[11:12] ok, so going to be setting up amazon right now, but I plan to actually use for different regions. Is that possible with the gui? Do i just need to change my region in between each change? [11:12] er, each deploy [11:19] MACscr: you deploy the gui into a bootstrapped environment so which region you use it up to your own juju environment config [11:20] rick_h: but i need to use 4 different ones [11:20] MACscr: yes, currently. The gui needs to be *in* your environment in order to see what's in there. [11:21] man, juju is not efficient at all [11:21] MACscr: so you can use --to to deploy the gui along with another service, or you can start/stop it. [11:23] MACscr: how so? the environments are meant to be isolated and so they are. You can delpoy multiple services on a machine with --to, and you can launch/tear down the gui as you need in each environment. [11:25] i just thought the point of juju was to have a central source to manage and deploy your environments. Now i find out i have to have a separate controller for each cloud (local, maas, amazon, and openstack), then now i have to have 4 more for just amazon [11:29] MACscr: well it's using one set of charms and config across them. The same script you use to deploy one thing can work in each cloud. It's 'one' tool. The gui is a helper in that. [11:31] how do i pick what size instance it creates on amazon? [11:31] MACscr: using machine constraints https://juju.ubuntu.com/docs/charms-constraints.html [11:32] MACscr: the gui is growing support for it deploying with constraints currently (/me is working on it right now actually) [11:33] im assuming i can set defaults somewhere though? [11:33] yes, it defaults to a small I believe [11:33] right, but where can i set it though? [11:33] i want to set it to tiny [11:34] then you can use the contraints cpu-cores=1 cpu-power=0 and I think it works out to tiny. === gary_poster|away is now known as gary_poster [12:01] hi y'all! [12:02] hi gary_poster, welcome back! [12:02] thanks frankban :-) [12:12] gary_poster: howdy stranger [12:13] hey rick_h :-) [12:14] The moment of truth: how many unread emails do I have? /me waits for Thunderbird to reveal the awful truth... [12:14] gary_poster: no way, you didn't check email at all while away? [12:14] gary_poster: I hope you purchased a new computer while gone [12:15] rick_h, lol. I skimmed subjects occasionally. None of them had a "GUI is on fire" subject line, so I only read the end-of-week summaries. Otherwise I tried to be pretty strict with myself. [12:16] mm, looks like it is going to be on the order of 2700... [12:17] I'm going to ask for a quick one-on-one catchup from everyone. I have to take boys to first day of school soon though (in 20 min or so), so won't start for an hour [12:19] sounds like a plan [12:31] 2835. Thunderbird is willing to let me at them now... [12:38] see ya thurs gary_poster :) [12:38] rick_h, you out today-wed? [12:38] oh [12:38] no, meant the 2800 emails [12:38] I see :-)\ [12:38] sorry, irc delay [12:38] yeah [12:39] * benji notices that he doesn't have his IRC client running. [12:40] * gary_poster bets benji did that before he reported on it [12:45] * gary_poster takes boys to first day of school [12:46] heh [12:46] hatch: when you're around would love to chat on normalizing this constraint representation stuff. [13:11] benji, bac. Do you gentlemen have time to review our bundle work this morning? I have unfinished business with featured, popular, and new bundles that needs discussion [13:12] sinzui, benji: certainly [13:22] sinzui: you want to have a hangout in a bit? [13:23] bac, I do. I just left three comments doc [13:23] Who you calling 'doc'? [13:25] bac: I think I meant to type "in the gdoc" [13:25] * sinzui finds more coffee. [13:34] Excellent news. The power key is not attached to the unity confirmation panels. One small over-reach will power everything off in 2 seconds [13:35] heh [13:36] * gary_poster notes that Francesco's mp needs two reviewers, and then marks the emails as read: https://codereview.appspot.com/12741051/ [13:46] yes, thanks Gary, guihelp ^^^ anoyone available for reviews + qa of a python branch? [13:47] hey frankban. you available for a quick one-on-one in guichat? [13:47] gary_poster: sure [13:47] ty [13:47] frankban: yes [13:47] bac: thanks! [14:07] morning all [14:11] rick_h: it's already been normalized. you just need to use the service-constraints-viewlet.partial template [14:12] hatch: yea, I think I'm in good shape now. I've gotten things adjusted. The whole 'adjust things to Go format' was a lie [14:13] hatch: the problem is that there's no clear object/format