[00:09] hatch: This bug still exists: https://bugs.launchpad.net/juju-gui/+bug/1321558 [00:09] <_mup_> Bug #1321558: Destroying a service leaves inspector visible [00:12] hatch: Also, your caching branch is lovely! [00:21] huwshimi: hmm, that was marked as fixed. The card for that was done by Makyo [00:21] rick_h_: Was the one for hiding the ghost inspector on deploy? [00:22] *that the [00:22] huwshimi: no, that bug you linked above on the destroying leave the inspector open [00:22] huwshimi: moved the card into urgent maint. [00:22] Makyo: ^ [00:22] rick_h_: Oh, I mean, didn't Makyo do the one about hiding the ghost inspector on deploy [00:23] huwshimi: not sure, I just did a search on the bug number in the kanban board and found the card [00:23] oh I see [00:23] I'll have to chase down the branch that was submitted and see what is up [00:24] huwshimi: but please make sure to note that one and the card in an email summary of the QA results please [00:24] huwshimi thanks :) [00:24] rick_h_: Will do, I have some small visual cleanups that I'll do myself today [00:24] huwshimi: awesome thanks [00:28] is comingsoon very slow for you guys? [00:28] oh maybe it's just me [00:28] hatch: no [00:28] huwshimi in your bug, how do you reproduce it? [00:29] can you add steps because I am not able to reproduce on comingsoon [00:37] hatch: I've added them to the bug: https://bugs.launchpad.net/juju-gui/+bug/1321558 [00:37] <_mup_> Bug #1321558: Destroying a service leaves inspector visible [00:38] huwshimi thanks, the bug only happens with ghost services not deployed ones [00:38] should be an easy fix the url isn't changing so just needs to fire a changestate event [00:39] good catch though [00:40] hatch: Ah yes, you are right. I thought I had reproduced it on a deployed service, but can't seem to now. [00:41] yeah, just looked, huwshimi if you have time today you could probably bang out a fix for it real quick-like :) [00:41] think you might have time? [00:41] if not I can probably do it while watching some tv tonight heh [00:43] huwshimi here is the fix (sans tests) https://gist.github.com/b85f7d0ebdbf05ee0632 [00:43] :) [00:44] hatch: It's ok I'll do it [00:44] hah! just used the chrome dev tools "copy a cURL" feature. sweetness [00:45] hatch: Unless you get to it first :) [00:45] rick_h_ haha cool, I've always wanted to use that :) [00:45] huwshimi true, although I have to start cooking supper right away so it won't be for a few hours for sure [00:46] needed a QA bundle file so went to the gui, went to code, loaded the bundle file and copied the curl command from the network tab to download it :) [00:46] keep me posted if you do tackle it [00:46] haha nice nice [00:46] hatch: Yep, I'll let you know if I start [00:48] bwuhahaha BradCrittenden you rule. <3 this [01:00] 920563 [03:27] huwshimi I'm going to fix that bug now [03:27] hatch: OK :) [03:27] just watched some House of Cards and Orange is the new Black [03:28] need a break [03:28] lol [03:28] lol [04:08] huwshimi bug is fixed if you want to take a look for review/qa if you find some time [04:09] hatch: Ah sure, let me take a look [04:09] huwshimi I'm hopping offline, if alls good feel free to shipit [04:09] hatch: Sure [04:09] thanks, have a good night [04:10] hatch: You too [08:55] morning rogpeppe1: is juju/charm ready to be used? [08:56] frankban: yeah [08:56] frankban: i'm just about to push a branch that changes juju-core to use it [08:57] rogpeppe1: cool [09:05] rogpeppe1: could you please review https://github.com/juju/juju/pull/73 ? [09:07] frankban: looking [09:07] thanks [09:09] frankban: the PR doesn't seem to update dependencies.tsv - is that because it's been done in a previous PR? [09:09] rogpeppe1: yes, http.go was moved to external testing some days ago [09:10] frankban: LGTM then [09:10] rogpeppe1: thanks [09:11] rogpeppe1: I'll start the store migration right after you merge your current branch [09:11] frankban: cool. am currrently running the tests. [09:20] frankban: https://github.com/juju/juju/pull/74 [09:20] rogpeppe1: looking [09:21] rogpeppe1: could you please create a juju/store repo and give me permissions? [09:21] frankban: sure [09:22] frankban: i wonder if the repo should