[09:55] hi rogpeppe: would you like to take another view at https://codereview.appspot.com/8716043. It includes s/st.client.Call/st.call/ in all the client calls. [09:55] frankban: looking [09:55] rogpeppe: thanks [09:56] frankban: nicer when using .call(), isn't it? :-) [09:56] rogpeppe: absolutely, less code and repetition [09:57] frankban: thanks for making the changes [09:57] my pleasure [09:58] frankban: LGTM [09:59] rogpeppe: cool! merging [12:08] teknico, sorry I was in a rush I guess, but yes, your shrinkwrap docs were excellent and much appreciated. :-) [12:09] and hi [12:10] gary_poster, hi, thanks, it feels like getting docs right on first try is even less likely than code :-) [12:10] hiya bac. fix fakebackend addRelation is still in landing on kanban. I'm guessing that is an oversight? [12:10] gary_poster: yeah, i'll move it [12:10] teknico, :-) [12:10] thanks bac [12:13] rick_h_, hi. I marked "remove charm panel for the charm browser integration" as blocked on the charm browser being ready. Even if/when it is ready I'd like to hear what the plans are before we do them (my basic goal being the usual "does everything still work after this commit" [12:13] ) [12:18] * teknico lunches [12:18] gary_poster: rgr, makes sense [12:36] thx [12:40] gary_poster: I suppose the high priority cards in story 1 are the best to draw from, in particular the ones that require changes to juju-core. Is that right? [12:41] benji, absolutely (though I hope that none of those actually require juju-core changes; I don't think they do) [12:41] uistage is down [12:41] I'm trying to figure out why and get it back [12:42] gary_poster: oh, I mesread the "Implement resolved in the Go environment" bug. I don't think any have juju-core-side work. [12:45] benji, good. relatedly, my understanding of bug 1169167 (thank you for filing) is that we are currently getting constraints, but not config; and that we can work around that absence in the GUI by explicitly getting service data before rendering the pertinent pages. Are things better than I understood (config is also included in the deltas, as your bug implies) or otherwise different? [12:45] "currently getting" == "currently getting from the watcher deltas" [12:46] my understanding (from frankban, hi frankban!) is that both the config and constraint data is coming in via watcher, but neither is wired up yet [12:46] ok great [12:49] benji, gary_poster: looking to the ServiceInfo structure (which I suppose it is included in the delta), I see "Constraints constraints.Value", but not the settings. So I guess the constraints are there, not the settings, as Gary mentioned. [12:49] teknico, shrinkwrap is broken somehow on uistage [12:49] frankban, ack thanks [12:50] teknico, if you are around let me know and I will ask you to help me diagnose on guichat [12:50] rogpeppe: do you confirm? ^^^ [12:50] rogpeppe: re: constraints and settings in delta [12:50] gary_poster: I'll ammend the bug. [12:50] Shall I add another about adding config data? [12:50] gary_poster: any error from uistage make? [12:51] frankban: you don't get settings yet [12:51] (I told rogpeppe that we could work around the config being absent from the deltas for now; it is a nice to have) [12:51] rogpeppe, if we can get it in the next day that would be cool, but not counting on it :-) [12:51] rogpeppe: but we get constraints, right? [12:51] frankban: yes [12:51] frankban, yes make fails with npm error [12:52] gary_poster: npm version? [12:52] gary_poster: and nodejs? [12:52] error is Error: No compatible version found: chai@'chai@1.5.x' [12:52] which is odd because it then says that 1.5.0 is a valid install target [12:53] benji: time for a quick review? https://codereview.appspot.com/8768043 [12:53] frankban: sure [12:53] benji: thanks [12:53] frankban nodejs is 0.8.11, npm is 1.1.62. I will see if I can update [12:53] gary_poster: I had npm errors in trunk, resolved updating npm and node from the ppa [12:53] ack thanks [12:54] gary_poster: after updating I have $ npm -v [12:54] 