[00:13] huwshimi: morning, how goes? [00:25] rick_h_: Hey! Good thanks. Yourself? [00:26] huwshimi: party party, LUG meeting [00:27] huwshimi: I wanted to check in on your test stuff. I know I wasn't amazing helpful with Y.later, while it's in the code base I've not used it much so wasn't sure [00:27] huwshimi: but I know I had looked at how to do testing around the animations and had those couple of ideas for you, not sure if they helped? [00:27] rick_h_: Sounds fun! [00:28] rick_h_: The main problem is that it messes with other tests. I think I'll just go with an extra check to see if the node exists. I know that will only come into play in the tests, but I don't have a better idea :) [11:58] frankban: thanks for the notes. i'll put that machiner file stuff back. No reason to pull it especially in a pyjuju specific branch [11:58] rick_h_: +1 [12:00] rick_h_: do you have a chance later to review + intensively QA my branch? [12:00] frankban: sure thing. [12:00] rick_h_: thanks! [13:30] benji: good news! we now have two stuck baskets. progress. [13:30] heh [13:51] gooood morning [13:53] mornnig [13:53] or something [13:53] haha, hows it going?\ [13:53] well, ec2 hates me but other than that...ok [13:55] boo ec2 [14:06] frankban: I've definitely done something that's taken the test run time waaaay up. the test_deployer test runs forever compared to trunk [14:09] rick_h_: looking again at your new mp [14:09] rick_h_: is the deployed gui slow as well? [14:10] I'll try it again. I did a manually make deploy once to make sure it was working but didn't time it [14:10] what should a make deploy run time be about? [14:11] rick_h_: https://codereview.appspot.com/52440043/diff/1/tests/20-functional.test line 61 on the right: force_machine should be 0 [14:11] not None [14:11] rick_h_: so the functional tests are creating new machines, that takes lots of time [14:11] ahhhhh! [14:12] rick_h_: I think you can safely delete that line and just pass force_machine=0 to juju_deploy() [14:12] bah, bad if replacement there. Gotcha [14:12] rgr [14:12] cool [14:13] thanks, I hated that everything seemed fine but didn't want to land it with these long test runs [14:13] will do one more test run and will reduced the test time back to 40min if all goes well [14:13] rick_h_: +1 [14:13] jujugui afk for a sec while I go pick up new glasses. [14:16] rick_h_, google glasses ? :-) [14:38] hazmat: hah, not quite. More like "old-man glasses" [14:38] hazmat: I'm more a watch guy, waiting for that steel to ship :) [14:40] man, new glasses adjust period...begin! [14:41] shouldn't take toooo long [14:42] hazmat interesting bug report re a node api module [14:42] lady is telling me that "you had a 15 degree axis of rotation change that'll take some getting used to". Not sure wtf that means but it's strange. [14:42] hatch: +1 on that as well :) [14:43] it's quite a bit of work beyond face value because we actually use a modified api because of the guiserveer [14:43] but I can see that being fixed [14:43] right [14:43] but overall it's an awesome idea [14:43] agreed [14:43] <3 api libs [14:43] we'll just add that to the list of things to do.....lol [14:45] rick_h_ I'm going to review huw's branch just fyi [15:08] jujugui: i've stopped charm/basket ingestion on staging.jujucharms.com so i can analyze the existing failure state without looking at a moving target. let me know if that causes you any problems. [15:08] bac: rgr [15:08] rick_h_: so you've got astigmatism and it's rotating? [15:09] bac: I have no idea. I've always had just a light correction, -.25 and .5 or something [15:09] but this year one eye went to a full 1.0 [15:10] and never heard of the axis stuff until today [15:11] rick_h_, benji: it looks like the stuck baskets are caused by entries in basket-queue that are locked and have been there for a while. my theory is that they were locked when the process was abruptly stopped for some reason, though i cannot prove it. Q: any reason to not delete the incoming charm and bundle queues at start up? [15:11] rick_h_ I asked, it's possible but as the delta increases the likelyhood the dr screwed up does too :) [15:12] rick_h_, pebble steel is ships in two weeks. https://getpebble.com/steel [15:12] bac: hmmm, the worry is that something falls through the cracks [15:12] rick_h_: i think the axis is for astigmatism. other problems (near/far sight) are not rotationally sensitive but astigmatism is. [15:12] bac: I can't work up a case where it could atm though [15:12] hazmat: yep, order placed...just now impatient [15:13] jujugui: speaking of wrist-based gadgets, anyone interested in a slightly used fitbit? the original, not the one with the time display. [15:14] bac: hah, I can't give mine away [15:14] oh, really? [15:14] it's on its second person [15:14] yea, gave it to a friend that tried to give it back and made them give it to a friend [15:14] bac, rick_h_ i'd be interested re fitbit.. you mean the original pod or the wrist thingy? [15:14] i like it, but it was made obsolete by the motion chip on the new iphone [15:14] hazmat: original wrist [15:15] hmm.. pass.. doesn't work with a standing desk very well. [15:15] hazmat: the flex [15:16] rick_h_: i did a comparison to fitbit vs phone. and the logs were almost identical. not worth having an extra doohickey if i'm always going to have the phone in my pocket. [15:17] bac: yea, I got tried of having two 'watches' and it was so inaccurate [15:17] once told me I took 680 steps in a shower [15:17] that and does't work with my phone (though they did just update it this past week) so I got rid of it [15:18] ha. yeah, hammering confuses it too [15:30] is gary out today? [15:31] hazmat till friday [15:31] anything I can help with? [15:39] brb in 15, relocating [15:50] jujugui call in 10 [15:53] jujugui call in 7 [15:58] jujugui call in 2 [15:58] call in 1:48 [15:59] :T [16:07] rick_h_: come back! [16:07] benji: oh righ [16:07] right [16:20] 119kph wind right now [16:21] yow! [16:26] ooookk well the 119kph wind gusts have knocked out the power [16:27] well...knocking out... [16:56] yay power [17:03] yay [17:03] charge those batteries [17:10] you bet, all plugged in now [17:57] it's so quiet [17:59] ssshhhh [18:00] I'm trying to juju set juju-gui juju-gui-source="https://github.com/juju/juju-gui.git network-prototype" [18:03] :) [18:06] hmm, well it errors but it loaded the source [18:06] https://ec2-50-16-67-171.compute-1.amazonaws.com [18:06] hatch: should this work? [18:06] looking [18:06] getting juju.views.NetworkListView is not defined [18:07] rick_h_ it's possible that the file which has that view is not getting rolled up [18:07] k, will call that a "not my problem" problem atm [18:07] but it worked bwuhahahaha [18:09] I believe it was only ever used with make devel [18:09] yeah I'd say it's a success for the deployment system though :) [18:09] https://ec2-50-16-67-171.compute-1.amazonaws.com/juju-ui/version.js is good so I'll roll with it [18:12] rick_h_ maybe the source should be two values [18:12] so the repo as one and the branch as the other [18:12] and maybe taking it one step further, the hash [18:12] so that we can 'update' something to a new source version [18:15] hatch: yea, I started down that path but it's kind of a pita [18:15] going with space to specify a branch/sha [18:15] and @123hfehw for a sha [18:15] so right now I'm testing juju set juju-gui juju-gui-source="https://github.com/juju/juju-gui.git @c7ddd186" [18:15] yeah I bet - I'm just thinking of the story where we want to use a certain branch but then keep it updated when new changes are pushed to it [18:16] do you have to use the @? [18:16] yes, because it's the diff between a checkout and a checkout -b [18:17] can you not go `git checkout mybranch` and it'll just checkout the latest revno? [18:17] not if the branch doens't exist yet :) [18:17] you need the -b [18:18] if you have a local branch with that name, sure [18:18] you can't do git clone url branch [18:18] you have to git clone url && git co -bt branch [18:19] wel you CAN but that doesn't actually solve the issue of requiring some indicator that it's a hash [18:20] huh? [18:20] git clone -b brancho --single-branch [18:22] bah, this sucks. because we did a shallow clone you can't checkout a hash anyway. [18:22] have to unshallow it first [18:22] * rick_h_ checks git version on precise [18:23] hmm [18:23] it seems like there should be a way to pull down said hash [18:24] working on it since git fetch --unshallow is only in trusty :( [18:28] rick_h_ could we not say they need to specify a repo and branch name then a hash? [18:29] then we could use the single branch checkout [18:29] which should be fast....I think [18:29] but a hash has nothing to do with a branch. It's just a single git commit in a detached head state [18:29] right but if we already have the single branch pulled down we will have that hash [18:29] it's not ideal, but in order to support the widest range of systems [18:31] I don't know if this idea will work just thinking of alternatives :) [18:32] yea, sorry testing out here. --depth=20000 gets me caught up and I can git co hash [18:32] I know we *can* do a lot of things. Just trying to cover initial bases atm [18:32] been on this card for a week and trying to support everything git does will drive me nuts [18:33] well we don't need to support everything. The BIG thing is a branch [18:33] right, and branch works [18:33] that's the primary goal, I think a hash is a 'nice to have' [18:33] now, I can see wanting to go back to an old commit in that branch to see "does xxx work" [18:33] yeah for sure [18:33] or move forward because you have updated that branch [18:34] but even that can be done by the dev with "git checkout hash && git co -b test-feature && git push origin