[01:35] bac, fwiw, matt's card never got a review. tomorrow. [01:36] I mean, didn't get a review other than mine [06:51] Is there a howto to bring up a juju-gui instance? Can I just 'juju deploy juju-gui'? [12:22] swebb: we are working on a charm that will allow you to do just that. it's not quite ready yet. [13:14] teknico, in regards to title capitalization, a few things. First, I don't want to talk about this topic. It is too trivial. :-) I think we can all agree that we should be consistent, at least. Second, In American English the commonly preferred approach is title case, as your quote also showed. However, unbeknownst to me, apparently the sentence case is at least more common in British English, and we are technicall [13:14] y supposed to be writing that, despite not having any British people on our team, so that's a start on an argument. The "sentence case is more appropriate for open source documentation" argument is not very compelling to me--I don't see any kind of correlation myself--but I recognize that not everyone shares my opinions. Lastly, I'm not fond of dictating much of anything. If I can't persuade, maybe I'm wrong. If [13:14] you would like to raise this with the team and get a consensus so we never have to talk about this again, that would be fine with me. :-) [14:02] gary_poster, we don't have to talk about it if you don't want to [14:04] teknico, let's talk about it on our call in an hour :-) [14:05] gary_poster, ok === mobile is now known as bac__ [14:23] * bradc is having issues [14:33] * teknico is back in a few minutes === teknico_ is now known as teknico [15:27] gary_poster: that was not me [15:27] ok bac [15:28] this is me [15:28] :-) good [15:29] bac bcsaller benji frankban goodspud jovan2 Makyo teknico call in 1 [15:44] * gary_poster restarts computer === gary_poster is now known as gary_poster|away === gary_poster|away is now known as gary_poster === gary_poster is now known as gary_poster|away === gary_poster|away is now known as gary_poster [15:56] bac I keep on meaning to tell you that the notes you made for yesterday's meeting with goodspud, hazmat, and greg rocked. thank you. [15:57] If you want to see Mark's vision... it's here... in the middle of the whiteboard. [15:57] https://docs.google.com/a/canonical.com/file/d/0B1IM--9A1RkTX1hyOUQ5YUlzenc/edit [15:57] interpret it how you want to [15:57] Things are IN THE CLOUD! [15:58] Everything is in the could... except the stick figures [15:58] is that a cookie in the top right? [15:58] mm, chocolate chip [15:58] being cooked by demon fires beneath it [15:58] It's a sea sponge [15:58] lol [15:59] Believe it or not they are all different sorts of branches leading to collections of charms [15:59] Launchpad related [15:59] I like the chocolate chip cookie interpretation better, myself, but I'm partial to chocolate chip cookies [16:00] I'm with you on that [16:00] :-) [16:00] My friend in Aus sends me some every so often... they are the most AMAZING [16:00] She makes them herslef [16:00] *herselft [16:00] heh [16:01] believe it or not, on Italian TV these days there's an ad [16:01] by a shitty hosting provider advertising www.cloud.it [16:01] My wife made began chocolate chip cookies this week. except for the chocoloate chips--she had to have a bit of milk chocolate there. but anyway, they were um, surprisingly awesome. I'm not supposed to be surprised that the vegan recipes are awesome, right? [16:02] and it displays a stick figure throwing itself into a cloud :-) [16:02] ugh, s/made began/made vegan/ [16:02] heh [16:03] Probably awesome because of the oh so tasty non-vegan bit [16:03] Hmmm... chocolate [16:04] teknico, so they little video basically says, "life is too busy and complicated so it's probably best to go to heaven?" [16:05] goodspud, that may be the subliminal implication, I guess :-) [16:05] :-) [16:05] :) [16:06] I don't think it would be too hard to find vegan chocolate. === rogpeppe2 is now known as rogpeppe [16:07] no, there is vegan chocolate [16:08] the chips we've found don't taste as good [16:08] there's a local chocolate place that makes chocolate direct from cocoa beans [16:08] they are vegan and really yummy [16:08] but don't make chips [16:09] (and if they did I suspect they would be prohibitively expensive) [16:09] * teknico is rebooting a couple more times, just for fun [16:14] individually gold foil