be named "charmstore". what do you think? [09:23] rogpeppe1: it will also be a bundle store in the future. So I guess store can be ok, but that's not a strong opinion [09:23] frankban: bundles are bundles of charms, really [09:23] frankban: i'll raise it in #juju-dev [09:23] rogpeppe1: yeah [09:27] rogpeppe1: what about github.com/bmizerany/assert and deps? were they missing? [09:27] frankban: yup [09:28] rogpeppe1: LGTM === rogpeppe1 is now known as rogpeppe [09:59] frankban: i've created the repo and added juju-hackers as collaborators [10:00] frankban: github.com/juju/charmstore [10:00] rogpeppe: cool, thanks. I'll wait for your branch to land and then start mgrating the store [10:02] rogpeppe: you have a conflict [10:02] frankban: oh, darn === rbasak_ is now known as rbasak [11:01] rogpeppe: how is it going? [11:02] frankban: running tests again after merging [11:02] frankban: worker/uniter/debug just timed out [11:02] rogpeppe: :-/ [11:03] rogpeppe: some of the tests seems to have some weakness [11:03] frankban: but just passed when i ran it again [11:03] frankban: yeah [11:03] frankban: it took 0.2s to run the second time [11:03] rogpeppe: heh [11:07] frankban: my computer just decided to reboot randomly [11:08] :-/ [11:08] once upon a time if the kernel panicked, you'd see some kind of a message [11:08] rogpeppe: so what do you think about this plan: I will 1) migrate store 2) add README etc... 3) change core to use external store in the few places it is imported. You a) wait for 2 (add README etc) b) add to charmstore the server entry points (e.g. cmd/charmd/main.go), including README updates on how to start the charmstore server and c) remove those bits from core [11:14] rogpeppe1: ^^^ [11:14] frankban: i haven't seen anything, sorry [11:15] rogpeppe1: so what do you think about this plan: I will 1) migrate store 2) add README etc... 3) change core to use external store in the few places it is imported. You a) wait for 2 (add README etc) b) add to charmstore the server entry points (e.g. cmd/charmd/main.go), including README updates on how to start the charmstore server and c) remove those bits from core [11:17] * rogpeppe1 thinks [11:19] frankban: how about, instead, migrating store and cmd/{charmd,charmload,charm-admin} all together. then there's nothing in core that imports store [11:20] frankban: because those commands feel like they're more part of the charm store than juju-core itself [11:20] frankban: i will suggest that in #juju-dev [11:22] rogpeppe1: charm-admin depends on juju/cmd. That said, the result is the same, b) and c) are about doing that move. My proposal is just to proceed with incremental steps, so that mechanical branches are really mechanical [11:22] rogpeppe1: so, focus on the goal of having an external store, and then move the commands where they belong [11:23] frankban: ah, i see [11:23] frankban: yeah, your approach seems fine [11:23] rogpeppe1: cool [11:30] rogpeppe1: uhm: your merge failed: godeps: cannot parse "/mnt/jenkinshome/jobs/github-merge-juju/workspace/tmp.o7sFxEIu8x/RELEASE/src/github.com/juju/juju/dependencies.tsv": cannot find directory for "github.com/bmizerany/assert": not found in GOPATH [11:30] frankban: oh frick [11:30] frankban: thanks [11:31] frankban: i'll remove that dependency for the time being [11:32] rogpeppe1: +1 [11:34] morning rick_h_ === BradCrittenden is now known as bac [11:35] rogpeppe1: created the cards for the plan above [11:36] frankban: lovely, ta [12:08] bac: morning [12:08] hi rick_h_. what was your message from last night about? the discussing in vague terms my awesomeness. [12:09] bac: QA went well for me yesterday. Had one note where if the terminal window is small, and you go in to 'use' one of the environments, you have to scroll down with the down arrow before you can select buttons [12:09] bac: oh, just was so cool to deploy a bundle and such from OSX. [12:09] yeah, pretty sweet [12:11] rick_h_: is that term size issue os x specific or would that be the case on linux too? [12:12] bac: I think it'll be any terminal. I don't notice it because of the bigger display/terminals I use on linux vs OSX the default is so small [12:13] bac: so I was debating about filing a