MakyoGood luck00:02
=== jamie___ is now known as Guest51225
=== gary_poster|away is now known as gary_poster
hatchquiet here today13:21
rick_h_we're partying too hard13:21
hatchgary_poster: whenever you have a spare moment lp:~hatch/juju-gui/hide-on-relation create a relation and watch the inspector - I found 'hiding' it a little to jarring13:23
gary_posterhatch ack.  I showed the first half (hiding) to luca and got his ok, but not reappearing.  Will look after I finish an email (minute or two)13:24
hatchsure no rush, I'm going to go grab some breakfast13:24
gary_posterfrankban, LGTM with question13:33
gary_posterfrankban, second LGTM13:38
adeuringbac, benji: could one of please review this mp: https://code.launchpad.net/~adeuring/charmworld/more-heartbeat-info-3/+merge/195065 ?13:57
benjiadeuring: sure, I'll take a look13:57
frankbangary_poster: thank you! the -e argument can come before or after, I put that after just to make copy/paste easier for users which already have a default env, can switch the order if you think it's better14:11
gary_posterfrankban, cool, understood. Leaving as is makes sense then14:11
frankbangary_poster: cool, landing14:12
hatchhahaha http://eleks.github.io/js2js/ javascript to javascript compiler14:13
gary_posterhatch, relation behavior looks nice to me.14:15
gary_posterhatch, it is not what I showed luca though14:15
gary_posterI showed him just disappearing14:15
gary_posterI think what you have is better14:15
hatchland and ask for forgiveness later?14:16
lucahatch: gary_poster where do I see it?14:16
hatchhmm good question14:16
hatchluca: can you build/run gui branches?14:16
gary_posterluca, do you want to check out a branch, or do you want to look at it on  comingsoon, or do you want a hangout?14:16
gary_posterif you look at it on comingsoon that means it will have landed, to be clear14:17
lucagary_poster: hatch I can bar but I'm pressed for time. I can see it via a hangout14:17
hatchok cool14:18
lucahatch: send me a link for a hangout and I'll jump on14:18
gary_posterhatch, luca, https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.j0rk5d371ph8331ijtf48t2uj014:18
lucagary_poster: hatch looks excellent :)14:19
gary_poster:-) cool14:19
benjiadeuring: your branch looks good; I had a few non-blocking thoughts and suggestions14:24
adeuringbenji: thanks!14:24
benjimy pleasure14:24
* frankban biab14:34
hatch3 line test...30 lines of stubs14:42
hatchaww yeah14:42
gary_posterI hope everyone is enjoying these bug emails as much as I'm enjoying causing them to be sent out.15:04
hatchthey make my day15:05
hatchit makes me feel important...like getting snail mail15:05
hatchwhen you need to write 70 lines of stubs for two method tests there is a good chance that the method needs to be split up...lol15:10
lucagary_poster: spam spam spam15:12
gary_posterluca, I have some good deals I'll forward you as well.  Great prices!15:12
lucagary_poster: sweet!15:13
rick_h_miracle drugs unite!15:13
gary_posterMakyo, interesting: https://bugs.launchpad.net/juju-gui/+bug/124692815:18
_mup_Bug #1246928: Offers inexistent revisions to 'upgrade' (really, downgrade) to <juju-gui:Triaged> <https://launchpad.net/bugs/1246928>15:18
bachi frankban, have time for a quick chat?15:30
frankbanbac: sure15:30
bacfrankban: cool, i'll invite you15:30
bacfrankban: sorry, hangouts is taking forever to log me in15:34
bacfrankban: https://plus.google.com/hangouts/_/72cpi9b979kk5galo5ar2b5c1o?hl=en15:37
gary_posterhatch, I'll take inspector shrinking 15:49
Makyojujugui call in 1015:50
gary_posterhatch LGTM & QAOK with a request for a bit more of a comment15:58
gary_posterjujugui call in 215:58
hatchgary_poster: thanks!15:59
gary_posterthank you, very nice15:59
gary_poster0 New bugs15:59
gary_poster5 gazillion spam emails15:59
gary_posterbenji yoo hoo16:01
gary_posterhey hatch, should I wait for dropdown help widget to land before I start making a gui release?16:35
hatchif that's alright, I'm just fixing the tests now16:36
gary_postercool hatch sounds good16:36
frankbanbac: the new bundle url option in the Deployer API call is optional, correct?16:42
marcoceppiHey party people, I'm at re:invent and I want to pimp the gui so hard it hurts16:42
marcoceppiis there an easy way to list bundles in the GUI?16:42
bacfrankban: bundle_id, yes16:43
hatchmarcoceppi: sorry there isn't a 'bundle' category, a bug was just recently created for that though16:43
rick_h_marcoceppi: maybe bac  knows if there's a way to generate a search for that doctype?16:44