wrapped chocolate chips, delivered by unicorn [16:26] :-) [16:33] gary_poster: re: the notes yesterday, i just set up the template and hazmat went nuts. i did try to repackage it a bit to be more understandable [16:33] i'm curious if mark's classification of "the future" includes kapils categories of "science fiction" and "fantasy" [16:33] and then some, probably [16:34] bac, heh, cool. :-) sometimes getting the ball rolling is the hard part. and it is understandable, so success there as well. [16:34] scifi/fantasy: yeah, I suspect you are right [17:04] gary_poster: the doc refers to the `tarball` make target, which does not exist, i guess it should be something like ``tarball: $(RELEASE_FILE)``. Also, `make release` seems to be now called `make dist`. I think this is a WIP. [17:05] bac, you have a card on the board in "Needs Specification" titled "staging deploy: add 'make clean' and 'appcache-force' to cronjob" [17:05] can we remove that, bac? [17:05] gary_poster: yes. [17:05] done thanks bac [17:06] frankban, looks like doc was not updated. State is as follows, afaik: [17:06] - doc should be updated to change "make release" to "make dist" [17:07] - Makyo's branch will hopefully re-add a "distfile" target, which will be equivalent to the "tarball" target. Hopefully he will update the docs for that too. [17:08] If Makyo's branch has those changes, you could start from it as a precursor [17:08] Ah. hm. Can do! [17:08] Almost ready to push, if not submit. [17:08] gary_poster: ack, cool. [17:08] Thanks Makyo :-) [17:09] frankban, I suppose you could make a release from trunk as is even without those changes, if everything else looks ok [17:09] you would only need the make distfile target when you get to that part f the charm [17:10] I bet it will be ready for you tomorrow morning in trunk [17:11] gary_poster: I agree, I was just following the steps in the docs, just to get used to it. Also, IIRC, launchpadlib si installed by default in all our OS targets, right? [17:11] frankban, pretty sure but not 100%. A lot of Ubuntu relies on it [17:12] indirectly at least :-) [17:19] gary_poster: ok thanks [17:35] OK, fess up... who broke the demo? [17:35] https://docs.google.com/a/canonical.com/file/d/0B1IM--9A1RkTM2ZfS3RSalZQODQ/edit [17:39] I have vays ov making you talk [17:39] Actually, should I file this as a bug or should we not worry about it? [17:39] ... for the time being anyway [17:52] goodspud: that is hideous. how did all of those arrows get in there? [17:53] bac, :) The evil red arrow monster [17:53] goodspud: with your new found powers you just need to 'bzr pull' and 'bzr log | more' and you can figure out the likely culprit [17:54] goodspud: btw, i don't see the same thing. it looks good from here. [17:54] bac, I'm not used to these powers. [17:55] bac, only happens when I resize the browser. [17:55] With great power comes great responsibility, etc, etc. [17:55] :) [17:55] goodspud: oh, now that is good info [17:55] goodspud: ok i can replicate the bottom. [17:56] i don't see the artifact at the top of your screenshot, though [17:56] bac, that's the background... the canvas isn't resizing [17:57] I'll log it as a bug and then you can all fight over it. I'm not necessarily concerned about it at the moment but it's not pretty. [18:07] benji, gary_poster: either of you used debug-hooks lately? [18:07] bac: a little last week; but I couldn't remember half of it [18:08] the charm has an install error. i ssh via debug-hooks but when i manually try to run install i get an error from juju-log b/c the JUJU_SOCKET_AGENT isn't available [18:09] it is as though the environment isn't there....but that's what debug-hooks *did* i thought, as opposed to just ssh-ing in [18:10] that matches my understanding [18:18] bac, not exactly [18:18] wehn you have debug hooks on then you can look around [18:18] but when a hook is called by juju, it dumps you into the environment then [18:19] and you have all the expected variables [18:20] gary_poster: but can you re-run the hook? i [18:20] bcsaller, I wonder if the test you loosened actually was catching this regression or not [18:20] https://bugs.launchpad.net/juju-gui/+bug/1090046 [18:20] <_mup_> Bug #1090046: Resizing the browser causes the GUI to break

