/srv/irclogs.ubuntu.com/2014/03/13/#juju-gui.txt

benjiIRC clients work a lot better when I turn them on.12:19
rick_h_hah12:20
frankbanrick_h_: Robie is working on quickstart/MIR, see #juju12:35
* frankban lunches12:37
bachi rick_h_, benji: i'm doing QA against charmworld on staging regarding searches.  with my change from yesterday the number of items returned for a search for rabbitmq-server is down from 319 to 18.  very good.13:11
benjicool13:12
rick_h_bac: woot!13:12
bachowever, i'm seeing the amount of data transferred is really high.  it is as if staging is sending uncompressed but manage is sending gzipped13:12
rick_h_bac: yea, that's probably true13:12
bacdo you know of a config diff that does that?  apache front end?13:12
rick_h_bac: that would be in the apache config right? So we'd run a custom apache config for prodstack vs staging13:13
bacif so, i'm declaring it ready for roll out13:13
rick_h_bac: sounds good, benji is hoping to have his branch later today so we'll shoot for afternoon deploy if we can get one13:13
bacrick_h_: i would guess but haven't done the investigation as it is hard13:13
rick_h_bac: yea, I think it's part of the invisible bits IS runs13:13
bachard meaning, production is opaque13:13
rick_h_bac: we can ask during deploy time just to confirm13:14
baci think there may be caching issues too.  meaning prod is doing caching but staging is not.13:15
rick_h_cool yea13:15
frankbanguihelp: is anyone available for a quick review/QA of https://github.com/juju/juju-gui/pull/177 ? thanks13:31
bacfrankban: sure13:31
frankbanbac: thanks13:31
baclooks very good frankban.  finishing QA now13:50
frankbangreat thanks13:51
frankbanrick_h_: do you want me to update the browser's series in this branch as well?14:01
rick_h_frankban: well just worried about keeping that in sync and trying to find one good home for that info. 14:02
rick_h_frankban: there's a bug where we need to update the series to be a proper name so we can update thing in that card14:02
rick_h_frankban: so no, just spreading awareness for now14:02
frankbanrick_h_: yeah, where are the browser's one used?14:03
rick_h_frankban: it was orignially used on the filters UI14:03
rick_h_but we don't use that any more, so now I think we're supposed to use it for a mapping from series in the charm to a pretty formal name14:03
rick_h_which is the bug we've got on it, we shouldn't show 'saucy' but 13.10 Saucy Salamander or something14:04
rick_h_bac: 1-1?14:04
frankbanrick_h_: ack thanks14:04
bacrick_h_: sorry, lost track of time14:20
bacjuju-gui: when i just hung up on the hangout with rick my browser went to a page with a large clock, a lovely background, and my day's hangout schedule in a right side bar.  has anyone else seen that and know how to navigate to it directly?14:33
rick_h_I see it whenevery I got to a hangout room14:34
hatchbac that just started for me yesterday too14:34
rick_h_so normally all the calls I setup are hangout/rick and that url with no one in it gets that pretty schedule thing14:34
hatchmaybe people hangout more than we do? lol14:34
bacit is very nice but unfortunately leaves my webcam turned on14:34
bacrick_h_: if i go to https://plus.google.com/hangouts/_/canonical.com/rick it puts me directly in the hangout14:35
hatchthey might be slowly trying to do an 'always on'  thing14:35
baci don't even have the new fb ui though.  they've tracked me as a late adopter.14:36
hatchhaha then I must be a super-late adopter 14:36
hatchshouldn't that series select have capitalized names?14:43
rick_h_jujugui call in 10, kanban please14:50
=== benji_ is now known as benji
rick_h_jujugui call in 214:58
benjirick_h_: I don't see the hangout on the calendar, will you give me the link?15:01
rick_h_benji: https://plus.google.com/hangouts/_/calendar/cmljay5oYXJkaW5nQGNhbm9uaWNhbC5jb20.j0rk5d371ph8331ijtf48t2uj0?authuser=115:01
benjithanks!15:02
rick_h_hatch: ^15:02
bacbenji: is your branch going to be in review soon?  let me know if i can help.15:11
benjibac: I'm hoping to have it in review in the next two hours15:11