bug and wanted to see if anyone else ran across it during qa [12:14] rick_h_: i tested the interactive stuff when i first started the process but have not made it a part of general QA as the interactive piece has been static. so i'm glad you saw this issue. [12:16] bac: just tried it out here on linux and yep, it's just a general thing with the UI. I happened to have a terminal of *just* the right height where I could seee the buttons, but there was a small amount of scoll available [12:17] bac: so definitely nothing block about and not even sure it's something we can do something about [12:19] rick_h_: we can make the buttons always visible (like the status bar): I would consider this a minor/nice to have and not a release blocker. [12:21] frankban: cool, I'll file a bug then for now. [12:21] bac: but QA went well for me and was very exciting to try out after writing docs and emails all day :) [12:22] rick_h_: cool. i'll wait for kyle to have a go and then do the release if all is good. [12:22] bac: sounds good [12:41] frankban: if you have a moment could you review https://codereview.appspot.com/103310043 -- it has updates to the HACKING doc for new release procedures. You may want to read the whole section carefully, not just the changes in this branch. I've landed a few mods there self-reviewed. [12:42] bac: I will [12:42] rogpeppe1: could you please take a look at https://github.com/juju/charmstore/pull/1 ? [12:42] frankban: looking [12:43] frankban: LGTM [12:43] rogpeppe1: thanks [13:06] rogpeppe1: another one? https://github.com/juju/charmstore/pull/2 [13:06] frankban: looking [13:07] frankban: LGTM [13:07] rogpeppe1: thank you! [13:16] morning all. [13:17] morning jcsackett [13:17] jcsackett: can you assis with release today? QA/review Huw's branch and do some QA please. [13:17] assist that is [13:18] rick_h_: was already planning on it--saw your email. [13:18] jcsackett: ty much [13:18] i'll start with Huw's branch and then start whacking the GUI. we have specific areas we're worried about? [13:19] jcsackett: we're just making the inspector left release. So with no flags, just does anything in the GUI have an issue before we release and update jujucharms.com with it [13:20] no flags, got it. [13:28] rogpeppe1: last one for the store: https://github.com/juju/juju/pull/76 [13:28] frankban: cool [13:28] frankban: looking [13:28] thnaks [13:28] bac: review done [13:28] ty [13:29] frankban: LGTM [13:29] rogpeppe1: thank you, merging === tvansteenburgh1 is now known as tvansteenburgh [13:42] rick_h_: huw's branch is good; there's no reason to hold off on shipping it, is there? [13:43] jcsackett: nope thanks [13:44] bac: replied [13:59] github added a 'jump to file' button [14:00] kadams54 did you see my comment on your PR? [14:01] I did. [14:01] * rogpeppe1 lunches [14:02] frankban: yeah, that step was attempting to do two things: use pre-booted env and remote bundle. i'll beef up the description. [14:02] Here's what I think is happening: the machine view panel… er… view has at least one child view, which means the child view's container is connected to a child node. [14:02] So when container.empty() is called, the child node is removed [14:03] And when the destructor is called on the child view, it attempts to empty its container and throws an error because there's no longer an associated DOM node. [14:03] rick_h_: the date in the first sentence of https://docs.google.com/a/canonical.com/document/d/1PI4XPNxVUP3mXqce-jxtT5KgqdZoTuvJZ71W-mSiCZg/edit does not match the later date [14:03] getDOMNode returns null [14:07] kadams54 right, but there should be no child view rendered into the createMachine view [14:07] if there is then something is wrong [14:07] create machine view should be managing it's own child views (if there are any) [14:09] I think there are quite a few child views… machine headers, service scale up, create machine, etc. [14:09] The one that started erroring when I changed to container.empty() was service-scale-up-view.js [14:09] yeah....but create machine and scale up are siblings [14:09] not parent/child [14:09] Maybe, but not in the DOM [14:10] well then the DOM structure is incorrect [14:11] you can't have sibling views that