=== _mup__ is now known as _mup_ === rogpeppe3 is now known as rogpeppe [12:20] hi luca__, i've got a branch that implements the feedback box. do you have a moment to have a look? [12:23] I increased limit to story 1 to 6 [12:38] frankban, thank you for fixing the test problem introduced by the new charm cookie functionality. AFAICT the most recent failure is different: the appflower search didn't ever actually execute on FF...when it had just worked before. :-/ are you looking at that as well? would you like to pass the new error to someone else? [12:39] gary_poster: the feedback form translates google-supplied content to spanish for me. The last checkbox of the first question is 'otros' and the button is labeled 'Enviar'. [12:40] gary_poster: is it looking at the live data for that? [12:40] gary_poster: because appflower isn't in the 'interesting' data atm and that will change over time [12:40] bac, huh. PR? [12:40] gary_poster: yeah, but i presume it would do it for any non-english region [12:41] rick_h, live data, but we do an explicit search (we type it in and press enter) [12:41] gary_poster: morning, yes the new error is different, and unexpected. Is there a 30m timeout for jenkins tests? [12:41] i'm always battling google to speak english to me [12:41] gary_poster: ok, charms can be deleted. If you're doing an explicit search I'd suggest using something we use/want to keep around like mongodb or apache or something [12:41] bac, uh. ok. any suggested actions? [12:41] frankban, yes there is [12:42] gary_poster: not at this time [12:42] gary_poster: perhaps now the suite takes more and we could just increase it? [12:43] bac, ok thanks. I already asked luca to investigate whether we had a surveymonkey account we could use, because the lawyer told me that Google's TOS means that they can use our data too, or at least it is unclear. [12:43] teknico: does google render any of this in italian for you? https://docs.google.com/a/canonical.com/forms/d/1Me4macMBX6KydCYyCC6YEjKnpCKakQCYlGlGxIB6Z0w/viewform?pli=1 [12:43] rick_h, noted, good pointm thanks. [12:45] frankban, ok I increased it to 45. That looks like a different problem though, yeah? when I watch the video the search never executes. it does appear to be intermittent though :-/ : the re-run succeeded. [12:46] gary_poster: last jenkins run succeeded in 27 minutes. the previous one failed after 29m. if jenkins quit the driver on timeout a stale element error could be related [12:46] frankban, ah ok. thanks, good possibility. [12:47] hey world, these branches need reviews: [12:47] https://codereview.appspot.com/10401043/ [12:47] https://codereview.appspot.com/10397045/ [12:48] hey luca__ I will triage your bugs from Monday as high. Huw has a question for you on https://bugs.launchpad.net/juju-gui/+bug/1191868, though [12:48] <_mup_> Bug #1191868: There is no dark grey hover over on alerts [12:51] matsubara, when you get back to tarmac, please increase the tarmac jenkins job timeout from 30 minutes to 45 minutes (assuming you have set one). How is tarmac going? I suspect you have a lot of projects going at once [12:55] gary_poster, yes, lots of things to do. I think I should have something to test later on today (in any case I'll do my tests in the jujugui-trunk-merger job to not mess with your workflow) [12:55] thanks matsubara ! [12:55] gary_poster, did you change the job to run under 45 on jenkins? [12:55] matsubara, I changed the job we use but not the tarmac job (didn't look for that one and don't have bookmark :-P ) [12:56] gary_poster, thanks. the other one is disabled so should be ok [12:56] cool thanks matsubara [12:57] np [13:00] gary_poster: I'll take a look now [13:03] luca__, thank you. also, I'm not clear on how drastic the changes will be to the masthead from the current round of redesign. should I ask for your three bugs to be fixed, because subsequent changes will also be relatively small; or should I wait until the full masthead changes are settled? [13:04] bac: in Chromium everything is in English. In Firefox the last checkbox of the first question is named 'Altro', the submit button is labeled 'Invia', and there's more Italian text below the button. [13:05] teknico: that's what i see in chromium (but in spanish). in safari all is english === BradCrittenden is now known as bac [13:05] gary_poster: considering that we've all decided to implement the new design I think they should be classed as low priority. Alerts will no longer live in the masthead, for example. However, If Huw has time then it might be good to fix just in case we can't implement the new stuff. [13:05] luca__, ok. I'll put them at low priority for now. I'm optimistic. :-) [13:05] (that we can implement the new stuff) [13:06] thanks [13:17] orangesquad: appliance guy supposed to arrive in 20min so might be a few