rick_h_marcoceppi: searching for jorge brings up all his bundles which is a series of them16:44
marcoceppiI just need a few that I can search for and deploy16:44
marcoceppirick_h_: awesome16:44
hatchyeah that's probably the best way16:44
rick_h_marcoceppi: yea, search for jorge then you can demo those handful16:44
marcoceppirick_h_: good enough, I'll also demo export import16:45
frankbanbac: ok, in that context I'd suggest "BundleId" <shrug>, it's good to have it optional, in the next future also quickstart can use that param16:45
hatchmarcoceppi: great!16:45
bacfrankban: there is a card for quickstart already16:46
hatchMakyo: I'm proposing the help-dropdown branch now, then i'll give it a once over and let you know....are you going to be able to qa in...~20mins?16:46
frankbanbac: ic, cool16:48
hatchnote: chrome slow? Check if you left the simulator running all night and have 1000's of services running in the background :D16:48
rick_h_hatch: lol16:48
rick_h_jujugui also heads up that the latest chrome dev edition is a mess, stay away if you're using it. Never had it this borked in a upgrade. 16:49
Makyohatch, yep16:50
hatchMakyo: ok it's ready....phantom didn't crash this time so it took a lot less haha https://codereview.appspot.com/25390043/16:51
gary_postermarcoceppi, search for ~gary: two good ones for demo16:53
gary_posterbac, I guess ~ fix is not deployed to charmworld yet16:54
bacgary_poster: should be16:54
bacwas the other day16:54
gary_postermarcoceppi, http://comingsoon.jujucharms.com/bundle/~gary/demo/2/instantBigDataNoSQL and http://comingsoon.jujucharms.com/bundle/~gary/demo/2/mixAndMatch/16:54
bacwfm http://manage.jujucharms.com/search?search_text=%7Ebac&op=16:54
gary_posterbac, http://comingsoon.jujucharms.com/sidebar/search/?text=~gary16:55
bacgary_poster: oh, that fix was in charmworld16:55
gary_posterbac, ack. thought would be in api too16:56
rick_h_gary_poster: oh hmm, we might need to url escape that search input perhaps. Maybe we don't16:56
gary_posterrick_h_, https://manage.jujucharms.com/api/3/search?text=~gary16:57
rick_h_gary_poster: heh, yea boom16:57
rick_h_bac: did you want me to ask webops for a traceback from the logs or should we just file a bug that api search is hitting a different path not correcting the ~ ?16:58
gary_posterhow would you expect that to be escaped? %7E doen't cut it16:58
rick_h_gary_poster: I think it's just probably a code path to search that didn't get covered in the other branch, my guess at least16:58
gary_posterk cool16:58
bacrick_h_, gary_poster: yes i think that's the case.16:58
bacrick_h_: so i don't think the traceback would tell us more16:59
rick_h_bac: right, well it'd tell us the code path16:59
rick_h_bac: cool, will leave it be16:59
baci'll make a card17:00
frankbangary_poster, rick_h_: anyone else: charm branch ready for review: https://codereview.appspot.com/26130043 . My branch is merged, so, thanks to Kapil, no forks in there.17:05
gary_posterI'll take it frankban 17:06
frankbangary_poster: thank you17:06
adeuringbenji: regarding your suggestion to record the time each (of possibly more than one) server was started: What should we use as the keys? the hostname caomes immediately to mind -- but the hostnames may change over time. Right now, it's for example "juju-staging-machine-3" for staging. Could become "juju-staging-machine-4" when charmworld is deployed again.17:07
adeuringthat would leave cruft in the mongo DB17:07
benjiadeuring: yeah, hostname was what I was thinking too, but I guess in this cloudy world hostnames are somewhat transient17:08
benjithinking out loud: we could use the UUID of the root partition17:09
benjiwe could just roll with it and know that very old hostnames are probably cruft17:10
benjiwe could expire very old server names, say after 30 days17:10
adeuringbenji: right, that would be an option. Another option would be to use the juju unit name -- but that does not work in test environments...17:11
benjiin the charm we have root so we could pull the ID out of the BIOS17:12
benjiI'm leaning towards expiring old name/startup times when we add a new startup time.17:13
adeuringbenji:question here is how to distinguish between machines that are really gone and those that are simply not restarted for some time.17:14
benjiadeuring: that suggests either the "human factor" solution of just realizing which are old and ignoring them or adding some kind of liveness indicator17:15
