[00:10] jujugui, quick review of small deployer export fix if anyone is up for it: https://codereview.appspot.com/14725044 [00:10] gary_poster: paring with huw atm. Will be reporting a bit late tomorrow fyi since you're around :) [00:11] rick_h_, cool thank you :-) [00:21] * huwshimi owes rick_h_ beer [00:21] gary_poster: that stupid error catching promise thing got us big time. More than just syntax error, any error got caught. We'll have to kill that or find a better way in a hurry. [00:21] huwshimi: all good, you're doing work so I don't have to :) carry on! [00:21] rick_h_, :-( ack [00:22] rick_h_: At the very least not call everything a API error :) [00:23] *an [11:55] hi rick_h_ [12:00] bac: morning [12:00] hey rick_h_ could you review this mostly mechanical branch? https://codereview.appspot.com/14486046/ [12:01] bac: sure thing, behind this morning due to a late start but have it opened up to go through. [12:01] thanks [12:02] ah, my old friend 'old chunk mismatch' :/ [12:02] really? weird [12:02] let me repropose [12:02] bac: all good, it's small enough to read ok on LP [12:03] yeah, but yeah, but i wanted to try 'rvsubmit'! [12:03] bac: lol, then resubmit away :) [12:06] bac: put a LGTM on it. [12:06] rick_h_: reproposing didn't help. [12:07] rick_h_: on LP? [12:07] bac: yea, I've tended to have to push to a new branch name and submit under the new branch [12:07] bac: no, on reitveld [12:07] ugh. ok, thanks [12:07] bac: so you should be good to go to try rvsubmit right? [12:07] yes, i think [12:07] bac: cool, yea just LGTM'd and noted I read it on LP [12:07] ty [12:13] rick_h_: seems to be happy. note plugin is rvsubmit but command is 'bzr rv-submit' [12:15] rick_h_: huh, your LGTM made it through to LP but not as an approved vote. will be interesting to see what the lander does with that. [12:20] rick_h_: worked fine [12:36] bac: cool, good to know [12:46] rick_h_, Is the message on the onboarding cr for me? [12:47] antdillon: that's me updating the branch with your changes from yesterday [12:47] sorry, getting a late start to the day [12:47] rick_h_, Ah cool, just checking your not waiting on me [12:48] antdillon: no, I'm getting a late start to today [12:48] the only thing left is to get with you on testing [12:48] we appreciate the work and I'd be happy to add tests to the code. [12:48] antdillon: but if you're interested I'd also be happy to help wallk you through or pair program on them as well [12:48] antdillon: your choice on how you are for time and interest [12:49] rick_h_, That would great, I'd like to see how test work [12:49] rick_h_, Dont have much today just QA'ing ubuntu.com for tomorrow [12:49] antdillon: ok, I've got to change locations here in a sec and will ping you back in 20ish [12:49] rick_h_, Awesome thanks [13:12] antdillon: https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.rsa4k6d5ht2nreuas4g8694668?authuser=1 when you want to get started [13:29] rick_h_, do you remember why we don't like mod_deflate/gzip encoding on server? I don't, but I'm pretty sure this has something to do with operational concerns [13:39] btw frankban, I meant to say that we probably ought to make a separate project for the quickstart. If you want me to do it, let me know. [13:40] gary_poster: well my question was, is that due to the explotation of https + gzip that came out o the sercurity conference earlier? [13:40] gary_poster: I've not ping'd IS to see what their feeling is on that stuff [13:41] gary_poster: yes thanks, last time I didn't have all the required permissions to complete the task myself. re gzip: http://breachattack.com/ ? [13:42] hey rick_h_, if you have time could you have a look at huw's new branch? i see there is a problem (logged it on rv) but don't know the solution. if you have thoughts that would help him it would eliminate a 24h cycle. https://codereview.appspot.com/14739045/ [13:43] guihelp: I need two reviews + one QA for https://codereview.appspot.com/14441074 (python, quickstart). Thanks! [13:44] frankban: i'll look [13:45] thanks bac [13:47] I'll take the other one, frankban . luca__, I am trying to work through a reply to your machine view email. I'll get it to you today, but if you want a faster discussion, let me know and we can have a hangout [13:49] bac: looking [13:49] gary_poster: I don't mind having a hangout, if your