test-feature" [18:34] juju needs a 'trigger' config option :) [18:58] dimitern are you still around? [19:33] rick_h_: I think I tracked down the cause of https://bugs.launchpad.net/charmworld/+bug/1263120. Let me know when you have a minute to chat about it. Thanks. [19:33] <_mup_> Bug #1263120: self related services in a bundle fail proof [19:34] benji: sure thing, your hangout or mine? [19:34] rick_h_: I'll make one. [19:35] rick_h_: https://plus.google.com/hangouts/_/7acpi7cc7l9l3r0bdfembikccg?hl=en [19:45] * Makyo runs out quickly to grab MiniDP->VGA adapter. Kind of important for presentations D: [20:22] yay, me, i'm sort of back [20:23] lies! [20:27] hah the power keeps dropping out here [20:27] it's like there is a switch that gets tripped when the power lines get stretched to far lol [20:42] hatch i think my ISP is on your power grid. not sure how that happened. [20:43] haha - ours is usually really stable, I guess 100+km winds are around it's limit [20:44] bac https://twitter.com/nheagy/status/423513314446962688 [20:44] hatch: you should use that three weeks in the summer when the ground isn't frozen to bury those lines [20:45] haha, they are except for the major powerlines...or something like that [20:45] wow, that's a twisty mess [20:45] heh yup [20:46] yxe? [20:46] Saskatoon [20:46] airport code [20:47] umm, ok. No idea wtf the y and x and e came from, but I'll take your word for it [20:49] rick_h_ well the Canadian airports start with y. But I have no idea how they decide where the other 2 come from [20:49] y are you going to canada? [20:50] lol [20:50] it's X-tremely E-...fail [20:50] E-lectric [20:51] I thought it was rather un-electric currently :P [20:52] lol [21:21] jujugui: anon [21:21] c-ya [21:28] So, juju-quickstart [21:29] this is awesome! [21:29] :) [21:31] it deploys the gui by default too? [21:31] oh man [21:31] yea, on node 0, pre-packaged so it installs super fast [21:31] should auto log you in with your admin secret as well [21:31] at least that feature landed, not sure on released/what version you're using [21:32] I'm running trusty, but i digress - this is good stuff [21:32] :) [21:33] we'll let frankban know you like his baby [21:34] i file bugs against juju-quickstart right? [21:34] sure [21:34] Ah, probably better to put it in the merge request [21:35] lazypower: are you testing out the branch I sent jcastro ? [21:35] indeed [21:35] When the local machine takes longer than the deployer expects to come online it fails the quickstart [21:35] lazypower: ah, then yea. Please put them actually over in the code review page link. Look for the 'codereview.reitveld' link in the MP [21:35] oh, that's a bug and different that the MP [21:36] file that as a real bug then please [21:36] mainly I wanted to get some others to help qa that interactive ncurses mode stuff [21:36] but yea, it'll hit 1.0 in this branch so want to make sure it's solid pre-1.0 and that seems like a bug to hit. [21:37] no problem i'll keep working through the ncurses stuff, i've got creds to do it. [21:37] cool, appreciate it [21:41] dude it works with the mouse too? [21:41] yep [21:41] i'm a fan [21:41] ship it [21:41] good thing i'm not your final quality gate [21:47] rick_h_, i've configured everything but openstack. Looks solid to me [21:47] thumbs up [21:48] lazypower: cool, thanks === rogpeppe2 is now known as rogpeppe [21:55] *note the form validation was an excellent touch [21:55] rick_h_: it's related to this, https://bugs.launchpad.net/juju-core/+bug/1269519 kapil knows about it [21:55] <_mup_> Bug #1269519: Error on allwatcher api

[21:55] if you guys find a solution to it, please let me know [21:56] marcoceppi: k [22:11] Morning [22:19] mornin huwshimi [22:20] hatch: Hey. Thanks for the review! [22:20] np, hope my comments make sense, I'm also open to pushback on the widget vs view thing :) [22:21] hatch: It all makes sense and seems straight forward [22:22] excellent [22:30] hatch: Are you all settled into home timezone yet? [22:30] huwshimi when I came back I actually went snowboarding in another timezone and just got back late last night haha [22:30] hatch: hah! [22:30] 3 timezones in 1.5weeks [22:30] too much [22:30] ouch [22:37] is there a GUI bug about making multiple relations to the same service? [22:37] I can't seem to find it [22:37] my search foo might be failing [22:58] hatch, ping. [22:58] hey [22:59] I think my current card is related to double-dispatch and removing full-screen, but having a hard time digging into it, because it only shows up in prod. [22:59] I don't want to leave it hanging, and am getting nowhere. [22:59] Interested? [22:59] Need to review frankban's branch before EoD/vacation. [23:00] yeah I can pick it up [23:01] hatch, cheers. I've tracked it down to somewhere in subapp/browser/browser.js, one of the state functions (it inits, but doesn't complete) [23:02] alright I'll pick it up first thing