[00:00] hatch: I don't think we have an asset for that ellipsis yet, so maybe you can crop it out of the mockup for now [00:07] hatch, for the source view, "Nice tests. LGTM with additional text I mention. qa good. Thank you!" [00:08] huwshimi, I think I don't have any bundle tasks for you today. Are you choosing one from your backlog? [00:08] gary_poster: I'm just adding the icon to the bundle charm details panel and trying to get showing/hiding working and then I'll move to the backlog [00:08] gary_poster: thanks :) [00:09] huwshimi: damn why won't you guys let me use UNICODE!!!!!! … … … … … [00:09] lol [00:09] huwshimi, sounds perfect. thanks! hatch, notice huwshimi is trying to do that part of what we talked about, but there's still other bits [00:09] hatch, huwshimi, running away. ttyl and thanks! [00:10] enjoy! [00:10] gary_poster: Have a good evening :) [00:10] hatch: You can use unicode if you can make it look like the mockup :) [00:10] lol [00:10] deal! [00:10] hatch: (hint, you won't be able to) :) [00:10] ok I'm going to be taking over from where you leave off on that panel so don't forget to push your changes up :) [00:10] haha [00:11] hatch: Yep, nearly done here [00:15] oh haha cool [00:20] hatch: What does scene inside events mean "events: { scene: {"? [00:21] hatch: This is inside topology/bundle.js [00:23] huwshimi: event bindings [00:24] bcsaller: how does that differ to our usual bindings which just live inside 'events' without the scene? [00:24] they are mapped to d3's event system [00:25] part of the topology system, the docs on d3-components go over this, should be in the docs dir [00:25] bcsaller: Ah I see [00:25] I'll take a look, thanks [00:30] bcsaller: So how do I get the event from a scene event? [00:30] huwshimi: you mean in the handler? [00:31] often you don't need to, 'this' is the DOM element and the 1st arg is the bound model that was triggered, but if you need the event d3.event should be correct [00:31] bcsaller: yeah, I want to do a "[event].halt()" [00:32] d3.event.halt() then [00:34] bcsaller: Forgive my ignorance, the d3 in that case is the first function parameter of the handler? [00:34] d3 is a global [00:35] d3.event is also global, js is single threaded so there can only be one event at a time [00:35] bcsaller: Heh, yeah, I just figured that out :) [00:36] bcsaller: Thanks for that, this is working now :) [00:36] great [00:46] hatch: That branch is now proposed, so once that lands you are good to go. [11:12] hi rick [11:12] rick_h_: i mean [11:14] rick_h_: when you're around could you review the changes that benji asked for in https://code.launchpad.net/~bac/charmworld/update-branch/+merge/190229 . he's marked it approved but i think the lander will want the new revision approved too. on interesting change is adding 'lint' to 'make check' [11:47] bac: looking [11:50] bac: cool approved [11:50] bac: the new landing from jenkins doesn't care about the new revs ime but cool to peek [11:51] bac: marked approved and watching staging whee [11:52] thanks rick_h_. i'll write a commit message and then it should go. [11:52] oh oops, hah [11:52] done [12:12] bac: hmm, CI isn't picking it up. Are you a member of juju-jitsu? [12:13] looks like you are [12:14] abentley isn't around yet to ask about it. He was changing stuff with it the other day [12:37] rick_h_: do you have any insight into the lander? or just been watching the MP? [12:38] bac: I was watching jenkins [12:39] bac: the way things are working now jenkins pulls the branch on its own [12:39] bac: so since it's not loading it, I'm guessing it's due to recent changes. It was working for my branches earlier this week [12:39] abentley should be around in the next hour to ask about it. [12:40] he'll probably want to test something :) [12:40] rick_h_: ok. abentley said it runs every minute so we'll just have to wait [12:40] for him to appear and enlighten us [12:40] yea [12:41] gary_poster: around for one-one or bypass this week? [12:41] rick_h_, sorry trying to wrap up [12:41] gary_poster: np [12:54] hi abentley, when you have a chance could you look into why the lander is not picking up my branch, reviewed and approved at https://code.launchpad.net/~bac/charmworld/update-branch/+merge/190229 ? [12:54] rick_h_, ok now :-P [12:54] benji, will be late :-/ [12:55] gary_poster: np [12:55] bac: I will. Thanks for pointing it out. [12:55] thanks [13:07] gary_poster: added "Charm details link not working correctly" to urgent [13:07] frankban, thank you [13:09] bac: That issue turned out to be a config issue where a list was expected but a string was supplied. [13:10] bac: The previous code was using "if x in y", so it didn't notice that y was a string, not a collection. [13:14] bac, benji, abentley: could one of you review this mp: https://code.launchpad.net/~adeuring/charmworld/use-charmtools-tarball/+merge/190361 ? [13:14] adeuring: I'll take a look. [13:14] thanks [13:19] abentley: so is the lander working on a queue of branches? i see gary's is up now. [13:20] bac: It's not an explicit queue, it just finds the first thing it can land and tries landing it. [13:21] abentley: oh, i see my MP is marked merged. cool. [13:35] abentley: does the lander update the charmworld revno on staging automatically now? [13:36] bac: yes. [13:36] oh, cool. i was just about to do it manually. hey is the process documented somewhere so i can quit bugging you with dumb questions? [13:37] abentley: ^^ [13:38] OTP [14:00] benji: Your canary work is preventing download-cache from being updated, which means that adeuring's branch cannot land: http://162.213.35.27:8080/job/charmworld-autoland-lxc/28/console [14:01] :( [14:03] bac: So, the first question wasn't dumb at all-- I had misconfigured the lander as part of the change I announced yesterday. That's a one-time thing. [14:03] benji do you need to address that or can we talk now? [14:04] bac: What sort of documentation would be helpful? [14:04] gary_poster: I don't know. abentley: do you want me to address that issue? [14:04] benji: Could you please discuss it with adeuring, figure out a solution, and then one of you can implement it? [14:05] abentley: sure [14:05] gary_poster: I'll ping you when I'm available (or we can postpone) [14:05] anthonydillon: sorry for the delay. If you get back around let me know and can walk you through it [14:05] gary_poster: our call is now. or are you talking to benji first? [14:05] benji: Thanks. [14:05] benji, ack [14:05] benji: well, I'd like to land my last branch sooner or later ;) It's not a matter of minutes, of course [14:05] bac, are you available now? [14:06] i am [14:06] bac, let's do it [14:06] abentley: does the lander update a long-lived checkout or build a new one every time? [14:06] benji: It updates a long-lived checkout. [14:06] :( [14:07] It would be better to do fresh checkouts every time (not just for issues like this, but it's better hygiene in general). [14:07] benji: Makefiles wouldn't be necessary if we didn't use long-lived source trees. [14:08] * benji ponders that assertion. [14:08] benji: If you were always building from scratch, you wouldn't want a dependency system, you'd want a traditional script. [14:09] true, but no one ever *always* builds from scratch [14:11] benji: So ideally, build-from-scratch and incrementally-update-and-build should both work. We could test both. I'm not convinced one is better than the other. [14:13] yep both should definately work; build-from-scratch must work, otherwise we just have software that only exists in a usable form on particular had disks scatered across the planet :) [14:14] from my perspective incrementally-update-and-build is a nicity that -- given the limitations of make -- can never be truely achieved [14:14] benji: But incrementally-update is trickier, so I think that's the one worth testing. [14:15] I disagree. From-scratch must work, so it must be tested. [14:15] (if it's not tested, it doesn't work, and all that) [14:17] benji: I disagree. We have a vcs, so we can incrementally build every revision starting with revno 1 if necessary, so build-from-scratch is no more necesary than incremental. [14:18] I've already stated my rebuttle to that argument, and I need to work on the issue at hand. [14:20] benji: I don't see a rebuttal. I just introduced a new argument. [14:20] I may have misunderstood. We can talk about it once I've addressed the immediate issue. [14:28] gary_poster: is there a url flag so we can see Bundles and