hatchmy latest blog post http://fromanegg.com/post/79462590516/self-documenting-code-is-not-enough15:22
hatchsomehow a traceback for one of these ghost tests is showing a loc that's 80 lines outside of the test :/15:29
frankbanhatch: good article. looking forward to reading the next "well-documented code is not enough", and then "well-tested-and-documented-code-is-not-enough", escalating up to the point we all realize the philosophical assumption that nothing is good enough for developers15:41
rick_h_:)15:41
hatchfrankban lol thanks15:41
Makyohatch, ping15:42
hatchThere are so many 'self help' books I think I could write one...."You're not good enough: the secret to not sucking"15:42
hatchMakyo yo15:42
rick_h_I think "You're not good enough: the secret to sucking well"15:42
rick_h_would be a better read ;)15:43
hatchhaha much better title15:43
MakyoGot a second for a hangout?15:43
hatchMakyo just gona go get the dogs inside bbiab15:43
MakyoSure thing.15:43
Makyohttps://plus.google.com/hangouts/_/calendar/cmljay5oYXJkaW5nQGNhbm9uaWNhbC5jb20.j0rk5d371ph8331ijtf48t2uj0?authuser=1 when you're back.15:43
* rick_h_ has visions of hatch checking various snow forts outside looking for the dogs watching some TV in an igloo15:44
hatchlol15:45
hatch""Often delayed by 30+ min. This flight is delayed 50% of the time"" lol wow16:06
benjibac: does charmworld use the "bzr rv-propose" thing now? (or whatever it's named)17:26
bacbenji: yes, bzr rv-submit17:26
baci have an alias "winnebago"17:27
benjibac: is there a doc on how to install it?17:27
bacnot really, but i should17:27
benjiheh17:27
hatchlol17:27
bacbzr branch lp:rvsubmit $HOME/.bazaar/plugins/rvsubmit17:28
bac+/-17:28
bacbenji: but that's only for submitting -- still use 'lbox propose' for proposing17:29
benjiah!17:29
benjibac: any idea how to install lbox on trusty?17:30
bacbenji: i think it was just there when i upgraded17:30
benjidarn, I'll have to figure it out17:31
bacbenji: i have the package from quantal installed and it works17:32
frankbanbenji: go get launchpad.net/lbox ?17:33
benjionce I get go installed, I'll try that, frankban 17:33
bacbenji: https://launchpad.net/~gophers/+archive/go17:34
bacbenji: yeah, just getting it via go might be smartest17:35
hatchjujugui looking for two reviews/qa on https://github.com/juju/juju-gui/pull/178 thanks!17:35
Makyohatch, on it17:36
hatchthank yaz17:37
MakyoMy new notification sound works, woo.17:39
rick_h_bac: with the release today can we import a bundle with tags=,,,,?17:44
rick_h_bac: just the gui, not charmworld17:45
bacrick_h_: tags are not supported until we get the comma conversion done17:45
bacer, i don't understand that last part17:45
rick_h_bac: ok, so the gui can't take a dropped bundle with tags still?17:45
baci think not17:45
rick_h_bac: if I have a bunle with tags that I've exported, can I drop it in the gui and have it deploy?17:45
bacno17:46
rick_h_k17:46
bacthe 'tags' constraint will be rejected17:46
rick_h_ah right ok17:46
hatchlunching17:49
benjibac: want to review the release blocker? https://codereview.appspot.com/7552004417:51
benjirick_h_: since bac seems to be away, do you want to review my branch? https://codereview.appspot.com/7552004417:55
rick_h_benji: rgr17:55
bacbenji: i was just getting a refreshing beverage17:55
benjirick_h_: sorry, I jumped the gun, bac will do it17:56
bacbenji: what is a 'reversed relation'?17:56
benjibac: a relation in the opposite order than is expected, the code used to expect requires -> provides, but we want to be able to handle either way17:57
rick_h_benji: bac ok, going to go back to the gui release then. Let me know if you need me to peek at anything17:57
bacbenji: in proof.py you removed lines 92-94.  how did that code ever work?  where is 'reversed' defined?18:01
bacwhy did it pass lint?18:01
benjibac: that code never could have worked the way it is; either it was broken from the start or broken along the way and no one noticed (and the test was flawed too, so the test didn't show the brokenness)18:02
