[12:46] morning benji [12:47] morning bac [12:47] you all get some snow? well, benji [12:47] benji i'm about to ask for an update to manage.jujucharms.com. should i wait on anything from you? [12:47] https://twitter.com/Earth_Pics/status/412390440936415232 is crazy [12:47] bac: nope [12:48] rt [12:48] rick_h__: nope, no snow here [12:49] rick_h__: ha, at first i thought those were snow pyramids in TN based on your question to benji. [12:49] heh [12:49] heh, well you know how they love their TN landmarks [12:49] we do have one pyramid: http://en.wikipedia.org/wiki/Pyramid_Arena [12:49] hah! [12:52] benji: great, government subsidized bass pro shop. [12:52] they forgot to ask me whether or not they should do that [12:53] benji: not quite as grand a failure as http://en.wikipedia.org/wiki/Ryugyong_Hotel [12:53] heh, yeah [12:54] maybe pyramidal development is a bad idea; the originals were tombs, after all [12:54] benji: i've heard BPS are quite the bullies, usually demanding free land from shopping centers since they are such a draw [12:54] good deal if you can get it === gary_poster|away is now known as gary_poster [13:04] morning benji: thank you for the great review, I updated the branch accordingly, would you like to take a look at the changes/comments? [13:07] frankban: good timing, I'm looking at it now [13:08] cool thanks [13:12] frankban: explanations and changes all look great, thanks! [13:12] (I commented on the review too) [13:13] benji: great! thank you [13:17] benji: hmm, we haven't updated the mongo charm on manage.jjc in several revisions ... so now they don't want to update it b/c the diff is too big and scary. [13:18] bac: hmm; can we update it one revision at a time? (we'll have to test that particular combination to make sure it's cool) [13:20] benji: they introduced changes after 29 that were not pushed back upstream, so now the merge has trivial conflicts which cause panic. i'm going to replicate what they've done and upgrade staging from my local charm. [13:21] k [13:37] jujugui, smallest possible review that has an actual change, and very quick QA? https://github.com/juju/juju-gui/pull/23 [13:37] gary_poster: i'll do it [13:37] thanks bac [13:40] gary_poster: sorry but i don't know how to get this version for qa. where would i find the url to pull from? [13:41] bac, I use the instructions in the hacking doc--see "Helpful Git tools and aliases". In this case, given that I've set up "juju" as my remote for the official trunk, I would use ``git qa-pr juju 23 qa-gary`` after having set up that alias [13:42] Setting up the "juju" remote is also in that doc, bac: "git remote add juju git@github.com:juju/juju-gui.git" [13:44] wow, not only can I delete the comments of others, I can actually edit them. That is *awesome*. [13:45] gary_poster: yea, everyone is an 'admin' level so I think we've got super powers [13:45] all reviews of my code from here on out will mimic bac's comment: "Code looks superb. LGTM." I might occasionally choose different superlatives [13:45] dealing with trying to fine grain it would be a pain otherwise because the permissions aren't super fine grained [13:45] :-) cool rick_h__ [13:45] gary_poster: those are limited to one line deletions [13:46] bac, not if I can edit everyone's review, they 're not! [13:47] oh, excellent [13:47] gary_poster: ok, 'git status' shows that i have an unpushed version from friday. i thought my branch landed and got merged. now i wonder if something didn't make it through. [13:48] bac, uh. I only pretend to know what's going on because I can paste things from the hacking doc. Ask rick_h__ ? :-D [13:48] bac, though... [13:48] you may not need to worry if it is in the main branch [13:48] juju/juju-gui develop [13:49] bac: there's another alias I sent around friday that I've not put in the docs yet to help with that. [13:49] bac: it means that your branch landed in remote juju, but in your own remote (your username) it's not been pushed [13:49] rick_h__: looking at github i can't find a changelog. [13:49] so you can correct it with a `git push origin develop` [13:49] bac: it's the 'commits' heading [13:50] # Update the develop from juju [13:50] juju-sync = "!f() { git checkout develop && git pull juju develop && git push origin develop; }; f" [13:50] doh [13:50] bac: ^ will give you a new command `git juju-sync` which will pull any updates to the main repository and then push them up to your own develop fork [13:50] assuming your remote is 'juju' (unlike hatch) [13:54] rick_h__: i ran that and got conflicts which seems unlikely. [13:55] bac: ok, did you hve uncommitted changes? [13:55] rick_h__: i mean it is just supposed to update my repo from the version where i made some post-review changes [13:55] if you look at that command, it changes your current branch to develop [13:55] fiik [13:55] which if you've got uncommited changes could cause a conflict between WIP and the updated branch [13:55] i'm almost certain i do not [13:55] * rick_h__ ponders if it should do a git stash first or something [13:56] bac: sorry, then I don't have enough info to diagnose. What branch are you sitting on, is the conflict in something that was changed recently? [13:57] rick_h__: http://paste.ubuntu.com/6583528/ [13:57] rick_h__: i was in my fix-stuff branch from friday. after running the juju-sync it shows i'm in develop [13:57] bac: you've got the wrong project as your origin. Sorry [13:57] it's looking at the jenkins-github-lander repository, not juju-gui [13:58] From github.com:juju/jenkins-github-lander [13:58] it should be what gary_poster posted: git remote add juju git@github.com:juju/juju-gui.git [13:58] rick_h__: how'd i do that? [13:58] https://help.github.com/articles/removing-a-remote [13:59] bac: maybe I did a bad paste? [13:59] sorry, I've got both in my history jumping between projects lately [13:59] rick_h__: from HACKING [13:59] git clone https://github.com/juju/jenkins-github-lander.git [13:59] or I've got a typo in the docs :/ [13:59] yeah, i meant to ask about that but just trusted it [14:00] it's updated [14:00] yea, that was fixed last week. The HACKING docs are updated for it. Sorry about it [14:00] bac: so use the github link ^ to remove the remote and re-add it [14:00] bac: then you should be able to `git juju-sync` [14:04] rick_h__: i think i need to first revert the changes from that juju-sync [14:06] rick_h__: 'git revert' seems to want options that i can't make sense of [14:06] bac: `git reset --hard HEAD` [14:06] bac: git revert is about generating a diff to revert a file to an earlier rev usually [14:09] bac: do you know anything about charmwolrd and proof? Charmworld has a config for proof port but (apparently) doesn't bother to actually start anything on that port. [14:09] benji: it's used for the ingest process. So that it can talk to localhost vs the root port 80 url [14:09] yeah [14:10] localhost chamrworld runs on 6543, but when you hit it via the cli tool it's the apache proxied proces on 443 (https) [14:10] rick_h__: yep, I got that [14:14] benji: so I'm guessing it's not set because the ingest process gets it from the currently running app config? the port in the ini that the app runs on? /me hasn't looked in a while [14:15] thanks rick_h__, that cleaned up my mistakes [14:15] bac: sorry about that :/ I need to hand out 'thank you for beta testing' stickes to everyone [14:15] stickers === benji_ is now known as benji [14:21] rick_h__: fwiw HACKING is missing a closing quote on the qa-pr definition. [14:22] bac: thanks, will update with the docs card on deck. === _mup__ is now known as _mup_ [14:26] gary_poster: qa done. the deleted header does not get rendered. congratulations. [14:26] you can delete with the best [14:26] bac, my pride knows no bounds. Thank you. :-) [14:26] sorry it took me the better part of the morning to get there [14:27] :-) np [14:29] jujugui small test update review please https://github.com/juju/juju-gui/pull/24 [14:29] jujugui note this requires a manual upgrade of mocha-phantomjs which is installed globally in the HACKING doc and will not auto update. `sudo npm install -g mocha-phantomjs@3.2.0` [15:06] bac: I'm not seeing the missing quote, did you mean on another line? https://raw.github.com/juju/juju-gui/develop/HACKING.rst [15:07] rick_h__: it must've been in a prior version [15:07] bac: ok cool. thanks for double checking [15:08] rick_h__: you could s/stick-headers/sticky-headers/ while you're there. just noticed it. :) [15:08] bac: thanks, will do [15:10] jujugui one quick note, you can often tweak docs straight from the webui. Go to https://github.com/juju/juju-gui/blob/develop/HACKING.rst for instance and notice the 'edit' link next to 'raw/blame/history'. It allows you to edit inline for a quick drive-by without going through review/landing. === _mup__ is now known as _mup_ [15:42] marcoceppi: I thought my recent branch had been merged into the 1.2 branch but I don't see it there or on trunk [15:42] benji: let me check, it was merged [15:43] I'm looking at bzr+ssh://bazaar.launchpad.net/+branch/charm-tools/1.2/ [15:45] benji: OH I remember now. There was a discrepency between the repos, I had meant to merge after getting the git stuff sync'd with bazaar again but it must have slipped my mind [15:45] benji: I re-organized the project on the way back from a conference a month or so ago, it just never made it in to bazaar === makyo_ is now known as Makyo [15:54] jujugui call in 6 [15:54] * rick_h__ is lost without the normal warnings [15:54] * gary_pos` is having issues === gary_pos` is now known as garyposter [15:55] * bac is a-ok [15:55] benji: your changes will be merged and release with 1.2.6 [15:56] marcoceppi: cool, thanks [15:56] * garyposter is having MS hate flashbacks, while trying to get password fixed [15:56] account.live.com says it needs to authenticate me by emailing or texting or calling [15:57] I have it text me [15:57] it gives me a code [15:57] "That code didn't work. Check the code and try again." [15:57] I have it call me. [15:57] it gives me a code [15:57] "That code didn't work. Check the code and try again." [15:57] * garyposter so happy === garyposter is now known as gary_poster [15:58] jujugui call in 2 === hazmat` is now known as hazmat [16:23] benji: before the netsplit i offered to pair on the stuck bundle if you want. going to eat now, though. [16:25] luca__, hi. thank you for relation line image. When someone clicks on relation in box (fourth example down) what happens? Note we can take them to a relation inspector page for either side, so we may want to allow directing to both sides. Also, remember that a single name is actually insufficient to describe the relation...and we need to show relations erroring on one side or the other [16:25] luca__, maybe worth a quick call to expand on that? [16:27] rick_h__, lgtm [16:27] Makyo: thanks, sorry, almost self-reviewed but meh [16:29] bac: If we don't care to diagnose the issue we can just run bin/dequeue to get rid of it. [16:29] I have a copy of the data that is stuck if we want to attempt a local repro (which is what I would try if we want to diagnose the issue) [16:30] benji: if that's the case why not fix manage.jjc now and then diagnose locally now or if it happens again? [16:30] s/or/or later/ [16:31] bac: that would be fine with me; the only reason not to would be if we really want to know why it is happening we need to be sure we can repro before clearing it [16:31] benji: i'll let you mull it over while i eat lunch. [16:31] and it won't be liver [16:31] heh [16:32] or tripe [16:32] I don't think its worth tracking down now, so a simple dequeue is good with me [16:32] two great tastes that are better together [16:32] darn, now I'm hungry [16:38] gary_poster: If you are free for a call now I can answer your questions, also, we need to talk about the stuff we've discussed with Mark [16:38] luca__, cool, let's do it. https://plus.google.com/hangouts/_/7acpienfehsrht2ggbiqk9i0og?hl=en [17:02] Makyo: gary_poster one thing to think about with that networking, I'd keep in mind the possibility of people trying to do poor man's cross env relations that way. One big environment, multiple networks, working on an HA distributed setup [17:37] gary_poster: or frankban avail for a pre-imp on the charm change for updating to pull from git? [17:37] * gary_poster not [17:37] calls calls calls :-) [17:37] wheee [17:38] rick_h__: sure [17:39] frankban: https://plus.google.com/hangouts/_/76cpim4h6nlsceotisgs317kdo?authuser=1&hl=en [17:56] rick_h__: at this point I'm considering wiping and re-indexing ES, is that 1) insane 2) easy? [17:57] benji: 2) [17:57] benji: sec, I can get you the RT that had the commands used to clear things out [17:57] the options weren't mutually exclusive, but that's good to hear :) [17:57] that'd be great, thanks [17:58] benji: well, I'd go for 3) :( [17:58] rick_h__: 3? [17:58] benji: 1) insane 2) easy 3) :( [17:58] ah [19:02] rick_h__: I'm working on the "make CI run the local browser tests" card and I've looked over the jenkins-github-lander code a bit but I don't see the obvious place to make my changes. [19:04] benji: it should just be in the jenkins job config [19:04] ah [19:04] benji: so login and go to http://ci.jujugui.org:8080/job/juju-gui/configure [19:04] and add steps to the "Build" shell commands that run [19:05] the idea is that the command must exit non-0 and then the rest of the system should kick in. [19:05] it would need it in both the juju-gui and juju-gui-merge builds [19:06] benji: for testing, there's a link on the page once you've logged in for "Build with parameters" to manually trigger a build. It just needs some git reference to know what you want to build [19:06] benji: just putting 'develop' in there or a sha string should work [19:06] cool [19:06] rick_h__: would you mind if we add a "ci-check" make target? It is nice that the actual test running bits stored in Jenkins are straight-forward and I'd like to keep it that way. [19:07] benji: yea, works for me. Maybe a saucelabs specific target? [19:07] or you want to ci-check vs check? [19:07] I was thinking we'd keep the same make check but then after make check