/srv/irclogs.ubuntu.com/2013/11/18/#juju-gui.txt

bachey benji, check it out http://staging.jujucharms.com/bundle/~jorge/mediawiki-simple/mediawiki-simple12:16
rick_h_downloaded!12:20
rick_h_frankban: ah, thanks for the bug update. Makes sense. it was a drive-by when I was testing out the instructions in the bundle deploy tab in the gui and didn't have juju-core installed yet12:42
frankbanrick_h_: np12:43
bachey lbox, would it be too much to retry my google credentials, especially at the end of a 10 minute process?12:59
=== gary_poster|away is now known as gary_poster
bacrick_h_: your merge destroyed my branch...  :(13:18
bacyou did warn me...13:18
gary_posterhey bac, lp thinks you have a merge conflict in https://code.launchpad.net/~bac/juju-gui/send-bundle-id-to-deployer/+merge/19559813:18
gary_posteroh you are talking about that :-P13:19
rick_h_bac: :(13:25
rick_h_bac: let me know if there's anything I can do to help 13:26
bacrick_h_: thanks, but i don't think so.  just have to re-qa now.  no biggie.13:27
rick_h_bac: cool13:27
gary_posterfrankban, doing review/qa13:39
frankbangary_poster: thanks! guihelp: anyone available for another review of https://codereview.appspot.com/28250044 ? thanks13:40
rick_h_frankban: sure thing, can peek at it in a few13:40
frankbanrick_h_: great thank you13:40
gary_posterfrankban, while qa-ing, idea: would be nice to have a quickstart option that exposes webbrowser options chrome, chromium, firefox, and (would require registering extra options in webbrowser module) chrome incognito and chromium incognito.  Would be convenient for qa plus for people who don't default to a browser we support13:55
frankbangary_poster: +113:55
gary_posterfrankban, also, this idempotent functionality is cool :-)13:56
frankbangary_poster: :-) yes it makes quickstart more useful13:57
=== adeuring1 is now known as adeuring
hatchmorning, how was everyones weekend?14:05
gary_postergood 'nuff.  yours?14:11
hatchdecent until last night14:13
gary_posterwhen?14:13
hatchfound a slow water leak at 12:00am14:13
hatchhad to fix it before sleeping14:13
hatchlol14:13
* hatch tired14:13
gary_posterugh14:14
gary_postersorry14:14
hatchyeah sheesh quick breaking my pipes!14:15
hatchlol, you'd make a good Canadian ;)14:15
hatchfrankban: I put my slides for my talk up this weekend - sorry no screencast yet but coming soon http://fromanegg.com/post/67170468715/intro-to-go-from-a-javascript-developer-barcamp-yxe14:19
frankbanhatch: cool, will take a look. how was your talk?14:21
hatchfrankban: it went really well, lots of good questions about juju/go/js and there was about 100 people there, so fairly large crowd14:24
frankbanhatch: nice!14:24
hatchyeah - and like always everyone is super impressed with Juju :D14:25
hatchwell, and the GUI since that's how I demo'd it haha14:25
frankbanhatch: oh! you also demoed the gui. cool, looking forward for watching the screencast14:26
hatchfrankban: heh yeah, it was just as an asside because I had some extra time but people were really impressed14:30
hatchone of the questions "how do you configure bundles before deploying them?" cc/ gary_poster ;)14:31
gary_posterhatch, heh :-)14:31
gary_postercool14:31
rick_h_frankban: I'm not following why the bootstrap check can't be the results of juju status? It fails if the env doens't exist, isn't bootstrapped, and you get a known res if it is. You'll see machine 0. So it's a single positive check?14:32
hatchrick_h_: before i forget, did you end up filing that bug about the guiserver returning the wrong info?14:34
rick_h_frankban: also, side note. Can we get a terminal wipe/clear after the "Created new window"14:34
rick_h_hatch: oh! thanks. It's on my notepad here. 14:34
rick_h_frankban: time for a chat actually? some questions on stuff from friday 14:34
hatchheh same here and I didn't see the email so I was curious :D14:35