min late to standup while i point him at our dryer. FYI [13:22] Hey guys, do you need to do anything special to deploy juju-gui on 0.7? [13:23] marcoceppi, pyJuju 0.7? [13:23] gary_poster: yup [13:23] someone's getting the "connecting to environment" circle screen on 0.7 in #juju [13:24] iirc, last time i deployed against pyjuju it just worked(tm) [13:24] not to our knowledge marcoceppi . our primary tests are still with that. I'll head over to #juju [13:24] gary_poster: thanks! [13:33] jcsackett: ping for standup [13:33] abentley: ack, was searching for headset. [13:43] jujugui, is there anyone who is willing to try to help a smart user on #juju who is trying to diagnose a problem like https://bugs.launchpad.net/juju-gui/+bug/1180095 ? right now he wants to look at haproxy logs, and I don't kbnow where they are, and don't have time to dig in. [13:43] <_mup_> Bug #1180095: GUI charm may have difficulties working with Juju on MAAS [13:43] he is mixired [13:44] maxired === BradCrittenden is now known as bac [13:46] gary_poster: do you want to land the feedback pointing to google form or wait for resolution on the alternative? [13:46] bac land it, unless you disagree [13:47] gary_poster: ok, i wasn't sure how creeped out you were about G's TOS [13:49] bac, not enough to worry about it ATM. I think we should start collecting if we can. [13:49] rt [13:54] rick_h: I could use some YUI help if you have a minute [13:59] benji: definitely [13:59] benji: guichat? [13:59] rick_h: +1 [14:00] benji: taken [14:03] jcsackett: It looks like The Verge lets you share articles on G+. The first click is a +1, and then when you hover, you get a box to share the article. http://www.theverge.com/2013/6/19/4443588/pindrop-acoustic-analysis-fights-phone-fraud-social-engineering [14:04] abentley: thanks! [14:05] np [14:09] abentley: so, it looks like what that's doing is +1 the g+ post verge makes for the article. [14:10] jcsackett: Oh, so not so applicable. [14:10] jcsackett: Sorry. [14:10] abentley: yeah. [14:10] abentley: no, still good to see. it's possible that's a road we'll want to go down. [14:11] i mean, it's not one *i* want to go down... :-P [14:11] jcsackett: I'm envisioning a bot that posts a new story every time a charm is updated... [14:12] abentley: we can just bolt that into ingest, right? ;-P [14:12] :-) [14:12] * abentley really doesn't get why G+ won't release a write API. [14:13] because people will spam it [14:13] it'll end up worthless like twitter [14:21] jcastro: abentley might know a good way to reset the MP to a non-mimms state. [14:21] abentley: jcastro was just asking about the bug about the missing proposals ^ [14:21] abentley: I showed him the bug with your comment. Maybe you can help him un-fubar the missing ones? [14:21] jcastro: Just request ~charmers as a reviewer. [14:22] Hi. I installed ppa:juju/devel and the arrows between the services keep on changing positions... [14:29] maarten__, hi. I'm not clear on what you are reporting, I'm afraid. You installed pyJuju from the PPA, and you installed the GUI charm, I'm guessing, and then what services/arrows are you referring to? [14:30] I installed gojuju and when I use the juju gui, the position of the arrows and the services changes and they are not connected. [14:32] maarten__, thank you for the report. Is this something you can screenshot? [14:38] gary_poster,just send you the screenshot [14:45] maarten__, thanks for screenshot. we have not seen this before. automatic service position changes are supposed to happen, but only if someone else connected to the same environment moves the service (positioning is shared). In that case, of course, relations should maintain positions with their associated services. Does anyone else have access to the GUI who might be moving service positions? [15:00] maarten__, sent you a reply and filed bug 1192596. Moving services is expected in a shared environment, & a big surprise otherwise. [15:00] <_mup_> Bug #1192596: Relation endpoints can get misaligned (after annotation update changes service positions?) [15:01] Makyo, if you have availability could you add this bug to maintenance in kanban, and if you have further availability could yo uinvestigate? [15:01] gary_poster, sure. [15:02] thank you Makyo [15:04] jujgui, after daily call I would like to talk to people about some "bundle" design decisions. If you are at all interested, please take a look at the definitions and the Q & A of https://docs.google.com/a/canonical.com/document/d/1gDspsjnYf1NqDvgyTqRCxtek5JWULMwSfTLwiZn12Nk/edit# [15:04] before the meeting, as a prep [15:05] core question: am I right that using/incproprating the juju-deployer is the right decision? [15:05] incorporating [15:08] rick_h, hatch, if you want to attend shared web meeting https://plus.google.com/hangouts/_/cd640721e4194224f0768c6ca181bb8ba92e7b1a [15:13] luca__: i'd really like you to look at my implementation of the feedback box. the review is at https://codereview.appspot.com/10407044 [15:14] Who is doing "find all tests that change URL and stop them from doing so" in maintenance? [15:14] jujugui ^^^ [15:14] Makyo: oops, it's me [15:14] Makyo: teknico had a branch for that [15:14] teknico, I'll put your face on it. [15:14] teknico, thanks, too. That's annoying :) [15:14] too late :-) [15:15] (meaning, I just did it) [15:15] sorry about that [15:15] rick_h: I can't figure out how to get the charm tokens to be draggable out of the side bar. [15:24] benji: otp, back in a sec [15:30] luca__: got a bit to chat on g+? [15:30] jcsackett: he's otp with us atm [15:30] jcsackett: currently in a call :) [15:42] orangesquad: Could you please review https://code.launchpad.net/~abentley/charmworld/ignore-oneiric/+merge/170386 ? [15:44] abentley: looking [15:45] hatch: you around? [15:46] sort of :) [15:46] what's up? [15:46] I have one single test that's not running anymore, and I cannot locate it: lp:~teknico/juju-gui/test-interdep-cleanup [15:47] benji: so I think you can only drag within the 'container'. I wonder if you make the container the if it might work? [15:48] benji: or maybe there's some tweak to that we can ask around in #yui or #yui-hatch :) [15:48] rick_h: nope, I tried that [15:48] hatch: there should be 950 tests but only 949 run in make test-server, and make test-prod does not output the number of tests at the end, as it does when some of them did not run [15:48] benji: ok /me goes to check css overflow rules on that sidebar div. [15:49] rick_h: I removed the overflows and it had no effect [15:49] benji: ok, one step ahead of me then. [15:49] :) [15:49] teknico it's building [15:49] hatch: thank you [15:49] benji: yea, then that's the only top of my head stuff I can think of. I'd hit up #yui and bug hatch for his previous DnD experience. I've not used it enough to say anything else without getting really into it [15:50] jujugui call in 10, kanban now. [15:50] * benji reluctantly heads over to #yui. [15:51] benji: do you have 'constrain2node' set? [15:52] hatch: not knowingly [15:52] ok you'd know heh [15:53] I've never seen it [15:53] it should set it to position absolute and then you're gtg [15:58] jujugui call in 2 [16:00] benji: let me know when the call is up. I have one more idea to check [16:00] teknico: I thought that your branch stopped these GET requests? [16:01] hatch: which ones? [16:01] well when you run the tests in the console there is a TON of GET requests to get the json data [16:02] Application start up sets the window.flags data: expected undefined to deeply equal {} [16:02] ^ teknico [16:02] there is your failure [16:02] abentley: r=me. [16:05] benji: the css rule on .bws-content is interfering with it. That position absolute there in the rule for: #subapp-browser #bws-sidebar .bws-content, #subapp-browser .bws-view-data, #subapp-browser #bws-sidebar, #subapp-browser #bws-fullscreen [16:05] lol that's a lot of nested id's [16:06] hatch: well it's several paris of ids :P [16:06] pairs [16:06] haha yeah [16:06] they say pair id's are bad because of the css indexing [16:06] hatch: I see it now, thanks [16:07] benji: so if I uncheck that absolute on the .bws-content div, and manually set a .yui3-charmtoken to position: absolute with left/top of 200px it moves and works just peachy [16:07] are you dragging a proxy or the actual element? [16:07] rick_h: cool, I'll try that in a minute [16:07] imho you should drag a proxy [16:09] but I suppose that's implementers choice :) [16:11] hatch: yea, but first stab is an element. [16:12] yeah I suppose [16:12] so with that, lunchables need to be fetched [16:12] I wonder if the proxy will have the same issue [16:12] I don't really know how the proxies work under the hood [16:16] rick_h: you want/can you jump in on the deployer/bundles meeting in guichat? i have to run shortly. [16:36] rick_h: nm. [16:40] jcsackett: k [16:48] luca__, guichat really quick? [16:50] gary_poster: sure [16:50] thx [17:11] rick_h: I was (almost) able to reproduce your results following the procedure you gave, but it seems that it relies on having the original css for inital layout and then swichting those styles. (when applied from the begining not only is the sidebar all a jumble, but the dragged item is still constrained to the container) [17:13] benji: ugh and ugh. [17:14] benji: I assumed that messing with that css would fubar up some things and send us chasing for a different way to do it. I thought/hoped that the issue was one of pure css vs JS functionality though [17:14] I'll look at it harder after lunch. [17:15] benji: k, I'm trying to get this branch up for review and maybe can see if I can help some. Let me know [17:17] teknico, are you finishing up your maintenance branch before you leave? [17:17] Makyo, it turns out that bug from maarten__ has nothing to do with sharing. The GUI release is broken in Firefox :-( [17:17] gary_poster: much as I'm trying, it looks like I will not :-/ [17:18] teknico, ok, I will mark this other issue as critical so we can break WIP [17:18] gary_poster, Will investigate trunk. [17:19] Makyo, to dupe, simply start uistage on Firefox, add two services, and move them around and try to add relations [17:19] thansk Makyo. We should make a 0.7.1 afterwards [17:19] gary_poster, check [17:23] matsubara, is there still a chance that we get you soon for regular weekly qa, or is that no longer an option? [17:23] gary_poster, should be an option, yes. How about a round of ET this Friday? [17:24] matsubara, sounds great. If possible, please use Firefox, Chrom[e/ium], and IE 10. [17:24] thank you! [17:24] ok [17:33] gary_poster: I caught the freakin' pest :-) so I'll be able to at least propose the branch shortly [17:35] orangesquad along with hatch, Makyo, or others interested: code review request please on related charm display: https://codereview.appspot.com/10416044 [17:40] everyone interested in tests (that is, everyone ;-) ) is welcome to have a look at https://codereview.appspot.com/10383047 Thank you! [17:51] jcsackett: What was scan_repo for / is bug #1111708 still relevant? [17:51] <_mup_> Bug #1111708: scan_repo and reindex need to have their own jobs/scripts [17:55] teknico, :-) great [18:03] teknico: looking [18:03] me too [18:04] teknico: and while you wait, i've got one that needs review. :) no squid pro quo, but thought you might be interested [18:05] bac: I'm not exactly waiting :-) I'm past my EOD and having dinner [18:06] ah, perhaps you'll have a break between your main and dessert! :) [18:06] jk, of course [18:08] bac: of course, but I did it anyway :-) [18:50] * gary_poster tried using Gmail over web interface for half a day. Not my thing. :-) Back to buggy, clunky thunderbird [18:57] jcsackett: Here's another site that seems to allow sharing a URL on G+: http://news.nationalpost.com/2013/06/19/bob-rae-resigning-as-an-mp/ [19:17] gary_poster: in kanban there is an lkk@admin.canonical.com marked as IS. is that a real person? [19:17] bac, fascinating question. [19:18] he and/or she is breaking lp2kanban [19:18] no idea [19:18] it should be fixed to continue gracefully ignoring people with no LP id [19:18] i've never seen the admin subdomain [19:19] bac, I could try to see if we can disable this account...but I suspect they are supposed to be there. would you like me to investigate, or do you think we should just plan to fix lp2kanban, or...? [19:20] i think it'll be an easy fix [19:20] cool [19:20] thx [19:20] s/assert/continue [19:20] :-) [19:58] gary_poster: my branch got two reviews but luca indicated he had some changes that he'll describe tomorrow, so the branch is stalled [19:59] ok bac. thanks for heads up. on call [19:59] abentley: that is awesome! that totally works. and with that to work from, finally able to find documentation. [20:00] jcsackett: Great! [20:11] abentley: responding to your much earlier question: scan_repo was used to re do the scan job on things already stored in the DB, iirc. as it's not come up, i don't know that it's something we still need. [20:12] jcsackett: I'm inclined to think we don't need either scan_repo or reindex. [20:12] abentley: i would be inclined to agree. [20:32] hi Makyo, do you have any info on the card 'juju go sandbox: next methods'? [20:33] bac, not immediately. I suppose I'd suggest expose/unexpose, or maybe destroy? [20:34] Makyo: so i guess that card should spawn another after picking one of those... [20:34] bac, yeah, until go-sandbox has at least everything py-sandbox has. [21:19] if we do have a sprint in raleigh at least we have a nice airport now. http://www.cnn.com/2013/06/07/travel/airport-architecture-paul-goldberger/index.html (photos 9 & 10) [21:29] hey, look at us with our #9 in the world! [21:29] * gary_poster has to run [21:29] have a great evening [23:04] evening [23:32] party [23:33] hatch: you see the new pr release from yui. good looking stuff [23:33] yeah it's looking awesome [23:33] we REALLY need to upgrade heh [23:33] :) [23:33] why can't we? [23:33] it's finally coming into it's own [23:34] bcsaller: too hard to change those 4 characters and re-make :P [23:34] lol [23:34] can we hit CI without it landing in trunk? [23:34] not until tarmac is up and running [23:34] ideally we'd just upgrade and push it up to CI and see what it said. If all's well then land it [23:34] buut you could pull down your branch and run from your machine [23:34] so yes [23:34] sortof [23:36] honestly I think the only reason we haven't done it is because noone has taken the time