bacbut why didn't lint complain about variable used before set?18:02
bacisn't that a thing?18:02
benjibac: because of the built-in "reversed"18:06
bacbenji: yep, i just saw that!18:06
bacbenji: code looks fine.  didn't do any qa18:12
benjithanks bac; so how do I install rv-submit?18:12
rick_h_bac: benji can you qa to make sure that the bundle jorge had in that bug proofs ok?18:13
benjirick_h_: sure, I'll take a look 18:13
rick_h_I assume the test verifies it I guess, but should be able to hit the api with the bundle to check without needing to worry about ingestion. At least ignoring 'charm does not exist' errors18:13
bacbenji: bzr branch lp:rvsubmit $HOME/.bazaar/plugins/rvsubmit18:13
rick_h_oh nvm18:13
rick_h_you'll have to ingest to get the provides/requires data to check 18:13
benjirick_h_: ah18:13
bacrick_h_: i'd wait and do it on staging18:13
rick_h_bac: rgr18:14
rick_h_sounds like plan, sorry. typing got ahead of thinking18:14
hatchthis is a new failure http://ci.jujugui.org:8080/job/juju-gui/550/console18:25
benjibac: branch submitted for landing18:26
rick_h_hatch: yea, was looking at that. Curious if it happens again. I'll try to find the code. 18:26
hatchI'll re-run it18:27
rick_h_hatch: is this from pr177?18:27
hatch178 is my branch18:27
rick_h_k, it's running now18:27
rick_h_so we'll see what happens18:27
bacbenji: jenkins saw it and is working.  yippee.18:28
rick_h_it did that http://ci.jujugui.org:8080/job/juju-gui/547/ here first18:28
rick_h_hatch: ^18:28
rick_h_then passed, and then failed18:28
benjicool18:28
hatchwhy did it run so many times?18:28
rick_h_hatch: I'm doing release. So I pushed updates to develop, master, develop again18:29
rick_h_hatch: each time a push hits it triggers the tests18:29
hatchohh ok18:29
hatchso now I wait 18:30
rick_h_hatch: yea, your test run is in progress. I'm out of the way18:30
hatchso we are doing a gui release now?18:30
rick_h_yes18:30
hatchdid anyone qa it? lol18:30
rick_h_and jujucharms deploy as soon as this finishes18:30
rick_h_yep, I did some qa during this release process. functional charm tests running now18:30
rick_h_hatch: but we've done a lot of qa this week during inspector changes and such18:30
hatchI hope so, I'm a little nervous that I introduced some hidden bug but I suppose you're right18:31
rick_h_hatch: but today is our release window for monday. 18:31
rick_h_hatch: yea, this next cycle I want to try to concetrate on getting the qa days back in order. 18:31
* Makyo is away: Lunch18:33
* Makyo is back (gone 00:07:43)18:40
rick_h_hah, a bit exact but cool18:41
MakyoXChat, you're weird.18:43
MakyoOh well.18:43
MakyoIf that gets annoying, I'll turn it off.18:43
hatchlol18:44
hatchdon't be late!18:44
hatchor are you planning on banking your time? haha18:45
hatchrick_h_ looks like the same CI failure18:45
rick_h_hatch: :( k will have to look into it18:45
hatchdamn ci :)18:45
MakyoWas just seeing if I could be more clear about when I'm here/afk.  That might be too much.18:45
hatchhmm I wonder if textual has something like that18:46
hatchhmm I can do /away18:47
hatchbut it doesn't say anything18:47
benjibac: my review is still marked "Pending" for "Charmworld Developers"18:47
bacbenji: yeah, that seems like a glitch.  it has merged, though18:47
Makyohatch, yeah, this was an option in xchat about making /away and /back be explicit18:48
benjiah, I hadn't noticed18:48
hatchcan you mention me now? I have /away set18:48
bacbenji: are you going to try jorge's example bundle on staging?18:48
benjibac: sure... once I remember how to point proof at staging18:48
bacbenji: charm-proof?  oh, it isn't easy18:49
bacbenji: i thought you'd do it via juju-gui18:49
benjiI hadn't thought of that, good idea18:50
baci think that'll work18:50
rick_h_hatch: failure is around setting up the ondemain test run. :/ not a code failure but seems saucelabs related18:50
bacbenji: i started to add options to charm-proof to make specifying the server easy, but it got too convoluted18:50
hatchlooks like it happens with safari18:50
benji:\18:51
