: yes [12:59] bac: hold on, i'll paste a slightly simpler script which should be equivalent [12:59] rogpeppe: cool, can you do our ci then? i'll just have it email you. :) [13:00] s/trusty-slave/roger-slave/ [13:01] bac: one thing that might be affecting it is that you really want to put $GOPATH/bin at the start of $PATH, in case there's a godeps elsewhere (unlikely i guess) [13:02] that's a good change but highly unlikely [13:04] rogpeppe: if you'll paste your simplification i'll try it on jenkins [13:05] bac: just making sure it works [13:09] bac: http://paste.ubuntu.com/7736692/ [13:10] bac: it's possible that bzr doesn't like running concurrently [13:10] bac: hence the -P flag [13:11] bac: but i don't understand the `unrecognized import path "labix.org/v2/mgo/..."` error [13:11] rogpeppe: yeah, i'll bet that helps. also the trap is nice. [13:11] bac: just like Go defer :-) [13:11] rogpeppe: line 11 won't cause my revision of charmstore to get updated will it? [13:11] bac: there is no revision of charmstore at that point [13:12] bac: so it'll get the latest [13:12] yes there is. the one i copied [13:12] i don't want the latest [13:12] ah twats [13:12] bac: i just realised what you were doing :-) [13:12] sorry, i didn't explain well [13:13] i think the -P might be enough to test [13:13] bac: can it make a difference which way the merge happens? [13:14] rogpeppe: i don't understand your question [13:14] which merge? [13:14] bac: the merge that's being done with git pull [13:16] rogpeppe: in http://paste.ubuntu.com/7736539/ everything before line 20 is done by jenkins. my script starts at 20. [13:16] bac: jenkins knows about GOPATH? [13:16] ah, i see [13:17] you mean the actual script that will be given to jenkins starts at line 20 [13:17] bac: is it possible that the bot doesn't have permissions to access some third party hosts? [13:17] rogpeppe: yes. prior to that is jenkins checking out the revision that it saw had changes. [13:18] rogpeppe: i don't think this machine has any outbound restrictions [13:18] bac: so jenkins sets up GOPATH for your script? [13:18] rogpeppe: with -P 1 godeps succeeded. [13:18] it is very odd it can't find the specified revisions [13:19] bac: ok, that's interesting. i wonder why it would fail when run concurrently. [13:19] rogpeppe: no, that GOPATH biz is done in my script [13:19] bac: so... some stuff before line 20 isn't done by jenkins? [13:20] rogpeppe: are you looking at http://paste.ubuntu.com/7736539/ [13:20] i wonder how two bzr processes running in two entirely separate repositories can interact [13:20] bac: ah no, i was looking at your shell script [13:20] bac: i understand now [13:20] gotcha [13:21] rogpeppe: yay, it worked: http://paste.ubuntu.com/7736744/ [13:22] rogpeppe: i'm a bit concerned that godeps isn't finding the exact revisions specified in dependencies.tsv [13:22] bac: in that latest output? [13:23] rogpeppe: yes, like line 36 [13:23] rogpeppe: all where it says "trying to fetch new version" seem wrong to me [13:24] bac: so, the default behaviour of godeps is to update whatever repository is already in place [13:24] why didn't it find the specified version [13:24] bac: but in this case, the repositories are not downloaded yet [13:24] rogpeppe: i thought the idea was you wanted to peg it to a specific version [13:24] bac: so it tries to update; that fails (because it doesn't exist), then it downloads and tries again [13:24] bac: it does that [13:24] rogpeppe: but why doesn't it get the version i asked for? [13:25] so when it says "trying to fetch newer version" it is really getting the one is specified in dependencies.tsv? [13:25] bac: it's fetching the newest version it can find [13:25] bac: then it updates that to the version it needs (see line 46) [13:26] oh [13:26] bac: the error message is perhaps misleading [13:26] oh [13:26] i understand now [13:27] given that -P 1 fixed it, i think i'll have to make that the default [13:27] a pity really, because it speeds things up a lot [13:28] yeah, it is quite slow. is it just bzr that has issues? [13:38] hmm, maybe there's a bug in my code; some of those errors look highly suspicious [13:59] rogpeppe: hey the 'trap' worked great. i've learned something new and useful today. [13:59] bac: cool [13:59] does anyone know what a url patter of /inspector/service/charm/true corresponds to? i see url generation for it and tests for it but can't actually seem nagivate to anything even *kinda* like that. [14:09] jujugui ^ [14:11] who runs azure.archive.ubuntu.com? [14:17] jrwren: not sure, on some clouds we run servers on their network to provide those close archvies [14:17] jrwren: I'd ask in #cloud-dev perhaps [14:17] jrwren: in that link you posted a MS eng says it was corrected? [14:19] jcsackett: so in theory we're supposed to have /inspector/servicename/charm to open the inspector for a service in the environment and /charm would open the charm details for that popout [14:19] rick_h__: ah, so it *is* meant for this card. [14:20] rick_h__: that was my assumption. [14:20] dispatching details, much like /unit seems to be for unit details. [14:20] jcsackett: right [14:20] huzzah. i'm not actually colliding with anything. [14:20] jcsackett: so the idea is that I should be able to share a link from my environment to you that has the inspector in the sidebar opened to what I'm looking at [14:21] rick_h__: right, that i know--inasmuch as that's dispatching details. i was just curious why the /charm pattern was already part of the url stuff when we weren't doing that yet. [14:22] jcsackett: ah, yea because it was ToBeImplemented [14:22] :) [14:22] * jcsackett laughs [14:29] get back home from vacation...power is out....no biggy.....open fridge....it's warm.... :/ [14:30] oops :( [14:30] guys drilling in fiber drilled through the cable bundle in many peoples backyards on Monday [14:30] so I'm hotspotting yay [14:30] heh [14:31] oh and I have a big above ground cable running to the box in the yard for power [14:34] hah [14:37] jrwren: how goes the bug reproduction steps? [14:38] not great. [14:38] I think I picked 2, effectively not functional, providers. [14:39] imo azure and joyent don't work, or at least don't work with 1.19, I didn't go back to 1.18 to see if that matters. [14:40] jrwren: ok azure we're aware of and there's work going on around that in 1.19 but not in the release [14:40] joyent we need to check for bugs on then and if we can't find them file them [14:40] different issue though, AFAIK. [14:40] i'm not hitting the azure ssh issue, which I see fixed in master. [14:40] jrwren: but you had talked about doing an ec2 test on the bootstrap takeover issue [14:40] i am hitting the azure.archive issue [14:41] jrwren: right, there's a thread right now about not running apt-get upgrade on bootstrap and such [14:41] yup, i've not yet tried ec2. i'll do that [14:41] do you mean http://bugs.launchpad.net/juju-core/+bug/1316185 ? [14:41] <_mup_> Bug #1316185: juju bootstrap hangs in slow environments

[14:42] jrwren: yes [14:42] yeah, that is a different azure issue. [14:43] jrwren: that's the one that bac hit on azure around slow/timeout related bootstrap [14:43] jrwren: ok, want t join the standup hangout early and let's chat out what's up and what we need to file bug-wise? [14:43] sure. [14:50] jujugui call in 10 [14:57] rogpeppe: this is pro forma, but would you look at this trivial PR so i can exercise the landing? https://github.com/juju/charmstore/pull/8 [14:57] rogpeppe: add a +1 if you would [14:58] bac: LGTM [14:59] jujugui call in 2 [15:19] kadams54, where you looking for assets for gui? [15:20] kadams54, I have Spencor on the design team to provide any assets you need [15:23] jrwren: so we're set from our call and have good stuff to move forward on from here correct? [15:23] yup [15:23] jrwren: ok cool, so once we get these bugs filed let's go to ec2 and try to have juju work a few times before you give up on us :) [15:23] ok. [15:24] i don't mind finding these bugs. it just makes juju better [15:24] rick_h__: if a jenkins -merge job fails, can you re-trigger another from github? i added a second :shipit: but it is ignored. [15:24] jrwren: definitely [15:24] bac: youhave to remove the comment from the lander that says " merge job accepted" [15:24] rick_h__: ty [15:24] bac: that's the trigger it uses to make sure it doesn't try to rerun a merge over and over [15:24] rick_h__: is a second marker required? [15:25] doubtful [15:25] bac: no, just remove the marker and it'll retry [15:25] assuming by marker you mean that comment [15:25] no i meant :shipit: [15:25] bac: oh no [15:25] it'll just retry [15:25] cool [15:26] rick_h__: we still have a call in five minutes, right? [15:26] antdillon: so we're looking for something for the uncommitted indicator for service blocks [15:26] bac: yes [15:26] antdillon: in the visuals/etc we've got things for the MV UX, but trying to find one for the service view and the blue circle indicator