[00:00] no problemo [12:46] guihelp: I need two reviews for https://codereview.appspot.com/13890044 . Anyone available? thanks! [12:46] hi frankban, i'll look [12:46] writing emails, but will look soon if no one ha claimed [12:46] bac, gary_poster thanks [13:39] gary_poster: re bug 1229939: it's a little more generic than juju-gui-debug. I think at the moment the GUI is not able to change any settings option in juju-core (we are sending an incorrect API request). I'd mark that as critical and start working on it. [13:39] <_mup_> Bug #1229939: Cannot set juju-gui-debug from GUI

[13:42] frankban, that was my assumption as well. The card is in urgent on kanban. have at it, and thank you! :-) [13:43] oh I see, yes, ok thanks [14:00] frankban, bac and I both gave you code LGTMs but with similar suggestions. bac is still running tests [14:01] frankban: my tests had to be restarted. my new environments.yaml file had problems with ec2 [14:05] gary_poster, bac: thanks for the review. re splitting the test, I am +1 on testing errors separately. I'd like to keep the two deployments story in a single test. I intended that test as a story indeed, but I agree with you that test separation is something we should always pursue [14:05] frankban: +1 [14:06] gary_poster: did you run frankban's tests? [14:06] bac no; on call [14:06] ok [14:06] frankban: i'm seeing failures. since gary didn't run them i'll investigate [14:07] bac: can I help? [14:07] frankban: http://paste.ubuntu.com/6154594/ [14:09] bac: are you using juju-core trunk? [14:09] frankban: yes. [14:10] frankban: should i retry with /usr/bin/juju? [14:10] bac: is /usr/bin/juju the last juju-core release? [14:11] 1.14.1-saucy-amd64 [14:12] bac: cool, yes, that could be better, but at this point, if you agree, I'd suggest you to wait until I divided the test as suggested in the reviews [14:12] frankban: ok [14:12] bac: thanks [14:14] bac: fwiw, I am experiencing problems with juju-core trunk, and I reverted to r1750. Also filed a bug two days ago: bug 1229286 [14:14] <_mup_> Bug #1229286: debug-log and boolean options are broken in trunk [14:15] ok [14:28] yeah I think there is something broken in general with the settings [14:29] I couldn't get them to work for the debug switch either on the gui [14:29] hatch: yes I know, I am working on the bug you filed [14:29] oh awesome thanks [14:30] I'd be interested to know the issue [14:30] once you're done of course :) [14:31] hatch: quite easy, we are sending Params.Config where juju-core expects Params.Options. While I was there I decided to switch from ServiceSet to ServiceUpdate too. The former will be eventually be deprecated. [14:32] ohh darn, that was a fail :) [14:32] hatch: the weird think is that juju-core seems to just ignore the unexpected parameter and return a ok-response rather than an error [14:33] s/think/thing [14:33] yeah that sounds like a core bug [14:34] anyway, config, options, settings, we use too many terms for the same concept [14:35] ^ truth! [14:35] at least we now have a single model :D [14:41] gary_poster: what files can be imported to the GUI? [14:41] luca__, config.yaml for charms, and bundle yaml [14:42] ok [14:53] hey benji, you free to chat? [14:54] bac: sure [14:54] your hangout or mine? [14:54] i'll start one [15:01] wow - there are a lot of versions of the GUI on the charmstore [15:21] bac: if you haven't gotten into your next task, here is a short review of the stand-in bundle token branch: https://codereview.appspot.com/13910043 [15:22] ok [15:23] benji is there a way to qa? [15:23] gary_poster: nope, it is not wired up enough to do that yet [15:23] (that's our next step) [15:23] ok cool benji thx [15:27] benji: did [15:28] frankban, is the fix for bug 1229939 something that you can get done today, or will you need to hand it off? Also, I want to talk to you about bug 1229215 but don't want to distract you on this task. maybe tomorrow. [15:28] <_mup_> Bug #1229939: Cannot set juju-gui-debug from GUI

