2014/03/13/#juju-gui.txt

[12:19] <benji> IRC clients work a lot better when I turn them on.
[12:20] <rick_h_> hah
[12:35] <frankban> rick_h_: Robie is working on quickstart/MIR, see #juju
[12:37]  * frankban lunches
[13:11] <bac> hi 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.
[13:12] <benji> cool
[13:12] <rick_h_> bac: woot!
[13:12] <bac> however, i'm seeing the amount of data transferred is really high.  it is as if staging is sending uncompressed but manage is sending gzipped
[13:12] <rick_h_> bac: yea, that's probably true
[13:12] <bac> do you know of a config diff that does that?  apache front end?
[13:13] <rick_h_> bac: that would be in the apache config right? So we'd run a custom apache config for prodstack vs staging
[13:13] <bac> if so, i'm declaring it ready for roll out
[13:13] <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 one
[13:13] <bac> rick_h_: i would guess but haven't done the investigation as it is hard
[13:13] <rick_h_> bac: yea, I think it's part of the invisible bits IS runs
[13:13] <bac> hard meaning, production is opaque
[13:14] <rick_h_> bac: we can ask during deploy time just to confirm
[13:15] <bac> i think there may be caching issues too.  meaning prod is doing caching but staging is not.
[13:15] <rick_h_> cool yea
[13:31] <frankban> guihelp: is anyone available for a quick review/QA of https://github.com/juju/juju-gui/pull/177 ? thanks
[13:31] <bac> frankban: sure
[13:31] <frankban> bac: thanks
[13:50] <bac> looks very good frankban.  finishing QA now
[13:51] <frankban> great thanks
[14:01] <frankban> rick_h_: do you want me to update the browser's series in this branch as well?
[14:02] <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 card
[14:02] <rick_h_> frankban: so no, just spreading awareness for now
[14:03] <frankban> rick_h_: yeah, where are the browser's one used?
[14:03] <rick_h_> frankban: it was orignially used on the filters UI
[14:03] <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 name
[14:04] <rick_h_> which is the bug we've got on it, we shouldn't show 'saucy' but 13.10 Saucy Salamander or something
[14:04] <rick_h_> bac: 1-1?
[14:04] <frankban> rick_h_: ack thanks
[14:20] <bac> rick_h_: sorry, lost track of time
[14:33] <bac> juju-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?
[14:34] <rick_h_> I see it whenevery I got to a hangout room
[14:34] <hatch> bac that just started for me yesterday too
[14:34] <rick_h_> so normally all the calls I setup are hangout/rick and that url with no one in it gets that pretty schedule thing
[14:34] <hatch> maybe people hangout more than we do? lol
[14:34] <bac> it is very nice but unfortunately leaves my webcam turned on
[14:35] <bac> rick_h_: if i go to https://plus.google.com/hangouts/_/canonical.com/rick it puts me directly in the hangout
[14:35] <hatch> they might be slowly trying to do an 'always on'  thing
[14:36] <bac> i don't even have the new fb ui though.  they've tracked me as a late adopter.
[14:36] <hatch> haha then I must be a super-late adopter 
[14:43] <hatch> shouldn't that series select have capitalized names?
[14:50] <rick_h_> jujugui call in 10, kanban please
=== benji_ is now known as benji
[14:58] <rick_h_> jujugui call in 2
[15:01] <benji> rick_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=1
[15:02] <benji> thanks!
[15:02] <rick_h_> hatch: ^
[15:11] <bac> benji: is your branch going to be in review soon?  let me know if i can help.
[15:11] <benji> bac: I'm hoping to have it in review in the next two hours
[15:22] <hatch> my latest blog post http://fromanegg.com/post/79462590516/self-documenting-code-is-not-enough
[15:29] <hatch> somehow a traceback for one of these ghost tests is showing a loc that's 80 lines outside of the test :/
[15:41] <frankban> hatch: 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 developers
[15:41] <rick_h_> :)
[15:41] <hatch> frankban lol thanks
[15:42] <Makyo> hatch, ping
[15:42] <hatch> There are so many 'self help' books I think I could write one...."You're not good enough: the secret to not sucking"
[15:42] <hatch> Makyo yo
[15:42] <rick_h_> I think "You're not good enough: the secret to sucking well"
[15:43] <rick_h_> would be a better read ;)
[15:43] <hatch> haha much better title
[15:43] <Makyo> Got a second for a hangout?
