[00:02] now to try and solve these ci issues [00:03] Makyo also thanks for the link re the h4n - it has some great stuff there [00:04] rick_h_ if you get a second could you take a look at these errors http://ci.jujugui.org:8080/job/juju-gui/457/console I don't see how they could be caused by my changes so I'm thinking maybe it's a CI problem? [00:04] the CI has no problems :P [00:05] hatch: hmm, that's in the functional tests [00:05] yeah and they pass properly locally [00:05] I'll re-run it [00:05] somehow ended up on the install juju page? [00:05] but I was hoping you have seen those before [00:05] did you click on something in a test on accident? [00:05] look at https://saucelabs.com/jobs/c38c27334e2846d3a2a6187f2dbdfa1d [00:05] it's on juju.ubuntu.com vs the test suite [00:06] woah what the [00:06] heh [00:06] something sent it over there? [00:06] nope it's all functional tests [00:06] no, not seen it before [00:06] that I added [00:06] yea, gotcha, well not seen it. It's in FF which has been trouble free and you can try to build trunk and see if it shows the issue [00:07] but yea, retry, but last time we let a temp issue through it caused us CI issues so I'd like to make sure we follow through with any hints of issues [00:07] there's not currently any intermittant failures in CI I know about [00:07] yeah I'm retrying...I hate that it's failing because it's a broken CI test to begin with :D [00:19] hmm nope looks like a real error [00:19] interesting [00:23] ahh the link is 'get juju' when it shouldn't be [00:26] now how the heck did I break that [00:47] rick_h_ even develop fails CI now.... [00:47] bleh, I'll look into it in the morning [00:48] to try and track it down (unless someone else beats me to it [00:48] hatch: rgr, thanks for the heads up [01:12] I'm fixing it now [01:15] rick_h_ https://github.com/juju/juju-gui/pull/143 I'm going to land this so that I can land my branch and continue working....We should figure out a better way to have jujucharms.com not show 'log in' on prod builds [01:15] hatch: what's the issue? [01:15] showGetJujuButton [01:15] hatch: how is that button getting triggered though? [01:15] I mean this is hacking the html issue out of the tree [01:15] did this file change recently? [01:15] * rick_h_ wonders if he did it on accident in the release [01:15] you changed it when you set the version back to unreleased [01:16] oh I thought it was intended :) [01:16] no, that was changed for QA [01:16] but should not have been ocmitted [01:16] so that explains why it's failing all of a sudden [01:16] yep [01:16] * rick_h_ goes to look at the diff from me setting release [01:16] ok cool then we don't need to chat :) [01:16] once this CI passes I'll :shipit: [01:16] no, this was a failure on me during release to set flags to qa and not setting them back [01:16] ok coolio [01:17] no problem [01:17] ok, so that was the only config change there [01:17] thanks for fixing that [01:17] sorry it hit your branch and caused you grief. Completely my bad. [01:17] yeah no problem, I got frustrated with Hearthstone so I figured I'd try and get a win [01:17] lol [01:17] I'll add a task to the release process to retrigger a build after updating post-release [01:17] to try to catch that [01:18] oh good idea [01:18] actually, if it's not landed yet and you wanted to add a quick bullet point to it with the copy/paste git branch to use that'd be above awesome [01:18] right after that "set back to unreleased" and git push origin develop [01:19] step in process.rst in the docs [01:19] oops [01:19] too late :) [01:19] * hatch twitchy fingers [01:19] all good, I'll add it. I'll be done with dishes in a minute and will be working on that email and such ayway [01:19] good catch [01:20] oh yeah supper [01:29] -49C outside right now [01:29] holy bonkers it's cold [01:59] that's a bit chilly [01:59] just -10 here [10:43] hi dimitern, thanks for your review [10:46] dimitern: re: filepath.Dir(filePath) != bundlePath, I am not sure that would work, since filePath can point to a subdir, e.g. bundlePath/hooks/install [10:53] frankban, i'm sure there's a way to do it with filepath, rather than strings [10:54] dimitern: there is filepath.HasPrefix but it's deprecated [10:55] frankban, take a look at findArchiveRootDir - it does something similar [10:55] frankban, using filepath.Split [10:57] dimitern: I should split the path multiple times in a