frankbanrick_h_: status can fail for a lot of reasons. and even if we have a bootstrap node, the agent can be down. moreover I prefereed to just bootstrap in order to not waste extra-time for the very common case when the environment is not bootstrapped14:35
rick_h_frankban: ok14:35
gary_posterfrankban, rick_h_, -1 on wipe/clear, but understand goal, I think.  Would be interesting in hearing other solutions14:35
gary_posterbut scrollback--including password for now!--too nice to have to wipe14:36
rick_h_gary_poster: well, not wipe, but there's a terminal code I think that moves the line correctly so that you don't get $prompt $message14:36
gary_poster+1 on that, right14:36
* rick_h_ remembers having to do something a while ago searches memory14:36
frankbanguihelp: I don't remember if the following is in the guiserver docs: when deploying a bundle, some info about what's going on can be found in two places: 1) https://<gui-server-address>/gui-server-info and 2) /var/log/upstart/guiserver.log14:36
bacjujugui: review, please, old skool style (lbox hates me this morning): https://code.launchpad.net/~bac/juju-gui/send-bundle-id-to-deployer/+merge/19559814:37
hatchbac on it14:37
hatchbac: is this the 4XX error you're seeing again?14:37
gary_posterfrankban, ack.  Could you add to...maybe both quickstart and charm docs?  Somewhere. :-)14:37
bachatch: i just went through the QA for the second time today.  do it if you want or pass, up to you.  (it takes a while)14:38
bachatch: yeah, 404 failure on tests from lbox14:38
bachatch: plus dumb google auth error.  just decided to skip it this time.14:38
hatchjust make sure someone does it :)14:39
frankbangary_poster: sure14:39
rick_h_frankban: let me know when you get a few min to chat on the bug/issue from friday14:39
frankbanrick_h_: now is ok14:39
hatchthese diffs are sure hard to read14:39
* hatch spoiled14:39
rick_h_frankban: https://plus.google.com/hangouts/_/72cpimpnpmmm1v3304ph2cjoa8?hl=en14:41
hatchbac: so the bundle object data is always required but the id is optional - can't we pull the id from the bundle data then?14:43
bacno14:43
bacnot in the yaml14:43
hatchso if someone drags the yaml in, we pass the id in?14:45
hatchI don't think that we do14:47
hatchso in that case it would be undefined anyways14:47
gary_posterRhetorical question: Why does it seem slower to load the GUI from jujucharms.com than from a Juju deployment running the real tool? :-/14:50
hatch3 layers of firewalls14:51
hatchoh sorry, Rhetorical14:51
hatchlol14:51
gary_posteryeah, was thnking something along those lines :-/14:51
gary_posterthe firewalls, that is ;-)14:51
rick_h_gary_poster: london 14:51
rick_h_gary_poster: jujucharms.com crosses the ocean, ec2 is a EST DC14:52
gary_posterah, yeah, maybe so14:52
gary_posterI mean, that's true :-)14:52
rick_h_plus the three fireealls :)14:52
gary_posterbut maybe that's the explanation14:52
gary_posteryeah14:52
rick_h_err firewalls14:52
rick_h_yea, there are times I think hosting all our stuff out of a central London location hurts us from a performance perspective. In my US is the center of the world view of things 14:53
hatchhehe14:53
gary_poster:-)14:53
hatchhttp://www.dell.com/us/business/p/xps-13-linux/pd.aspx14:54
hatchlooks like they come with touch screen, haswell, ubuntu14:54
hatchcredit...card....burning....14:55
rick_h_no trackpoint :(14:55
hatchyou're clingin to the past....man....no need for a trackpoint when you have touchscreen.....man14:56
rick_h_ugh, I don't believe in touchscreen laptops14:57
hatchI didn't either14:57
rick_h_already have a fit when my wife points at things and touches it on accident14:57
hatchhaha14:57
gary_posteripad + keyboard = great experience14:57
hatchlying in bed browsing the internet a touchscreen would be nice14:57
rick_h_I've got my N10 + keyboard and it's cool for some things, but I get annoyed at touching the screen. I'm too terminal/vim/keyboard all the things! I guess14:58