Ice-cream? [14:32] bac: http://staging.jujucharms.com/charms/precise/juju-gui/json looks good and complete! [14:32] bac: so I think we'll call this case closed with your branch fix? [14:33] benji: I think the syntax on "then : bzr up download-cache" may be malformed. I think the colon is preventing execution of the "bzr up". [14:33] cool rick_h_, OTP [14:33] bac: k [14:33] guihelp: I need 1 review + 1 QA for https://codereview.appspot.com/14439054 . Is anyone available? thanks! [14:33] abentley: hmm, I'll take a look at that [14:33] frankban: I can [14:34] hatch: thanks [14:34] frankban, had a good conversation about quickstart with rick_h_ . should share with you [14:35] maybe after our daily call [14:35] gary_poster: sounds good [14:36] rick_h_: just looked, yep it does look good. [14:39] abentley: I think you're right. [14:40] benji: I know that colon is a do-nothing in some contexts. e.g. ": echo foo" does not echo. [14:41] benji: , abentley: yes, that seems to be the problem. I'll remove the ':' in my branch. [14:41] yeah, it is for evaluating expressions for their side-effects [14:42] that's how I'll make my millions: an easy way to test make files [14:44] so... this works in dev because pip actually downloads the file from the Internet; we should disable that so dev matches the IC/prod enviroment [14:44] gary_poster: ok, I'm ready for a call whenever [14:45] thanks benji. arranged it in calendar [14:46] k [14:48] I retract that. pip didn't download from the internet. [14:52] frankban: still spinning up core to qa - will hopefully have it all tested in 20minss [14:52] hatch: cool [14:54] ORDER PLACED: Oct. 1 2013 DELIVERY ESTIMATE: Friday Oct. 25 2013 - Friday Nov. 1 2013 by 8:00pm [14:54] NOOOOOOO amazon why do you hate me so [14:55] hatch: lol [14:55] hatch: gotta go prime baby [14:55] haha I'm guessing the book isn't in stock [14:55] that's even slow for their usual Canadian dogsled delivery team [14:56] I was thinking of going prime, but I can't remember the last time I paid for shipping [15:03] abentley: my branch failed again. Do I understand the report right that lint errors are considered bad too? [15:06] adeuring: that was a recent change bac included [15:06] frankban: I'm hoping it's still switching to your branch - but it's been rejecting my connections for a while now [15:06] adeuring: make check now checks lint as well [15:06] and of course [15:06] there it goes [15:06] lol [15:06] hatch: it takes a while [15:06] hatch: I ran into that yesterday with Makyo's branch [15:06] rick_h_: thanks. I was just not sure if missed something [15:06] adeuring: naw, we're shifting the sand beneath your feet :) [15:07] ;) [15:07] adeuring: That's how it looks. I'm just running a make target. [15:07] adeuring: yes, that just landed. i should've announced it but it came out of a discussion we had yesterday, after your EOD [15:07] adeuring: I'm just running "make check". Whatever's in that target will run. [15:07] hatch: on ec2 "juju debug-log" can help. if you are using the local provider, then you can just "tailf ~/.juju/local/log/unit-juju-gui-0.log" [15:08] ahh I should do that from now on [15:08] yeah using local provider [15:09] done [15:09] sorry forgot to type LGTM ;) [15:09] :-) [15:10] hmm the card I was working on is gone [15:10] ohh huw stole it [15:13] there is a new fitbit out that allows you to check the time of day, right on your wrist! revolutionary! [15:13] bac: HAHA I was JUST thinkign the same thing [15:13] lol [15:13] i sure hope they patent that wrist-mounted time display [15:14] I think it would clash with my watches [15:14] hatch: How many watches do you wear at a time? [15:15] one - but none are digital or silicone :) [15:16] rick_h_: You said bootstrap-dropdown isn't working. I remember the tools dropdown working. How long has it not been working? [15:16] abentley: oh, I just know we had JS that was never wired up. I didn't realize that it was working tbh [15:16] * rick_h_ shrinks my window really small [15:16] gary_poster: should I add a card in urgent to remember us to make a new charm release after we release the GUI? [15:17] frankban, yes thank you [15:17] abentley: no idea. If it's meant to work then I'll try to track history on it and see why it's 