loop until the base == bundlePath. it seems more expensive than the current approach [11:08] dimitern: interesting thread: https://groups.google.com/forum/#!topic/golang-dev/InMjKTqZjQo also, strings.HasPrefix seems to be used elsewhere in the code with the same semantic, e.g. in environs/sshstorage/storage.go [11:13] frankban, sshstorage is being deprecated btw [11:13] frankban, i'm trying a quick playground example, and will paste when I get it to work [11:14] dimitern: cool thanks [11:22] frankban, we're actually discussing your CL on our standup now [11:23] frankban, and some ideas came out of it [11:23] frankban, I'll post a follow-up review [11:23] dimitern: ack [13:20] hi rick_h_, thanks to you and benji for getting that search branch done. reviewing the changes i see the thing that hung me up was the size specification on the search. [13:21] rick_h_: did the deploy to manage.jujucharms.com go smoothly without cowboyed changes? [13:28] bac: yes, it went well. [13:28] bac: so hopefully that's all solved now [13:28] hurrah [13:29] yea, no kidding [13:29] bac: were you here, we had elastic search go nuts and finally figured out why [13:29] oh? [13:30] bac: mojo runs a CI test in the same network as prodstack. It brings up a new ES instance using the deployer file used for prodstack deploy of charmworld [13:30] when that ES came up, it tries to join the cluster and ends up taking out ES on prodstack charmworld [13:30] wth is mojo? [13:30] ES has these auto search, find, and join cluster stuff [13:30] it's some tooling around the deployer they use for managing prodstack deploys [13:30] so *our* ES is not isolated [13:30] and in this case is used to run some sort of CI tests to make sure things are deployable and they run them semi-regularly I guess [13:31] bac: right [13:31] oopsie [13:31] bac: and darn near impossible for us to see what's up [13:31] bac: so we should be good now. They're changing the cluster name for the CI environment that's tested [13:31] rick_h_: anyway, i'm annoyed i was felled by that pre-existing 'size=0' setting. i assumed zero meant 'get me everything' and never questioned it. [13:31] which should prevent this in the future, but it's something to be very aware of if ES bombs out [13:32] bac: heh, all good. I appreciate the work and benji took the branch and ran with it. Works out very nicely. Just slow, but it's mainly a demo tool that can be prepped for now [13:32] and now you know, thanks for looking into what was done so you know [13:33] off to the coffee shop for the morning, brb [13:33] rick_h_: before you go, suggestions on next things to do? [13:43] bac: so you've got your head on the document trusty updates card if you want to finish that off [13:44] sure [13:44] bac: and we can chat on the other guiserver card and see if that's something we still need to address [13:44] ok [14:31] ahh deleting a ton of old merged branches [14:32] feels good to have only a single entry in `git branch` [14:34] heh [14:34] so rumour has it that in addition to my hack fix of my wifi settings, also trying a different router can remedy the MBP wifi issue until a fix comes out......bahahaha Apple release a fix? [14:34] these guys are funny [14:35] gardening is good for the sould. I've been trying to make sure to use the delete branch button when things land [14:35] I can't even play a mp3 from my NAS :/ [14:43] rick_h_ https://gist.github.com/hatched/b17094aff0abe6462bb7 lol [14:43] latency? what latency? [14:43] hatch: lol [14:43] type slower [14:45] from 84wpm to 20wpm should be good? haha [14:47] * hatch is going to try and write this new inspector with a Y.View.....oooo [15:11] rick_h_ https://gist.github.com/4cb5b88c3a84330b1c34 from a device which is connected via eth0 instead of wifi [15:11] only a marginal difference? lol [15:11] hatch: it's all just a percentage :) [15:12] 100000000% but still just a percentage [15:12] hatch: time to build a really long network cable for use [15:13] haha, well I ran cable throughout my whole house so I have a plugin 3ft from the computer....but the MBP's don't have an eth port without an additional dongle [15:14] maybe that's why they have these bugs....so you have to buy accessories as workarounds [15:14] lol [15:14] yea, I almost bought one yesterday because in my failed ubuntu install I've got no wifi [15:14] and need to be able to update the kernel to try [15:14] * rick_h_ actually goes to look at that dongle again [15:37] they are pretty expensive for what they are [15:37] yea :/ [15:50] jujugui call in 10 [15:51] * bac wonders why ntp doesn't work on mavericks [15:52] bac, i wonder why ssl doesn't work on mavericks ;-) [15:53] oooh low blow [15:53] bac, hazmat, I wonder why [insert technology here] is broken on mavericks :) [15:53] hazmat: zing [15:55] the estate of Edsger Dijkstra should send Apple a big "Told You So" [15:56] the fact they didn't have tests for SSL is a little concerning.... [15:56] not to mention the lack of linting [15:58] jujugui call in 2 [15:58] jujugui call in 2 [15:58] Hah [15:59] Makyo: is a bot, that is all [15:59] Not yet. I have plans for one, just not the will to follow through :) [16:09] switching back to busted-net [16:11] my phone gave me a warning yesterday because I went over 6GB, apparently when I was at the lake I used a lot of data hah [16:12] hah [16:13] rick_h_, added video links to the post thing. The YT link is unlisted, and U1 is borked, so the download link is personal server, not for publication. [16:14] Makyo: rgr, thanks [16:14] * rick_h_ goes to chromecast up some Juju UI fun [16:17] that is sexy on the tv [16:18] Haha, HD quickstart? [16:18] yea [16:19] little chopped on the left. Will try agin non-chomecat to see if it's just that [16:22] yeah it wasn't cropped on youtube [16:22] cool [16:22] Makyo: oh, I added a card about the subordinates issue maarten brought up. I'm not really sure how it did work. Can you verify? [16:23] Will check - thought we were getting rid of that functionality. [16:23] ok, I'm not sure on the plan there. I don't know enough of the back/forth history [16:28] rick_h_, Yeah, I don't really have any information beyond the diagram and conversations with Luca, tbh [16:28] ok, good to know thanks [16:28] man this git and gist plugin for sublime is pretty rockin [16:28] definitely want that juju-gui plugin I was dreaming up [16:29] juju gui plugin? [16:29] rick_h_ so I downloaded a git and gist plugin which allows me to do git/github functionality using keybindings [16:30] so I was thinking a jujugui 'make' plugin which allowed me to do the same [16:30] cmd+j+l [16:30] could lint [16:30] for example [16:30] or cmd+j cmd+l [16:30] more likely [16:31] type that ten times and it can be considered a DDOS [16:32] haha [16:32] Yeah, at that point, my computer really would catch fire. [16:32] I can't seem to find a good plugin which embeds a terminal into sublime so this is the next best thing [16:32] Makyo lol [16:34] I figure once I get all these key bindings it'll be like vim without the suck.... [16:34] * hatch runs [16:34] :P [16:57] jujugui anyone want to have a pre-imp on the design/layout of this local charm upgrade inspector? [16:57] hatch: available here if you need [16:57] sure ok, switching to hotspot then I'll make link [16:57] rgr [16:59] rick_h_ https://plus.google.com/hangouts/_/7ecpi9mumqnf5ah03qgnfgjifo?hl=en [17:06] dimitern: updated the MP in https://codereview.appspot.com/67750045 with the requested change (do not expand the archive) [17:19] frankban, looking [17:19] dimitern: thanks [17:40] frankban, reviewed [17:41] dimitern: cool thanks [17:42] dimitern: using sublime3 with GoSublime I assumed gofmt is run on each save, but will take a look. [17:42] frankban, thanks [18:21] hatch: up for a quick test? [18:22] sure what's up? [18:36] * bac reboots to install ssl patch. yay. [18:55] hmm he has been gone a while [18:55] maybe the patch doesn't work out well [18:55] lol [18:56] heh, guess I'll hold off on that path until he comes back [18:56] haha yeah I was thinking the same [18:57] jujugui the large bundle I have renders in prod/devel/on comingsoon. Were there other specific problem bundles? [18:58] I haven't experienced this bug at all sorry [18:58] Makyo: the big ones were the brad ones [18:58] Makyo: I'd just do a search for 'bundle' and walk through them [18:58] and see if you can get one to not pull up as you move down the side [18:59] shoot [18:59] Makyo: so right now brad's "wiki:wiki" won't load for me [18:59] hatch: ? [19:03] rick_h_ oh I just ran into a spot where switching the seriesSelect viewlet to a view would save me a lot of time moving forward [19:04] but that sets me back... [19:04] cool [19:04] oh :( [19:04] yeah so trying to decide the direction to go [19:04] maybe I can land what I have... [19:15] and