rick_h_every time to move my hands off the keyboard I feel like someone's going to bring a rule down on my hand :)14:58
rick_h_tablet for that14:58
hatchmouse --move --left 10 --click14:58
rick_h_(the bed browsing) I do <3 the N10 for a lot of that casual stuff14:58
bachatch: missed your question.  yes, if the yaml is deployed from a local file then the bundle has no ID and it is not credited in charmworld.15:00
hatchbac: right, so we can 'always' parse the id out of the object, if no id, then no problem15:00
hatchso you don't need to deal with passing it in externally15:00
bachatch: what line are you refering to?15:01
hatchwell it could be implemented in env.bundleImport()15:01
rick_h_frankban: ok, submitted under https://bugs.launchpad.net/charms/+source/juju-gui15:02
hatchthen you wouldn't need all of the extra data flying around15:02
hatch(plz correct me if I just15:02
hatch"don't get it")15:03
hatch:)15:03
bachatch: first, i don't know what env.bundleImport is.15:03
hatchgot a second for a quick call?15:03
rick_h_hatch: it's moved remember15:03
bacsure15:03
hatchhttps://plus.google.com/hangouts/_/72cpi0qnq012rqp4mojqnq42kg?hl=en15:03
frankbanrick_h_: thanks, so this means that it's an error only when you deploy that bundle, right? It is not the interaction between two subsequent deployments. What happen if you deploy a working bundle aftre a failing one?15:04
rick_h_frankban: if I deployed the bundle on it's own it works fine15:05
rick_h_frankban: so it only died out when I first had the failing one15:05
gary_posterfrankban, LGTM with small and trivials; QA good, though I added a second note about something that didn't go as your QA instructions prescribed.15:05
frankbangary_poster: thanks, re your QA note, in the previous step I added "juju deploy -e ec2 juju-gui" after the service removal...15:07
gary_posterfrankban, oh, sorry!15:07
frankbangary_poster: np, I did that, it works ;-)15:07
gary_posterfrankban, cool :-)15:08
gary_posterbac, you don't need another review/qa, right?15:09
bacno15:09
gary_posterk15:09
gary_postergood, 'cause I wasn't going to give one. ;-P15:09
frankbangary_poster: great stuff in you review, thanks! only one question: how should the core bug be? "please don't change your error message"? a --format for an error seems weird to me, but maybe I am wrong15:13
frankbangary_poster: maybe "expose a safe way to know if an env is bootstrapped"15:14
frankbanrick_h_: so discourse does not fail unless another bundle failed before?15:16
frankbanrick_h_: or discourse bundle intermittently fails, and when it does, the error is empty?15:16
gary_posterfrankban, heh.  Minimally, we can describe the problem to them and ask for ideas. As an idea, though, a --format that affects both normal output and error output seems reasonable to me FWIW--you are asking for a machine-readable response, whatever the response is.  What you suggested is similar to my "minimally" bit; if you go that way, I would s/safe/safe, machine-readable/ or something15:17
rick_h_frankban: jenkins does not get the failure status unless discourse was run before hand and failed15:17
frankbangary_poster: ack15:19
frankbanrick_h_: ack, I was looking the wrong bug: https://bugs.launchpad.net/charms/+source/juju-gui/+bug/125229515:21
_mup_Bug #1252295: guiserver bundle deployment error is empty <juju-gui (Juju Charms Collection):Triaged> <https://launchpad.net/bugs/1252295>15:21
rick_h_frankban: ah ok. Yea filed those as two parts. One is the error being empty which can happen on its own. 15:21
rick_h_frankban: the second was the first failure making the second one register as failed15:22
frankbangary_poster: rethe already_bootstrapped argument. perhaps I can make deploy just take a "to" argument and make the check in "run"15:22
frankbanrick_h_: cool thanks15:22
* gary_poster looking/thinking15:23