[15:28] <_mup_> Bug #1229215: Confusing unit retry/resolve UX [15:28] gary_poster: did huw's branch get merged? [15:28] bac yes [15:28] gary_poster: is that not the one that is in active coding? [15:29] bac I thought so but that branch looked more like a prep branch. I can make a new one instead and rewrite this one. will do so. [15:29] gary_poster: I am proposing the fix now. Then let's do a call when you want [15:29] cool thanks frankban! [15:31] frankban: is there any way I can provide my current GUI branch (which is on launchpad) to my local GUI charm? [15:31] the path is lp:~hatch/juju-gui/close-inspector-1228410 [15:31] when I Set the juju-gui-source it appears to work....but doesn't :) [15:31] hatch it takes about 10 minutes to build [15:31] and wfm as of last week [15:32] hatch: it should work, yeah, you have to be patient [15:32] ok so while it's building the 'old' GUI is still functional? (I reloaded and everything is back to normal) [15:32] yes [15:32] ahh ok - so I wonder how we can indicate that something is happening [15:32] gary_poster: in the inspector the typeface seems to be wrong, it's bolder than the visuals. [15:33] Juju nor the GUI has any indication that anything is happening [15:33] hatch: juju ssh juju-gui/0 tailf /var/log/juju/unit-juju-gui-0.log [15:34] hatch, yeah juju doesn't give us the visibility. It's where the log idea came from--letting charms say what files should be exposed [15:34] luca__, is this a regression? [15:34] gary_poster: don't think so, though I've only just noticed it [15:35] Hmm I think I have an issue with my local setup juju ssh juju-gui/0 -e local throws error: environment has no access-key or secret-key [15:35] luca__, ack. Could you shoot huw an email with details, cc'ing peeps? I don't see it on the face of it, but I'm sure he will if you point him in the right direction [15:36] gary_poster: sure, will do [15:36] hatch: try to put "-e local" after "ssh" [15:36] thanks luca__ [15:37] frankban: same error - typically I add the -e local to stop this error [15:37] lemme check my yaml file [15:37] hatch: anyway, with "juju switch" you can avoid passing -e each time [15:37] hmm I definitely have an admin-secret [15:38] hatch: you should be able to just ssh into the juju-gui lxc, user ubuntu [15:43] frankban: when I try that it says Permission denied (publickey). [15:44] weird, what command are you using? [15:44] I just tried to ssh into the ip of the machine [15:46] user ubuntu? [15:46] hatch: ^^^ [15:47] oh darn I typo'd [15:47] haha woops [15:47] you wrote redhat? [15:48] lol [15:48] ubnutu [15:48] guihelp: anyone available for one QA + one review of https://codereview.appspot.com/13900044 ? [15:48] frankban, on it [15:48] thanks gary_poster [15:50] jujugui call in 10 [15:51] ok now I can say it definitely did not work [15:51] I'll try from the console see if that triggers it [15:51] hatch: oh, you changed the juju-gui-source from the GUI? [15:52] hatch: if so, that's the same bug you reported for juju-gui-debug. changing options does not work at all in juju-core [15:53] doh! [15:53] :) [15:53] :-) [15:57] now it looks like it's running [15:57] gary_poster: hum... I am reconsidering my decision to switch to ServiceUpdate: using that we no longer support older juju-core releases... If that's a problem, I have no problem quickly set up another branch that just fixes ServiceSet [15:58] frankban, ack, we can talk after call [15:58] jujugui call in 2 [16:08] oh [16:12] change the setting in the console.....watch it change in the gui....oh I love it [16:13] gary_poster: while I wait for these things to spin up - is there another task you'd like me on? [16:13] hatch on call, will thank after [16:13] sounds good [16:24] hatch how much time do you have do you think? [16:25] hey luca__ do you have a few minutes to talk about bug 1229215? [16:25] <_mup_> Bug #1229215: Confusing unit retry/resolve UX [16:26] luca nm I have a call :-P [16:26] luca__, I will shoot you a note [16:27] gary_poster: ok, Jamie is finalising all of the buttons in the GUI at the moment [16:27] gary_poster: sorry missed the ding - well I can start on anything - if this QA passes then i'll be ready to propose [16:27] gary_poster: so that will fix most of that issue [16:28] umm I resolved a service in error and it destroyed the service? [16:29] luca__, oh ok cool. will that include confirmation messages for removing and resolving? If so, I have some input on text for the confirmation dialog in the error removal case. basically, we need to clarify that we are resolving the error and removing the unit, so people familiar with juju CLI are not too surprised [16:29] hatch that typically means that you tried to remove it first [16:29] juju core waits until the error is resolved [16:29] which I did...:D [16:29] haha [16:29] which is exactly what I was talking about and one of the points to the bug I mentioned :-) [16:30] gary_poster: I didn't know we needed confirmation messages for those buttons, so it won't include them but if you send us more information we can