require one to be rendered into the other [14:11] machine-view-panel.js, line 848 - the view passes a node as the container for ServiceScaleUpView [14:11] That node is a child of MV's container [14:12] right [14:12] so why does another child of machine-view-panel destroy that node? [14:13] Another child doesn't destroy it [14:13] machine-view-panel itself destroys it, when container.empty() is called in machine-view-panel's destructor [14:13] kadams54: you still willing/have time to do the quickstart QA this morning? [14:13] bac: Yeah, I started on that last night [14:13] oh, cool [14:14] kadams54 no the empty() is called in 'createMachineView' not 'machineViewPanel' [14:14] kadams54 https://github.com/juju/juju-gui/pull/375/files#r13618728 [14:15] bac: ty, updated [14:16] hatch: Ah, that's my problem :-) For some reason I changed line 901 in machine-view-panel.js :-) [14:17] lol oh boy [14:17] rick_h_: one minor bug that i believe is linked to inspector left https://bugs.launchpad.net/juju-gui/+bug/1328932 [14:17] <_mup_> Bug #1328932: "tutorial" link under "help & feedback" sends you to /sidebar and does nothing [14:17] not remotely done looking yet. [14:17] jcsackett: rgr, ty looking [14:18] rick_h_ jcsackett I'll take it [14:18] hatch: rgr ty [14:29] frankban: here is a new section http://paste.ubuntu.com/7628821/ - i've tested that the steps work. [14:31] hmm looks like onboarding can never BE rendered any longer, good catch jc [14:39] bac: looks good [14:39] rick_h_ low priority but should probably get some new assets from UX https://bugs.launchpad.net/juju-gui/+bug/1328936 [14:39] <_mup_> Bug #1328936: onboarding now puts the inspector in the incorrect location [14:40] hatch: oh ouch, yea. Not sure I agree on low priority though [14:40] luca: ^ [14:41] rick_h_: ah yeah [14:41] luca: I feel like that should block our release today. [14:41] rick_h_: if I remember right, they are just images that Spencer made. [14:42] hatch: what's the file name? [14:42] rick_h_: is that correct? [14:42] luca: correct [14:42] I'm trying to find the one we're using right now [14:43] umm, I broke it, lemme get it back working and grab the file name :) [14:43] juju-ui/assets/images/non-sprites/onboarding/3-inspector.png [14:44] rick_h_ luca the issue we will have though is that the charmbrowser will be rendered underneath the new inspector image [14:44] so it's more work than simply changing the image [14:44] hatch: oh, but we can work with this image [14:44] hatch: we just have to relocate it and cover the sidebar [14:45] it'll only cover part of it [14:45] which may look worse :) [14:45] rick_h_: Spencer can create a new image of the inspector [14:45] right, but I mean cover the sidebar (entirely) and then put this image on there [14:45] rick_h_: he’s busy doing the google mocks but once he’s done he can send one over [14:45] luca: ok, but we'll have a height issue regardless [14:45] We have to solve this better than updating the image [14:45] the sidebar is 100% height, so that image will need to be like 3000px tall to accommodate all screens [14:46] hatch: I see :) [14:46] right, thus we use a smaller image that's the 'top part' of a demo inspector, cover all of the sidebar and have that kindn of floating in place [14:47] heh, a mask in a mask [14:47] :) [14:47] I foresee problems in out future [14:47] our! [14:49] huw can do it :) [14:50] well we could create a fake service and show the inspector for that service [14:50] that's probably going to be the end goal [14:53] rick_h_: hatch do you need me to action anything? [14:54] rick_h_ I'm not sure we want to try and do this pre-release [14:55] I think there will be dragons here which may end up causing us to push release [14:59] jujugui call in dos [14:59] jujugui call in Un [14:59] rick_h_, hatch: i've got another one: https://bugs.launchpad.net/juju-gui/+bug/1328941 [14:59] <_mup_> Bug #1328941: Deploying a local charm shows a compacted inspector [14:59] jujugui on the phone with ramm at the moment, might be late [14:59] ok I can run it [14:59] jcsackett ok can you make a card in urgent plx [14:59] hatch: yup, marking the bug as high too. unless we feel "critical" is better. [15:00] jcsackett high is