frankbangary_poster: sorry, s/already_bootstrapped/env_type15:25
gary_posterah!15:25
frankbangary_poster: well, maybe it's not a great idea. and +1 on s/already_bootstrapped/check_preexisting or similar15:25
gary_posterfrankban, I like the idea of replacing  env_type with to.  Keeps the local function simpler, and the arguments more self-describing15:26
frankbangary_poster: cool, ok15:27
gary_posterfrankban, fwiw, when you said already_bootstrapped at first, I was trying to make it work, :-) and thought you meant something that seems maybe interesting.15:27
gary_posteryou would split up deploy service and deploy unit into two functions15:27
gary_posterand run would get the status and decide what to do (perhaps from a third function)15:28
gary_posterfrankban, shrug, might make code simpler to read?  it would keep the deploy functions only about deploying15:29
gary_posterwhich might be nice15:29
gary_posterjust an idea from trying to make sense of a miscommunication, so feel free to ignore :-)15:30
frankbangary_poster: hum... it seems interesting, but then you have to pass service_data and unit_data, or put the logic in run... something to investigate in a future branch?15:31
gary_posterfrankban, +1, or not :-)15:32
frankbangary_poster: cool ok :-)15:32
bachatch: finally!  https://codereview.appspot.com/2844004315:36
hatchhah yay!15:37
hatchbac: lgtm'd with a comment as we discussed15:39
frankbangary_poster: filed bug 125232215:42
_mup_Bug #1252322: Expose a safe/machine readable way to check if an environment is already bootstrapped <juju-core:New> <juju-quickstart:Triaged> <https://launchpad.net/bugs/1252322>15:42
gary_postercool frankban thanks15:42
rick_h_frankban: LGTM'd and qa ok. Thanks for the update! This should allow me to use it on lxc now as long as I pre-bootstrap yay15:43
frankbanrick_h_: not yet ;-) we still have an explicit check for local providers (and the bits in deploy is actually a yagni). but I planned lxc support to be my next quickstart card15:44
rick_h_frankban: ah right, but the deploy a bundle doesn't worry about the local provider. That's mainly what I was try to get at15:45
* rick_h_ looks again, maybe I'm not in my lxc happy place 15:46
Makyojujugui call in 1015:50
hatchhey Makyo did you ever find out a good set of apps for doing a screencast on Ubuntu?15:51
MakyoRecordMyDesktop + Audacity + kdenlive15:51
MakyoThough I'm still working on finding the best transcoding rate for kdenlive to get rid of artifacting.15:52
MakyoRMD records at 15fps which kdenlive doesn't like.15:52
hatchhmm - I wonder if I can use http://www.telestream.net/screenflow/overview.htm in OSX but have it cast the Ubuntu vm15:53
MakyoThis happened last time you asked, too :)15:53
hatchyeah I'm indecisive because it's a $100 gamble haha15:54
hatchhttp://www.omgubuntu.co.uk/2013/04/lightworks-enter-public-beta this was also just released15:54
gary_posterjujugui call in 215:58
benjijujugui: well, I guess that was the wrong window to close16:12
hatchlol16:12
gary_posterlol16:12
rick_h_benji: drops the mic and leave "I'm outta here"16:12
benjiheh16:12
benjias a live sound guy, I will kill anyone that drops a mic on purpose16:13
rick_h_just know who to give that fisher price mic to 16:14
benjiSM58s are indestructable16:14
rick_h_gary_poster: http://techcrunch.com/2013/11/18/surveymonkey-enterprise/ is interesting fyi16:14
gary_posterrick_h_, huh, cool, thanks.  will forward to Luca and Charline16:16
hatchhey look at that I got the YUI is not defined error16:27
hatchfirst time I think16:27
hatchwell at least now I know what the issue is16:27
hatchjujucharms.com is not serving the all-yui.js file at all16:27
hatchit 404'd16:28
gary_posterhatch, I get it16:28
hatchoh really? yours still downloads -and- throws the error?16:28
luca__gary_poster: do you have a hangout?16:28
gary_posterluca__, https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.nm61lc93dnv0a4s18q2tt3650c16:29