1.2.18 [12:54] [master]frankban@syd76:~/devel/juju/ui/bug-1168449-releation-names$ nodejs -v [12:54] v0.10.4 [12:54] yeah, same here locally [12:55] back [12:55] gary_poster, getting there [12:56] I think I'll add "LGTM" to my email signature. [12:59] :-) [13:10] uistage is back thanks to frankban and teknico [13:11] yay! [13:18] hi gary_poster [13:18] hi bac [13:19] gary_poster: in reviewing my destroy service branch makyo suggested that the corresponding units and machines should also be removed [13:19] bac, I saw [13:19] gary_poster: i agree about removing units and have done that [13:19] gary_poster: but what about machines? if they need to be removed should removeUnits take care of that? [13:19] currently it doesn't [13:21] bac, interesting question: AIUI pyJuju keeps machines around and jujucore does not. Given our current goals, I think an XXX with that comment, with the expectation that we will add a "removeMachines" flag to removeUnits when we implement the Go side of the sandbox, is a good way forward. [13:22] gary_poster: good idea. doing that and landing my branch. [13:22] awesome, thanks [13:25] rogpeppe: I am confused about "ResolvedResults" in params.go: I think it is not used anywhere, and the Resolved call just returns an error [13:25] frankban: looking [13:25] good morning all - looking for 2x reviews -> (at least one familiar with relationships) https://codereview.appspot.com/8735043/ [13:26] frankban: looks bogus to me. i think it can go. [13:27] hatch, I'm doing one (and marked it on kanban) [13:27] and good morning :-) [13:27] alright thanks :) [13:27] benji: in case you already investigated, does goenv.resolved need something back from the api backend? [13:28] frankban: I don't know that yet. Shall I look? [13:28] benji: I noticed that you used Y.Object.each() on an array in your last landed branch, why use that vs Y.Array.each? [13:28] Or even Y.each [13:29] hatch: I changed it to Y.Array.each as suggested. Did that message not get sent from Rietveld? [13:29] benji: no rush, please let me know when you figure it out [13:29] benji: oh I just looked at the merge and it showed that Y.Object.each was still being used [13:30] hatch: what's the difference between Y.Array.each and [].forEach? [13:30] that's because "lbox submit" doesn't update the review. It should. [13:30] * gary_poster guesses browser support [13:31] frankban: if the browser supports it....nothing :) [13:31] well it executes the callback under the proper context [13:32] that would have to be manually defined (which isnt' really an issue) if going straight native [13:32] hatch: do you know if forEach is supported by all the browsers we support? [13:32] see http://yuilibrary.com/yui/docs/api/files/yui_js_yui-array.js.html#l96 [13:32] frankban: it is [13:33] I think all of the array methods we use are [13:34] benji, hatch, we should be using Y.Array.each in app/views/service.js on the env.genericConstraints. Pre-existing problem to benji's branch [13:35] gary_poster: I'll make up a quick fix branch and throw in a couple of little things I have lying around [13:35] cool thanks benji [13:35] branch stew I call it [13:35] :-) [13:35] ahh yeah I see that on line 652 [13:43] frankban: I created a ticket to discuss switching to use native methods https://canonical.leankit.com/Boards/View/102529849 feel free to comment on it if you would likke [13:44] hatch: cool [13:45] jcsackett: https://code.launchpad.net/~rharding/juju-gui/update-urls is the branch I'm working on [13:45] rick_h_: need review on 8768044? [13:46] hatch: sinzui was going to grab it in a sec [13:46] alrighty [13:46] hatch: but jcsackett will need a second one in a few min [13:46] * sinzui is reading it [13:46] hatch: once lbox burns enough cpu cycles [13:46] heh - I noticed it doesn't put both cores into 100% on my machine [13:46] I have been meaning to investigate [13:48] hatch: that'd be cool if it ran test-debug/test-prod at once in diff cores [13:49] hehe I agree! [13:49] rick_h_, do "we" dislike the "unless" handlebar