404'ing then. [15:22] rick_h_: Not very important functionality, and if it's been disabled for a while, no rush to restore it. [15:22] abentley: I think this was work when huw moved the ui to be with ubuntu style guidelines [15:23] abentley: all of bootstrap is removed so removing the file. [15:31] bcsaller: i see export is now using blob urls. did that just happen? sadly safari cannot handle them. [15:32] bac: or benji got a sec for a tiny review please? https://code.launchpad.net/~rharding/charmworld/jc-bundle/+merge/190396 [15:32] rick_h_: yes [15:32] ty [15:34] gary_poster: is where huw left the bundle topology details popup where we want to leave it for now? [15:35] hatch, hey. yeah, meant to highlight that to you. I think we could add next/prev, but lower priority. later [15:35] ok cool - so I'll do the charms panel now? [15:35] bac: I believe that has been there for some time. There are very few options for generating downloadable files in the browser. When did safari make the supported list? [15:36] bcsaller: it hasn't but there have been discussions. [15:36] bcsaller: it was more of a sad-face comment than something than a bug. [15:37] bac: bcsaller there was a thread about supporting it recently though I don't think it made the official list [15:37] safari that is [15:37] bac: with the server doing exports now we can generate a url there and get more support, but with fakebackend and so on I think the blob stuff is still a good option [15:39] hatch: those icons in the bundle looking purdy [15:40] rick_h_: I know right? For the amount of crap I give you about the tokens it was sure easy to implement :P [15:40] but don't tell rick_h_ I said that [15:40] it'll go to his head [15:40] hatch: :) yea hopefully you're finding that once you get past learning the 'rules' the browser stuff isn't as horrible as it seems [15:41] hatch: if not then shush no one asked you anyway :P [15:41] lol [15:41] did you get a chance to look into that attribute issue last night? [15:41] rick_h_: done [15:41] I didn't :/ [15:42] hatch: no, I was busy trying to implement chrome 'add to homescreen' to my own app. :P [15:42] bac: ty much kind sir! [15:42] rick_h_: hah - hard to do? [15:42] hatch: no, it just hates me and won't give me the pretty icon on the home screen I wanted [15:43] hatch: and then started redoing the chrome extension without any framework bits [15:43] ahh cool cool [15:43] so no, I didn't play with the ATTR setter thing at all. off time and all that :) [15:44] haha [15:44] those triangles still look weird to me [15:44] hatch: +1 [15:44] http://comingsoon.jujucharms.com/precise/juju-gui-77/#bws-related-charms [15:44] on the details page they even look worse [15:45] they looked good in the mockups [15:45] I don't know why not here [15:45] hatch: can you find the mockups? I was looking trying to find some wireframe to see what we're missing but all the wireframe/visuals i can find in google docs have the * [15:45] hmm [15:46] hatch: because I'm kind of with you. I've never been a big fan but I don't remember it looking this 'off' [15:48] rick_h_: https://drive.google.com/a/canonical.com/#folders/0B7XG_QBXNwY1NEtGaHJYZGM4enM the button states png [15:48] I think it's because on the bundle token it had more 'stuff' to fill in the whitespace [15:48] hatch: yea, and not really in context for things like the related charms/etc [15:48] many things look ok in isolation, but in context... [15:49] hatch: the background colors of the tokens isn't right either. [15:50] jujugui call in 10! [15:50] hatch: oh nvm [15:50] hah! [15:50] jujugui call in 10 [15:50] lol [15:50] hatch: we don't have the hover color change [15:50] Darn :) [15:50] Weekly Numbers: Makyo - 3; gary_poster - 1 [15:50] my shining moment [15:50] :P [15:50] :-) [15:51] hey benji, did you have the call with mark b? did he get back to us? I've been a bit swamped [15:51] gary_poster: he never got back to us (that I saw) [15:51] ok [15:52] too bad [15:53] hey jujugui, somebody give the second review of bcsaller's https://codereview.appspot.com/14485046/ so he can land it! [15:53] on it [15:53] thanks hatch [15:54] Makyo, that branch should replace some but not all of my branch that I shared with you yesterday evening . We can talk