[15:43] <hatch> Makyo just gona go get the dogs inside bbiab
[15:43] <Makyo> Sure thing.
[15:43] <Makyo> https://plus.google.com/hangouts/_/calendar/cmljay5oYXJkaW5nQGNhbm9uaWNhbC5jb20.j0rk5d371ph8331ijtf48t2uj0?authuser=1 when you're back.
[15:44]  * rick_h_ has visions of hatch checking various snow forts outside looking for the dogs watching some TV in an igloo
[15:45] <hatch> lol
[16:06] <hatch> ""Often delayed by 30+ min. This flight is delayed 50% of the time"" lol wow
[17:26] <benji> bac: does charmworld use the "bzr rv-propose" thing now? (or whatever it's named)
[17:26] <bac> benji: yes, bzr rv-submit
[17:27] <bac> i have an alias "winnebago"
[17:27] <benji> bac: is there a doc on how to install it?
[17:27] <bac> not really, but i should
[17:27] <benji> heh
[17:27] <hatch> lol
[17:28] <bac> bzr branch lp:rvsubmit $HOME/.bazaar/plugins/rvsubmit
[17:28] <bac> +/-
[17:29] <bac> benji: but that's only for submitting -- still use 'lbox propose' for proposing
[17:29] <benji> ah!
[17:30] <benji> bac: any idea how to install lbox on trusty?
[17:30] <bac> benji: i think it was just there when i upgraded
[17:31] <benji> darn, I'll have to figure it out
[17:32] <bac> benji: i have the package from quantal installed and it works
[17:33] <frankban> benji: go get launchpad.net/lbox ?
[17:33] <benji> once I get go installed, I'll try that, frankban 
[17:34] <bac> benji: https://launchpad.net/~gophers/+archive/go
[17:35] <bac> benji: yeah, just getting it via go might be smartest
[17:35] <hatch> jujugui looking for two reviews/qa on https://github.com/juju/juju-gui/pull/178 thanks!
[17:36] <Makyo> hatch, on it
[17:37] <hatch> thank yaz
[17:39] <Makyo> My new notification sound works, woo.
[17:44] <rick_h_> bac: with the release today can we import a bundle with tags=,,,,?
[17:45] <rick_h_> bac: just the gui, not charmworld
[17:45] <bac> rick_h_: tags are not supported until we get the comma conversion done
[17:45] <bac> er, i don't understand that last part
[17:45] <rick_h_> bac: ok, so the gui can't take a dropped bundle with tags still?
[17:45] <bac> i think not
[17:45] <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?
[17:46] <bac> no
[17:46] <rick_h_> k
[17:46] <bac> the 'tags' constraint will be rejected
[17:46] <rick_h_> ah right ok
[17:49] <hatch> lunching
[17:51] <benji> bac: want to review the release blocker? https://codereview.appspot.com/75520044
[17:55] <benji> rick_h_: since bac seems to be away, do you want to review my branch? https://codereview.appspot.com/75520044
[17:55] <rick_h_> benji: rgr
[17:55] <bac> benji: i was just getting a refreshing beverage
[17:56] <benji> rick_h_: sorry, I jumped the gun, bac will do it
[17:56] <bac> benji: what is a 'reversed relation'?
[17:57] <benji> bac: 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 way
[17:57] <rick_h_> benji: bac ok, going to go back to the gui release then. Let me know if you need me to peek at anything
[18:01] <bac> benji: in proof.py you removed lines 92-94.  how did that code ever work?  where is 'reversed' defined?
[18:01] <bac> why did it pass lint?
[18:02] <benji> bac: 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] <bac> but why didn't lint complain about variable used before set?
[18:02] <bac> isn't that a thing?
[18:06] <benji> bac: because of the built-in "reversed"
[18:06] <bac> benji: yep, i just saw that!
[18:12] <bac> benji: code looks fine.  didn't do any qa
[18:12] <benji> thanks bac; so how do I install rv-submit?
[18:13] <rick_h_> bac: benji can you qa to make sure that the bundle jorge had in that bug proofs ok?