helper? [13:50] sinzui: no opinion that I know of but it's not come up. [13:50] thank you [13:50] sinzui: if you're not a fan it'd be great if we can land but add a card/bug for a followup drive by branch [13:51] I am not a fan of an empty if-block, but the else-block has code [13:51] "unless" is used for if-not and huw might not know that [13:52] sinzui: yea, I've got that on a 'bring up to him' but I wanted to look over the file in full more carefuly and pull notes together I send him at EOD here. [13:53] sinzui: so we're not having the monday sync call today correct? We had it friday. [13:54] rick_h_, I have it on my calendar still. Antonio was looking to move it. [13:55] sinzui: ok, I wasn't sure if friday replaced it this week [13:55] gary_poster: if possible could you QA the relation code? Just so we can be sure there isn't any other odd issues like last time? [13:56] hatch, will do. I'm being picky in the code too, because I don't understand some of it at first reading. [13:56] sure no problem [13:56] but trying to be helpful too :-) [14:07] hatch, rick_h_: if you have the time https://codereview.appspot.com/8612045 [14:07] lbox has burned its cycles. :-P [14:07] jcsackett: will do [14:14] done [14:24] hatch quick chat in guichat? [14:24] 5 mins? I just have some eggs in the pan :) [14:25] hatch sure [14:31] ok egged up [14:31] gary_poster: guichat? [14:32] hatch come on by :-) [14:58] gary_poster: little branch here https://codereview.appspot.com/8770044 [14:58] hatch: you might be interested in reviewing this branch, it quiets the linter down: https://codereview.appspot.com/8770044 [14:58] sure I'll check it out [15:00] jcsackett: the api guides use 'term' for the search phrase [15:01] jcsackett: so I was wondering if you mind doing a s/text/term/g [15:01] rick_h_: i would rather update the api guides. [15:01] rather than landing another branch against charmworld just to get this rolling. [15:02] do you really have a strong concern about term/text? [15:02] jcsackett: just hate 'text' as it's so generic and was looking to sync with the api docs. I didn't realize the server side wasn't term as well :( [15:02] rick_h_: i don't really know that "text" is generic...it is, in fact, a "text" search, right? [15:03] and yeah, the api is already implemented with text rather than term. [15:03] see the link in my reply. [15:03] jcsackett: then yea match the server and I've got a good card for 'slack time' one day :) [15:04] jcsackett: the last thing was yuo were doing extra work in the test than just create the data source and update it in the instance. [15:04] jcsackett: but I guess that was to get the url updated to check? [15:04] jcsackett: if that's the case ignore me [15:08] rick_h_: the extra calls are to change sendRequest into something that captures data in a way the test can interrogate. [15:08] jcsackett: right, you wanted more than the res of makeRequest. I folow now. Just seemed extra work from the test I'd done before [15:08] jcsackett: ignore me, all good. [15:09] rick_h_: cool, thanks. :-) [15:09] rick_h_: you want me to update the API guides? [15:09] or leave them alone and you'll do your slack card one day? [15:09] jcsackett: :/ I guess. No sense having bad known info out there. It'll bite me down the road when I go to look at how to generate an api req by hand for that [15:10] rick_h_: when you say api guides, you mean the "Api sketch" doc in g drive, yeah? [15:10] jcsackett: rgr [15:13] rick_h_: actually, reading the api sketch, term is used as a parameter when looking for autocompletion. [15:14] well as per above, ignore me [15:14] so term=apa will return interfaces, owners, charms, etc, that you might want to autcomplete too. there's no mention of term used for full-text; just a mention of text=apache2 as an example in some searches. [15:14] rick_h_: just updating you, so you know what's in there. :-) [15:15] let me know how wrong I am please kthx :) [15:15] less letting you know how wrong you are, more letting you know you have more work to do in