about details later if you want [15:54] gary_poster: to QA just drag the deployer file? [15:54] gary_poster, alright, just going through the annotations stuff now. It is annotation, but I need to make sure which ones need changing. [15:55] cool thanks Makyo [15:56] hatch, I think verifying that export has x-y annotations again is the core qa element. bcsaller, am I right? [15:56] hatch, drag deployer file might not yet work still until we have my change--or at least my change makes it more reliable [15:57] gary_poster: actually its a little trickier than that for a reason that caught me, the gui-x/y didn't make it back to the client models before, but even once it started working I couldn't tell at first because we delete them as soon as we apply them in the draw code [15:57] yeah it fails [15:57] I spent a while on that in the "this should be working stage" [15:57] can't dd the wp-deployer.yaml [15:57] that worked for me last night, testing [15:58] and the tests import that all the time [15:58] Object {err: "[object Object]", DeploymentId: undefined} [15:58] ohh [15:58] the way we handle next in the go sandbox is broken without my branch [15:58] manage.jujucharms.com doesn't have mysql-26 or wordpress-15 [15:58] jujugui call in 2 [15:59] hatch try benji's [15:59] might work [15:59] but gui will always miss the first delta returned to it until my branch lands [16:04] rick_h_: i just put both of our faces on the release card. first one to it tomorrow wins. [16:04] bac: cool, see you at noon :P [16:04] :) [16:19] hatch: the dnd import has issues with the charmworldv3 flag, works without, seeing if there is a simple resolution [16:24] cool thanks - yeah that should probably be fixed heh [16:28] hatch: it appears to be an issue with not having backfilled charms in api3, if I set the revision in the import to current it works fine, but it gets 'no such charm' for older version in the bundle [16:29] ohhh ok cool [16:29] I'll qa without the flag [16:32] bcsaller: yep, benji is working on that so hopefully made better soon [16:33] bcsaller: so I drag it to the canvas, it lays out in a certain 'layout' then it snaps to a different layout on the next delta [16:36] bcsaller: so I add 3 services, relate them, lay them out, export, refresh, import - layout is different [16:37] does that mean that QA is no good? [16:38] hatch: There is an issue with export, its using the client db, and we delete the position annotations, when we apply them. It should either use an env call to get its YAML or we can't delete env. Either of those are non-trivial changes [16:38] ok so as long as the export file has the annotations we are good? [16:39] hatch: yes, but it won't the way its setup now. [16:39] the server and the client both get the annotation [16:39] thats what this branch fixed [16:39] but then the client deletes the position [16:39] so more changes are required [16:41] oh ok so there is no way to qa this [16:41] because the export doesn't contain the xy annotations [16:41] the export should have position, that isn't what I was verifying before [16:41] hmm [16:41] maybe that should be another branch, but I could do it here [16:42] yeah it's definitely not in there [16:42] yes, are you reading what I'm saying? does it not make sense? :) [16:42] hatch: lol anthonydillon got bit by double dispatch. "Yep, we know about that one" :) [16:42] lol [16:43] itit gets set on the client, applied and then removed by the client [16:43] and then we use the client db for the export [16:43] which no longer has it [16:43] rick_h_, Good to know :) [16:43] ok then because you just said it should have it haha [16:43] so I was confused [16:44] it does have it, but then it removes it when used. so it has it only till it draws it in the right position [16:44] no the file does not have the xy positions [16:44] ahh, right [16:44] which makes sense from what you're saying [16:44] I think we were just talkign around eachother haha [16:44] yeah, the wire protocol was broken before [16:48] hatch: so this is a real fix and I'd like to land it. The solution to the other problem should be another branch. Ideally that would be the server doing the export but that really means core support. Because we don't have that I need to