gary_posterhatch, error is intermittent16:29
hatchyeah it's working again - I'm just saying that the issue is not related to anything clientside - the server is closing the connection to all-yui.js16:30
hatchdumb question but what is #!/bin/bash called ? I'm trying to figure out if sublime can enable syntax highlighting based on it16:34
rick_h_shebang16:34
* hatch turns on safesearch16:34
* hatch searches for shebang16:34
hatchlol16:34
rick_h_http://www.sublimetext.com/forum/viewtopic.php?f=3&t=1465616:34
rick_h_hmm, https://sublimetext.userecho.com/topic/20806-syntax-highlight-based-off-shebang-lines/ says it should be doing it 3yrs ago16:36
hatchheh yeah - it definitely doesn't16:36
hatchjust configuring the applysyntax plugin16:37
hatchoh that might have been for sublimetext 116:37
frankbangary_poster: re charm URL check (if a pre-existing service is found in the env): is it ok for you to make another card?16:48
gary_posterfrankban, on call but ok.  you do it? :-)16:49
frankbangary_poster: sure :-)16:49
gary_posterthx16:49
hatchblarg16:59
hatchis there a juju hook api somewhere?16:59
rick_h_hatch: heh, it was about 10 sessions at the last sprint16:59
rick_h_nick is getting the first api for a hook written up currently :)16:59
hatchthis just reading charms for the various commands is driving me nuts17:00
rick_h_yea, those meetings were painful imo. I like clear documented standards, but a large group of people want the flexibility/don't hold it back stuff in place now17:00
hatchI asked in #juju hoping that someone has something17:02
rick_h_good luck :)17:02
hatchhah17:02
hatchactually....17:02
hatchI do have 1 q you may be able to help me out with....time for a quick hangout?17:02
rick_h_sure17:02
hatchhttps://plus.google.com/hangouts/_/72cpj792njti59826rubg1v7ls?hl=en17:03
hatchoh wait17:03
hatchcancel that17:03
hatchhttps://plus.google.com/hangouts/_/76cpilhst0i5ii6btp3crpe6e4?hl=en17:03
hatchthat one17:03
frankbanhazmat: ping17:22
hazmatfrankban, pong.. in meetings all day fwiw. watsup?17:29
frankbanhazmat: ok, feel free to read this later. re unit placement in machine: in the code path followed when the GuiEnvironment is used, an int value is never converted to a string. Either the Deployment object or the GuiEnvironment itself can make that conversion. What do you prefer? If you want to avoid that conversion to be done in a lower layer (the Deployment class) I can quickly set up a branch which does it in t17:31
frankbanhe GuiEnvironment.17:31
hazmatfrankban, oh... yeah.. this is the issue from friday17:41
frankbanhazmat: yes17:41
hazmatfrankban, i think service or deployment is the most appropriate17:42
hazmatfrankban, the service already has get unit_placement property that would be a good spot for the conversion and a test17:43
Makyohttp://eleks.github.io/js2js/17:43
frankbanhazmat: sounds good, do you want me to work on this?17:45
hazmatfrankban, if you need it the next day, that's probably best, else i can cover it.17:48
hazmater. within the next day17:48
frankbanhazmat: it's not that urgent, I believe end of this week can be ok17:49
hazmatfrankban, cool17:59
frankbanhazmat: great thank you17:59
frankbanguihelp: I need 1 quick review for a documentation only branch: https://codereview.appspot.com/2849004318:04
frankbanthank you18:04
Makyofrankban,  on it18:04
hatchsure18:04
frankbanMakyo: thanks18:04
hatchoh darn18:04
bacdoh18:04
hatchok18:04
hatch:)18:04
hatchughhhh18:07
hatchwhen requestin a boolean value from `config-get` it returns 'False' which is not a boolean value according to bash18:10
hatchbenji: you kickin around?18:27
benjihatch: yes... I think18:27
hatchbenji: http://bazaar.launchpad.net/~hatch/charms/precise/failtester/trunk/view/head:/hooks/handler do I have to quote the $hook and $fail_* to get them to compare as strings?18:28
* benji looks18:28
hatchline 51: False: command not found18:28