fine, we'll do it right away either way [15:01] jujugui call now [15:04] jcsackett: ^ antdillon [15:17] kadams54: thanks for your help doing the QA. please ping me when you're done. [15:17] bac: will do [15:20] the tow truck drivers are driving around and honking the horn in front of peoples cars to get them to move it before they tow it....that's pretty nice [15:20] I mean...the big yellow signs have only been here for 5 days.... [15:20] but still nice :) [15:34] jujugui lf a review/qa for https://github.com/juju/juju-gui/pull/379 plz and thx [15:36] hatch: looking [15:36] thanks [15:39] jcsackett looking at your screen shots remind me of what internet was like in the 90s :P [15:41] hatch: yeah; it's something weird with shutter. [15:41] i get great screenshots, and then the moment shutter does anything with them, they turn into horribly compressed jpegs without any chance of recovering them. [15:41] hatch: when I visit a charm details page and then close the panel, the onboarding is shown. Is that the expected behavior? [15:41] i haven't figured out what's going on--it didn't used to have this issue. [15:42] frankban if you dismiss it once then that doesn't happen, correct? [15:42] hatch: yes [15:42] frankban tough to say.... luca ^ [15:43] frankban I'm pretty confident that that is how it should stay... [15:43] hatch: ok [15:44] you don't want to necessarily miss out on the onboarding just because you clicked a link to view some charm details [15:44] at least thats my reasoning [15:44] hatch: frankban you mean that when you visit Juju for the first time and you’ve been linked to a charm details and you close the details the onboarding shows? [15:45] luca no if you click a link to charm details [15:45] but you've never been to juju before [15:45] we show the details w/o the onboarding [15:45] but when you then close those details, we show the onboarding [15:46] luca, hatch: I think you are describing the same path... [15:46] oh.....yeah, I read it wrong [15:46] and you are both right [15:46] English is hard... [15:46] lol [15:47] hatch: rofl [15:47] frankban: hatch that is expected behaviour [15:47] haha yay [15:47] cool [15:47] hatch: then QA is good [15:48] sweet shippin! [15:48] frankban actually, can you comment on the PR THEN i'll ship :) [15:48] there it is! [15:49] hatch: done [15:51] hmm this local inspector business is probably gona be some work [15:51] poo [15:53] https://bugs.launchpad.net/ubuntu/+source/juju-quickstart/+bug/1317939 [15:53] <_mup_> Bug #1317939: New upstream release 1.3.2 is available

[15:54] can I get some eyeballs here so we can get this fixed in utopic so we can get it backported to trusty? [15:57] jcastro: I'm your guy. I've been working on this for a while. [15:58] jcastro: there have been process issues getting in the way, and I've been working on resolving them permanently. [15:58] jcastro: right now the blocker is that juju-core ftbfs on utopic on powerpc, and I need this to land in utopic to fix juju-quickstart in utopic, and with that I'll be able to SRU to Trusty. [16:14] rick_h_, QUESTION: do you plan to make it possible to deploy a bundle with containing local charms in juju-gui? [16:14] rick_h_, I will ask you during the wrap up [16:28] luca do the local charm inspectors also go 100% height in the left column? (thinking so, just want to confirm) [16:28] hatch: they do [16:29] juju-gui: anyone QAing on safari? [16:29] luca thanks [16:30] jcsackett giver' [16:31] hatch: huh? [16:32] jus' giver' ? [16:33] i have no idea what you're saying. [16:33] dialects written in text tend to fail hard. :p [16:33] go hard? [16:33] you can do it? [16:33] ok. thanks. [16:33] :) [16:33] it's all yours [16:33] please sir, may you do the safari qa [16:34] * jcsackett laughs [16:34] :D [16:34] i've never heard "giver'" before...is that "go for it", i'm guessing? [16:34] yeah - we say it all the time when doing something stupid [16:34] "ok dude giver'" [16:35] hatch: yea, that's pure canadian nonsense :P [16:35] haha [16:36] I looked up the urban dictionary definition....yeah there is some messed up stuff in there [16:39] juju-gui: if someone else has safari, can they please open the gui in coming soon and verify that scrollbars