that I recall having there [15:28] luca: what is the final decision on uncommitted and service blocks? [15:28] luca: we talked about a lot of designs but can't find a final one that matches the grey line for relation. [15:29] luca: and what we need to indicator on the service block. Did we ditch the blue circle? [15:29] rick_h__: no [15:29] rick_h__: sec i’ll find image [15:29] luca: ty, kadams54 is working on it and needs the resrouces [15:29] rogpeppe: are you free for a call in 1min? [15:29] rick_h__: sure [15:29] rogpeppe: k, will link you in a sec [15:30] https://plus.google.com/hangouts/_/canonical.com/jaas-store?authuser=1 rogpeppe [15:30] rick_h__, Sure I just wanted to put kadams54 and Spencer in direct contact to speed up asset creation [15:31] antdillon: rgr thanks === urulama is now known as urulama-away [15:45] Makyo did you have a reviewer for your branch? I just updated on develop forgetting about that bug and now I'm blocked lol [15:45] hatch, need a QA [15:45] ok on it [15:49] Makyo qa ok - but I don't understand the fix....why did destroying the inspector early cause it to not dispatch? [15:49] hatch, It left the state pointing at the temporary id. [15:49] hatch, and we wanted it pointing at the deployed ID. [15:50] ahhh it didn't fire the serviceDeployed event [15:50] got it, thanks, shipit!!! [15:52] Makyo looks like your avatar is no more on github [15:54] What happened? [15:54] Shows up for me.. [15:58] Makyo https://github.com/juju/juju-gui/pull/414 you see it here? [15:58] I had to remove/add mine again when it did this a couple weeks ago to me [15:59] Yep. [16:00] hmm intersting [16:01] Makyo build failed [16:01] looks like gh error [16:02] * Makyo froths [16:02] Mocha timeout. [16:02] Trying again. [16:05] something tells me series: precise machine mixed with cs:trusty/juju-gui-3 won't work. am I right? [16:07] jrwren the gui charm is identical for precise and trusty [16:07] but juju may complain about the tools missing [16:07] :) [16:07] excellent. [16:12] I want to write a android wear app......but Java...... [16:16] Makyo looks like another fail.... [16:16] CI borked? [16:17] Yeah, can't reproduce locally. Not sure what's up. [16:17] rick_h__ any insight into the failures of Makyo's branch? [16:17] mocha is failing to start [16:19] well Makyo I'm just going to pull your branch into mine so I am not blocked and hope for the best with this landing :) [16:20] Alright, sounds good. [16:21] hmm apparently pulling it in creates a merge entry [16:21] not a FF [16:21] Makyo is your branch up to date with develop? [16:23] hatch, looks like there's one branch ahead of it. [16:28] Let me see if updating helps. [16:36] jcsackett so this branch I'm working on requires me to modify the state object to select a tab in the inspector so we are likely going to conflict [16:37] jcsackett so I can put this branch on hold, any ETA on yours? (although it doesn't look like we can actually land anything anyways) [16:40] hatch: looking [16:45] rick_h__ thanks, I'm also going to move onto another card as mine and jcsackett's will conflict pretty badly in a few files [16:46] hatch: wtf, mocha init timeouts? [16:46] right? I have noooo idea [16:46] no idea honestly. Quick google seems to have some version unhappiness [16:46] rick_h__ is mocha instlled? [16:46] installed [16:46] but we've had good branches land [16:46] yea, if it wasn't installed that would be a diff error I'd think [16:49] hatch: quick call please? [16:49] sure [16:49] standup? [16:50] yup there [16:54] * rick_h__ goes to get food for today biab [16:58] bac could you ssh into the ci box and see if mocha is already running? [16:59] bac the process is mocha-phantomjs [17:04] some of the reports seem like there is a mocha plugin missmatch which is causing our issues [17:12] juju-gui doesn't load in safari eh? :) [17:13] jrwren there is an https issue with safari [17:13] you need to pull some strings to make safari show it [17:13] safari may be fast and good on battery usage but just like everything apple if you don't do it their way.... [17:14] it loaded for me, but sits there spinning at Connecting to the Juju environment. must be XHR https issues or something? [17:14] wss [17:14] ah. i see. [17:14] yeah - we have a card/bug to investigate/fix [17:14] that is windows sharepoint services, right? :p [17:15] haha - yeah that'll be the day [17:17] jrwren: you have to give it explicit permission [17:17] jrwren: see the bug in the maint lane with a link to the bug report to make it work [17:18] jrwren: if you're interesting in helping fix that as the bug report for the final mission let me know. It'd be a good one to fix up our charm around that [17:18] https://launchpad.net/bugs/1322596 is the bug report with the instructions for clicking on 'show certificate' first and then it will work/load [17:18] <_mup_> Bug #1322596: deploying to 1.0.2 to live environment fails in safari