benjifor example, probabalistically updating an "I'm alive" timestamp in MongoDB on each request17:15
adeuringbenji: ok, a helper script "remove_stale_start_times" or somesuch (started manually) would work.17:16
benjisomething like this: on each request we get a random number and if that number is > 0.99 we put the current timestamp in the DB in this server's slot17:16
benjiadeuring: yeah, I would keep it simple initially17:17
adeuringbenji: random updates of "start time" are probably not what rick_h_ had in mind ;)17:18
rick_h_adeuring: what are you guys talking about?17:19
benjiadeuring: heh, no this would be for the liveness indicator, in other words, the start time is valid as long as the machine is still "alive" and it is alive if its liveness timestamp has been updated in the last N hours/days17:19
* rick_h_ is confused what this is all in reference to17:19
adeuringrick_h_: your suggestion to show the time a charmwolrd server was started on the heartbeat page17:19
hatchgary_poster: sorry I'm trying to submit this branch but npm keeps failing for whatever reason...17:19
adeuringbenji: ah, ok, another indicator, once we identify machines17:20
gary_posternp hatch.  doing qa and collecting changelog and so on.17:20
rick_h_adeuring: oh, we can't just do a time = datetime.utcnow() in the app's __init__.py?17:20
rick_h_adeuring: and have the heartbeat diff it?17:20
benjiadeuring: well, it could be another indicator, but it is to address the problem of telling the difference between a machine with an old start time because it just hasn't been restarted in a while and one that has gone away and been replaced17:21
rick_h_adeuring: it's not a huge memory hog and if the thing gets restarted the timer restarts?17:21
* rick_h_ didn't think it was his suggestion anyway, but ok. 17:21
adeuringrick_h_: I have already a branch ready; uses a mongoDB record instaed (I don't like global varoables...)17:21
rick_h_adeuring: yea, but getting persistant storage involved is ugh because you have to track/clear/etc as you're running into. 17:21
rick_h_adeuring: I know there's an aversion to global variables, but it's tiny, config is already global, and it auto resets when the app restarts17:22
benjiplus, keeping the value in the DB lets us deal with having multiple servers (which we very much should be considering)17:22
rick_h_benji: I guess, but we care about the server we're talking to right?17:22
adeuringrick_h_: see benji's remark ;)17:22
rick_h_benji: I guess it goes back to the 'why' did we want it. 17:22
rick_h_adeuring: yea, catching on slowly. 17:22
rick_h_adeuring: I don't think this was something I asked for, I don't see the usefulness off the top of my head tbh. The revno is good enough for things I can think about caring about. 17:23
rick_h_so you guys have fun with it :)17:23
gary_posterfrankban, LGTM and QAOK.  Thank you!17:27
frankbangary_poster: cool thank you! I am investigating a problem reported in #juju and then land it17:30
hatchlooks like NPM is down I won't be able to submit this until it's back :/ I'm getting 500's on all npm calls17:33
frankbangary_poster: quick hangout?17:44
gary_posterfrankban, https://plus.google.com/hangouts/_/7ecpj64sk94egc6grifl8ko0b4?hl=en ?17:46
Makyofrankban, time for a quick chat after this?17:46
* hatch hopes npm is back up now17:47
* Makyo restarts for updates.17:57
benjiI have a charmworld branch up for review: https://codereview.appspot.com/26140043/18:02
frankbanMakyo: I am ready when you want18:02
Makyofrankban, https://plus.google.com/hangouts/_/7acpjtb7d5f9efjcrvqtpc5cd8?hl=en18:03
hatchok they say they are back up, trying again18:07
Makyogary_poster, have a moment? If not, can wait for our call.18:11
gary_posterMakyo, sure.  have a hangout or should I make one?18:11
MakyoHave one already: https://plus.google.com/hangouts/_/7acpjtb7d5f9efjcrvqtpc5cd8?hl=en18:12
gary_posterMakyo, can't hear or see you18:12
hatchwelp still broken18:18
gary_posterhatch, "npmjs: i feeling fine again" ?18:29
hatchstill not working for everyone it seems18:30
gary_posteroh: "i was feeling better briefly on sick hardware so i'm moving myself to another machine."18:30
hatchthey should use Juju18:30
gary_posterheh, yeah18:30
gary_posterhatch, why is this hurting us?  I thought cache was our savior in situations like this18:31
rick_h_gary_poster: you have to have an initial pull to cache18:31
hatchyeah that18:31
rick_h_gary_poster: so when you lbox, into a tmp dir, it's empty18:31