rick_h_hatch: oh hmm, wonder if the safari/os versions got updated?18:51
hatchnot here...18:52
hatchour travel agency is being DDOS'd by us right now I guess lol18:54
rick_h_hatch: heh yea18:54
rick_h_jujugui released gui v1.0.1 and new charm. Will wait for that to get pulled into the store and try to get the jujucharms upgraded shortly. Heads up :)18:55
hatchkewlio!18:55
Makyo\o/18:56
benjirick_h_: the bundle referenced in that bug is now a 40418:59
rick_h_benji: ah crap18:59
rick_h_jcastro: ^18:59
rick_h_benji: let me look on my laptop, might still have a copy18:59
jcastrowhich bug?18:59
benjijcastro: https://bugs.launchpad.net/charmworld/+bug/129043919:00
_mup_Bug #1290439: proof of bundles errors when order is reversed <charmworld:Triaged> <https://launchpad.net/bugs/1290439>19:00
jcastrooh let me repush, sorry!19:00
rick_h_jcastro: just pastebin the bad version19:00
rick_h_jcastro: we just want to qa and make sure it validate correctly now19:00
jcastrohttp://paste.ubuntu.com/7086358/19:00
rick_h_thanks jcastro 19:01
* benji looks19:01
hatchlooks like we will NEVER be able to zip charms up in the browser :'-(19:02
hatchhttps://code.google.com/p/chromium/issues/detail?id=343571 https://codereview.chromium.org/190533005#msg219:02
hatchwahhhhhhhhhh19:02
rick_h_hatch: :(19:02
hatchI'm going to respond to the bug and say not to close this one but to address the issues with restricting symlinks within the folder being dropped19:03
benji:(19:05
benjihatch: yeah, if they only allow "downward" symlinks it would comply with their worries and still let us drop charm directories19:06
bacbenji: let me know when you're qa-happy.  got the RT ready for the charmworld release.19:07
benjibac: well, I can't get the bundle to fail when dropping it on the GUI19:07
rick_h_benji: no, it's purely a lint issue19:07
bacthat's good, right?19:07
rick_h_benji: so it won't ingest in the live env19:07
benjibac: well, if I had pointed the GUI at staging yet, so (I assume) it is hitting prod19:08
rick_h_benji: you should be able to hit the api with that bundle on manage. and staging. and have the error on manage and pass in staging19:08
rick_h_benji: but it's nothing to do with deploy-a-bility. Purely ingestion to be in the store19:08
benjirick_h_: do you have a quick way of doing that?19:09
bacrick_h_, benji: or have it fail/pass with charm-proof, right?19:09
rick_h_benji: sec19:09
benjibac: right, and we're back at having to hack proof to get that to work (which might not be too hard)19:10
rick_h_benji: http --form POST http://charmworld:2464/api/3/bundle/proof deployer_file="`cat bundle.yaml`"   using httpie python library19:10
bacbenji: i'll give you a patch19:10
rick_h_benji: updated url and such19:10
bacor, what he said19:10
* benji tries19:10
* benji remembers he doesn't have httpie installed any more19:10
rick_h_I'm sure it's curl-able but dont' have that command handy19:11
bacbenji: here is a charm-tools patch: http://paste.ubuntu.com/7086424/19:12
benjithanks19:12
bac'pip install httpie' might be faster, though your machine seems to be made with bailing wire19:12
benjimy machine is just fine thanks19:13
bac:)19:13
rick_h_lol19:13
benjirick_h_: that command timed out19:13
rick_h_benji: did you adjust to the manage. url? https?19:14
benjiI forgot to remove the dev port19:14
rick_h_cool19:14
benjirick_h_: QA passed!19:14
rick_h_benji: woot19:14
benjithe bundle proofs on staging (but not on prod)19:15
bacbenji: yay19:17
bacbenji: i see the same with charm-proof:  http://paste.ubuntu.com/7086443/19:18
benjicool19:18
bacrick_h_: requesting charmworld deploy.  no vanguard.  :(19:19
rick_h_bac: booo, thought lamont would be around19:19
bacrick_h_: he seems around but not listed as vanguard19:20
baci'll wait to see if anyone bites19:20
rick_h_bac: k, ping'd will see. I asked about it earlier and told he'd be around today and would help19:20
bacrick_h_: what do you mean "once juju picks it up"?19:21
rick_h_bac: the juju store isn't picking up the new release yet :/ https://store.juju.ubuntu.com/charm-info?charms=cs:precise/juju-gui19:22