your future slack time. :-) [15:21] hatchm I'm having to get my laptop on raring working. almost done, I think... [15:21] hatch ^^^ [15:21] ooo living on the edge :) [15:21] :-) on the laptop, at least [15:22] I usually wait a couple months so others can deal with the driver issues lol [15:23] I don't think I can upgrade my laptop to 13.04 because it was originally installed with wubi and 13.04 doesn't have wubi [15:23] I could be wrong with that though [15:26] I should probably ask [15:26] hi frankban, i'm having trouble accessing a gui charm launched on juju-core on ec2 [15:27] frankban: i set the juju-gui-source but juju status still shows it as a cs: charm [15:27] frankban: password of 'passwd' is rejected. has something changed i missed? [15:29] frankban: nm, i see admin-secret now works... [15:29] bac: auth is now enabled on juju-core api, you should use your admin-secret [15:29] heh [15:30] frankban: i'd accidentally cut-n-pasted the secret-key, not admin-secret. thx. [15:42] Thank goodness I don't have that 53 mile commute anymore. http://ubuntuone.com/3OEL5eifsK690J9p0dRzX7 [15:43] Makyo: awesome! [15:43] overnight? [15:43] In the last six hours. [15:44] Still coming down like crazy., [15:44] right on! I wish I was there [15:46] frankban: to answer your question from earlier: no, resolved doesn't return anything (other than an error indicator) and the front-end doesn't need it to, either [15:47] benji: cool, thanks [15:49] If anyone pinged me in the past 15 minutes, I didn't see it, so please re-ping [15:49] 20, even [15:50] and meanwhile... [15:50] jujugui, please update kanban board. I'll review soon [15:50] hatch, I'm in guichat. Things are kinda sorta working. :-/ [15:58] jujugui call in 2 [16:01] Makyo, starting [16:15] hatch, https://plus.google.com/hangouts/_/02bb45411739e441fe107c9f66e2a8cc36ba4ba7?authuser=0&hl=en ? [16:22] gary_poster: are the bugs I'll be filing "High-Priority Maintenance"? [16:22] benji, good question. at this stage in the game they are all kind of flowing together into "things we have to fix in order to deliver." Yeah, I think that sounds right [16:23] k [16:23] benji, but please put the high priority marker on them in the card [16:23] k [16:23] * gary_poster will think of some better name for that lane in the future :-P [16:23] or someone will tell me of one :-) [16:24] gary_poster: I suggest "Story Alpha" [16:25] benji :-P there actually is intended to be some meaning to the "maintenance" part of the name :-) [16:25] Just ribbing you on the "Story 1"/"Story A" thing [16:25] :-) [16:28] rogpeppe, you available for a quick call? I was writing you another novel, and I think talking might be better :-) [16:29] gary_poster: definitely. give me a few minutes and i'll ping you [16:29] thanks rogpeppe === deryck is now known as deryck[lunch] [16:41] rick_h_: avail to chat now? [16:43] hatch: sure [16:43] ok in guichat [16:45] Really need to shovel the deck. Back in a few. [17:00] gary_poster: ok done the calls so whenever you want to continue with that relations stuff [17:01] cool hatch. My call now is cancelled. I'm kind of hungry though. Maybe call in 20? [17:01] sounds good [17:01] cool tty then [17:01] restarting the CI - it hung [17:03] gary_poster: jcastro ping'ed me asking about the gui on the tablet and the uistage url is showing the circles but not the squares/etc. [17:04] gary_poster: did that get working? He's asking if yuo could email a status update for the gui on the tablet to robbie and cc him. [17:04] rick_h_, tell him to reload, and give him Makyo as contact [17:04] rick_h_, sure will do. [17:04] gary_poster: thanks [17:04] Makyo, I'm going to forward your email to Kapil, plus a note about the reload for circles -> squares [17:05] gary_poster, ack. Hoping it's just a cache thing. [17:05] hah just got as text to nvm it works!! [17:05] gary_poster: Makyo ^^ [17:05] :-) [17:08] rick_h_, sent Jorge and Robbie the email anyway. Makyo, cc'd you. [17:08] rick_h_, Hah, yeah, probably the cache, then. Should be fine on