[00:02] hi huwshimi [00:03] bac: Hey [00:14] happy new year huw. hope all is well in taz [00:18] bac: Thanks Brad. Things are good. How's life in Puerto Rico? [11:58] frankban: morning, if you get time this morning have a couple of questions on my branch to remove the pyjuju stuff https://codereview.appspot.com/51620043 [12:03] rick_h_: lunching in 30 mins, we can chat now if you want (or after lunch) [12:03] frankban: can wait until after. I've still got one of the functional tests dying on me on ec2 I'm debugging as well [12:03] rick_h_: sounds good, ok [13:01] jujugui useful link on git foo http://sethrobertson.github.io/GitFixUm/fixup.html [13:02] I like the choose-your-own-adventure structure. [13:02] yea, it's a cool approach [13:13] benji: ha [13:34] uhm... subscribing the free trial for azure also means receiving a phone call from ms people wanting to help you... [13:35] lol [13:41] heh [13:47] benji: every time i have to link a branch to a bug report in LP and it suggest correctly, i think of you fondly. [13:48] s/suggest/suggests/ [13:48] heh [13:48] it is my favoritest feature [13:48] That is probably my single biggest contribution to all of LP. [13:48] well, it is the most strongly identifiable [13:49] and completely unsolicited, if i recall [13:51] frankban: did the MS people speak italian? [13:51] bac: yes [13:51] rick_h_: you going to be able to do that review this morning? [13:52] bac: loading it up now [13:52] ty [13:58] bac: feedback inbound [13:58] rick_h_: thanks [14:00] rick_h_: so you think the inheritance work-around is sane? if so i'll add some tests. [14:01] bac: the work around? everything looks sane there. The monkey patch on charm tools is a bit :( but since we lock our version of charm tools it hopefully won't bite us [14:02] rick_h_: i think it is less gross than maintaining our own fork. [14:02] yea :( is all [14:19] frankban: how's the day looking? [14:20] rick_h_: I spent some time debugging a weird juju-core error, we can chat in 10 if you are available [14:20] frankban: cool thanks [14:28] and the GUI runs in all its glory on azure: https://juju-azure-wy2xql47ef.cloudapp.net/ [14:29] https://juju-azure-wypgw0upb1.cloudapp.net/ :) is the quickstart gui on azure that runs CI [14:32] :-) [14:42] rick_h_: ready when you are [14:44] frankban: https://plus.google.com/hangouts/_/76cpjpka9o5dhd2u0ami5rqbl8?authuser=1&hl=en [15:25] frankban: yay tests ran in 62min [15:26] rick_h_: weird, they are still running here [15:27] well just happy they pass and the changes don't break things [15:27] except I introduced some sort of time warp into the tests I guess :/ [15:44] frankban: updated the MP and the agent/zookeeper stuff. Care to review? I'm going ot walk away for a few and then do a 'fresh eye' run through. https://codereview.appspot.com/51470044/ [15:45] rick_h_: I'll take a look in a few minutes [15:45] guihelp I didn't give lbox enough permissions when I auth'd it to the launchpad api and now I can't update my MPs. I don't see the token in my oauth tokens under my lp account. Any ideas on how to revoke/reset that? [15:46] rick_h_: what's the url for looking at lp tokens? [15:46] i'll see if mine is listed [15:46] https://launchpad.net/~rharding/+oauth-tokens [15:46] https://launchpad.net/~makyo/+oauth-tokens ish bac [15:46] I see a few, but I just set this up this week and no tokens from 2014 listed [15:46] duh, lbox should have stored it. Should be able to remove it there [15:47] rick_h_: is it using a system-wide token? can you revoke it and have it prompt again? [15:47] bac: I'm trying to see where it would have stored it. [15:48] rick_h_: keychain [15:48] or seahorse or whatever we call it [15:49] oh bah [15:50] hmm, nothing in seahorse [15:50] jujugui call in 10 [15:51] rick_h_: i don' t recall specifically authorizing lbox, so i suspect it is using the system-wide auth, at least on machine [15:51] bac: yea, that makes sense [15:52] are those listed in LP somewhere else then to revoke? Seems I should have an oauth token in LP for this host [15:53] rick_h_: my test run has been killed by timeout [15:54] frankban: at he 40min? [15:54] frankban: or the 80min? [15:54] rick_h_: 40m [15:54] yea, /me should do a run on trunk to make sure my branch isn't doing it [15:54] I'll go set that up to run [15:55] cool thanks [15:56] rick_h_, ~/.lbox afaicr [15:58] rick_h_, actually ~/.lpad_oauth [15:59] oops, jujugui call in 1 [16:00] benji: yo [16:05] benji: soon the kids won't get that clock thing [16:05] heh, indeed [16:06] well, I get "you can lead a horse to water..." but I have never lead a horse anywhere [16:08] frankban: heh, first try to run tests hit the 5min timeout to bootstrap :/ [16:08] frankban: I must be in the wrong ec2 zone right [16:08] * rick_h_ keeps hammering ec2 hoping for different results...insanity coming [16:08] rick_h_: :-/ [16:09] benji: you can lead a person with a question to google.com, but cannot make them search? [16:09] letmeleadthathorsetowaterforyou.com [16:09] lol [16:39] frankban: ok, got it to run on trunk in 30.0 minutes that time [16:39] frankban: so maybe I did do something that's killing functional test run time [16:39] rick_h_: or maybe now ec2 works properly :-/ [16:39] frankban: yea, rerunning the tests in my branch one more time to see [16:54] Can we update comingsoon? I'm looking at a card I can't reproduce locally. [16:55] Er, let me rephrase. Is comingsoon being updated regularly? [16:56] Makyo: it should, check the