hatchfor some reason it won't cache -all- modules for me for some reason18:31
hatchno idea why18:31
rick_h_*cough*download cache*cough*18:31
gary_posterI thought even then we had a user-level cache18:31
* rick_h_ zips up his .npm and emails it to hatch :)18:32
hatchyeah - for whatever reason it never deploys from there18:32
hatchright now I'm getting a shasum match failing18:33
hatchmaybe someone else will have better luck landing for me?18:34
bacbenji: has anyone started your review?18:34
rick_h_hatch: sec, I can try. 18:34
benjibac: not that I am aware of18:36
bacbenji: ok, i'm looking18:37
frankbanguihelp: quick fix for the manual provider bug: https://codereview.appspot.com/2619004318:39
frankbangary_poster: thanks18:40
gary_posternp, ty18:40
rick_h_hatch: -cr went through, doing submit now18:43
gary_posterfrankban, LGTM with no changes, doing QA...18:45
bacbenji: code looks great but i cannot QA right now as i have a charmworld running that took a long time to set up.18:46
benjibac: that's fine; I have QAed quite a bit and the potential failure modes are pretty benign (famous last words)18:47
hatchrick_h_: thanks lemme know18:47
gary_posterrick_h_, btw do you know about lbox -adopt?18:48
rick_h_gary_poster: kinda, I tried it once and got it wrong so just go the long way around 18:49
gary_posterheh ok18:49
rick_h_hatch: landed18:49
hatchrick_h_: awesome thanks18:49
rick_h_gary_poster: yea, this kind of sucks to dupe, but always works for me 18:49
bacfrankban: you around?18:51
frankbanbac: yes18:51
bacfrankban: does this look like what i should be doing with the async client?  is the last line necessary?18:52
frankbanbac: looks good from a quick look. the last line must be removed, the io loop must be started once per thread, and the app already does it.18:55
gary_posterfrankban, the relationship does not occur in the bundle until after the services have started, right?18:56
bacfrankban: ok, i'm trying to test it interactively in isolation so that's why it is there.18:56
bacgary_poster: that is what i've observed18:57
rick_h_gary_poster: yep, takes a bit18:57
frankbanbac: if you really need to check the response, I usally prefer to used gen.coroutine and yield, so that it is sequential and you don't need to write callbacks18:57
gary_posterk cool18:57
gary_posterfrankban, "LGTM and QAOK.  Thank you!"18:57
bacfrankban: i actually don't but am doing it here so that i can figure out why it is not working18:57
frankbangary_poster: to your previous question, yes, that's how the deployer works. thank you for the review!18:58
gary_posterfigured. welcome!18:58
frankbanbac: cool18:58
* frankban landing and then disappearing18:59
gary_posterthank you18:59
gary_posterjujugui, I'm making a gui release...while on calls :-P .  I may ask for help.  Anyway, Please hold off on landing to gui branch for a bit19:00
rick_h_gary_poster: rgr19:00
hatch:) gary_poster in hangout, come whenever19:00
hazmatout of curosity has anyone else had slowness issues with leankit boards loading slowly?19:18
rick_h_hatch: just tried it and not fast, but not unbearable. 5-10s?19:18
rick_h_err hazmat 19:18
hazmatrick_h_, i was seeing minutes, just related it directly to chrome-beta though.. chrome stable and its fine19:21
rick_h_hazmat: I'm on chrome dev channel19:21
rick_h_hazmat: but the latest dev channel release is full of fail, so not surprised19:21
hazmatrick_h_, specifically the zendesk.js asset would take minutes to load and that would hold up the whole page19:21
rick_h_hazmat: I only see a zenbox.js at 40ms from a cleared cache19:22
rick_h_hazmat: so network latency?19:22
rick_h_well, network fall-over19:22
hazmatrick_h_, no.. it would never show up19:22
hazmatjust timed out always19:22
hazmati could wget the file but chrome would never load it19:23
rick_h_gotcha, yea not seeing it here. 19:23
rick_h_chrome 33, cleared cache, etc19:23
hazmatno worries, i'll stick with stable for now19:24
rick_h_yea, been back/forth between ff nightly and chrome dev a lot the last couple of days19:24
Makyogary_poster, I can still hear you.  Just a sec, it;'ll come back19:51
MakyoThis network...19:56
arosalesgary_poster, hello20:25
gary_posterhey arosales 20:26
arosalesgary_poster, could you recommend anyone to talk error messages with evilnickveitch  and juju-core in terms of documentation?20:27
gary_posterarosales, which error messages?20:27
arosalesin general error messages generated by the gui, and from juju-core we would like to document for users20:27