for constraints (well there's two) and you have to know when to use/build the two [14:13] hatch: I was wishing for something that was ConstraintsTemplateFormat ConstraintsDataFormat this morning [14:14] well the getConstraints method returns the format needed for the template [14:14] hatch: right, but that's the *wrong* format for the back end which is where I got tripped up [14:15] ohh - yet another instance of too many formats for the same data :) [14:15] hatch: kinda, and there's no clear names/objects with defined structure to lean on. [14:15] just 'helpers' [14:16] morning hatch! ok, who is my next victim...benji, you around for a one-on-one? [14:16] guichat if so [14:16] gary_poster: sure [14:16] thanks benji [14:17] wb gary_poster! [14:17] thanks hatch :-) [14:37] kicking off another CI manually, hope it works else I'll have to look deeper === marcoceppi__ is now known as marcoceppi [14:42] ugh, tests don't use the config-debug or anything do they :/ [14:43] looks like CI is up and running again [14:43] rick_h: the CI tests do [14:44] hatch: yea, was just trying to write up tests for the constraints UI but it's dependant on backend and tried to switch and see how to verify which was used [14:44] hatch: ends up defined as DEFAULT_BACKEND, so tests go boom when you switch. [14:44] so maybe I won't try to land this switching the default config. Just raise it as a topic to go through [14:45] well you can specify the backend when creating your mock app instance [14:45] hatch: yea [14:45] that's true, maybe I should add a test on each one then. [14:45] it's a huge hassle - we should really create a utils method to mock that up [14:46] creating jujugui instances that is [14:49] jujugui could I get two reviews on a single line fix and an IE QA plz https://codereview.appspot.com/12942044/ [14:49] hatch, on it, got the IE bit. [14:49] thanks, it's trivial so maybe a single review is ok :) [14:49] frankban: doing qa your script has been stuck on "Sending request #10: Next..." for quite a while. is that to be expected? [14:50] frankban: nm, i was impatient. it just made progress. [14:51] bac: yes, it unblocks when the second deployment starts [14:58] * gary_poster has another call now [15:02] jujugui 2 reviews and qa please. https://codereview.appspot.com/13224043 [15:02] rick_h, on it. [15:02] ill take one [15:02] hatch: do you know if we have any UX directly on the collapsing header/etc? [15:02] hatch: directly/direction [15:06] rick_h: nope but it would be nice to have the sticky headers with the collapsing headers [15:06] but collapsing headers are trivial to do [15:06] hatch: yea, just something to get the final bits of this 'working' and releaseable [15:06] did you end up putting constraints on the top like we talked about? [15:07] hatch: trying to think if something more accordian (one or the other), collapsing, indicator it *can* collapsce, etc [15:07] hatch: yep [15:07] awesome [15:07] hatch: only way to really see/use them atm on charms with any decent config [15:07] right - I was actually thinking over the weekend that even with collapsing headers, it isn't really discoverable unless you collapse them [15:07] they could 'start' collapsed though I guess [15:07] hatch: yea, that's why I was trying to think on any UX guide on this. [15:08] hatch: because collapsing doesn't get you much really [15:08] hatch: I almost wonder if we really need the icons/tab-like view we use in the normal inspector [15:08] at least then it's consistent with iconography [15:08] and easier for discoverability [15:08] hatch: yep [15:09] good idea [15:09] fire off an email to peeps and luca about it? [15:09] hatch: and the wireframes don't really do anything on the ghost-inspector, just 'there is one' [15:09] hatch: yea, that'll work. /me forgets email sometimes in the world of immediate irc-ness [15:09] haha right :) [15:10] seen that photo of a girl chatting with a guy saying scientists should invent something where people can like talk into their phone and the other people can talk into theirs to reply? [15:10] lol [15:10] lol [15:12] I love the one going around recently. bah and now I can't find it [15:14] ain't that how it happens :) [15:19] sinzui: I'm trying to figure out where the charm version api is; can you point me to it? [15:19] * sinzui asks browser to remember what bac landed [15:21] oh, charm, not bundle. [15:21] * sinzui reads routing [15:23] benji, this is the url format to get the json for a versioned charm: http://staging.jujucharms.com/~abentley/precise/charmworld-27/json [15:23] note the - separating charm name and revision [15:24] sinzui: thanks! [15:29] jcsackett: you around? [15:29] * hatch is about to break the 62hour lock by jcsackett [15:30] going once... [15:30] twice [15:30] broken! [15:31] hatch: huh? [15:31] Unable to obtain lock held by jcsackett@bazaar.launchpad.net on taotie (process #22513), acquired 62 hours, 13 minutes ago. [15:31] <_mup_> Bug #22513: during install, in console: fsync failed [15:31] ah [15:31] that's the second time I've run into that [15:31] very odd [15:32] ugh I can't even break it for some reason [15:32] trying again [15:33] third time's the charm right [15:34] Makyo: hi, what's the current status of the go sandbox? [15:35] frankban, complete except for perhaps import/export, which I haven't touched. [15:35] rick_h: oh I bought a new backpack on Friday http://www.thenorthface.com/catalog/ca_ecom/en/sc-gear/equipment-daypacks/router.html should allow me to pack all my stuff in there and not require a checked bag [15:35] Makyo: cool, but not released, correct? [15:36] frankban, I don't believe a release has been made with it, correct. [15:37] Makyo: thanks, my next charm branch will enable go sandbox support in the charm, and make it the default sandbox mode (i.e. the go sandbox is used even if the environment in which the gui charm is deployed is pyjuju). How does it sound? [15:39] frankban, when will that land? We may want to do a release beforehand so that the user isn't swamped with notifications of errors. [15:40] Makyo: it is not the charmers charm, it will land the next time we update jujucharms.com, presumably at the end of this week/beginnig of next one [15:40] frankban, so long as some gui release of some sort happens between now and then, I'm all for it! [15:41] Makyo: yes, I'll consider a release a pre-requisite for the deployment [15:41] hatch: replies inbound. Did some, not all. reasons listed [15:41] frankban, cool, sounds good. [15:41] frankban: so I was looking at doing that in my branch as I was working between constraints and such. [15:42] jujugui current CI failure an actual code issue with deploying appflower - investigating [15:42] frankban: if we do that, I think we should change the default env in env.js and config-debug [15:42] frankban: but that breaks some tests which are coded to assume python env [15:42] frankban: was going to bring this up on the call today [15:42] Makyo: ^^ [15:43] rick_h, oh, hmm [15:43] rick_h: thanks looks good [15:43] hatch: cool [15:43] Makyo: yea, I like the idea of switching, but it's a couple of steps that should be coodinated from what I can see [15:44] rick_h, yeah, for sure, just like the switch to inspector. [15:44] Makyo: yea, the big thing here is failing tests when you switch the default env [15:44] gary_poster, https://bugs.launchpad.net/juju-gui/+bug/1214821 [15:44] <_mup_> Bug #1214821: new units added to the canvas overlap old ones [15:44] Makyo: in my new tests I did one test for each env and manually build the one I wanted [15:44] hazmat, ^ [15:46] rick_h: do you think this is a problem for the charm too? I mean, the config is generated by the charm, and my idea was just allowing sandbox=true in juju-core environments, and, when that's set, always change the env to "go" in the generated config.js [15:47] frankban: well, I just think if the charm deploys something, we should be testing/deving against it [15:47] frankban: so more of a 'good practice' like our deploy/dev environment issue [15:48] rick_h: don't submit your branch yet plz, current CI issue looks related to the constraints stuff [15:48] rick_h: that's right, so, is it a tests problem/lack for the go environment in general or just for the go sandbox mode? [15:48] hatch: ugh, rgr [15:49] bac: thanks for reviewing and QAing my branch, I believe QA took some time [15:49] frankban: just that the tests default to python env and the tests are coded to that expectation. I didn't chase down the hole far, but changing that default results in 56ish test failures. [15:49] frankban: so if the charm default is go, and we put that in front of users, I'd also like to see us change config-debug and the env default to go. [15:49] frankban: yes it was slow, but your instructions made it possible [15:49] however tests go boom [15:50] rick_h: https://gist.github.com/hatched/3de377a72e02a6c419a2 when trying to deploy appflower on rapi on trunk [15:50] hatch: any hints? e.g. will the fix of the constraints code in this branch help fix the issue? The old branch is generating bad data to the back end [15:50] hatch: k, I'll run rapi and see if my branch fixes that then [15:50] thanks [15:51] rick_h: could you please file a card for switching the default environment to go and fix