free. [13:49] thanks gary_poster [13:50] frankban, rick_h_ right breachattack.com, thank you both [13:51] frankban, ok will make project after I review your branch and other stuff. luca__ I have a call in 39 but if we can move fast then let's do it. I'll make a hangout [13:52] luca__, https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.rsa4k6d5ht2nreuas4g8694668 [13:56] bac: it looks like there's no default rule that .sprint.normal is display: block in the css [13:57] rick_h_: sprite.normal ? [13:57] bac: he sets .sprite.normal to display: none on hover/active, but it never gets set back to block [13:57] bac: in that diff, in the .less file [13:57] bac: so the issue is that the 'clickable' icon is not visible to click on the second time around [13:57] rick_h_: ok. will you add that to the rv? [13:58] bac: sure thing, did you have the branch pulled down? A one line diff and you can test it out [13:58] rick_h_: i do [13:59] bac: in stylesheet.less http://paste.mitechie.com/show/1046/ [13:59] bac: ok, not one line, but you get the point. See if that fixes it please [14:00] * bac trying [14:03] rick_h_: no affect [14:03] bac: ok, I'll pull it down in a sec and try to see if I can figure it out. Quite the nest of CSS there [14:04] yeah [14:04] gary_poster: antdillon is on his way to becoming a JS testing guru. :) [14:04] rick_h_, antdillon , awesome :-) [14:17] rick_h_, Thanks! [14:17] rick_h_, Feel like I've leveled up ... on level 0.01 now :) [14:17] antdillon: :) you'll be cursing at failing tests in no time [14:18] rick_h_, Already am [14:22] * frankban bbiab [14:25] bac: it looks like a browser issue. They don't always clear :active when you click. As the link is still the 'active element' It gets set, and something else must clear it. Kind of like :blur [14:25] bac: so doing this in css is probably a non-starter unless we ditch the :active version [14:25] i see [14:25] bac: after all, how long would you see it in the course of a single click [14:26] it would flash the :active state if that [14:26] bac: will write up something for the MP [14:26] rick_h_: thanks [14:33] bac: ok, replied. Let me know if that is muddled/confusing please [15:04] hey jcastro, if we want a google doc for a blueprint do we just make one in the 14.04 specs folder? Do I link it to the blueprint as well? [15:05] I think so? [15:05] I guess I just put it in the spreadsheet [15:05] ok cool thx jcastro [15:05] trying to sort out summit right now sorry [15:05] this whole blank schedule with less than a week to go is flippin' me out [15:08] rick_h_: it wasn't muddled at all [15:09] jcastro, I bet. :-( at least the software is super snappy... [15:21] bac, could you do qa for frankban? I'm swamped. I will get to his review now, but I don't want him waiting on me beyond that if I can help it. [15:21] gary_poster: sure [15:21] thank you very much bac [15:22] jujugui does anyone here use lastpass? any luck getting it to work in saucy? [15:22] bac, I do, and it ha worked fine without fiddling [15:23] gary_poster: huh. i see different problems in firefox and chromium with neither working. [15:25] gary_poster, bac: just re-proposed with the change bac requested, and fixed a typo. [15:25] frankban: thanks. shouldn't affect my qa, ;) [15:25] bac: no [15:26] :-) [15:29] frankban, need another review? [15:29] Or has someone else claimed it? [15:29] Makyo: no thanks, gary_poster is doing it [15:29] Okay, cool. [15:29] frankban: how long do you estimate the qa of quickstart should take/ [15:29] Will look anyway, to keep up with qs [15:30] Makyo: yeah, thanks [15:31] bac: ~10 minutes for each successful quickstart run. in my qa instruction, I forgot to mention that you have to destroy the environment between the last two steps [15:31] bac: but if you don't, quickstart should still exit gracefully with an error [15:31] great [15:32] Makyo: I presume there is a way (manage.jc.com?) to retrieve the URL of the last revision of the gui charm, this is something we should do in quickstart [15:32] frankban, yes, request with version HEAD [15:33] Er.. [15:33] That will give the charm info. [15:33] Fromwhich I suppose you could check out the charm from lp [15:34] Makyo: http://manage.jujucharms.com/charms/precise/juju-gui/json ? [15:34] But just juju deploy juju-gui will get latest from charmstore, right? [15:34] http://manage.jujucharms.com/api/2/charm/precise/juju-gui-HEAD [15:35] frankban: api right, use api v3 if you can. It's feature flagged and will be the default once we release bundles [15:35] Makyo: you can leave off -HEAD and it'll get latest [15:36] or not...wtf [15:36] Oh, alright, cool. Just remember a bug where doing that would try to get the juju charm with version 'gui' [15:36] rick_h_: Internal Server Error on http://manage.jujucharms.com/api/3/charm/precise/juju-gui [15:36] with or without HEAD [15:36] http://staging.jujucharms.com/api/3/charm/precise/apache2 works so we've got a bug then [15:36] http://manage.jujucharms.com/api/3/charm/precise/apache2 also works [15:36] bac: known bug with juju-gui? ^^ [15:43] frankban, LGTM with ignorable trivial. Thank you! [15:44] gary_poster: thanks [15:46] gary_poster: you aere right in your comment. I think, instead of adding a comment, I could use "juju status --format yaml": it's the default, but this way it's explicit and it should prevent the program to stop working if they change the default [15:46] frankban, ah, sounds nice [15:48] gary_poster: I cannot use the jujuclient there: I do that step to ensure we have a state server which exposes an API to which then jujuclient will connect [15:48] frankban, ah! of course, makes sense. should have realized. [15:48] rick_h_, https://bugs.launchpad.net/charmworld/+bug/1240576 is kinda critical, yeah? [15:48] <_mup_> Bug #1240576: charm without a version and a - in the name throws an error. [15:48] gary_poster: yea, just creating cards for that one and one to bundles as well [15:48] gary_poster: they're kind of related so fix might go together. [15:49] jujugui call in 11 [15:49] Makyo: cheater! lol [15:49] rick_h_, ok thanks. you mean bundles can't have hyphens? [15:49] Makyo, lol [15:49] gary_poster: #1229179 [15:49] <_mup_> Bug #1229179: Revisionless bundle requests raise ValueError [15:49] ah, ok thanks rick_h_ [15:49] 0:) [15:49] lol [15:49] those should be critical b/c i have to room with sinzui next week [15:50] lol [15:50] I concede today's game to Makyo [15:50] It's my day to run it anyway, heh [15:50] :-) [15:53] frankban: qa was a-ok [15:53] bac: great thank you [15:54] Makyo, is https://bugs.launchpad.net/juju-gui/+bug/1238190 fixed? I think t might be [15:54] <_mup_> Bug #1238190: while launching a service, the canvas re-centers the deploying charm repeatedly [15:54] gary_poster, I think so, but it' [15:54] It will fall under today's work if not. [15:55] ok [16:00] jujugui call now [17:01] jujugui need a second gui review of antdillon's onboarding branch https://codereview.appspot.com/14700043/ [17:01] behind the feature flag onboard [17:21] rick_h_, I'll try to do it quickly [18:45] hey rick_h_, did you have a discussion with ant about not showing the onboarding if the app starts in a fullscreen mode? This seems a bit tricky [18:45] I mean, the desired behavior seems a bit tricky [18:45] It feels like the desired behavior is a bit like this: [18:46] "If, the first time we go to the sidebar mode, we are at the root, then show onboarding" [18:46] but even that has subtle things that make me unhappy [18:48] I think I'm going to propose "If we are in sidebar mode at the start then do the onboarding" as another constraint [19:02] just made comment to rv wth my suggestion, including patch [19:10] gary_poster: yea, good point. Originally it was looking for / (before feature flag) and so it was only tied to showing in sidebar [19:10] gary_poster: so you're right, will have to get at url, check things out [19:10] gary_poster: maybe even move it to part of the subapp/browser.js code so that it can just check viewmode and such. [19:11] * rick_h_ had a thought about that yesterday, but wen tmeh [19:11] rick_h_, cool. patch I landed does good things except for knowing a bit too much about viewmode [19:11] I mean [19:11] I didn't land it [19:11] I put it in rv [19:11] gary_poster: right [19:11] thanks! [19:12] thanks rick_h_ :-) ttyl [19:39] so...problem #1 is small: my new compression appears to be excluding uncompressed source for some reason. I need to investigate. [19:39] problem #2 is more interesting. [19:39] not all charms are being obtained [19:40] by the whole endpoints thing [19:40] and I'm guessing the promises are swallowing the problems [19:40] the symptom is that the gui falls over when trying