< https://launchpad.net/bugs/1090046 > [18:21] I'll add the card to the secondary story [18:21] bac, yes [18:21] looking [18:21] gary_poster: that's when i see failures when the hook calls juju-log [18:21] bac juju resolved --retry [18:21] gary_poster: that bug looks to be related to the charm store filter picker. but i'm pretty sure browser resizing worked when i landed it [18:22] is that what you wanted? that's what we used [18:22] gary_poster: gah! that's what i'm missing [18:22] cool [18:22] so you do that before debug-hooks, right? [18:22] or inside? [18:22] bac, before, on your local machine [18:22] righto [18:26] gary_poster: no difference. :( [18:27] bac, it didn't retry and dump you into the hook in a separate session? weird. supposed to, pretty sure. hunting around... [18:28] gary_poster: actually this page https://juju.ubuntu.com/docs/write-charm.html says you do d-h and then from another shell resolved --retry [18:28] bac, right, on your host machine [18:29] ohm, sorry, I mislead [18:29] np [18:30] https://juju.ubuntu.com/docs/hook-debugging.html should definitely mention resolved --retry [18:30] and the dance [18:58] Makyo: around? [18:59] i'm seeing juju-gui charm failing on make due to app/assets/javascripts/d3.v2.min.js being missing again. shouldn't that be fixed? [19:00] things have been changing in the Makefile bac, so it may be broken again :-/ [19:00] bac, maybe see if teknico's branch fixes it, by chance? [19:00] getting... [19:01] bac lp:~teknico/juju-gui/improve-makefile any better? [19:01] or worse :-P [19:02] * gary_poster will be glad when we are done with Makefile [19:02] gary_poster: i'll try it now [19:02] Looking. Sorry. Not quite 100% :P [19:02] * gary_poster also sympathetic to benji's call for Makefile tests, even though he is worried about what those would look like [19:03] Makyo: nm, i've got a lead. [19:05] bac, okay. make all in a lightweight checkout ran okay.. [19:06] But that was just local. [19:06] don't see why it should be diff :-/ [19:08] gary_poster: just got back from grabbing lunch, the regression might have been caught by the test or caused by other changes in the refactor, I imagine a few minors will crop up and then again as we move things to modules [19:08] agreed bcsaller, no worries, just something to address [19:09] one problem we have is none of our clean targets don't get rid of the links in app/assets/javascripts so initial builds look different from subsequent ones [19:09] s/don't// [19:09] s/don't// [19:10] ah [19:10] yes [19:10] that should be fixed [19:11] mm [19:12] bac, if you develop a target for that in anger that you think would be reasonable, let me know. I have an evil plan in that case: if teknico's branch helps for you, get your review of it, branch it, add your change, and land it. I will be happy to do this evil if so desired. [19:13] gary_poster: ok. haven't yet gotten teknico's branch to try [19:13] bac, is that likely to happen, or should I go ahead and make a branch/card? [19:13] ok [19:13] I'll make a bug/card then [19:29] gary_poster: where do we chat? [20:08] benji, *probably* not pertinent to the problem you are trying to debug, but just in case, https://bugs.launchpad.net/juju-gui/+bug/1090046 [20:08] <_mup_> Bug #1090046: Resizing the browser causes the GUI to break

< https://launchpad.net/bugs/1090046 > [20:08] IOW there might be a real problem that the pixely things are complaining about [20:08] * benji looks [20:08] I know I tried to write pixely tests about this [20:08] they were nasty but better than nothing, i thought [20:09] not good enough though :-) [20:09] :) [20:09] I am pretty sure the problem was two fold: the test that fails in prod is just a bad test (it doesn't test exactly what it thinks it tests) plus there is /some/ difference in CSS between dev and prod [20:10] ah [20:10] ok [20:10] I can demonstrate an unrelated style difference between test and prod, but I am not certain what is causing the difference that exposed the bad test [20:10] weird [20:11] I am reworking the test to just test the pixel calculation instead of the accedental details and will call it done, but there is still something broken somewhere [20:11] I will file a bug with what I have on that front. [20:22] thanks [20:32] gary_poster: ready when you are [21:19] Going to take the dog to the vet, then lay down, will be on email. [22:30] Q: I got the juju gui going, but it doesn't seem to know about my juju cluster (or environment). How do I point the GUI at my env?