tests? If we are lucky, maybe it's only a matter of explicitly specifying what environment to use and where... [15:51] frankban: definitely [15:51] rick_h: thank you [15:51] jujugui call in 9 kanban now [15:52] ty [15:54] I have to run someone to the airport in a few minutes, I'm going to have to miss the meeting today, sorry [15:54] hey frankban, would you please remind me to talk to you after the call about whether we can cleanly push GUI server deployer integration into the deployer codebase? Kapil is keen on doing that, including tests, which would be great by me in order to hopefully reduce our chance of being broken by deployer changes [15:54] bcsaller: be sure they tip you well [15:54] ha [15:54] bcsaller, ok. please ping me when you are available. I'm hoping to have a quick one on one with everyone today [15:54] gary_poster: sounds good, should be back in under 45m [15:54] gary_poster: sure, sounds good [15:55] cool thanks bcsaller. [15:55] thanks frankban [15:55] bcsaller: what, you can't drive and use google hangouts on your phone? [15:55] * gary_poster tries to relocate before the call [15:57] hatch: I keep getting http://paste.mitechie.com/show/1005/ when I run rapi? Look familiar at all? [15:57] yeah that's fine [15:58] that happens around the auth stuff [15:58] jujugui call in 2 [15:59] hatch: new bug :/ http://stackoverflow.com/questions/11300694/chrome-20-websocket-handshake [16:34] gary_poster: I think i'd like to be part of that onboarding call - sounds like it might help me complete/further one of my allhands tasks [16:34] hatch, ack cool [16:40] benji, bac: hangouts died [16:40] But with your approval, I am happy to take on featured then maybe new today [16:44] sinzui: are you rejoining us? [16:44] trying [16:44] ok [16:51] gary_poster: Do we still want to show a warning for hatch yes [16:52] oops [16:52] * hatch reverts [16:52] :) [16:52] hatch: got a fix. Giong to land it with my branch [16:52] rick_h: awesome thanks! [17:02] http://connect.microsoft.com/IE/feedback/details/787208/ie10-waiting-for-domains-url-text-occasionally-gets-stuck-on-tabs-title-only-after-refreshing-it [17:02] UGHHHHHHHHHH [17:03] closed as won't fix??? how is this not an issue? [17:03] seriously [17:03] /rage [17:03] now how do I document this so we don't waist time trying to debug it in the future? [17:03] hmm [17:04] hatch, file a bug and mark as won't fix or invalid or something [17:04] +1 [17:04] can do! [17:07] gary_poster: let me know when you are able to catch up [17:09] hatch: ok, landed. Giong to fetch food and will look for CI report shortly hopefully. [17:09] hatch: thanks for hunting into that! [17:10] sinzui, do you know if jujucharms.com searching looks at the readme file? [17:11] rick_h: cool, thanks for fixing it :) [17:26] jujugui lf two reviews and a IE QA on removing the ie warning https://codereview.appspot.com/13229043/ [17:27] hatch: looking [17:27] thanks [17:28] thanks bcsaller. will ping later. [17:31] rick_h, gave you #2 LGTM [17:31] bah sorry hatch [17:32] thanks! [17:33] arosales, it does not. Description and summary are the only text that is indexed. [17:36] sinzui, ok, thanks for the info [17:39] hatch, do you happen to know if the ghost inspector not showing the charm icon is a relatively new regression or something that we simply haven't gotten around to? I notice that the issue is there in jujucharms.com and comingsoon, so it is likely to be something we haven't noticed/gotten around to [17:39] gary_poster: we haven't implemented that feature yet [17:40] hatch cool thx [17:40] hatch is it tricky? [17:40] I'm not sure, I haven't looked into it :) [17:40] I can thought [17:40] though* [17:40] shouldn't be. I can look at that next since I can't get UX feedback on the constraints today due to holiday [17:40] hatch if you wanna. I'll make a card for someone [17:40] ok rick_h can take it so I can do the IE release? [17:41] hatch: sounds like a plan to me [17:41] cool thanks [17:41] rick_h: thought you would find this funny https://plus.google.com/116181126985457119832/posts/XYmPkzu7Cj7 [17:41] hatch: lol, yea my boy will fetch a minnow from the minnow bucket, but when I caught the pike this weekend he ran to the other end of the dock and watched from there [17:41] haha awesome [17:48] sinzui: ping, got a sec? [17:49] hi rick_h [17:50] sinzui: got a quick sec in guichat? [17:50] rick_h, trivial CSS, but we should make ghost inspector fields white [17:50] sure rick_h [17:50] constraints I mean [17:50] gary_poster: ok [17:50] gary_poster: yea, hoping to do a css cleanup as part of getting things setup/ux cleaned up. [17:50] but make sure that config