=== imbrando1 is now known as nodnaebmi [12:07] morning. [12:08] hi bac [12:08] frankban: hi. reviewing your branch now [12:08] cool, thanks bac [12:09] bac: it seems lp2kanban sync stopped working [12:09] frankban: ok. i suspect the lane renaming killed it. [12:09] i'll look after i send your review. === rogpeppe1 is now known as rogpeppe [13:22] teknico, "Autoset version number in project docs": land with Francesco's changes, though I did have one idea in review you might find interesting. [13:22] (but completely up to you what you do with it) [13:31] g'morning gui folks [13:31] mornin hazmat [13:33] morning [13:33] gui charm panel review @ juju-ui hangout [13:33] now, hazmat? [13:34] gary_poster, yeah [13:34] just starting [13:34] k [13:38] goodspud, there's a "daily standup" for all of gui today right after our daily standup. Stephanie Danevoy created it. Do you happen to know why we have that second reservation? (It conflicts with another meeting proposal I received) [13:38] gary_poster. No idea. I'll go have a chat with her if you want? [13:38] I was wondering about that. [13:38] sure, thank you goodspud :-) [13:45] gary_poster, seems to have been a technical issue caused by sunspot activity [13:45] goodspud, lol ok thanks for checking [13:46] She's "removed" the offending item. [13:46] :-) cool thank you [13:53] gary_poster: re charm, I am implementing user/password options. do we want staging=True to automatically skip the login screen (passing admin/admin)? [13:56] frankban, yeah I was wondering about stuff like that--similarly whether staging should change login_help (if we didn't do what you proposed) [13:56] gary_poster: yes [13:57] frankban, sorry on call so trying to multitask, which I suck at :-P [14:00] frankban, the only reason not to do something like that is if someone wants to be able to show improv without password...mm, if you like that idea then I'm +1 on it (staging=True sutomatically sets password) [14:00] with the understanding that I have .25 brain for this :-P [14:01] gary_poster: np, my proposal is: 1) staging = false: if both user and passwd are provided, pass them, otherwise login screen. 2) staging = true: if both user and passwd are provided, pass them, otherwise pass admin/admin. Anyway, I am not sure about a use case where you may want in the future to set user and passwd for staging. If there aren't, maybe we could always ignore user and passwd in case 2) [14:05] hazmat, around? [14:05] therve, yes.. 5hrs of meetings. regression on trunk for openstack.. [14:05] hazmat, https://pastebin.canonical.com/82165/ fwiw [14:06] pad.lv/1100281 [14:06] therve, pastebin.ubuntu.com pls ? [14:06] * hazmat reunites with his 2fa device tonight [14:06] therve, or chinstrap if sensitive [14:07] frankban, +1 on proposal. I am ok with either approach for #2, and lean slightly towards ignoring [14:07] gary_poster: cool, thanks [14:08] hazmat, http://pastebin.ubuntu.com/1537716/ [14:09] therve, yeah.. that's the regression.. test suite didnt catch nor did functional test on ec2.. i'm in meetings for 5hrs but will fix [14:10] today.. there's a branch in review to fix .. [14:10] ok cool [14:10] tell me if I can do anything [14:11] the relationship between a boy and his 2fa device is a beautiful thing [14:13] therve, the branch in review needs a unit test if you so inclined [14:13] https://code.launchpad.net/~ev/juju/fix-1100245/+merge/143483 [14:13] let me try that :) [15:28] bac bcsaller1 bcsaller2 bcsaller3 :-) benji frankban goodspud hazmat Makyo teknico call in 2 [15:28] lol [18:03] gary_poster: you going to do surgery on the kanban lane names? [18:14] done bac [18:19] gary_poster: can you look at https://pastebin.canonical.com/82211/ [18:20] gary_poster: do you think any of those other automatic lanes should be set? [18:20] the lanes are very lp-project specific [18:23] agreed lp-project specific... [18:27] bac, I'm not very clear on the effect of them. Quick call in juju-ui? [18:30] gary_poster: ok [18:52] gary_poster: lp2kb has been updated with those lane definitions. please keep an eye out for any funny business. [18:53] ack, thanks bac [19:10] Makyo, land with trivial changes for https://codereview.appspot.com/7141043/ , thank you. bac bcsaller1 benji, if one of you could claim the second review for that branch soon, Makyo could land it and we could make room in the secondary story. [19:10] * bac looks [19:10] gary_poster, thanks. [19:11] Makyo, I'd like to get your thoughts on bug 1099921 before you go. Do you think you could do some preliminary digging, and update the bug with notes on what we might need to do to investigate further, or even fix? [19:11] <_mup_> Bug #1099921: Dragging services fails intermittently < https://launchpad.net/bugs/1099921 > [19:12] gary_poster, sure. [19:12] Makyo, beyond that, if you were to actually start in and not