hatchis the error I'm getting18:28
hatchwhich means that the $fail_start command is trying to execute as "False" or something18:28
benjihatch: you can use double brackets and not have to quote18:30
hatchoh ok so change them all to [[ ]]18:30
hatchthanks, I didn't know that18:30
hatchI wish that juju returned real boolean values18:32
hatchfor bash18:32
hatchTrue/False I think are -only- booleans in Go18:32
benjiright, double all the square brackets and you should be good to go18:33
benji(double brackets are a bash extension that is much saner than the Posix single brackets)18:33
hatchI really need to study bash more18:34
hatchthe tutorials are.....thick18:34
hatch:)18:34
hatchhere we go, 4th time is the charm, then this charm is done18:35
hatchcharm18:35
hatch:)18:35
hatchsweet works18:41
hatchthanks18:41
hatchnote to self....dont try and create a bunch of lxc's at once18:59
rick_h_hatch: :)19:00
rick_h_hatch: yea, I think my limit on my desktop is around 2019:00
rick_h_over that and things lock up for a long while, IO bound 19:00
hatchyeah the box has raid 1 too so it's really bound19:01
hatchheh19:01
hatchDart 1.0 was released19:05
* hatch starts to port the GUI19:05
gary_posterheh19:13
rick_h_hatch: come on https://twitter.com/mitsuhiko/status/40109674419875840019:16
hatchsee he has the right idea19:16
rick_h_surely you're ready to take that port and go to typescript19:16
hatchI write dart in typescript19:16
rick_h_then we can check out rust and see if it offers anything :P19:16
hatchlol19:17
hatchI don't know who this guy is but I hope that his comment is joking19:17
rick_h_Armin?19:17
hatchyeah19:17
rick_h_he's famous python dude 19:17
rick_h_though I'm not sure how much python he does these days19:18
hatchI just hope he realizes that Dart may compile down to js like typescript but it is also it's own language19:19
rick_h_yes, I'm pretty sure he's aware19:20
hatchyou never know on these internets...you never know19:20
hatchhave I mentioned how much I love the gui?19:20
* hatch frames a picture of it on his desk19:20
hatchI just wish I could deploy a specific url from the GUI19:21
bacthanks for the note bcsaller19:25
bcsaller:)19:25
* hatch shakes juju "let me deploy from launchpad"19:29
hatchjcastro: is discourse.ubuntu.com supposed to be mobile only css?19:30
hatchlooks fixed now :)19:31
=== marcoceppi_ is now known as marcoceppi
* hatch is taking lunch19:42
* gary_poster switches computers, because he needs to sit outside, because it is gorgeous19:55
=== gary_poster is now known as gary_poster|away
=== gary_poster|away is now known as gary_poster
hatchrick_h_: doesn't charmworld ingest every 15?20:33
rick_h_hatch: rgr20:34
hatchhmm20:34
rick_h_hatch: 00 15 30 4520:34
hatchhttps://localhost:1111/sidebar/search/~hatch/precise/failtester-3/?text=failtester is on v320:34
rick_h_hatch: then it takes a few min to process20:34
hatchand v5 was pushed like an hour ago20:34
rick_h_hatch: well, it's meant to fail right? does it pass proof?20:34
hatchoh lol20:34
hatchactually it should...20:34
rick_h_well try it out, if it fails proof then it won't ingest20:35
hatchcan I just deploy from launchpad somehow?20:35
rick_h_hatch: sure, bzr branch .... && juju deploy --local ....20:35
hatchok so no :D20:35
rick_h_:)20:35
hatchkind of a crazy limitation20:35
rick_h_meh, why? it's one extra command to pull it down local20:36
rick_h_and you get safety/checks and such if you use the store based stuff, which we want to encourage charm authors to get their charm into and users to look for latest/best charms20:36
rick_h_seems kind of natural to me20:36
hatchright, but it's an aritificial limitation20:37
hatchevery other package manager allows you to point to any location20:37
rick_h_yep, for the good of the users20:37
rick_h_apt-get install random_url ?20:37
rick_h_:)20:37
rick_h_mac store install random_url ?20:37