reloads, since the cache will be populated. [17:09] cool, thanks guys [17:09] Should look into gzipping the svg assets. [17:09] sounds like they're having fun at the conference [17:09] :-) cool === matsubara_ is now known as matsubara-lunch [17:17] semantic difference between go- and python-back-ends: the go back-end resolves the unit as a whole while the python back-end can resolve just a unit or a unit/relation pair [17:52] hatch: https://codereview.appspot.com/8679045 is the up and ready for destruction :) [17:53] haha alright do I need to qa it? [17:53] hatch: hmm, you can. You have to hit each url manually. I can give you a list [17:53] oh actually I need to add the commit message on there still doh === matsubara-lunch is now known as matsubara [17:54] hatch: I think just the code for now will be cool and QA once the missing bits get added. [17:54] alright sure [17:54] it's a bit of a regression for us atm. [17:54] but urls ftw [17:55] CI hung again it looks like it was only able to requisition a single machine === deryck[lunch] is now known as deryck [18:09] Makyo: Might you have time soon to help look over my charm branch? [18:21] rick_h_: done - few things to chat about in the review [18:21] hatch: looking now [18:21] bcsaller, These docs are getting out of hand, but I'm at a good stopping point now. [18:23] Makyo: up to you. lp:~bcsaller/charms/precise/juju-gui/sandbox-charm/ has the current. I'm testing my latest updates now. I'd also be happy to walk you through the changes [18:24] hatch: did yuo want to chat then? [18:25] sure [18:26] guichat [18:26] rgr [18:34] bcsaller, looking through diff over lunch, give me a few, then I'll QA best I can. [18:36] Makyo: happy to answer questions, there is both new code and new features [18:41] gary_poster: back from lunch yet? :) [18:42] hatch, sorry, yeah, been back. [18:42] no problem, I'm at a good point to get back onto the relation stuff if you are available [18:43] hatch, cool,let's do it. see you in guichat in 2 min. [19:01] bcsaller, looks fairly straight forward. What backend should I test first? [19:02] Makyo: sandbox: true, staging: true is the config set we want right now [19:02] Makyo: you're the first to look at the new backend code, that looked sane to you? [19:08] bcsaller, yeah, I think so. I wasn't expecting the chain/merge stuff, but it makes sense when I read throughit. [19:09] Makyo: yeah, composition patterns rather than inheritance [19:09] bcsaller, yeah, makes sense. [19:15] * bac reboots [19:44] bcsaller, Testing's going okay so far, minus EC2 problems. [19:49] If I was doing core dev right now I think I'd be working on a Docker provider, local provisioning of units should only take seconds and the way it is now is driving me crazy [20:09] A pox on all self-signed certs. [20:09] I'm getting "SyntaxError: JSON.parse: unexpected end of data" now, though, which is new. [20:10] I haven't seen that one [20:10] I'm trying to figure out how to make firebug break on errors, so I can figure out what it's talking about. [20:11] Oh, derp. [20:20] Makyo: any update? [20:21] bcsaller, Trying to dig into the uglified source. [20:24] bcsaller, from the look of it, sandbox isn't being set in config.js [20:24] Making sure it's not just a cache problem. [20:25] Yeah, not cached. [20:25] So it's trying to connect to pyjuju's ws, which isn't running. [20:27] Makyo: it should be set, that would be a real issue, I haven't seen that not work. The improv server doesn't run in a sandbox though, the ws faked on the client [20:28] * benji records in the Big Book of Things About Working from Home... Danger 27: chatty pest control guy [20:29] bcsaller, yeah, not seeing a sandbox attribute in config.js [20:30] Should note that I forgot to specify the juju-gui-source and changed it later with juju set, I'm sure that regens the config.js file, but I'm chcking [20:34] benji: chatty pests are worse [20:34] lol [20:35] benji: i really need a pest control guy. we've got a mouse who has eaten/contaminated about $200 of food last