gary_posterarosales, ack.  we probably would want to include quickstart in that too.  I think hatch (around now) and frankban (not around now) would be good choices.20:29
* hatch perks up20:30
arosalesgary_poster, thanks. I still need to find a good juju-core person. So it will probably be morning us time. I'll reach out to hatch and frankban to coordinate20:30
gary_posterfrankban in particular would be good for bundle error messages, juju-core messages we expose, and quickstart.  hatch can help with where we might expose them.20:30
gary_postergreat arosales, sounds like a good initiative20:30
hatchactually right now I'm working on juju-core unit errors for the GUI hah20:31
arosalesgary_poster, np, thanks for the help.20:31
hatchtrying to track down all of them20:31
arosaleshatch, works out nicely :-)20:31
gary_posterhatch, nice relation shrink/expand does not work on saucy FF for me :-/ .  Filing bug.  not a showstopper20:38
hatchhmm, do other css transitions work?20:39
rick_h_want me to test on FF nightly? steps to repo?20:39
rick_h_repro ugh20:39
gary_postertrying to remember where others are20:39
hatchedit a config field20:39
hatchthe 'save' thing will slide up20:39
gary_posteroh right.  yeah that works20:40
hatchhmm darn20:40
hatchI'm trying on qantal, see if I can repro20:40
gary_posterI'm trying on IE1020:40
rick_h_reminds me, I saw IE11 was released to win7 and such. Do we need to qa/check that out? 20:41
* rick_h_ isn't sure how the rules on IE-land are written20:41
hatchrick_h_: I already created a card for Friday :)20:41
gary_posterrick_h_, hatch, yeah we need IE11. :-/20:41
rick_h_hatch: ah, good stuff20:42
hatchI'm ON IT!20:42
gary_posterhatch, shrink/expand works on IE1020:42
hatchyeah I tested in IE10 but not in FF20:42
hatchlooks like it doesn't work in quantal either :/20:42
gary_posterAnd Mac FF is broken in other ways I forget20:42
hatchcan't drag and drop20:43
gary_posterEverything seems to work pretty nicely in IE10 actually.  bundle deploy was good.20:44
hatchyeah IE10 is surprisingly good in 'most' areas :)20:44
gary_posterDnD in IE10 doesn't get positioning right but I am overlooking that :-P20:44
hatchpositioning....what positioning :P20:44
hatchfyi it's caused by the center 'drop here' thing20:44
gary_postereven when it is not there?20:45
gary_posternot visible, sigh20:45
hatchhmm I thought it worked properly when it wasn't there20:45
hatchit's been a while since I looked into it20:45
gary_posterseems to put everything off to the left20:45
hatchyeah darn ok20:46
hatchI think the issue with firefox is a height calculation issue20:47
hatchsourcemaps aren't working though20:47
hatchwill have to look into it later20:47
gary_posterthanks for quick triage20:47
hatchlooks like benji 's bug 1251014 is a case I didn't anticipate20:48
_mup_Bug #1251014: The "Select relation type" dialog can be hidden by the inspector. <juju-gui:Triaged> <https://launchpad.net/bugs/1251014>20:48
hatchshould be easy to fix though20:48
hatchjust move the event firing20:48
benjiYay for exploritory QA20:49
hatchbenji: is this your first? lol20:49
benjihatch: almost20:49
BradCrittendenbenji: i just saw your question about the password.  you still need it?20:52
benjiBradCrittenden: yep20:52
gary_posterjujugui finished releasing GUI.  trunk is open.  Working on charm now...20:54
=== BradCrittenden is now known as bac
* bac walks jojo. bbiab.20:58
rick_h_hatch: around?21:48
hatchrick_h_: yup22:21
hatchsorry I missed the ding22:21
hatchMakyo: what vm did you use for your air?22:23
MakyoVirtualBox, running Raring server.22:38
* Makyo dogwalks, btw.22:38
hatchMakyo: cool thanks22:39
hatchhmm apparently the dinging in my irc app is broken22:40
rick_h_hatch: the mac folks I talk to are more parallels/fusion folks. Guess it's supposed to work better on OSX23:04
rick_h_hatch: all good, I'll check in with you tomorrow. Want to make sure that the way this communication with the WSS and suck goes it's blocking or waits on rpc call or not23:04
hatchrick_h_: yeah but parallels doesn't appear to support the latest ubuntu23:06
hatchI just twitted at them to ask if they have an ubuntu support matrix23:06
hatchso I'm going to give virtual box a go on the mrs computer23:06
hatchI'll need more information on your wss q23:07
hatchsaucy server with 3.11 installed with virtual box on the air23:25
hatchlooks like it's working well so far23:25

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