[17:19] juju quickstart is so awesome [17:19] even if amazon is taking forever to give me new instances [17:35] rick_h__ whenever you return I'll need some extra insight into this because it works just fine when I create a new instance myself....I'd need ssh access into the ci machine to investigate further [17:56] hatch: rgr, getting ready for our call. We can jump in early [17:57] just grabbing some headphones [18:21] Makyo hey can you do me a favour to see if your CI can land? [18:21] in test/index.html can you increase the timeout to 200,000 from 100,000 [18:22] rick_h__ lol we were so lagged, you left the room and then started talking again [18:23] lol [18:23] yea, gotta love tech until the bandwidth drops down [18:24] yep, using the latency of lte and video conferencing just doesn't quite pan out [18:30] jrwren those are some funky steps for safari :) [18:32] it took me a while to even read them. [18:32] I looked, and said "huh" and looked again, and said "huh" again, and finally read again and understood. [18:33] I don't think showing a dialogue to the user in safari to follow those steps is a very good UX haha [18:34] ha! no. [18:35] rick_h__ do you have a preference for the cards I work on for the next...5 days? I can start on the scale up journey [18:35] hatch: sounds good to me :) [18:37] redir: ? who let him in :P [18:37] * redir hides in the shadows [18:37] kadams54: jcsackett I cancelled out 1-1s because I'm out tomorrow. If you want to chat I've got 1 hr before I EOD and crawl to the couch for a nap [18:38] haha, hey redir [18:38] bac: sorry, same for you as well if you want to chat this week [18:38] hi hatch [18:39] how goes things? [18:39] rick_h__: i think i'm good. [18:39] but just wait until next week....amirite bac? amirite?? [18:39] uh, sure. [18:40] bac: cool, I see successfull charmstore runs woot! [18:40] hatch: busy:) finally home after 3 weeks of travel [18:40] hey rick_h__, i'm keeping my eyes on the hurricane, though. there is a chance we may have to change our travel plans. if we do i'll work friday and swap to next week. [18:40] bac: oh right, I heard about that heading up the east coast [18:40] bac: ok, well stay safe [18:41] redir oh right, all of that vacationing [18:41] and don't tell mramm about that [18:41] :P [18:41] hatch: 2 wseeks vaca one for work [18:41] it is supposed to get to around norfolk friday 8am. we land there friday 2pm [18:41] hi redir. [18:41] bac: /me crosses fingers for you [18:41] hola [18:41] bac: are you set to work on other charmstore cards then tomorrow/monday? [18:42] rick_h__: haven't looked really. but yeah, i'll pick up something [18:42] bac: especially keep an eye with roger on the api v4 stuff [18:42] rt [18:42] bac: ok cool then. Sounds like a plan. [18:42] rick_h__ so the scale up journey card has a bug associated with it for unit scaling....is this just for reference or... [18:43] hatch: yea, it was filed as a bug from luca if I recall [18:43] * rick_h__ looks at the bug [18:44] hatch: oh, well I guess let's make sure while we're doing scale up journey that's fixed :) [18:44] it might be drive-by-able or fixed as part of updating the scale up journey [18:44] sounds good, and do you have reference to the latest scale up journey ui stuff? [18:44] was it just what was in those new inspector mockups? [18:45] wherever those are.... [18:45] looking [18:46] I don't think a lot of these files were put into the drive, just shared... [18:47] https://drive.google.com/a/canonical.com/#folders/0B7XG_QBXNwY1NEtGaHJYZGM4enM [18:47] yea, the shared stuff is in drive [18:47] it's shared from there [18:47] oh, I need to pin the proper directories or someting heh [18:48] yea, search for the juju gui directory [18:48] and in there is a design folder [18:48] and in there are assets/wireframes for the things [18:48] hatch: all else fails bug ant [18:48] and spencer [18:50] Makyo: looks like your branch landed yay [18:55] rick_h__: http://ci.jujugui.org:8080/job/charmstore-merge/ [18:56] big blue dot [18:56] rick_h__: i have had to install