bacok19:22
rick_h_hazmat: the store pulls in every 15ish right? I remember using juju publish to look for an issue for not loading up. Anything else I can sanity check?19:22
rick_h_maybe it did pull it in. 19:23
rick_h_isn't there a juju command to download a charm?19:25
rick_h_not seeing it in juju help commands19:25
hatchrick_h_ can I help get CI back up somehow? I'd like to land my ghost inspector branch19:25
rick_h_hatch: sure, if you can try to run the CI commands manually. pdb in browser.py to see what the data is at the point of failure19:26
rick_h_hatch: and look into if there's something upstream that's changed we need to adjust to19:26
rick_h_hatch: I'd fire off another run of trunk to see if it'll pass19:26
hatchI don't think I've ever used pdb before...this could be interesting :)19:26
rick_h_hatch: it's strange that pr 177 failed once and passed once and we've had landings today19:26
rick_h_hatch: :)19:26
rick_h_MelJun!~#19:29
rick_h_hatch: and see if landing works?19:29
rick_h_I don't see any failures in the -merge job19:29
hatchok shipping19:29
rick_h_oh, it's charm get, from charmtools doh19:30
hatchumm19:31
hatchso apparently I can't use juju because of clock skew19:46
MakyoWhaaat is happening?19:46
hatchthat's a new one19:46
hatchrick_h_ looks like merging failed as well19:47
benjiI had an android tablet that wouln't install new apps and it took me 30 minutes to realize that it was because the certs wouln't validate because the clock was set into the past, before the certs were signed19:49
rick_h_hatch: huh?19:50
hatchrick_h_ the same error when trying to land my branch 19:50
rick_h_oh, that merging19:51
rick_h_hatch: ok, can you create a branch that removes safari from CI and see if it lands?19:51
rick_h_hatch: sorry, not lands, but tests19:51
rick_h_hatch: and if that is so, let's check the safari version info we're using vs the saucelabs list of browser combinatinos19:51
hatchyeah I can't actually use Juju right now....ubuntu thinks that it's an hour behind right now...19:52
hatchstupid dst19:52
hatchlooks like the browser data is correct19:54
bacrick_h_: it doesn't look like ingest is running yet on mjc. not sure why19:56
hatchok safari will boot up normally in sauce labs19:58
hatchok I was finally able to get bin/test-charm to run locally20:01
hatchnow to see if it actually passes20:02
hatchmaybe...20:08
hatch:/ hung at bootstrapping20:08
hatchman I'm having no luck here heh20:08
rick_h_hatch: huh? what are you doing?20:09
rick_h_hatch: to test out CI you shouldn't need a juju env20:09
hatchwe don't use bin/test-charm anymore?20:09
rick_h_jujucharms.com updated woot20:09
rick_h_thanks benji and bac for the charmworld updates20:10
hatchoh I guess not20:11
rick_h_hatch: want to hangout?20:12
bacrick_h_, benji: it looks like ingest on mjc started up about 15 minutes after update.  that surprises me.20:12
hatchsure20:12
bachttp://manage.jujucharms.com/heartbeat -- happy20:12
hatchrick_h_ https://plus.google.com/hangouts/_/76cpjal06fs1756gl0akdhk4no?hl=en20:12
lazyPoweryay!20:22
lazyPowerbookmarked heartbeat20:22
rick_h_hah20:25
bacbenji, rick_h_: the search for rabbitmq-server on prod now returns the trimmed down version. but 1) it is not compressed and 2) is huge.  most of it for the bundles.20:26
rick_h_bac: yea, I noticed that. Going to hold onto the bug20:26
rick_h_I think other searches are better, but yea that request time is still looooooong20:27
bacit is requestig gzip20:27
rick_h_bac: got a sec to hangout?20:27
bacsure20:27
bacpaste an url20:27
rick_h_bac:  https://plus.google.com/hangouts/_/72cpio8fijd1hv74kjstcsthmk?hl=en20:27
hatchFormula 1 starts in 3 days20:30
hatchlots of cool new tech this year20:30
hatchMakyo everything going ok with the inspector stuff?20:52
* rick_h_ runs away for a while. Thanks for the great work today everyone!20:52
Makyohatch, lots of digging, but your branch is helping.  Will be dogwalking soon, may ping later/tomorrow if there are questions.20:52
hatchok no problems20:53

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!