[00:07] bcsaller, final branch is up for review, still haven't been able to get the review diff right.. it looks fine by hand [00:14] bcsaller, revno 52 has the complete diff, all the other revs are on trunk.. [01:17] hazmat: did a review on lp for you [03:16] bcsaller, thanks for the review.. commented back. surprised your coverage report didn't pickup the test, was it only being run against a particular test module instead of the full suite? [03:17] hazmat: it was 'make coverage' [03:17] odd [03:18] there's a test that explicitly checks for the backoff content in the logs in test_session [03:19] ie. test_managed_client_backoff [03:19] your version of coverage shows it being trigged in manage then? for me managed only shows 83% but most of that is missing calls to error logging [03:19] coverage locally reports its covered [03:20] for the package i have 94% for the managed module 89% (mostly exception handling) [03:20] hmm, maybe I have an old coverage package [03:20] well for the test to pass it has to hit those lines, which is what's od [03:20] odd [03:23] hazmat: I'll look into it, I'm guess its something with my environment === rogpeppe2 is now known as rogpeppe [13:06] teknico, feeling any better about the bug today? If not, why don't you ask to pair with Ben or Matt this morning? [13:07] gary_poster, having lunch now, but maybe a quick call? [13:08] sure teknico. I'm free for 53 minutes, then we have our regular call in another hour. Can do it whenever you like within my available time [13:09] gary_poster, I'm in jujugui [13:39] gary_poster: when you get a second I would like to discuss the UX of the browser compatability warning [13:39] ack benji [14:22] guihelp - anyone free to review my css branch? [14:32] good morning [14:32] morning, hatch [14:40] morning === teknico___ is now known as teknico [14:42] bac, on it [14:43] gary_poster: cool. won't take long [14:43] cool === teknico__ is now known as teknico [15:07] thanks gary_poster [15:10] welcome === teknico_ is now known as teknico [15:13] gary_poster: here is a bit of a rambling explanation: http://dalelane.co.uk/blog/?p=2222 [15:14] the reason i suggest we add this is it keeps IE from behaving differently when it sees the server is on the same LAN. seems like a recipe for disaster to have it work one way in development and then another when deployed. [15:15] bac +1. Just add a link to the blog post maybe? Not sure [15:15] thank you [15:15] +1 [15:16] * bac looks for something more authoritative === teknico_ is now known as teknico [15:31] bcsaller_, benji frankban gary_poster call now? [15:31] baf [15:31] sorry [15:31] thanks [15:32] benji frankban hazmat like Makyo said :-) [15:32] goodspud too :-) [15:45] guihelp: is the "failfast" behavior of the GUI test suite a bug or something we precisely decided? [15:46] frankban, bug I think [15:46] gary_poster: ok thanks [15:48] frankban: I think there is a 'bail' flag in mocha that is being set [15:50] bcsaller_: I suspect we can pass bail: false to the mocha init [15:50] talking to myself now, great [15:51] bcsaller_, that's ok, as long as you listen to yourself [15:51] all about good manners [15:52] :-) bcsaller_ : it works [15:53] gary_poster, bcsaller_: do we want failfast? otherwise I'll add "bail: false" in my branch. [15:54] :-) I vote for not failfast, but don't feel strongly about it frankban [15:54] frankban: I don't mind fail fast on the cli tests as we are mostly doing a sanity check at that point, I do the actual debugging in the browser [15:55] but I don't feel strongly either, usually with the cli tests though a single failure would prevent a lbox submit [15:56] mm, lbox is a good point === jovan2_ is now known as jovan2 [16:15] guihelp re cs:precise/juju-gui anyone know why it would result in 2013-02-07 11:10:16,812: hook.output@INFO: E: Unable to locate package python-charmhelpers [16:16] hazmat, that used to happen if person was not using juju from ppa. supposed to be fixed now, because we add ppa at start anyway [16:17] is `guihelp` a string that I should set to ding in my chat? [16:17] yup [16:17] its a group broadcast [16:17] hatch, yes, guihelp (among ourselves) and jujugui (on #juju channel for user help) [16:17] that's a very smart idea :D [16:18] gary_poster, is that fixed in the ~charmers version only the ~juju-gui version? [16:18] s/only/or only [16:18] hazmat, ~charmers, pretty sure. verifying. [16:19] hatch, hi, you might also want to configure your IRC client so that it shows your first name and last name too [16:20] hazmat, ~charmers. bootstrap_utils.install_extra_repositories('ppa:juju/pkgs') in install hook [16:20] teknico: in my whois information? [16:20] hatch, yep [16:20] oh cmon I need some kind of anonymity ;) [16:21] frankban, you happen to have any experience with what hazmat mentioned above in charm? I don't see why/how it would happen [16:21] hatch, do