hatchnpm install <url> go get <url>20:37
rick_h_that's a library, not a software package 20:38
rick_h_that's for developers and in some languages requires extra flags and such20:38
rick_h_pip install http:// fails20:38
rick_h_you have to -e git@....20:38
hatchcharm proof is broken20:38
hatchdarn20:38
rick_h_since by default it looks for things that it knows are buildalbe in the pypi 'store'20:39
rick_h_how is charm proof broken?20:39
hatchit's forcing me to add a revision file....which is deprecated20:39
rick_h_update it recently? 20:39
hatchjust installed it20:39
hatchheh20:39
rick_h_hmm, I thought it was a warning?20:39
rick_h_it's still darn handy because there's a -u flag on juju deploy that'll auto increment the version and help force an update over the cache20:39
hatchW: No icon.svg file.20:40
hatchE: no copyright file20:40
hatchE: revision file in root of charm is required20:40
rick_h_ah, cool20:40
hatchso the docs or proof is wrong20:40
hatchwonder who would know which is which?20:41
rick_h_https://bugs.launchpad.net/charm-tools then I guess. Get it a warning. 20:42
rick_h_it is useful in charm dev ime until they add a no-cache or something flag20:42
hatchhmm this is interesting20:43
hatchI've never actually 'cared' while pushing a charm20:43
rick_h_no, proof is run in charmworld during ingest20:43
rick_h_proof is up to you. You don't need to get your charm in charmworld20:43
rick_h_it's just encouraged20:43
hatchwell I do if I want to deploy it without pulling it down20:43
rick_h_and a bzr push doesn't run proof since you just installed it20:43
rick_h_sure, but you can push it, and pull it down on another computer and use it20:44
rick_h_and share with others, etc20:44
hatchyeah - I'm not sure that's a good experience20:44
rick_h_right, it sucks...so that you'll proof and fix things to get the good experience20:44
hatchmaybe it's because the documentation isn't explicit20:44
rick_h_I know the publishing your charm docs talk about proof/etc20:44
hatchlike maybe on the 'deploying charms' page if it said "this is why you can't deploy from launchpad, here is what you should do"20:45
hatchright now it says <repository>:<series><service>20:45
hatchwhich kind of implies that you CAN use other repositories20:45
rick_h_#4 https://juju.ubuntu.com/docs/authors-charm-store.html20:46
rick_h_so I think the juju-core store can point at different repositories20:46
rick_h_charmworld only looks at the official juju-core store off LP20:46
rick_h_and that's just an artifact of 'technically you can do' vs 'what you should be doing'20:47
hatchyeah I think the docs maybe need to be a little bit more explicit maybe20:47
hatchhonestly I have no idea how up to v3 got into charmworld20:48
hatchit's never had a revision or copyright file haha20:48
rick_h_well I wonder if there's something else there then. Maybe we ignore those errors in charmworld? 20:48
hatchis there any way I can get a report of the charmworld ingest?20:49
hatchjust to see what the issue is?20:49
rick_h_not really :/ http://manage.jujucharms.com/~hatch/precise/failtester20:51
rick_h_shows the lint status, but if it pulled in 3 revs it must ignore those errors20:51
rick_h_hmm, and rev 4 diff seems innocent enough20:51
hatchand the lint check results are different than my local lint check results heh20:52
rick_h_hatch: ah, it's not in the juju-core store https://store.juju.ubuntu.com/charm-info?charms=cs:~hatch/precise/failtester20:52
hatchok I don't know what you mean20:52
rick_h_hatch: yea, that'll probably be a revision thing. mjc will only update proof when told to manually20:52
hatchohh20:52
rick_h_hatch: yea, basically it's not your fault and not charmworld's fault. It's in juju-core's hands20:52
hatchso.....should I file a bug somewhere?20:53
hatchI guess I don't really understand the problem20:53
rick_h_yea, it's the black hole. We've hit issues here before but I've not been sure either. 20:53