[18:13] <benji> rick_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' errors
[18:13] <bac> benji: bzr branch lp:rvsubmit $HOME/.bazaar/plugins/rvsubmit
[18:13] <rick_h_> oh nvm
[18:13] <rick_h_> you'll have to ingest to get the provides/requires data to check 
[18:13] <benji> rick_h_: ah
[18:13] <bac> rick_h_: i'd wait and do it on staging
[18:14] <rick_h_> bac: rgr
[18:14] <rick_h_> sounds like plan, sorry. typing got ahead of thinking
[18:25] <hatch> this is a new failure http://ci.jujugui.org:8080/job/juju-gui/550/console
[18:26] <benji> bac: branch submitted for landing
[18:26] <rick_h_> hatch: yea, was looking at that. Curious if it happens again. I'll try to find the code. 
[18:27] <hatch> I'll re-run it
[18:27] <rick_h_> hatch: is this from pr177?
[18:27] <hatch> 178 is my branch
[18:27] <rick_h_> k, it's running now
[18:27] <rick_h_> so we'll see what happens
[18:28] <bac> benji: jenkins saw it and is working.  yippee.
[18:28] <rick_h_> it did that http://ci.jujugui.org:8080/job/juju-gui/547/ here first
[18:28] <rick_h_> hatch: ^
[18:28] <rick_h_> then passed, and then failed
[18:28] <benji> cool
[18:28] <hatch> why did it run so many times?
[18:29] <rick_h_> hatch: I'm doing release. So I pushed updates to develop, master, develop again
[18:29] <rick_h_> hatch: each time a push hits it triggers the tests
[18:29] <hatch> ohh ok
[18:30] <hatch> so now I wait 
[18:30] <rick_h_> hatch: yea, your test run is in progress. I'm out of the way
[18:30] <hatch> so we are doing a gui release now?
[18:30] <rick_h_> yes
[18:30] <hatch> did anyone qa it? lol
[18:30] <rick_h_> and jujucharms deploy as soon as this finishes
[18:30] <rick_h_> yep, I did some qa during this release process. functional charm tests running now
[18:30] <rick_h_> hatch: but we've done a lot of qa this week during inspector changes and such
[18:31] <hatch> I hope so, I'm a little nervous that I introduced some hidden bug but I suppose you're right
[18:31] <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. 
[18:33]  * Makyo is away: Lunch
[18:40]  * Makyo is back (gone 00:07:43)
[18:41] <rick_h_> hah, a bit exact but cool
[18:43] <Makyo> XChat, you're weird.
[18:43] <Makyo> Oh well.
[18:43] <Makyo> If that gets annoying, I'll turn it off.
[18:44] <hatch> lol
[18:44] <hatch> don't be late!
[18:45] <hatch> or are you planning on banking your time? haha
[18:45] <hatch> rick_h_ looks like the same CI failure
[18:45] <rick_h_> hatch: :( k will have to look into it
[18:45] <hatch> damn ci :)
[18:45] <Makyo> Was just seeing if I could be more clear about when I'm here/afk.  That might be too much.
[18:46] <hatch> hmm I wonder if textual has something like that
[18:47] <hatch> hmm I can do /away
[18:47] <hatch> but it doesn't say anything
[18:47] <benji> bac: my review is still marked "Pending" for "Charmworld Developers"
[18:47] <bac> benji: yeah, that seems like a glitch.  it has merged, though
[18:48] <Makyo> hatch, yeah, this was an option in xchat about making /away and /back be explicit
[18:48] <benji> ah, I hadn't noticed
[18:48] <hatch> can you mention me now? I have /away set
[18:48] <bac> benji: are you going to try jorge's example bundle on staging?
[18:48] <benji> bac: sure... once I remember how to point proof at staging
[18:49] <bac> benji: charm-proof?  oh, it isn't easy
[18:49] <bac> benji: i thought you'd do it via juju-gui
[18:50] <benji> I hadn't thought of that, good idea
[18:50] <bac> i think that'll work
[18:50] <rick_h_> hatch: failure is around setting up the ondemain test run. :/ not a code failure but seems saucelabs related
[18:50] <bac> benji: i started to add options to charm-proof to make specifying the server easy, but it got too convoluted
[18:50] <hatch> looks like it happens with safari
[18:51] <benji> :\
[18:51] <rick_h_> hatch: oh hmm, wonder if the safari/os versions got updated?