jenkins-github-lander on the slave but i have not made it self-updating yet. does it really need to be? what about just a daily cron that does a pull? [19:00] bac: to update the jenkins-github-lander tool itself? [19:00] rick_h__: yeah [19:00] bac: to date I've just manually updated those whenever we landed an update to the lander tool [19:00] so I'm fine either way [19:00] rick_h__: ok. just need to be aware it lives on the slave too [19:00] fortunately it's a quick manual step we could automate with juju run [19:00] bac: rgr, thanks for the heads up [19:00] you and 'juju run' [19:01] well I got thinking a lot of the stuff we're doing could be scripted [19:01] especially like this, since we can target the same command at many services/units [19:01] rick_h__: enabling future go projects should be much easier now... [19:01] did a bit of face-palm that I didn't think of doing that earlier [19:02] bac: awesome, thanks for fighting that down into shape so we've got some good problems solved for us going forward [19:03] rick_h__: i've created a bundle, and a branch with all of the required bits including the current .jenv file. i'll push to a private location at ~yellow shortly. [19:03] bac: awesome [19:05] hatch, https://github.com/juju/juju-gui/pull/416 [19:06] Makyo thanks :) [19:06] rick_h__ was the "please test this" something added int he latest iteration of ci? [19:07] Makyo hopefully this works heh, although I can't believe a 100s timeout is not enough lol [19:07] hatch: hmm, yea it should just be auto running [19:07] ahh ok then yeah something broken :) [19:07] it doesn't auto-run any longer it doesn't look like [19:07] hatch: hmm, did you trigger the current running test run? [19:08] nope, Makyo is landing his latest branch [19:08] oh hmm, maybe that triggered it [19:08] but that should be the merge [19:09] hatch: right [19:11] rick_h__: naw, i'm good. [19:14] jcsackett: ok cool [19:14] hatch: flipped a couple of switches. Will see if it runs test runs [19:14] sounds good thanks [19:23] rogpeppe: juju/charmstore now has jenkins landing. like our other projects, after approval use :shipit: to trigger a final test and landing. [19:27] bac: he's EOD, can you send an email to peeps maybe? [19:29] wow, didn't realize how popular a name uros was until looking in lp for uros [19:42] jujugui I'm out. Time to get the camper loaded up and ready for the trip. I'll be around via email or hangout if you need anything. [19:42] have a good holiday rick_h__ [19:42] I'll try to not break anything..... [19:42] try [19:42] woot woot [19:42] :P [19:42] :D [19:42] keep an eye on that shifty jrwren character :P [19:43] and make sure to help out with any juju fun [19:43] you bet [19:47] imma see what I can break. [19:49] plz don't let it be ci again......plz not ci :P [19:51] !np [19:52] i'm really NOT going to see about breaking anything. With my luck lately, things will just happen. [20:17] hah === tvansteenburgh1 is now known as tvansteenburgh [20:37] oh blarg this viewlet manager [20:52] good evening gui-peeps [20:53] cya bac [20:57] jcsackett ok I'm totally with you on this viewlet manager thing [20:57] er rick_h__ ^ [20:57] it's gota go [20:57] gotta [21:06] hatch: \o/ [21:07] not that we have time to rearchitect the whole thing right now, but i'd say bye-bye viewlet manager is right up there with bye-bye slots. [21:08] yeah, the problem that it solves still exists but it needs some work so that there can be nested views [21:21] hatch: yea, it was an interesting idea to managing chunks of UI but bit off a bit more than it can chew at this point. Frameworks and patterns that last are darn hard. [21:22] yeah I'm not sure how this is going to work properly without fixing viewlet manager.... [21:22] working on it [21:22] databinding will just not work for a while :) [21:22] :/ we just can't rip the thing out at this point though. That's a bigger chunk of work and should have more input from devs gone. [21:23] oh yeah no definitely can't go [21:23] it'll be a ton of work to get rid of it [21:23] I'm hoping we can fix it [21:23] hatch: if it's going to be that bad then put the card back and work on the other ones until Makyo and I get back and we can get together and put more brains at it [21:23] hatch: I think it'll take all our vision on it to be honest. There's got to be a path in there somewhere [21:24] hatch: but we can work on other bits. The destroy stuff, the config changed stuff that should be unblocked at this point right? [21:24]