week. [20:35] bcsaller, damn, config on install wasn't applied, bet it was a typo on my part. Trying with set. [20:36] Makyo: I might properly have triggered the change in the case you're looking at either though, been checking into it [20:37] bac: yow (I think the previous owners of this house (the ones that had it forclosed and taken away from them) had a bad rodent problem, judging from all the mummified ones I found when I tore out the kitchen; makes the fees seem worth it) [20:38] i got a good whack at him the other night with a broom. missed but he ran away and hasn't been back yet [20:38] of course there is no food out for him now. that may have more to do with it [20:39] Isn't this the Future with a capital 'F'? Why do we still have rodent home invasions? [20:40] Makyo: I pushed an update that may help in that case, but calling set might do it as well. [20:44] bcsaller, Alright. This thing's being stupid anyway. [20:50] gary_poster: the config/settings bug is marked as blocked. any info? [20:51] bac, yes, Roger might magically fix it in the watcher between now and tomorrow evening. If he does, then this card will mean work A, otherwise it will mean work B. [20:51] Had to resolve a unit to destroy a service due to hook failed, then had to terminate the machine, now the whole environment's hosed. [20:51] So, since we can do either after the go env is frozen, we are waiting for roger, bac [20:51] Or, maybe just status? [20:51] gary_poster: "A" being adding it to serviceInfo handler? [20:52] bac, right [20:52] ok. [20:54] * bac does QA while waiting for two reviews, he said in his most aggressive passive voice. [20:54] heh [20:54] jujugui ^^^ [20:56] Running into https://bugs.launchpad.net/juju-core/+bug/1168155 again, sorry this is taking a while bcsaller [20:57] _mup_, not a fan? #1168155 [20:57] :( [20:57] Makyo: I understand about it taking a while, I appreciate you looking it. I fought with this over the weekend and each iteration takes far too long [20:58] Makyo: you might change the hash on your ec2 control-bucket or delete the s3 bucket directly [20:59] bcsaller, Yeah, starting over anyway. [21:00] gary_poster: ping [21:00] hey rogpeppe [21:01] gary_poster: i've just discovered a slightly annoying wrinkle [21:01] gary_poster: to the service config watching stuff [21:01] bcsaller, "each iteration takes far too long": I want to explore that with juju core for 13.10 [21:01] rogpeppe, uh-oh? [21:01] gary_poster: which is that watching the settings collection doesn't actually tell us all the current values of the settings [21:02] gary_poster: because the settings collection only holds values that have been set explicitly [21:02] as opposed to defaults? [21:02] gary_poster: yup [21:02] hm [21:02] rogpeppe, I don't think that's too bad [21:02] gary_poster: so... you do *theoretically* have all the info you need to do the defaulting [21:02] rogpeppe, right, we always work off of a model [21:03] and the model isinstantiated with data from call to get service info [21:03] which I am pretty sure has the settings [21:03] gary_poster: you can get the defaults from the charm metadata [21:03] * bac -> dogwalk. bbiab. [21:04] rogpeppe, ok, we already have that too, and mutate the service on the basis of it for the endpoints [21:04] so we can set up the initial values from the charm at that point [21:04] gary_poster: so the question is: is it easier for me to do the "right thing" than for you to cope with the wrong thing at your end? [21:05] rogpeppe, where the right thing is to include defaults? [21:05] gary_poster: yeah [21:05] gary_poster: although... i'm not totally sure [21:05] gary_poster: it's actually unnecessary network traffic [21:06] rogpeppe, if you are contacting me, I suspect it is not trivial on your side. I think it will approach trivial on our side, and I don't mind the semantics in the abstract. Not 100% convenient, but not horrible. I wouldn't mind if you only sent us config changes, in fact, and might think about this approach as an approximation thereof [21:07] gary_poster: ok, that sounds good [21:07] cool thanks rogpeppe [21:07] gary_poster: i'll go with that for now [21:07] cool [21:07] gary_poster: the implementation is basically done, i think - "just" the tests to go [21:07] rogpeppe, :-) great! [21:18] bcsaller, ip-10-212-73-147:2013/04/15 21:17:21 INFO worker/uniter: HOOK ValueError: Unable to use staging with go backend [21:19] Makyo: thats true :) [21:20] no improv and we only implement the pyjuju protocol for sandbox currently [21:20] bcsaller, alright. Should I test something other than staging: true, sandbox: true? [21:21] I suppose the next would be to verify a normal python install on ec2, I've done it locally but not there yet. If that works then the charm is still valid for its primary use-case [21:21] Makyo: ^^ [21:22] bcsaller, alright, will do. [21:22] thanks [21:34] bcsaller, looking good. [21:41] Makyo: thank you :) [21:52] bcsaller, I have to run. where are we with charm? [21:52] Checks out on python and go (before I messed it up) [21:52] gary_poster, bcsaller ^^^ [21:53] Makyo, great! You tried the new features as well, with sandbox and the "look someplace else" functionality? prob hard to qa that last one [21:54] gary_poster, just straight deployments, but I can. [21:54] Makyo, thank you. sandbox should be relatively easy to test (within constraints of what sandbox privides [21:55] gary_poster, alright, on that one now. [21:56] Makyo, the other functionality might be best tested by verifying that deploying it fails, then logging in and verifying that no ppas are installed in etc/apt/sources.list.d. [21:56] maybe you or ben have better idea [21:56] running [21:56] back later [21:56] must write email about charm to is tonight [21:57] gary_poster: I think we can use the branch for now, and then do another rev once its code reviewed [21:57] cool thanks bcsaller. shoot me a mail with branch pls [21:57] ttyl [21:57] * bcsaller nods [22:06] am I supposed to be able to create a relation from wordpress to puppet in the UI? or do I need to do something fancy to get the subordinate relations? [22:07] gary_poster, bcsaller Good to go on sandbox. [22:07] :) [22:07] ohh [22:07] do we support deploying subordinates in the sandbox? [22:08] I don't get the subordinate UI in the sandbox after deploying puppet [22:08] hatch: if we don't its a bug [22:09] alrighty I'll file [22:09] hatch: I think we previously set a flag on the incoming delta that might not be happening in deploy. I expect there is that and not joining peer relations that might be issues with sandbox deploy at this time [22:10] alright - so because of that I can't create a subordinate relationship to QA [22:10] but normal relationships are good [22:13] https://bugs.launchpad.net/juju-gui/+bug/1169350 [22:14] Alright, dogwalking time. [22:14] hi Makyo, still here? [22:14] bac sure, for a minute or so [22:15] Makyo: sorry to keep you. doing QA and wondered if there was a bug for the trail of artifacts left on the canvas when dragging around a service? [22:16] bac, not seen that yet, but dragging a service only updates element attributes, not create new elements. Can you check in the DOM through firebug/JS tools? If there are extras, definitely worth a bug. [22:17] Makyo: will do. have a nice walk. [22:17] bac, cheers [22:17] hatch: updated per your review. Working on jcsackett's now. If you get a sec can you toss a LGTM or even a LGTM with the following XXX so I can finish/land it later tonight if possible? [22:17] hatch: in looking at it, removeUnit checks service.get('is_subordinate') but I think for service its just get('subordinate') (no 'is') [22:18] bcsaller: if you get a chance before EOD https://codereview.appspot.com/8735043/ [22:18] hmm [22:18] hatch: I can do it now, the charm is 'good enough' for now [22:18] rick_h_: sure thing lemme take a peek [22:23] rick_h_: which patch set should I be looking at? [22:23] 4? [22:23] hatch: yea, latest. [22:23] alright [22:23] hatch: no hurry. I won't get back to it for a couple of hours [22:26]