rick_h_last time hazmat did some poking at things for us. There was some revision issue that caused grief. I don't recall the details 100%20:53
hatchso do you think that adding the revision file should fix it?20:54
rick_h_hazmat: is there someone we can bug about the core-store not pulling in a revision for a charm? 20:54
rick_h_hatch: only incidentilly as it'll add a new revision that might be picked up by the store20:54
hatchahh - darn I hate black boxes20:54
rick_h_hatch: so basically charmworld checks that the store knows about your revision https://store.juju.ubuntu.com/charm-info?charms=cs:~hatch/precise/failtester 20:56
rick_h_hatch: if it doesn't we don't ingest it since you can't deploy it20:56
hatchahh ok that makes sense20:57
rick_h_hatch: juju deploy hits this store, not charmworld to do the actual deploy work20:57
hazmatrick_h_, use juju publish the charm is probably broken20:57
hazmatpublish will give some feedback, alternatively its a really simple go program to walk through the charm validation/errors from core20:57
rick_h_hazmat: rgr hatch ^^20:57
hatchcool I'll try that20:57
rick_h_hazmat: cool, thanks20:57
hatchohh21:02
hatchapparently you can't have a requires and provides with the same name21:02
hatchguess I should have assumed that21:03
hatchthanks hazmat rick_h_21:03
hatchrick_h_: your CHC meetups are they actually in coffeeshops?21:10
rick_h_hatch: yes21:10
hatchthey don't seem to mind?21:10
hatchor are there no 'presentations' ?21:10
rick_h_no, the last place had a meeting room in the back and we rented it out each week21:10
hatchoh that's cool21:10
rick_h_no, we've done a couple of 'let's check out YUIConf videos on the wall'21:11
rick_h_but it's mostly dedicated time to hack on something you care about of your own choice each week21:11
hatchcool - we have had a few small dev meetups but the issue is always the location21:11
rick_h_no matter how busy the week gets, I've got 2hrs of wed night to hack on my bookmark app, or read that article I was meaning to get to, or to vent, etc21:11
rick_h_yea, our shop got bought by another coffee chain and they opened up the room and then changed closing time from 10pm to 8pm21:12
rick_h_so I had to go hunting for a new coffee shop to hit up this weekend21:12
hatchoh darn21:12
rick_h_visited a couple of places21:12
bacjujugui: just deployed a bundle.  used the gui to destroy both services.  both removed from canvas.  but 'juju status' shows mysql still hanging around.  known issue?21:14
bacoh, it is marked 'dying' and has been for a while.21:15
hatchyeah it takes a long time to die sometimes21:15
hatchat least in my experience21:15
hatchwhat is 'a while' ?21:15
hatchive seen it take 30mins sometimes :/21:15
bacgah, really?21:18
bacit's only been 10-15 minutes.  i just picked another test bundle with no overlapping services21:19
rick_h_bac: yea, done that before as well21:20
rick_h_jenkins is a good alt21:20
hatchyay updated21:20
hatchrevno not matching the bzr val is odd though :)21:21
hatchman I'd really love to know why discourse.ubuntu.com keeps switching to mobile styling21:49
hatchand FINALLY I have finished testing my charm22:10
hatchboy that takes a long time22:10
hatchgood thing I could do other stuff hah22:10
hatchI'll be writing up an email to the juju mailing list but here it is if anyone wants to use it until then https://jujucharms.com/fullscreen/search/~hatch/precise/failtester-5/?text=failtester22:14
bacjujugui (i.e. hatch): quickstart review? https://codereview.appspot.com/28520044/  no rush, i'm sure frankban will get to it in his morning.22:55
Makyobac, on it.22:58
bacoh, hi, Makyo.  forgot you were still around.22:59
MakyoYeah, just got back from dogwlak.22:59
MakyoOr however you spell it.22:59
MakyoSun sets so early these days :/22:59
bacthat's how i always spell it22:59

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