=== mattuk1972_ is now known as mattuk1972 [13:14] bcsaller: I had a couple questions about your review of https://codereview.appspot.com/6909046; I put them in comments === benji___ is now known as benji [13:34] gary_poster: ping [13:34] yo, BradCrittenden === BradCrittenden is now known as bac [13:34] gary_poster: could you have a look at my review at your leisure? [13:35] sure bac, I'll do it now. [13:49] gary_poster: leave added to c.a. [13:49] bac I just gave you an approve but I meant to wait to approve until after I tried it out [13:49] ok will look bac [13:49] gary_poster: i'll wait [13:50] just let me know when the branch is ok [13:50] ok [13:50] bac, I accepted Dec 19 [13:51] ok [13:51] thanks [14:01] I am going to assume that bcsaller is not in this morning and that we can clear up any remaining issues with his review of https://codereview.appspot.com/6909046/ later and land my branch. [14:02] benji: I'm here, but its fine if you want to land it, nothing would prevent that [14:03] bcsaller: oh, did you see my message at 13:14 (UTC) [14:03] yeah, just got in, in time for the second message [14:04] benji: the jist was I thought we either needed more detail to track revno's in CHANGES or less and just record the current revno in a single file w/o the need to parse it, but I don't feel strongly [14:05] to what use would we put the current revno stored in a file? [14:06] benji: to what use is the description? I think they both honor the same intent, to help recover a significant point in time [14:06] oh, the current on [14:06] only [14:06] can't type today [14:06] umm, that would be the only thing that makes it onto the version page [14:07] w/o the need to grep and so on [14:07] the description is so end users can see what (materially) changed between releases [14:07] benji: and I was thinking in that case that being able to pull the associated revno would help in testing vs a particular revision [14:08] they are tearing down the scaffolding on the building across the street today, its pretty distracting, sorry [14:09] bcsaller: I think tags would be an easier way to handle that; we can accosiate the release version number with a particular revision directly and not have to worry about keeping up with revnos (and the race involved in doing that) [14:09] benji: agreed, if we switch to that its better all around [14:10] cook, I will make a card to "Add instructions on tagging releases to the release procedure" [14:10] s/k/l [14:19] hi all, I'm trying to make tests run under the production environment, per #1087107 [14:19] <_mup_> Bug #1087107: Tests should be runnable on the production build of the code < https://launchpad.net/bugs/1087107 > [14:19] there are 78 failures, apparently caused by incorrect test setup [14:20] I need some help from someone who's familiar with the way tests run, and what they need [14:21] w/o knowing any details I suspect that the deps are not being included in the proper order so some modules fail to resolve [14:22] that should show up in the dev console though [14:24] bcsaller, when I run "make test-prod" there are no errors in the Makefile output, and then none in the node server log [14:24] teknico: no, it would be in the browsers console [14:24] except for the usual 404 on favicon.ico (which is weird), and another one on /juju-ui/assets/slider-base/assets/skins/sam/thumb-x.png [14:24] oh, ok [14:25] if I'm correct its not that the work failed, its that it failed to do the right work [14:26] bcsaller, yes, there are hundreds of messages in the browser console, and the failures in the page itself also seem to hint to problems with test setup [14:27] teknico: yeah, I think the dep ordering was never right in prod, and thats what we needed to test was working properly [14:28] bcsaller, oh well, at least we have failing tests now :-) [14:28] indeed [14:29] I wonder whether I can land a branch with 78 test failures, though :-) [14:37] gary_poster: for some reason my emacs config went nuts recently wrt to tabs. i've tamed it now. [14:37] cool bac [14:38] gary_poster, hi, so you're there? you look away [14:38] I am here teknico [14:39] I see holiday requersts :-) [14:39] gary_poster, yes, all the end of year stuff [14:39] gary_poster, I'm kind of stuck wrt to ^^, what do you suggest? [14:40] gary_poster, namely, I enabled tests on the production environment, and there are 78 failures [14:40] they run though :-) shall I land that branch? [14:40] bcsaller or hazmat, goodspud asks "Can the GUI charm automatically run a script to launch the GUI in a browser?" My answer would be "AFAIK it would be difficult to do this productively, and quite surprising to a juju user." Would you agree? [14:41] gary_poster, we can do a jitsu command.. jitsu gui which does the write thing [14:41] right even.. [14:41] teknico, right, I saw that. I think you could land it, because our existing test command still passes, right? I wouldn't mark the bug as fixed [14:42] hazmat, yes. goodspud replied to your comment