[18:52] <hatch> not here...
[18:54] <hatch> our travel agency is being DDOS'd by us right now I guess lol
[18:54] <rick_h_> hatch: heh yea
[18:55] <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] <hatch> kewlio!
[18:56] <Makyo> \o/
[18:59] <benji> rick_h_: the bundle referenced in that bug is now a 404
[18:59] <rick_h_> benji: ah crap
[18:59] <rick_h_> jcastro: ^
[18:59] <rick_h_> benji: let me look on my laptop, might still have a copy
[18:59] <jcastro> which bug?
[19:00] <benji> jcastro: https://bugs.launchpad.net/charmworld/+bug/1290439
[19:00] <_mup_> Bug #1290439: proof of bundles errors when order is reversed <charmworld:Triaged> <https://launchpad.net/bugs/1290439>
[19:00] <jcastro> oh let me repush, sorry!
[19:00] <rick_h_> jcastro: just pastebin the bad version
[19:00] <rick_h_> jcastro: we just want to qa and make sure it validate correctly now
[19:00] <jcastro> http://paste.ubuntu.com/7086358/
[19:01] <rick_h_> thanks jcastro 
[19:01]  * benji looks
[19:02] <hatch> looks like we will NEVER be able to zip charms up in the browser :'-(
[19:02] <hatch> https://code.google.com/p/chromium/issues/detail?id=343571 https://codereview.chromium.org/190533005#msg2
[19:02] <hatch> wahhhhhhhhhh
[19:02] <rick_h_> hatch: :(
[19:03] <hatch> I'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 dropped
[19:05] <benji> :(
[19:06] <benji> hatch: yeah, if they only allow "downward" symlinks it would comply with their worries and still let us drop charm directories
[19:07] <bac> benji: let me know when you're qa-happy.  got the RT ready for the charmworld release.
[19:07] <benji> bac: well, I can't get the bundle to fail when dropping it on the GUI
[19:07] <rick_h_> benji: no, it's purely a lint issue
[19:07] <bac> that's good, right?
[19:07] <rick_h_> benji: so it won't ingest in the live env
[19:08] <benji> bac: well, if I had pointed the GUI at staging yet, so (I assume) it is hitting prod
[19:08] <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 staging
[19:08] <rick_h_> benji: but it's nothing to do with deploy-a-bility. Purely ingestion to be in the store
[19:09] <benji> rick_h_: do you have a quick way of doing that?
[19:09] <bac> rick_h_, benji: or have it fail/pass with charm-proof, right?
[19:09] <rick_h_> benji: sec
[19:10] <benji> bac: 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 library
[19:10] <bac> benji: i'll give you a patch
[19:10] <rick_h_> benji: updated url and such
[19:10] <bac> or, what he said
[19:10]  * benji tries
[19:10]  * benji remembers he doesn't have httpie installed any more
[19:11] <rick_h_> I'm sure it's curl-able but dont' have that command handy
[19:12] <bac> benji: here is a charm-tools patch: http://paste.ubuntu.com/7086424/
[19:12] <benji> thanks
[19:12] <bac> 'pip install httpie' might be faster, though your machine seems to be made with bailing wire
[19:13] <benji> my machine is just fine thanks
[19:13] <bac> :)
[19:13] <rick_h_> lol
[19:13] <benji> rick_h_: that command timed out
[19:14] <rick_h_> benji: did you adjust to the manage. url? https?
[19:14] <benji> I forgot to remove the dev port
[19:14] <rick_h_> cool
[19:14] <benji> rick_h_: QA passed!
[19:14] <rick_h_> benji: woot
[19:15] <benji> the bundle proofs on staging (but not on prod)
[19:17] <bac> benji: yay
[19:18] <bac> benji: i see the same with charm-proof:  http://paste.ubuntu.com/7086443/
[19:18] <benji> cool
[19:19] <bac> rick_h_: requesting charmworld deploy.  no vanguard.  :(
[19:19] <rick_h_> bac: booo, thought lamont would be around
[19:20] <bac> rick_h_: he seems around but not listed as vanguard
[19:20] <bac> i'll wait to see if anyone bites
[19:20] <rick_h_> bac: k, ping'd will see. I asked about it earlier and told he'd be around today and would help
[19:21] <bac> rick_h_: what do you mean "once juju picks it up"?