version? [16:56] Oops, forgot. Will check. [16:56] Makyo: thought we had that fixed up, but we needed to verify that the update does in fact work [16:56] so it's possible there's still an issue [16:58] rick_h_, comingsoon does a merge, right? Hashes don't match. [16:58] Makyo: oh hmm, will it never match? I guess not. [16:59] bah, guess hatch was right, we should stash and pop the config diff [16:59] rick_h_, Yeah. May be nice to put the commit subject in there too. [16:59] Or that, yeah [16:59] bac: oh friend oh pal with comingsoon access. :) [17:00] Makyo: can you help him with the git-fu? Right now comingsoon sits on a branch and merges from upstream develop [17:00] Makyo: I guess what it should do is to run develop with uncommitted changes that it stash and stash pops after a git pull? [17:01] rick_h_, I think so, yeah. [17:27] rick_h_: review done, your changes should not affect functional tests in theory, how is your run going? [17:28] frankban: still running [17:29] frankban: which isn't promising as I think I started it back at almost an hour agao [17:31] frankban: thanks for the review. Will go through it. Appreciate the extra set of eyes [17:31] rick_h_: yw [17:47] hey rick_h_, what's broken [17:48] Makyo: ? [17:48] bac: the issue is that we can't check what version is on comingsoon because of the sha is always different due to that commit of the config change [17:48] just read backwards. Makyo you have suggestions on fixing? [17:48] rick_h_: i'm all ears [17:49] bac, change the config file in clean develop. On update, git stash; git pull; git stash pop [17:54] Makyo: ok, i think we have a clean develop branch and a 'comingsoon' branch that has the required local change but it is checked in. [17:54] so i need to 'git checkout develop' and then (somehow) get the local changes [17:55] bac, correct. The changes are just the config file, right? [17:55] Makyo: yes, i thing app/config-prod.js [17:56] so how do i get it from comingsoon to develop? [17:56] s/thing/think/ [17:57] bac, okay. Would make those changes by hand and leave them uncommitted (say by copying the file elsewhere, then moving it into place after checkout develop). The update script should `git stash save; git pull; git stash pop`. [17:57] That will stash the changes, merge the updates into develop, then unstash. [17:57] ah, yes that makes sense [17:57] hold plese [17:57] s/plese/please/ [18:00] Makyo: it looks like my develop is unclean. i did a 'git pull' to bring it up-to-date and it opened an editor for a merge message. [18:01] bac, If we never push from comingsoon, that should be fine. correct me if I'm wrong, rick_h_ [18:01] (Otherwise, might be worth just wiping the repo and cloning from scratch) [18:01] s/repo/working copy [18:01] yea, I'd suggest re-cloning if we know what the changes to the config file are ? [18:01] yes, we ever push from comingsoon [18:02] rick_h_: i've copied the config-prod.js to a safe location, but i'd like to verify the diffs [18:02] bac: rgr, so http://stackoverflow.com/questions/3338126/git-how-to-diff-the-same-file-between-two-different-commits-on-the-same-branch [18:02] using git log to find the sha of the commit/prev commit [18:03] otp [18:11] ok, rick_h_ and Makyo i have this in develop: http://paste.ubuntu.com/6751741/ [18:12] you'll see it includes the local mods. i'm a bit concerned by the 'ahead by 37 commits' [18:12] yea same here [18:12] :/ [18:12] compare git logs to https://github.com/juju/juju-gui/commits/develop ? [18:15] rick_h: yeah, so it has hatch's #73 and then the lates is Merge branch 'develop' of https://github.com/juju/juju-gui into develop [18:15] s/lates/latest/ [18:16] bac, yea I'd wipe-re-clone and then copy over your config file [18:16] rick_h_: so how can i just blow away this develop branch and then get it again [18:17] oh...right because you don't want to re-make it all? [18:17] i just want to start from a known good place and go forward [18:17] i have no problem rebuilding [18:17] git fetch && git checkout master && git branch -D develop && git checkout -t develop [18:17] bac: or cd ../ && rm juju-gui [18:17] and then git clone https://github.com/juju/juju-gui.git [18:17] from scratch [18:18] hey that sounds like a winner [18:18] after you nab your config file you want to keep :) [18:18] yep, it is safe [18:18] or, git clone https://github.com/juju/juju-gui.git juju-gui-temp [18:18] and then mv juju-gui-tmp to juju-gui when built [18:19] sure [18:19] not sure what that gets you though [18:20] rick_h_: it keeps comingsoon.jujucharms.com alive while i rebuild, etc [18:20] bac: hmm, not sure if you can mv it after make is dont [18:20] done [18:21] * rick_h_ thinks some symlinks/paths might be off? [18:21] oh yeah, piss [18:21] is Gary away this week? [18:21] luca: yes, until friday [18:22] rick_h_: damn...is anyone is second in command? :D [18:22] luca: depends on what you need [18:26] Makyo: does this look right http://comingsoon.jujucharms.com/juju-ui/version.js [18:27] bac, yep! Thanks :) [18:27] cool. we still need to keep an eye on it as things land [19:33] rick_h_: i updated my branch. i assume you wanted a look-see since you didn't approve it originally. https://codereview.appspot.com/51650043 [19:34] bac: loading [19:36] bac: LGTM thanks! [19:41] benji: my branch that modifies bundle proofing for charmworld is landing soonish. you may want to merge it into your sooner than later. [19:41] shouldn't be big impact,though [19:41] bac: thanks for the heads-up [22:00] Morning