on the sidebar work? i'm almost certain this is my machine sucking, but we should double check. [16:39] looking [16:40] jcsackett working good here [16:44] hatch: awesome. well, downside, i can't QA on safari b/c my macbook is *oooold*. plusside, at least the GUI isn't completely hosed on safari. :p [16:45] haha, s'ok we have other osx'rs kadams54 bac frankban rick_h_ :D [16:45] hatch: if you've deployed an older version of a service, the upgrade section in the inspector should show the newer version, right? i'm not grossly misunderstanding 'upgrade' in this context, am i? [16:46] rick_h_ I propose we make a spreadsheet of our supported browsers/platforms and sandbox/real env and then people can put their name beside it when it qa's ok [16:46] jcsackett no, atm it now shows every version of the charm (And some which don't exist) [16:47] no, you're not misunderstanding it, it's just broken [16:47] :) [16:47] hatch: well, no, it seems to show every older version of the charm. [16:47] hatch: ok, as long as know that's broken. [16:47] yeah there are existing cards [16:48] hatch: is the upgrade path one then not something we're QAing? b/c there seems to be an inspector left issue in it as well if you actually click upgrade. [16:48] hmm..... [16:48] though triggering it is hard, since most of the time clicking "upgrade" doesn't seem to do anything. [16:48] I'm pretty sure that should still work. But it does list charms which don't exist [16:49] so the error may be caused by that [16:50] jcsackett it wouldn't hurt to create new bugs though then rick_h_ can go through and mark as duplicates if need be [16:50] oh sure :P [16:50] haha, well 2 bugs is better than 0 :) [16:50] hatch: all good, happy to help garden [16:51] destroying services leaving machines around is a juju-core issue, right? [16:51] hatch: I think the spreadsheet thing is something of formalizing the QA process we should do. I'm going to create a task for that. [16:51] cool [16:51] jcsackett: yes, and there's a branch up for review that changes that [16:51] rick_h_: cool. [16:51] hatch: but I think that's a bit much for the moment [16:52] s'ok! [16:52] ok, UOS session is done and out woot [16:54] jujugui UOS session if interested https://plus.google.com/116120911388966791792/posts/aLNHXFMW5DC [16:55] UOS? [16:55] ubuntu online summit, i think. [16:56] rick_h_ thanks, watching [16:56] what jcsackett said [16:56] I'm trying to train myself to not use UDS [16:57] we need to do a lot better marketing of these things, I only noticed it yesterday haha [16:59] hatch: i'm making a card in urgent for the upgrade bug i've found; if we decide we don't care for release feel free to alter bug and delete card. [17:00] ok, i need food. [17:00] jcsackett: ok, make sure you've got reproduction steps please [17:00] and I also need to find food now. Phew, busy morning [17:00] rick_h_: yup. [17:01] * rick_h_ runs away for a few min [17:01] rick_h_: i try to never file a bug without reproduction steps if i can avoid it. [17:01] * jcsackett runs away in the opposite direction [17:02] * rogpeppe1 is done for the day [17:02] g'night all [17:03] night rogpeppe1 [17:03] rick_h_: see ya [17:03] jujugui wife is comandeering the mifi for part of tomorrow for dr home visits...I might be up no-net creek without a paddle. Cannot coffee shop as I have to be here during install [17:04] gah! imagine a world without internet access! The horror! [17:04] lol [17:04] I'll get someone to be a meeting backup for the cross team call in case I have no net tomorrow [17:04] anyway, really off to lunch [17:05] enjoy [17:14] rick_h_: where's your real internet? [17:15] bac it's being upgraded to add more ips [17:15] business line or something [17:16] oh, yeah, that [17:16] i tune out people with their first-world internet woes [17:19] lol [17:25] bac: sent you a PM… going to break for lunch but wanted to chat about some things I was seeing. [17:26] kadams54: sorry i was at lunch and didn't see that on my return [17:26] np [17:26] kadams54: do you want to do a hangout? [17:26] Sure [17:27] kadams54: going into daily standup [17:28] bac: I'm there… [17:28] bac: :P definitely first world [17:29] kadams54: got kicked. trying again [19:01]