[19:22] <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-gui
[19:22] <bac> ok
[19:22] <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?
[19:23] <rick_h_> maybe it did pull it in. 
[19:25] <rick_h_> isn't there a juju command to download a charm?
[19:25] <rick_h_> not seeing it in juju help commands
[19:25] <hatch> rick_h_ can I help get CI back up somehow? I'd like to land my ghost inspector branch
[19:26] <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 failure
[19:26] <rick_h_> hatch: and look into if there's something upstream that's changed we need to adjust to
[19:26] <rick_h_> hatch: I'd fire off another run of trunk to see if it'll pass
[19:26] <hatch> I 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 today
[19:26] <rick_h_> hatch: :)
[19:29] <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 job
[19:29] <hatch> ok shipping
[19:30] <rick_h_> oh, it's charm get, from charmtools doh
[19:31] <hatch> umm
[19:46] <hatch> so apparently I can't use juju because of clock skew
[19:46] <Makyo> Whaaat is happening?
[19:46] <hatch> that's a new one
[19:47] <hatch> rick_h_ looks like merging failed as well
[19:49] <benji> I 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 signed
[19:50] <rick_h_> hatch: huh?
[19:50] <hatch> rick_h_ the same error when trying to land my branch 
[19:51] <rick_h_> oh, that merging
[19:51] <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 tests
[19:51] <rick_h_> hatch: and if that is so, let's check the safari version info we're using vs the saucelabs list of browser combinatinos
[19:52] <hatch> yeah I can't actually use Juju right now....ubuntu thinks that it's an hour behind right now...
[19:52] <hatch> stupid dst
[19:54] <hatch> looks like the browser data is correct
[19:56] <bac> rick_h_: it doesn't look like ingest is running yet on mjc. not sure why
[19:58] <hatch> ok safari will boot up normally in sauce labs
[20:01] <hatch> ok I was finally able to get bin/test-charm to run locally
[20:02] <hatch> now to see if it actually passes
[20:08] <hatch> maybe...
[20:08] <hatch> :/ hung at bootstrapping
[20:08] <hatch> man I'm having no luck here heh
[20:09] <rick_h_> hatch: huh? what are you doing?
[20:09] <rick_h_> hatch: to test out CI you shouldn't need a juju env
[20:09] <hatch> we don't use bin/test-charm anymore?
[20:09] <rick_h_> jujucharms.com updated woot
[20:10] <rick_h_> thanks benji and bac for the charmworld updates
[20:11] <hatch> oh I guess not
[20:12] <rick_h_> hatch: want to hangout?
[20:12] <bac> rick_h_, benji: it looks like ingest on mjc started up about 15 minutes after update.  that surprises me.
[20:12] <hatch> sure
[20:12] <bac> http://manage.jujucharms.com/heartbeat -- happy
[20:12] <hatch> rick_h_ https://plus.google.com/hangouts/_/76cpjal06fs1756gl0akdhk4no?hl=en
[20:22] <lazyPower> yay!
[20:22] <lazyPower> bookmarked heartbeat
[20:25] <rick_h_> hah
[20:26] <bac> benji, 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 bug
[20:27] <rick_h_> I think other searches are better, but yea that request time is still looooooong
[20:27] <bac> it is requestig gzip
[20:27] <rick_h_> bac: got a sec to hangout?
[20:27] <bac> sure
[20:27] <bac> paste an url
[20:27] <rick_h_> bac:  https://plus.google.com/hangouts/_/72cpio8fijd1hv74kjstcsthmk?hl=en
[20:30] <hatch> Formula 1 starts in 3 days
[20:30] <hatch> lots of cool new tech this year
[20:52] <hatch> Makyo 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] <Makyo> hatch, lots of digging, but your branch is helping.  Will be dogwalking soon, may ping later/tomorrow if there are questions.
[20:53] <hatch> ok no problems