my safari gets a thumbs up. https://gotofail.com [19:16] yay, we were worried about you [19:16] we're waiting for you to come back before we update [19:16] bac: oh hey, you use fusion for your VM? [19:16] bac: I can't get the shared drive stuff to work with it [19:16] rick_h_: i do use fusion but don't do much sharing [19:17] if i do need to move the occassional file i just use scp [19:17] yea, what I ended up doing [19:17] cool [19:17] rick_h_: update is 450M [19:17] rofl [19:17] ouch!? [19:17] for missing {} [19:17] ? [19:17] maybe they realized they should add some tests [19:17] haha [19:18] no, they just added it to their existing 10.9.2 that was pending [19:18] oic [19:18] so it includes lots of other silliness [19:18] hatch: hope they aren't shipping the tests in the patch [19:18] oh cool 10.9.2 apparently includes a fix for the wifi issues i'm having [19:19] "fixes an issue that may cause VPN connections to disconnect" [19:19] amazing how they refuse to ack that there are issues then release fixes for them [19:19] lol [19:38] Whoops: https://gist.github.com/makyo/9216139 [19:41] Makyo that renders it in another dimension [19:41] ;) [19:42] An eldrich one. Right angles from everywhere, etc. [19:49] Looks like the auto-positioning stuff isn't working for bundles. [19:49] Makyo: there's a couple of bugs on that we want to look at that came out of cape town [19:50] rick_h_, Yeah? [19:50] hmm, well really was more scaling though in the main canvas so maybe not related [19:50] Makyo: yes, on some larger deployments the initial load is zoomed in oddly and you have to initially zoom out a way to get to see what's there [19:50] Probably not related, but worth fixing, yeah. [19:50] rick_h_, Ah, okay, yeah. I've seen that one. [19:50] I was supposed to get a sample bundle from them to be a test case and never did [19:50] The zoom-to-fit was turned off. [19:50] ah, gotcha [19:50] rick_h_, http://comingsoon.jujucharms.com/sidebar/search/bundle/~makyo/openstack/2/openstack/?text=makyo [19:51] hah, just a little space there [19:56] Makyo: is there a hint as to what the issue is though? [19:56] Makyo: I mean were you able to reproduce? [19:57] jcastro_: dude [19:57] jcastro_: https://jujucharms.com/sidebar/search/?text=bundle for you, just be patient with it as it takes a while to load [19:58] rick_h_, I love you. [19:58] love bac he got it going [19:58] is there a way to get that non-sidebar? like full screen results? [19:58] jcastro_: charm works as well [19:58] no, full screen is dead :/ [19:58] we were forced to remove it [19:58] it'll come back in a future thing later [19:59] * jcastro_ nods [19:59] rick_h_: re: fusion, i was having boot problems with trusty. turning off plymouth helped. removing the virtual printer helped more. ymmv. [20:00] bac: ok, not had any issues yet but good to know [20:01] jujugui lf a review and qa for a wip branch https://github.com/juju/juju-gui/pull/145 so I can continue on with a breakout task [20:02] rick_h_, I can reproduce, yeah, It may be really quick to fix. If so, I'll just get it done. [20:02] Makyo: awesome [20:05] rick_h_ if I merge pr#145 into my current branch so that I can work now without waiting for review etc do you know if that merge will be there in this branch's pr? [20:05] or should I rebase it in? etc... [20:10] parsing that sec [20:10] hatch: the thing to do is to go to your branch for 145 [20:10] and git co -b from there [20:10] so that it keeps going off that branch in the tree [20:11] ahh good idea thx [20:11] then when 145 lands you rebase develop into your current branch and I think it'll just clean up [20:23] rick_h_, this URL isn't going away right [20:23] aka. I can document this? [20:23] jcastro_: yep, just note it's for demo purposes and such. We're not going to spend a bunch of time speeding it up for a while [20:24] jcastro_: but we wanted you guys to have it for talks and such [20:24] man, how do I even scroll this thing [20:25] oh, More, then scroll [20:25] you have to open more up first [20:25] yea [20:25] Good Enough [21:30] jujugui someone review my branch!!!!! :P [21:30] soon I'm going to have two landing lol [21:32] Alright, one sec. [21:32] Let me switch contexts. [21:37] haha thanks :) This one can't land until that one does so I don't want to create some weird out of phase backlog haha [21:37] hatch: will look tonight if you need [21:40] quite possitbly the