is still gray/white depending on "default config" toggle [17:50] cool [17:51] rick_h, btw agree with constraints being at top under current circumstances [18:05] hey bac, you available for guichat? [18:05] if not convenient, np! [18:06] hatch: yay ci happy [18:06] yay! [18:06] sinzui: are there diff keys for lcy01 vs 02? My key is named lcy01.key [18:08] rick_h, they are different [18:08] hey rick_h you available for guichat? :-) can try someone else if you are on a roll [18:08] gary_poster: sure thing [18:08] thanks [18:26] hatch you implemented a version of "replace", right? (I say "a version" because it has to be rickety given what's available) [18:27] thinking.... [18:27] thinking.... [18:27] nope the button isn't hooked up [18:28] hatch cool. Let's replace it with "remove" :-) [18:29] sounds like a plan to me - want that done for the IE release? [18:29] hatch no, just planning, sorry :-) [18:32] excellent :) [18:35] benji, sinzui, just to clarify and if this affects your plans, we will plan on having a version of the GUI that talks to version 3 if we use the bundle feature flag, and version 2 otherwise [18:36] gary_poster: that should be fine [18:36] thanks benji [18:37] gary_poster: BTW, it turned out that the charm version task wasn't done, so that's what I'm working on now [18:37] ok cool benji thank you [18:39] bcsaller, you available? np if you want to wait a bit. [18:39] gary_poster: now is good [18:39] cool, guichat is open bcsaller [18:47] sinzui: ping, ok, I can't get in and I don't think I'm supposed to be able to. [18:47] sinzui: got another sec to help me figure out the best way in? [18:47] rick_h, really? [18:47] * sinzui tries without orangestack [18:48] sinzui: yea, I've updated and gotten a lcy02.key, but I don't see how I can use that to get in. [18:48] since you started it yourself in the ~orange juju environment right? So I think I should have to use the ~orange creds at least. [18:48] rick_h, I can get in without sourcing orange or canoni stacks [18:49] sinzui: right, but you created it right? [18:49] I did. I used the orangestack env [19:22] rick_h: https://bugs.launchpad.net/juju-gui/+bug/1217060 any idea what introduced this error? [19:22] <_mup_> Bug #1217060: Cannot switch from unit list view to fullscreen browser [19:31] https://bugs.launchpad.net/juju-gui/+bug/1217062 [19:31] <_mup_> Bug #1217062: Sidebar won't open from minimized after viewing unit list [19:31] jujugui there are two blocking bugs for the release in the urgent lane ^ [19:35] thanks hatch. good qa. rick_h if you have any cycles it looks like you might be the only expert around today. [19:37] gary_poster: thanks - just starting on IE for the QA so hopefully those are all the bugs [19:46] benji, sinzui: could one of you review https://code.launchpad.net/~bac/charmworld/index-charms-in-bundles/+merge/182194 . it is surprisingly compact. [19:46] * sinzui looks [19:46] * benji is too late. [19:46] * benji gets a soda instead. [19:47] benji: dr pepper? [19:47] heh, nope; Pepsi today [19:49] https://bugs.launchpad.net/juju-gui/+bug/1217067 another bug from QA [19:49] <_mup_> Bug #1217067: fullscreen browser fails to open after viewing unit list in IE10 [19:52] bac: r=me [20:05] thanks sinzui [20:10] sinzui: could you mark the MP please? [20:12] bac, whoa! What did I review and approved then? [20:13] bac you available for a quick one-on-one? [20:13] gary_poster: i am [20:13] bac: I /think/ I reviewed and approved it this time. [20:14] bac, cool. guichat when you have a mo [20:14] * sinzui has no idea where the first approved + code went [20:37] hatch: looking [20:38] rick_h: I'm working on bug #1217060 right now [20:38] <_mup_> Bug #1217060: Cannot switch from unit list view to fullscreen browser [20:38] I know the issue but not sure of the best sollution - got a minute for a guichat? [20:38] hatch: sure thing [20:38] hatch: looks busy [20:51] sinzui: you seeing the 503 issues on mjc? [20:51] if I hit the urls manual it works [20:53] benji you available for a quick guichat with me and bac? [20:53] sure [20:53] thx [21:12] Makyo, hey. you available for a quick guichat? I should stop 12 minutes ago, so hopefully I won't keep you forever :-) [21:12] gary_poster, sure, now's best. [21:30] hatch, hey, you have a sec for guichat or are you consumed? [21:30] sure [21:31] thanks [21:33] * bac -> walk dog [22:39] do North American phones work in the UK? [22:40] I am thinking of picking up an HTC One tomorrow but not much point if it won't work :) [23:01] Morning [23:01] morning huwshimi [23:58] rick_h: any chance you are around? I have come up with a simpler sollution to this issue