about that in this way: "Tom Haddon (IS) has indicated they won't be using Jitsu because they can't rely on any of it's features being supported" [14:42] gary_poster, yes, "make test" passes (equals to "make test-debug") [14:42] hazmat (https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0AlIM--9A1RkTdFFyZ0JHZ2R2LTVoZFhidHpvYldIc1E&disco=AAAAAFJ5Z1g&pli=1#gid=0) [14:43] teknico, yeah, I think it is OK to land, but the bug won't be fixed until we actually have the tests passing. [14:43] gary_poster, well, the bug says "Tests should be *runnable* on the production build of the code", not "Tests should *pass* on..." :-) [14:44] teknico, :-P [14:44] teknico, next you will be sending me a printout of the production build of the code, with you running on top of them and holding a printout of the tests [14:45] gary_poster, :-D [14:45] gary_poster, isn't the idea to compartimentalize things? separation of concerns? divide and conquer? :-) [14:45] gary_poster, seriously, I do think it's a valuable distinction, even if it was not completely intended :-) [14:45] gary_poster, but ok, I'll leave the bug open [14:46] teknico, compartmentalize, yes! that's why I said it is good to land. I'm also ok with you filing a separate bug for the passing and marking this bug fixed. It brings some value--the value to work on the next bug. ;-) [14:47] gary_poster, ok, I'll do that then [14:48] cool [14:50] gary_poster, hazmat - I'm trying to find an easy way to launch the GUI and thought automating that process would be the nicer experience, but open to thoughts and suggestions. Tom said they couldn't rely on features in jitsu always being supported so they would find other ways to do things that it could do [14:53] ack goodspud. we could provide the jitsu command as a standalone script, I guess, but that is a first step towards recreating jitsu [14:53] goodspud, there will likely be juju implementation agnostic subset of jitsu supported in the future. but in the abesnce of that, .. introducing a new command to juju-core is possible, but no guarantees till core signs off and gives a timeline. [14:53] possibly a juju open unit_name command [14:54] or juju go unit_name [14:54] teknico, approved all your holidays, thank you. You get Jan 1 off too, right? That was not in your approval list FWIW [14:55] gary_poster, oh right, it's a national holiday, I'll input that too [14:55] thanks [15:19] hazmat, gary_poster - are the juju-core team open to bribes? :) I do quite like the "sound" of "juju go..." [15:19] :-) [15:19] I haven't actually heard of that approach tried yet... [15:22] bac, are you free after today's meeting to talk through the process of setting up juju/juju-gui on machines here with an up to date branch so we can help with testing and feedback? [15:23] goodspud: sure [15:23] Awesome [15:23] goodspud: how long until your EOD? [15:23] 2.5 hours, although I have a meeting in my diary for the last hour of that [15:24] Happy to talk now if you are free too [15:24] goodspud: ok. i may take 15 minutes after our meeting to grab a bite to eat before chatting with you. would that be ok? [15:24] bac, certainly would be [15:27] goodspud, sent an email to juju-dev requesting it.. tbd [15:29] hazmat, cheers! [15:30] Call now? [15:30] i'm there! [15:31] bah sorry [15:31] bac bcsaller benji goodspud hazmat Makyo teknico call now [15:31] let's all jump in GvR's time machine! [15:35] bac, can you do ssh-import-id hazmat [15:36] bac, on uistage [15:36] hazmat: done [15:37] bac, thanks [15:40] benji, can you please move your mic out of your breathing flow? thanks :-) [15:41] teknico: thanks for the heads-up; the world needs a solution to open mics on video calls [15:41] benji, indeed, as long as it is not from apple :-P [15:41] heh [15:42] I will be buying one of these at some point: http://whirlwindusa.com/catalog/black-boxes-effects-and-dis/miclineinst-switchers/mic-mute [15:43] goodspud: do a G+ hangout at the top of the hour? [15:43] benji, wow, pro footswitch :-) [15:43] bac. yes. [15:45] the PP model looks good; but what I would *really* like is a hybrid with a push-to-talk button for large groups and a latching switch for times when you are doing one-on-one talks with people [15:45] benji, nice, it's also latching, just like the switch on your average $20 chinese headset ;-) [15:46] benji, yes, using the headset latching switch as push-to-talk takes some state-managing attention :-) [15:46] yep :) [15:48] bac, the uistage stuff should be fast again with nginx i suspect [15:49] python simple http server was never meant for this use case imo [15:49] its pretty slow [15:51] bcsaller: where is the review for 1077067? No review is linked from the card and no branch is linked to the bug. [15:51] benji: https://codereview.appspot.com/6867072/ [15:51] thanks [15:54] bac, should be faster now.. just switched simplehttpserver for nginx [16:01] goodspud: hi [16:01] bac, 2 secs.... [16:02] goodspud: ok, spinning up the hangout [16:02] goodspud: https://plus.google.com/hangouts/_/2917c88c62c0938e080426c3ee1a2280dea330c1?authuser=0&hl=en === mattuk1972_ is now known as mattuk1972 [16:06] goodspud: you went away [16:16] goodspud: you went away, again [16:19] goodspud: hello...london...you there? [16:24] bac. sorry about that - computer said "no" and crashed. [16:24] oy [16:24] goodspud: https://plus.google.com/hangouts/_/2917c88c62c0938e080426c3ee1a2280dea330c1?authuser=0&hl=en [16:26] bac, hmmm, now you are invisible [16:31] goodspud: still there? [16:47] bac, apologies for the interruption. VMware keeps crashing my machine. I'll get that sorted then come back to you another time to get things up and running [16:48] goodspud: ok. you may be able to make progress on your own following the README [16:48] bac, I'll go through the instructions and let you know if I have any issues [16:48] goodspud: just let me know when you'd like to try again. tomorrow would be good sometime after your lunch [16:49] bac, will do [17:14] teknico, bac and I are trying to quickly review your branch so you can land before your EoD [17:14] gary_poster, great, thanks [17:20] gary_poster, re: the standard review markers we recently discussed, are they documented somewhere? I don't remember what they are :-) [17:20] teknico: done [17:21] bac, gary_poster, applying and landing, thanks [17:22] teknico, we should document them in docs/process.rst . They are currently recorded in https://canonical.leankitkanban.com/Boards/View/102529849 . I am on the hook for writing them up on a blog post or similar. [17:25] gary_poster, that's the kanban board, where within it? :-) [17:26] gary_poster: lp2kanban is currently down [17:30] Makyo, I wanted to list myself as reviewer of https://codereview.appspot.com/6905059/ , but cannot find the card [17:30] teknico, It's 'build-prod overwrites juju_config' [17:31] oh wow, that one, right there :-) [17:33] I figured a puzzle in the name of the bug would help...uh...something. Heh [17:42] teknico, sorry https://canonical.leankitkanban.com/Boards/View/102529849/103215444 [17:42] bac, ok. you working on it? [17:43] otherwise we can get someone soon [17:44] gary_poster: i don't understand. are you talking about teknico's branch? i approved it a while back [17:44] 13:20 to be exact [17:45] gary_poster: oh, you mean lp2kanban? [17:45] yes, i'm working on it after i get my branch landed. [17:45] bac, lp2kanban, cool thx [17:49] teknico, gary_poster, so uglify2 supports source maps.. ie. it could enable debug prod story [17:49] http://lisperator.net/blog/uglifyjs-v2-news/ [17:50] hazmat, does it have source maps for both js and less/css? [17:50] teknico, just js i'd imagine [17:52] hazmat, here's a nice survey of available source maps tools: https://github.com/ryanseddon/source-map/wiki/Source-maps:-languages,-tools-and-other-info [17:53] teknico, cool [17:54] hazmat, gary_poster, in order to do QA on uistage, how do I check the currently deployed revision? [17:57] teknico: not a satisfying answer but it is 274 [17:57] bac, yeah, thanks for the fish anyway :-) [17:58] teknico: is that the rev you wanted? [17:58] bac, yes, my branch is on 273 [17:58] excellent [18:04] possibly we could do some for of bzr var properties for revno. needs a plugin i assume [18:04] ala cvs or svn rev id substitution in a file [18:05] not sure if such a thing exists [18:05] we could just add it to the makefile to generate [18:13] teknico, hazmat, I think something exists now, thanks to benji's make dist branch. looking... [18:17] teknico, hazmat, benji, there's a build/juju-ui/version.js target which would include what teknico wants, but benji, it is no longer hooked up to anything. I think the base build ought to make that file and hook it into both debug and prod [18:18] hazmat, uglify2/source maps: sounds great [18:58] * benji returns from lunch and reads the backlog. [18:59] I guess we need tests for the Makefile [19:29] gary_poster: the green kanban board is causing an exception because there is no default card type. where is that set? [19:30] BradCrittenden, click the gear on the top right to configure the board. Click "Card Types" in the new window. look in the right column for a "default" checkbox === BradCrittenden is now known as bac [19:31] gary_poster: i don't see 'card types'. perhaps i don't have permission [19:31] bac on our board or theirs? [19:31] ours [19:31] oh, weird. looking [19:31] you should have the same powers as I... [19:32] gary_poster: what is our default card type set to? [19:33] bac, we are both Managers. Want to hang out on juju-ui for a sec and screenshare? our default type is "Task" (yellow) [19:33] gary_poster: sure [19:34] gary_poster: you making a hangout? [19:34] bac, juju-ui [20:48] * bac dogwalk [22:06] * Makyo dogwalk, also. [23:27] gary_poster, hazmat - if you have any comments or thoughts on the visuals for the login screen, add them to the image and I'll get Matt C to work through them in the morning. If all is good then I'll get him to provide assets