get it to completion, you could hand it off to someone else at your EoD. I think it would be great to have your expertise there, though. Thank you [19:13] bac, similarly, if you don't get to the end of bug 1099909 by your EoD, since you are out tomorrow it would be great if you could write a handover email [19:13] <_mup_> Bug #1099909: We are trying to talk on the web socket before we have connected < https://launchpad.net/bugs/1099909 > [19:13] gary_poster: ok [19:13] ty [19:14] heh, all my "fix released" lies have come back to haunt me thanks to lp2kanban :-) [19:17] lp 2 resurrection ;-) [19:22] benji: in env:login the comments makes claims about scheduling to call itself in the future. is that at all accurate? [19:22] bac: not any more I don't think. That should be removed. [19:22] benji: great. my eyes don't deceive me. [19:23] * benji slowly transforms into a cactus. [19:26] gary_poster, is the charm testing slowness still a significant issue for us? [19:27] gary_poster, i just talked through a significant speedup with jimbaker and mims.. just wanted to understand prioritization [19:28] gary_poster, on my side of jitsu test this speed is trivial; sounds reasonable in juju too [19:28] speedup [19:30] hazmat, I expect our charm work to be very small after the next week or two, so it is not a super high priority. So, I'd love to see a speedup, and it would help us, but.. [19:30] ...I think the wins in terms of encouraging broader use of charm tests might be more compelling arguments than ours, at this point [19:31] In sum: not a high priority for us, but definitely nice to have [19:31] (and for juju, a bug +1 generally) [19:31] s/bug/big/ :-P [19:31] gary_poster, cool, thanks [19:31] welcome [19:31] * hazmat drops into next meeting... [19:50] ok, I wanted to get these tests done before lunch, but that's not going to happen unless I postpone lunch until midnight; I'll be back in a bit. [19:52] Thanks for reviews gary_poster, bac. [19:56] welcome [19:58] An unsolicited review from the #juju channel about the charm: [19:58] gary_poster: Looks great! An excuse not to clean my office tonight. [20:05] :-) [20:07] therve, fix is in the ppa btw [20:20] so 'make prod' seems to be broken in trunk. can someone verify? [20:21] * hazmat tries [20:22] bac works okay for me.. did a fresh make clean && make && make prod [20:23] hazmat: hmm, thanks [20:23] tests pass in debug mode as .. 291 ok [20:23] as well [20:26] hazmat: tests pass for me as well. but 'make prod' and hitting localhost:8888 give me "GET /juju-ui/app.js HTTP/1.1" 404 - [20:29] bac.. oh.. i never actually used it ;-) [20:29] * hazmat double checks [20:30] bac loads ok for me, i get the login [20:30] hmmm [20:46] hey gary_poster, got a sec for a quick call [20:47] bac on call now but available in 15 or 30 minutes or so [20:47] ok [20:50] Going to full screen then back to windowed browsing leads to incorrect sizing of the viewport; doesn't affect just resizing a browser window. Worth reporting as low? [20:53] I don't know how much we are planning on thinking about that transition. It looks fine in full-screen, which would be nice for a display, granted. [21:26] bac can call now? [21:26] gary_poster: ok [21:26] Makyo, worth reporting, high IMO [21:26] we ought to fix it, which is all that means to me [21:26] Alright. [21:26] low means "when we get around to it" :-) [21:27] Makyo, any diagnostic progress on that bug, by chance? [21:28] gary_poster, Yes. When the database is changed at all, the view is redispatched. This winds up firing two more instances of update_annotations (one for the service that was just updated, with the same coords, one for the one being dragged currently). [21:28] Unless magic happens, I won't likely come up with a fix; if that's the case, I'll focus on a write-up. [21:28] (not ruling out magic yet) [21:28] Cool Makyo :-) thanks [21:34] Quick dogwalk to think about it, back in a few. [21:45] gary_poster, fwiw that testing improvement branch. https://codereview.appspot.com/7143043 [21:46] gary_poster, on ec2 it might only be 30s-1m per instance.. on other providers (hpcloud) multiple minutes saved per instance [21:46] cool hazmat, sounds nice [21:58] hazmat: you might enjoy the #testing channel; they even have a bot that will send you a private message of your specification so you can test those too [21:59] benji, so many channels.. so little time [21:59] :) [22:09] gary_poster: i've actually gotten something to work and put up a merge proposal. [22:09] must run now [22:09] * bac runs [22:11] hmm.. that branch feels like its missing something.. it short-circuits if its not connected, instead of setting up a attr/evt subscriber on connected [22:14] Dog -> dogcamp. Have ideas on branch, will write them up when I get back. === Makyo is now known as Makyo|out [22:52] crossing the streams is dangerous