[01:07] hi :-) [11:14] cat -> vet. might be a hair late to start but shouldn't be too bad. [13:06] bac: do you want an email brain dump about the backfill branch or a call about it? [13:06] benji: chat i guess [13:07] bac: ok, I'll create a hangout [13:07] ok, give me 2 minutes [13:08] bac: https://plus.google.com/hangouts/_/2db6f1f3bd754cfc2427148bdc58022605d3e851?hl=en when you are available [13:30] benji: i just lost that file. stupid hangouts. [13:43] hi gary_poster, you around? [13:44] bac, yeah, hey [13:44] cool, it was just awful quiet here. :) [13:44] ssshhhhh [13:44] gary_poster: i logged my swap day in c.a. from today to day after t'giving [13:45] for those of you with nothing else to worry about, BART is contemplating a strike. http://nyti.ms/1ekVjHd [13:45] bac, saw, had not gotten around to approving, but just did. thank you! [13:45] yea, and my wife is freaking out over the dry ice 'bomb' in LAX [13:45] gotta love travel [13:45] rick_h_: i don't know that one [13:46] http://r.bmark.us/u/2cabf797b26f4a [13:46] hmm, kind of did a sucky job parsing that one :/ /me files a bug [13:47] http://www.latimes.com/local/lanow/la-me-ln-airport-ice-bomb-20131013,0,1098000.story [13:49] so an "explosion" that did no damage but made a lot of noise. [13:49] yea [13:49] not really what you want in an airport, though [13:51] gary_poster: just noticed staging.jc.com is down. investigating. [13:52] manage.jc.com is happy [13:54] thanks bac [14:09] staging is happy again. just had to reboot the elasticsearch instance. curtis' "debugging staging" document to the rescue. [14:26] yay thanks bac [14:27] jujugui, charm branch for review: https://codereview.appspot.com/14425057 [14:27] gary_poster: i'll do it [14:27] thanks bac [14:54] gary_poster: code looks good. doing qa now. have one question about the release_expression regex. it seems to only be applied to the embedded release file. why do need to continue supporting tgz if that is the case? [14:58] bac, 'cause we have one now that is tgz. I could remove that from regex once the charm includes xz, as a separate branch [14:59] gary_poster: but why not swap them out in this branch? [14:59] gary_poster: doesn't matter, of course [15:01] bac, I could. I've thought about it. I'd have to switch to the 0.10.1 tag, apply the new Makefile, make a new dist, copy that over to the charm, and test. seemed like enough work that doing it separately made sense [15:01] bac, or waiting for 0.11 [15:01] ok [15:03] bac, could you do https://codereview.appspot.com/14494056 too? It's the gui diff you are already looking at [15:03] 1 line :-) [15:34] gary_poster: sure. (sorry i didn't see this earlier) [15:35] thx np [15:36] gary_poster: still qa'ing. slow, slow. [15:37] gary_poster: did you ever ask juju-core about the whether deploy downloads the charm locally? [15:41] bac, I did ask. they do. :-( [15:45] that's crazy. i wonder if it changed? [15:50] jujugui call in 10 [15:50] jujugui call in 10 [15:50] DAmnit [15:50] hah HAH! [15:50] oooh, gary_poster takes the early weekly lead catching Makyo getting coffee [15:51] I thought it was a text, so I looked at my phone rather than acting! Sigh. [15:57] hah, now we're going to have to watch for false alarms. gary_poster will start texting Makyo an hour before stand up to confuse him [15:58] mwa ha ha [15:58] jujugui call in 2! [15:58] Trying to figure out how to change the vibration pattern on my phone. [15:58] lol [16:07] Hey, is there a gui var that stores the path from the URL? [16:08] antdillon: window.location [16:08] antdillon: well I guess depends on what you're looking for [16:08] antdillon: the browser path? or the overall domain url, or something [16:08] rick_h_, That will work but just wondered it there was a system config var that would hold it [16:09] antdillon: which part of the url are you looking to get? [16:09] I just want onboarding to be rendered if for example the url is localhost:8888/ and not if there is any path after [16:10] rick_h_, I can use window.location but figured it better to use the initial config's if possible [16:11] rick_h_, So onboarding doesnt display if the url is: http://localhost:8888/fullscreen/ [16:11] rick_h_, Or http://localhost:8888/precise/juju-gui-77/ [16:12] rick_h_, Happy to use window.location though if that is how other views do it [16:13] antdillon: ic, hmmm. I'd