look at a) not deleting the position on application (which might be a big change) or b) polling the position on client generated export (which is best but another branch) [16:49] I guess that means the export deployer call can take an optional topology to ask for position info during the export [16:50] it's lgtm'd [16:50] hmm [16:51] well we really need a way to export the xy's - without it they won't be able to make layouts for the bundles [16:51] I guess I don't understand where we are deleting the annotations on export [16:52] hatch: not on export, on draw [16:52] but they aren't in the export either [16:52] how could they be, they were deleted from the model on draw ^^ [16:53] so we delete them from the model, waiting for the delta to re-populate? [16:53] I can explain it in a hangout if you'd like [16:54] yeah I think that's best haha [16:54] too much typing [16:54] :) [17:01] allll cleared up! [17:03] bcsaller, (b) is very high priority then, yeah? Because exporting x/y is very important [17:03] sounds pretty easy? [17:04] gary_poster: it is, and its my next task, should be about an hour [17:04] mm, no, middlin [17:04] ok awesome [17:12] hmm the inspector no longer shows the unit count on comingsoon [17:12] could someone else confirm this for me? [17:12] deploy a service with 10 units, the inspector after deploy will show 0/0 [17:13] shows 0 until I set 10 then it updates to 10 running [17:13] now I can't change it :/ [17:14] the units field got disabled? [17:14] so after you set the number of units, the field gets disabled [17:15] hatch: only doing it on the first one deployed to me [17:16] hatch: once I close it and deploy a new service the count works and doesn't show 0 [17:20] hatch: I'm qa'ing on juju trunk and not seeing it, but ther'es already a charm in the env (the gui charm) [17:20] hmm [17:21] this is a critical bug [17:21] Makyo: before release we need to fix ^ [17:21] oh wait [17:21] is comingsoon running an old version? [17:21] no idea [17:21] nope [17:22] frankban: your branch landed right? [17:22] hatch: yes [17:22] the handle for the side bar is under the charm details on comingsoon [17:22] hatch: the minimize tab? [17:22] I wonder if comingsoon needs a 'make clean-all && make' [17:22] yeah [17:22] hatch: yea, ther's a bug for that [17:23] hatch: filed that one the other day [17:23] frankban fixed it [17:23] rick_h_: do you know how to log into comingsoon and re-make it? [17:24] hatch: no, bac and gary_poster have access [17:24] hatch: what's up? [17:24] can you clean-all and make comingsoon ? [17:24] hatch: sure [17:24] it doesn't appear to be running the latest code [17:24] thanks [17:24] hatch: https://bugs.launchpad.net/juju-gui/+bug/1234321 [17:24] <_mup_> Bug #1234321: the minimize tab should have a higher z-index than the inspector charm details [17:25] hatch: that's not what he did, is this what you mean? [17:25] rick_h_: yes - frankban fixed that in his last branch [17:25] hatch: oh, didn't realize that. [17:25] rick_h_: ah, didn't know there was a bug [17:26] frankban: all good, updating now. Thanks for the fix! [17:26] yeah comingsoon is very broken compared to local trunk [17:26] but it does disable the unit scale still [17:26] hatch: yea, I'm filing that bug as my QA bug :) [17:26] haha deal! [17:27] make a critical card too plz [17:27] we can't release with that [17:27] hatch: rgr [17:27] hatch: noooooo https://bugs.launchpad.net/juju-gui/+bug/1236427 [17:27] <_mup_> Bug #1236427: Scale up input stops working after units are added/removed [17:27] frankban: beat me to it [17:28] lol [17:28] :-) [17:28] hatch: card added though [17:31] benji, https://plus.google.com/hangouts/_/55ecf5ce6f3c3ad26992f6b8bb2db173e572ffb6 if you can [17:31] gary_poster: coming [17:55] jujugui: reminder, monday is a US holiday: http://theoatmeal.com/comics/columbus_day [17:56] bac: lol that comic was funny [17:56] lol thanks bac [17:58] I don't know how much of that is true but if it is that's funny [18:09] bac: so which holiday are you going to enjoy? :) [18:10] hatch: dunno, i may work and take off the friday after thanksgiving. haven't decided. [18:10] our thanksgiving is monday [18:13] gary_poster: laptop battery died; back in a second [18:13] ok benji [18:13]