most irritating error in all of YUI [21:40] TypeError: 'undefined' is not an object (evaluating 'c._buildCfg') [21:43] hatch: hah [21:43] hatch: still need the review? [21:43] it means that it can't resolve a dependency [21:43] but there is no indication of which one [21:43] so it's wak-a-mole [21:43] unless Makyo picked it up [21:43] rick_h_, grab the next one? [21:43] hatch: well it's git diff time [21:43] to see wtf changes that broke it :) [21:43] Makyo: sure thing [21:46] loving https://jujucharms.com/sidebar/search/?text=bundle :-) [21:46] arosales: good stuff :) [21:46] sorry we were a bit late on getting it for last week [21:46] https://bugs.launchpad.net/juju-gui/+bug/1284843 any input on this? [21:46] <_mup_> Bug #1284843: Dragging and dropping a local charm does not create a ghost service icon [21:46] rick_h_: hey its there thats all the counts [21:47] rick_h_: you guys have prompted jcastro_ to make the policy and doc story better [21:47] hatch: honestly not yet. I want to go through the flow a few times [21:47] in regards to bundles [21:47] arosales: always love making more work for jcastro_ :P [21:47] well, I have been whining a bunch and you've been fixing things. :p [21:48] I have nothing left to complain about [21:48] hah, we are listening [21:48] lol [21:48] except of course I want `juju quickstart anyurlIwant` [21:48] but I get that you hate me. [21:48] give it time...give it time [21:48] you have that, at least it's supposed to work [21:48] jcastro_: so show me a url that doesn't work and we'll file it as a bug and make it work [21:48] I can do juju quickstart something on github? [21:48] well it's got to be to the raw yaml file in the repo [21:48] right [21:49] oh right, you want to just point at github and have it find the file for you [21:49] rick_h_, http://i.imgur.com/Ufbr5ej.gif [21:49] * jcastro_ finishes up an email and I will try it [21:49] lol [21:49] lol [21:50] +1 [23:22] must....finish....branch [23:22] rick_h_, [23:22] juju quickstart https://raw.githubusercontent.com/castrojo/mongodb-bundle/master/bundles.yaml [23:22] seems to be working! [23:23] all you need to do is make juju quickstart https://github.com/castrojo/mongodb-bundle DTRT. :) [23:25] jcastro_ so that would then deploy the master all the time? [23:25] (your second link) [23:25] sure [23:28] that would be pretty cool [23:31] ugh my slow internet is bonkin me drivers man! bonkin me drivers! [23:32] rick_h_ kickin around? [23:37] jujugui lf a review whenever you get a chance https://github.com/juju/juju-gui/pull/146 [23:38] Promises! *fistshake* [23:39] The problem is that the fakebackend doesn't know how to deal with revisionless charms, looks like. [23:39] So it returns undefined, so creating a service fails. [23:42] Makyo the large bundle rendering issue? [23:42] Yeah. [23:42] sorry I've been out of the loop on this a bit [23:42] It doesn't have to do with large bundles, it has to do with revisionless charms in bundles [23:43] Which is why large bundles exported from the gui work. [23:43] are revisionless charms valid> [23:43] ? [23:43] I think they are [23:43] I think so? Recent work made them so serverside, I think. Just blows up fakebackend. [23:43] So there's just a few more steps to get them totally working. [23:44] well I guess you know what your tommorrow-card will be :D [23:44] YEP [23:44] haha [23:44] the fakebackend stuff is confusing [23:44] we should straighten that out somehow...I always forget the order of operations in there [23:45] drop marble in fakebackend.....shake......marble comes out :) [23:45] Hah [23:46] Well I'm going to update to the latest patch....hope it doesn't brick this thig [23:46] :) [23:46] have a good night [23:46] Good luck _o/ [23:46] hah thanks [23:48] man [23:48] rick_h_, this mongodb bundle, makes me so happy. [23:49] hey so some more feedback [23:49] let's say I have a working environment [23:49] and someone sends me a bundle [23:49] if I try to run it since the bundle includes the gui it bails because it tries to deploy the gui [23:49] is there a way we could be smart about it? [23:50] or perhaps consider not exporting the GUI when I shift-d [23:53] jcastro_, Can probably do that incrementally. First step would be to use the GUI specified in the bundle by default (we already allow passing a GUI charm URL). Second step would be checking if there's already a GUI instance. [23:56] * jcastro_ nods