just start out with window.location for now I think [16:13] rick_h_, Cool thanks [16:49] gary_poster: my qa deploy still has juju-gui in 'pending'. seems like its been a really long time, though i didn't note the start. going to eat now, will restart after lunch if no progress [16:54] bac that's weird. it is fast for me [16:54] bac, faster than before [16:55] bac, I was using an ec2 environment fwiw [16:55] gary_poster: me too [16:55] :-/ dunno [16:55] i'll review what i did. pebcak a possiblity [17:15] bcsaller, I don't understand why the gui-x/gui-y annotations shouldn't be the right thing to export. You said we strip those annotations off? It seems like we ought to keep them [17:15] * gary_poster looks for code [17:16] gary_poster: they are, but the current draw code removes them on apply, this just polls the current position from the client and adds them to the export [17:16] gary_poster: the code path I'm working on now doesn't remove those annotations though [17:16] but the logic is different [17:16] bcsaller, ah, excellent, that sounds very much like the right thing to do [17:17] just hard to get right so far, getting closer though I think [17:18] bcsaller, cool, thx. so this initial patch from you is really a band-aid. I'll try to decide what to do with it, but I'll move to something else for now [17:18] (hoping that you get the better fix ready soon) [17:19] yeah, I hope so too, this change is more disruptive and will need more QA, that one just quickly gets export values out [17:19] ack [17:33] jujugui, filed https://bugs.launchpad.net/juju-core/+bug/1239783 fwiw. [17:33] <_mup_> Bug #1239783: CLI deploy command downloads charms to local system and then uploads to environment [17:34] I fear we may have awakened some eldrich horror with this bug http://ubuntuone.com/0Nki2dkLSmnKb2A0OvWLTs [17:34] Makyo: man, I'm having double dispatch flashbacks like that [17:34] and my sanity slips away [17:41] AHA [17:51] bac: got a sec, want a second opinion on this [17:52] rick_h_: sure, just a minute [17:54] gary_poster: my second attempt launched just fine, so qa ok. [17:54] rick_h_: what's up? [17:55] bac: quick hangout? [17:55] sure [17:55] bac: https://plus.google.com/hangouts/_/9f3a5023b9270776d0b5bac6934f305d4332cf96?hl=en [17:57] Makyo: what do you think about a 1sec timer after a drag event where we are still in a drag state, DRAG_ENDING, where we ignore annotation changes till it times out? [18:00] bcsaller, uh, give me a sec to think about it. Found a solution for bug-at-hand, want to get test down 1st [18:00] righto [18:03] thanks bac. [18:21] bcsaller, okay with another drag state set we send annotations and unset in CB. Then can check that in updateServiceNodes. [18:22] Thinking about timer. [18:22] Makyo: yeah, I basically did it that way, I think it helps with some of the drag, anno delta, drag races we see [18:26] Yeah [19:28] Hey Makyo, do you want me to take reviewing frankban's branch, or do you think you will get to it before your EoD? [19:29] gary_poster, proposing now, will review in ~ 1min. [19:29] great! thx [20:08] Makyo, still not out of the woods with positioning on your branch, unfortunately--have you deployed Benji's wiki bundle and then tried to drag services?--but can you give me qa instructions for what you did fix? [20:09] gary_poster, initial load from a real environment with position-annotated services. [20:09] I went for the bug I already knew about [20:09] Makyo, ok will try, thanks. [20:09] Reviewing, then onto the rest of the position bug. [20:10] cool [20:10] bac: review time today or in the morning? https://codereview.appspot.com/14441069/ [20:11] bac: going to head to EOD, put in lbox with reviewer notes [20:11] rick_h_: ok, either today or the morning [20:11] bac: np, thanks [20:12] rick_h_: but i don't understand your request about lbox [20:12] bac: just fyi, I putt the review in lbox and added reviewer notes (vs LP based review) [20:13] rick_h_: so you're saying i need to read the launchpad merge proposal too? [20:13] oooo,nm [20:13] bac: no, just normally we've been doing LP only reviews for charmworld [20:13] bac: but I got lbox going in my lxc and decided it'd be nicer/easier there. Link is to reitveld [20:13] rick_h_: ok, but you can't use lbox submit [20:13] bac: k, all good. [20:13] bac: just trying to make it easier for you to read :) [20:14] rick_h_: that'll clobber the lander and then aaron will clobber you [20:14] bac: understood [20:14] he has that plugin to help [20:14] aaron does [20:14] gary_poster: oh did that happen? [20:14] yeah [20:14] never got it [20:14] would be nice [20:15] maybe we can get him to show us next week [20:15] I think I blogged about it. was a while ago. [20:15] if it isn't obvious [20:15] yea, I remember him doing it. Didn't mess with it at the time because I was doing full gui work then [20:15] bac http://jujugui.wordpress.com/2013/05/24/thanks-to-diogo-matsubara-well-be-migrating-to/ [20:16] nice, i'll be trying that out! [20:28] Makyo, every time I add or adjust a second ghost on a real environment in your branch, the whole canvas moves around--seeming to center incorrectly. Is that a problem with trunk as well? I can't dupe in the sandbox of your branch or trunk. [20:28] I can switch to trunk if you are not sure [20:29] Makyo, also the initial position of the juju-gui service is off [20:29] that is, when I drag it, it is dragged from a different position [20:30] I can switch to a branch [20:30] I mean switch to trunk [20:30] gary_poster, Can't dupe the first, give me a sec on the second. [20:30] I'm running from local, sorry [20:31] Makyo np. it looks like, after you have placed the juju-gui charm, the first problem goes away. [20:31] I mean, manually placed, and encountered bug #2. Once you encounter bug #2, bug #2 and bug #1 cannot be duped anymore [20:33] oh weird, #2 is still a bug [20:34] I mean, I can't dupe #1 any more, but I can still dupe bug #2 [20:43] Makyo trying to dupe with trunk, using fresh env [20:53] Makyo both of those bugs are in trunk. qa ok for your branch [20:55] I would say "whew", but I'm not sure that's good. AT least this is a quick fix for one thing that can land :) [20:55] exactly [20:57] Makyo, should we have another brainstorm meeting? I'm tempted to try and dig into this further; it certainly is the only area I know of where we have showstoppers, and it is also affecting the bundle work. I might tackle things from the bundle deployment perspective. Are you looking at things from there? [20:58] Whether or not I tackle things is a separate question from whether you would feel a meeting was helpful) [20:59] Yeah, can we do a quick meeting? If there's a separation of things bundle/not-bundle, I can grab the not for now, if you want to get from the bundle side. Otherwise, just triage. [20:59] cool Makyo. https://plus.google.com/hangouts/_/10fe9c48d75b9ebbe273219cb03f4e53869ce4f7 [22:04] Morning [22:17] Makyo: have you seen the issue where initial placement after import is confused due to the annotations not being available on the service in its first delta? [22:18] hi huwshimi [22:19] bcsaller, haven't done much with import yet, unfortunately. However, the part I'm working on (deploying from CLI, which leads to service deltas with no annotations) may get that incidentally? [22:19] yes [22:19] I think the draw code is hitting the semantic difference in having annotations on the initial object and being in another call/delta [22:20] from the way PyJuju did it [22:20] I'm wondering about delaying the initial draw to see if we can get annotations later (next delta for example) [22:21] bac: Hey there. [22:21] otherwise we draw with pack and then update to the annotation [22:21] bac: Do you have a timeline for needing those import/export icons? I've currently prioritised them below some other work, but if you need them soon I can get on it. [22:23] we currently save the position after a pack placement which can now conflict with the annotation set by the import (as that happens later now) [22:29] Makyo, gary_poster maybe when you have a second you can run through some placement tests on lp:~bcsaller/juju-gui/exportXY/ and see if the behavior makes sense. There are fewer special cases in the code. [22:29] bcsaller, sure, give me a second to switch. [23:11] bcsaller, looks okay against a real env so far. [23:11] Wil lkeep poking, little slow since James just got home. [23:11] Makyo: ahh, thank you [23:56] bcsaller, qa looks fantastic to me [23:56] ahh, great, doing some cleanups to get all the tests passing, seeing some failures after I merged trunk [23:56] but I'll propose soon [23:56] and I think the code is cleaner [23:56] awesome [23:56] thanks bcsaller!