to draw an inspector [19:40] because it expects the charm to be there [19:41] in db.charms [19:41] and it is not [20:12] oops, my connection dropped and i didn't notice [20:26] gary_poster: ok, yea the inspector uses the store's promise charm method to get the charm [20:26] gary_poster: so that might be eating things there [20:27] rick_h_, yeah it is. Also problem appears to be (one problem appears to be?) that the API socket is falling over on the juju core side [20:27] rick_h_, from api server log: http://pastebin.ubuntu.com/6247538/ [20:27] ugh on that one [20:27] eyah [20:42] wow, app.onLogin can be called be called before app.initialize [20:42] weird [22:00] jujugui, if you didn't see on juju-dev, I filed https://bugs.launchpad.net/juju-core/+bug/1240708 . Thumper will be investigating after he fixes local provider again [22:00] <_mup_> Bug #1240708: API server falls over repeatably during AllWatcher Next, killing GUI [22:00] thx [22:01] gary_poster: so there was nothing special about omgponies' set up but scale? [22:01] Morning [22:01] hi huwshimi [22:01] bac, that's my current hypothesis, yeah bac. :-/ [22:02] gary_poster: that's, uh, ungood [22:02] yeah, that's how I felt about it [22:02] huwshimi: rick and put notes on your merge proposal about the import/export icons [22:02] hey huwshimi [22:03] * gary_poster runs away again [22:03] bac: Ah great, I'll take a look. [22:03] gary_poster: If you return I have a question :) [22:03] gary_poster: About what to actually display for the config/options of the charms on the bundle [22:04] huwshimi, I'd suggest merely the changed values, and then link to the full charm detail [22:04] if someone wants to see description and detail [22:04] huwshimi, does that make sense? [22:05] by which I mean, do you agree with that from a user persepctive? [22:05] gary_poster: OK, but we have two parameters on the charm, one called 'config' and one called 'options' and 'config' isn't populated for any of the wiki bundles [22:05] gary_poster: Also, do we need to display unit counts etc? [22:06] Maybe I need a real bundle example... [22:06] huwshimi, unit count: yes. to get a bundle example, build something in coming soon and export. I'll make a quick example [22:08] huwshimi, http://pastebin.ubuntu.com/6248029/ [22:09] gary_poster: is that just an environment export (is that the same thing as a bundle?) [22:09] huwshimi, envExport key becomes data key in the charmworld doc you are looking at [22:10] gary_poster: I'm sorry, I don't follow :) [22:10] huwshimi, yes to both questions. charmworld gives you a lot of extra information about the bundle (like charm_metadata and author and whatever) but the data key is this export, but in json format [22:11] huwshimi, this is the source of benji's bundle: http://bazaar.launchpad.net/~benji/charms/bundles/wiki/bundle/view/head:/bundles.yaml [22:11] he called his bundle "wiki" [22:11] in the export I showed it is envExport [22:11] gary_poster: Ah right, and can I just drag and drop that on the gui? [22:11] the charmstore converts benji's input [22:11] huwshimi, yes you can [22:12] gary_poster: Ah great. [22:12] huwshimi, now look at http://bazaar.launchpad.net/~benji/charms/bundles/wiki/bundle/view/head:/bundles.yaml [22:12] gary_poster: So it looks like we only want to display the 'options' data [22:13] if you scroll to bottom of that output you will see "data" key [22:13] gary_poster: I'm not sure what 'config' even is looking at that.... [22:13] that "data" key is the same as the source of that bazaar page I showed you [22:14] anyway, yes, options [22:14] and constraints [22:14] gary_poster: You linked to the same page :) [22:14] oh huwshimi sorry! http://staging.jujucharms.com/api/3/bundle/~benji/wiki/wiki [22:14] ah cool [22:14] scroll to the bottom of *that* [22:15] and data key contents == wiki key contents from bazaar page [22:15] :) [22:15] :-) [22:15] huwshimi, s'ok? [22:15] gary_poster: OK that's all great. I can fix that up now. [22:16] cool thanks huwshimi! [22:17] gary_poster: Rick and I talked about there actually not being much info to display per charm at the moment so we thought it might be easier to land a first run with just a table with the charm and details side by side. That way we don't have to worry about a hiding/showing interface yet... [22:17] huwshimi, +1 [22:17] gary_poster: Great!