succeeds do a second 'make saucelabs' or something so that it keeps with local dev practice [19:07] rick_h__: I was going to make ci-check first run check and then the in-browser tests [19:07] k [19:09] jujugui: feeling unwell with what seems to be two competing headaches. going to lie down. [19:09] bac: darn; feel better [19:09] bac: good luck die headaches die [19:33] gary_poster, let me know if you have a moment for a call sometime in the next little bit. [19:59] ack [19:59] almost [20:16] Makyo, ready when you are. lemme know/ [20:16] Now's good. [20:17] https://plus.google.com/hangouts/_/7ecpi5tmtfd0i12os0cp6tt6ro?hl=en [20:24] terminal in devtools http://www.html5rocks.com/en/tutorials/developertools/devtools-terminal/ === hatch_ is now known as hatch [20:33] someone was sitting on my nickname :/ [20:45] bac hey are you around? [20:46] hatch: what's up? [20:46] so I ended up picking up a new MBP but Ubuntu looks like total garbage, all blurry and stuff [20:47] How did you get yours to look good? [20:50] apparently windows 8 now has charms... https://www.evernote.com/shard/s219/sh/1ca90ff7-0b47-4bf2-a631-639873211aeb/fd8f3960433b917b405afc95d23bc0b3 [20:53] heh [20:54] hatch: you mean in a vm? [20:54] bac well right now I'm running 12.04 in a parallels vm to even see if it's possible, then once I get that up and running I'll attempt to install it on metal [20:54] is yours in a vm? [20:55] hatch: yes, i'm using fusion. and it does not support retina very well. it is usable but it ain't retina. [20:56] ohh I thought that you ended up getting the retina to work [20:56] at the high resolution it looks good but its soooooo small heh [20:56] and any smaller it just gets blurry [20:56] :( [20:56] hopefully Trusty will add high dpi support [20:57] i use 1920x1200 [20:58] rick_h__: I reversed course on the ci-check thing since I was able to make test-browser behave nicely enough to just add it to "check": https://github.com/juju/juju-gui/pull/27 [20:58] there was a high dpi discussion on warthogs. i think the outcome was not before unctious uakari [21:01] Hmm I'll try that res [21:02] lol [21:02] that would be funny if that's what 14.10 is called [21:02] basically 'greasy monkey' :P [21:04] well 1920x1200 still looks blurry but it's better than it was before :) [21:13] houston I mean rick_h__ , we have azure account liftoff [21:25] gary_poster, looks like we're go for within the quickstart session, but not outside of it. Will put in some text for instructions regarding that. [21:26] Makyo, great! sounds like a reasonable way forward [21:42] Makyo, what was that thing we needed to do to make Inkscape icons for charms show up in the GUI properly? It seems to have dropped off the instructions [21:44] benji: awesome [21:45] gary_poster: rgr working on creating an account for the list [21:45] evilnickveitch, save as optimized svg, check enable viewboxing. https://code.launchpad.net/~makyo/juju-core/enable-viewboxing/+merge/180354 [21:45] Makyo, thanks! I'll make sure it doesn't get lost this time [21:46] rick_h__, you mean on existing box? [21:46] gary_poster: so is the password just SSO? [21:46] gary_poster: well I got an email to verify my email address from MS so assumed it's related? [21:47] rick_h__, for azure? no, unfortunately not. shared creds. trying to send emails. [21:47] gary_poster: oh, that went to the list no me directly, gotcha [21:47] evilnickveitch, charmworld now does this automatically, per https://code.launchpad.net/~makyo/charmworld/svg-viewbox/+merge/180400 so I don't know if it's strictly necessary. gary_poster c/d? [21:48] jujugui, if you get an email today from MS about verifying your email, you probably don't need to. check that it is from the new list. If so, I am handling it [21:49] rick_h__, check mail [21:52] gary_poster: got it [21:53] cool [21:55] benji: installed x11-utils and rerunning your tests. [21:57] benji: hmm, it wants local firefox. so will have to chat on it [21:59] benji: or I'm mis-understanding. I thought this was to trigger saucelabs tests [21:59] benji and rick_h__ , goal is saucelabs when run from jenkins, right? [21:59] +1 rick_h__ [22:04] Pretty hard to push to lp when you moved your SSH keys aside :T [22:04] :-/ [22:04] * Makyo dogwalks briefly, will get tests cleaned up and propose after. [22:04] rick_h__ and gary_poster: I must have misunderstood "Hook up browser sandbox tests with Jenkins" but its not a big worry because making it run the tests in saucelabs should be a one-liner change. [22:04] yay! [22:05] benji: ok cool [22:05] I'll take a look at it tomorrow. Have a good evening all. [22:06] benji: night, thanks [22:06] I'm running away also [22:06] bye all === gary_poster is now known as gary_poster|away