look at it. [16:30] luca__, ack. I'll state as problems rather than solutions. [16:31] benji: did you see the ci failure....very odd.... [16:32] hatch: hmm, no; looking now [16:32] hatch: yeah, that is very odd [16:35] luca__, sent email with details [16:37] hatch ask benji if there is a bundle task you can take. bundle detail stub view looks good to me on the face of it, and then you can follow up with integrating Ben's bundle visualization, but check with benji [16:37] ohhhhh benji iiiiiiiiiiiiiii [16:37] yep, bundle detail stub view would be a good one [16:37] are there any bundles in the charm store? [16:38] I know very little of the bundle stuff so far [16:38] been on inspector duty for a while :) [16:38] yes, two at least. search in manage for "bac" [16:38] http://manage.jujucharms.com/~bac/bundle/wiki/wiki [16:39] the other I know about: http://manage.jujucharms.com/~abentley/bundle/wiki-bundle/wiki [16:39] sorry for all the pinging, everyone [16:39] :) [16:39] thanks [16:40] *sigh* the qa failed [16:48] * benji wonders where bac went. [16:50] ok I think I have a fix landed - now onto the bundle stuff [16:54] gary_poster: I was thinking when the GUI looses connection we should highlight the background of the canvas red or something so that they know that they are no longer receiving updates [16:54] hatch, +1, file a bug [16:55] cool will do [16:56] luca__: https://bugs.launchpad.net/juju-gui/+bug/1230410 to keep in mind :) [16:56] <_mup_> Bug #1230410: There is no indication in the GUI when the user looses connection to Juju [16:57] benji: do we have UX as to what this detail page is supposed to look like? [16:58] hatch: not that I am aware of, your current card is simply to create a stand-in [17:01] gotcha [17:05] I wish we would move assets out of app so my greps won't return stray hits from minimized third-party code all on one long line. [17:07] benji: grep harder! [17:07] heh [17:09] benji: ok so I've got a new dev branch here now how do I get the bundles to show up in the 'build' list ? [17:09] build charm list that is [17:09] hatch: you can't; not yet at least [17:10] oh - ok so with no data, and no design - what exactly am I supposed to write ;) [17:10] you'll either have to hack up some test data to be able to see the views in-browser or do test-driven development [17:11] gotcha :) [17:11] how far away are we from getting them in the search results? [17:13] gary_poster: here is the new MP: https://codereview.appspot.com/13917043 [17:13] hatch: bac is working on a card which is a prerequisite of getting them in the search results [17:13] oh ok cool [17:14] so I should be able to make a json query to manage to get the meta data though right? [17:16] frankban, ack will run with it. thank you! [17:19] bcsaller: have a second for a quick chat about deltas? [17:22] hatch: sure [17:22] calling [17:22] https://plus.google.com/hangouts/_/64aa4ab508ffc6cd26933070e433ddac3a5248de?hl=en [17:31] gary_poster, ping [17:32] yeah hey Makyo coming [17:32] Cool, np [17:37] bcsaller: we had a huge delay haha [17:37] I didn't notice until the end...lol [17:40] hatch: who knows what I agreed to in that case [17:40] rofl [17:59] gary_poster: in hangout, ready whenever [18:23] bac: I updated the charm branch and re-proposed, you can do the QA when you want [18:40] bcsaller: so the decision was that we will flag off of the 'life' attribute and when it changes to dying then destroy the inspector/menu instead of keeping it around and having a dying state [21:29] jujugui anyone around who can QA on juju-core ? https://codereview.appspot.com/13938043/ [21:37] does anyone know how to make a json request for bundle data? [21:37] https://manage.jujucharms.com/api/3/bundle/precise/wiki [21:37] I would have assumed that? [21:37] but no luck [21:43] ahh the url in the source is wrong but I found it [21:43] http://manage.jujucharms.com/api/3/bundle/~bac/wiki/wiki [22:10] hatch: where is the url wrong in the source? [22:10] bundle.js:22 [22:11] ^ bac [22:11] I've updated it in my local branch [22:11] I'm just trying to navigate through this browser code...I'm very lost - lol [22:13] http://manage.jujucharms.com/api/3/bundle/~bac/wiki/wiki works for me [22:13] yeah that's the proper url [22:13] that's not what's in the source [22:13] :) [22:13] well http://staging.jujucharms.com/api/3/bundle/~bac/wiki/wiki would work if it were answering [22:14] ohh [22:14] :) [22:14] not sure why staging is donw [22:14] down [22:15] will look in a bit [22:16] sure np [23:15] Morning