you? :-) [16:22] gary_poster: hum... no, looking at the ~charmers revision [16:22] frankban, http://bazaar.launchpad.net/~charmers/charms/precise/juju-gui/trunk/view/head:/hooks/install fwiw [16:24] hazmat, are they running on precise? [16:25] hazmat, or more pertinently, are they trying to run on raring. We only test on precise. the juju ppa does not have charm-tools for raring [16:26] python-charmhelpers are in the charm-tools version supplied by both precise and quantal [16:27] gary_poster, its all precise [16:28] gary_poster, the problem is actually ppa add because its a proxy environment [16:29] the add apt repo ends up adding junk the apt repo [16:29] ah :-( [16:29] :-/ [16:29] after that it fails because adding nodejs and we don't have a key for it.. [16:29] nodejs ppa [16:30] imo the charm shouldn't be downloading things it doesn't need for it current configuration [16:30] ie. i'm using a release i don't need nodejs/imagemagick/etc.. i'm not using staging mode.. i don't need java/zookeeper [16:31] hazmat, added bug for that yesterday on your comment then. triaged as low given that it was nice to have, but put it on kanban as slack task. if this is blocker for important things then we can reprioritize. [16:32] its a blocker for important things.. but the timeline is tight.. [16:32] i might just hack do a hack for it [16:32] sadly lacking access to fix directly [16:32] hazmat, right. is this a drop everything? The change would be risky IMO [16:33] you could hack and put it in your own branch? [16:33] gary_poster, yeah.. or just pass them to the hook [16:34] risky only in the sense that it would need to be tested. we would have to move downloading prerequisites into config-changed [16:36] gary_poster, understood, i'm testing a hack fix now (just commenting out all the unesc pkgs) [16:36] cool, sounds good [16:42] Makyo: quick hangout? [16:42] frankban, sure. [16:43] Makyo: http://tinyurl.com/see-emily-code [16:46] frankban, There. [17:14] odd the gui deploys and gets to started (no hook errors) without the extra packages but it doesn't actual run.. the nginx config gets odd [17:24] * benji tweaks the unsupported browser wording. It now reads "Your browser is bad and you should feel bad." [17:25] hazmat: what oddities are you seeing? [17:27] benji, is "now" before or after the tweaks? :-) [17:27] heh [17:27] we should replace all of our UI elements with memes [17:28] frankban, nginx is running but not serving traffic.. all the pieces look like there in place but i can hit the front end webserver (its exposed) [17:29] ie. api server running, files on disk [17:34] hazmat: any error in /var/log/upstart/juju*? [17:39] frankban, no.. afaics the only issue is the nginx config [17:40] the hooks all run to completion without error [17:41] nginx config : http://paste.ubuntu.com/1621578/ charm log: http://paste.ubuntu.com/1621582/ [17:47] hazmat: is "gui" the name of the charm you deployed? [17:48] frankban, its the name of the service charm [17:48] hmm.. i deployed ~juju-gui's charm version [17:49] with this diff http://paste.ubuntu.com/1621609/ [17:49] i meant to deploy the ~charmers version.. i'll retry with that to verify [17:51] hazmat: aha! ppa:juju-gui/ppa is required so that the ssl-enabled haproxy is installed, not the precise one [17:54] frankban, gotcha.. that's not in the ~charmers version yet right? [17:55] hazmat: right [18:04] jovan2, the new gui vision looks much improved [18:05] hazmat glad you like it [18:19] >:/ [18:37] Makyo, ? [18:38] gary_poster, Firefox and Flash problems. [18:38] gotcha [18:40] bac you available to review Matt's branch? When that lands I can move some more of the CI Jenkins stuff forward [18:41] gary_poster: i can. trying to get that hatch patch to work [18:41] bac - I've never actually tested that code so I would be interested to hear your results :) [18:42] hatch: not good so far [18:42] i expect the ie10 class to be added to but it isn't [18:42] hmm is the js executing? [18:43] not sure. [18:43] not too adept at using the ie tools [18:43] inside the script tags console.log('it works!'); [18:44] thank you bac [18:45] I need to close some chrome tabs (I'm out of ram haha) and then I can boot up a Win8 VM if you like [18:46] hatch: cool. i'm doing a review now but will be back at it in a bit [18:48] Makyo: done [18:49] bac, thanks [18:55] hatch: got it to work. [18:55] oh excellent - what was the issue? [19:18] hatch: there was a typo in the code i copied from you. missing a closing > before script. no biggie. [19:19] oh woops I should fix that in the gist [19:19] thanks [19:21] bac: I may be doing this review this wrong but I think you forgot the console.log() in the html ;) [19:21] ergh, probably [19:22] thanks for the catch [19:23] np and I updated the gist with the missing > :) [21:27] having trouble submitting my changes due to lbox.check [21:42] and then it worked