2014/02/05/#juju-gui.txt

[00:21] <Makyo> AUGH
[00:21] <Makyo> I can't reproduce these CI errors at all locally.
[00:39] <hatch> Makyo you are creating an environment then doing nothing with it in one of your tests
[00:39] <hatch> maybe that's it?
[00:39] <Makyo> Oh?  Hmm..
[00:40] <hatch> line 1001 in test_environment_view.js
[00:41] <Makyo> Oh, right.  Hmm.  I got yelled at by jshint for that.
[00:41] <Makyo> Er, for assigning it to a variable I never used.
[00:41] <Makyo> But I bet I can find a way around that.
[00:42] <Makyo> Oh, there's already a global x.x
[00:42] <Makyo> If this works, I owe you.
[00:43] <hatch> haha, guess we'll see
[00:44] <Makyo> Damn, I bet that's it, too.  Just causing weird cascading issues.
[00:50] <Makyo> ...
[00:50] <Makyo> ... ... ...
[00:50] <Makyo> Disregard that, I'm an idiot.
[00:51] <Makyo> It worked locally because of untracked templates.
[00:51] <hatch> lol
[00:51] <hatch> hey, at least you can now reproduce it locally
[00:51] <Makyo> That'll teach me to use commit -a
[00:52] <hatch> ahh yeah I always do git add -A
[00:53] <hatch> ran into the same issue with untracked files
[00:53] <Makyo> Yep.
[00:53] <Makyo> It was that.
[00:53] <Makyo> I'm an idiot.
[00:54] <Makyo> SIGH
[00:55] <Makyo> Would appreciate a review/QA tomorrow, if you get a chance, hatch 
[00:58] <hatch> Makyo yeah no problem
[02:04] <hatch> Makyo are you still around?
[02:08] <Makyo> Yep ^^
[02:08] <Makyo> What's up?
[02:09] <hatch> QA OK
[02:09] <hatch> buuuut
[02:09] <hatch> I keep getting relation change errors 
[02:09] <Makyo> Huh?
[02:09] <hatch> it happens in trunk, just wondering if you noticed them too?
[02:09] <Makyo> I get them with simulator on, yeah.
[02:10] <hatch> ok so you think it's a simulator bug? I haven't been able to pin down the trigger
[02:12] <Makyo> I thought we had intentionally added them to the simulator for testing the inspector.
[02:14] <hatch> well only if relations are created
[02:14] <hatch> so maybe the change is coming in thinking the relation is still there
[02:14] <hatch> ok well that at least makes some sense :)
[02:14] <hatch> it doesn't appear to have any effect on the app so I was just trying to pin it down
[02:14] <Makyo> Oh, hmm.
[02:14] <Makyo> Alright, yeah.,
[02:15] <Makyo> I mean, I'm in there plenty, of late, I'll keep an eye out,
[02:15] <hatch> sounds good, well review and qa are done
[02:16] <hatch> now I'm thinking on how to add databinding to the viewlet manager
[02:16] <hatch> I'm thinking composition of some how
[02:16] <hatch> it just wasn't written with that in mind
[02:17] <hatch> anywho...
[02:17] <hatch> just sitting on the couch with my laptop....probably a mistake hah
[02:17] <hatch> what work/life balance
[02:18] <Makyo> Hah!
[02:18] <Makyo> Go away!
[02:18] <Makyo> Go away.
[02:18] <Makyo> I have gin, that's my sign to go away.
[02:18] <Makyo> Will get to review tomorrow.
[02:24] <hatch> haha sounds good, have a good night
[09:33] <rick_h_> frankban: ping, have you seen Import from "ghost-charm-master.zip" failed.<br/>invalid request: invalid YAML contents: unacceptable character #x0003: special characters are not allowed in "<unicode string>", position 2 
[09:34] <rick_h_> in deploying a local charm? It souded like gary_poster|away had tested the ghost charm but hitting that and not seeing any unicode chars in the config/metadata files?
[09:34] <rick_h_> oh bah, I didn't use the latest core in this env
[09:34] <rick_h_> ignore me, will rebuild the env and try again
[09:35] <frankban> rick_h_: never seen. ok. FWIW, I am trying to release the charm
[10:02] <frankban> rick_h_: in manage heartbeat the Last ingest job seems to just return datetime.utcnow or similar
[10:14] <frankban> rick_h_: charm released
[10:29] <rick_h_> frankban: oooh cool I'll try that first then thanks
[10:35] <frankban> rick_h_: worked here with the ghost charm and core trunk revno 2295: juju boostrap && juju-deploy juju-gui && juju expose juju-gui && juju set juju-gui juju-gui-source=https://github.com/juju/juju-gui.git
[10:48] <rick_h_> frankban: thanks, just back from lunch and working on starting it now
[10:48] <frankban> rick_h_: cool
[10:49] <frankban> rick_h_: the local charm deployment is awesome
[10:49] <rick_h_> frankban: yea, I'm excited to show it off. 
[10:49] <rick_h_> I've got let hatch know it worked out well because the field guys gave a talk on "juju and tools issues"
[10:50] <rick_h_> and one thing that came out was lack of local charm deployment support and I got to say "We've got it this week"
[10:50] <frankban> :-)
[10:50] <rick_h_> which lead to Mark S saying "...you should ask for a unicorn in a month"
[10:50] <frankban> great
[10:50] <rick_h_> was cool to have that ready to go/say in front of everyone
[10:51] <rick_h_> now we just need a better offline/not connected story
[10:51] <rick_h_> but those were the two issues they mentioned
[10:53] <frankban> rick_h_: it would be cool to have the deployer file accept URLs to zip files and/or git branches, and the deployer to download the file (e.g. from git) and to use the API to upload and deploy the local charm
[10:53] <rick_h_> frankban: interesting. We had a talk on new symantics added to bundles today that would be useful info for. 
[10:54] <rick_h_> I think there's a mission to try to keep the deployer simple, so not sure that the git thing would work, but zips seems like something we could do. 
[10:56] <frankban> rick_h_: yeah, that would allow people to deploy their topology even if they are launchpad allergic
[10:59] <rick_h_> frankban: can you file that as a bug/feature request on the deployer?
[11:00] <rick_h_> hazmat: expressed some interest in getting feature requests to help discuss/go through use cases and I think it's a good one
[11:01] <frankban> rick_h_: sure
[11:10] <rick_h_> frankban: for that source you can just do "develop" got the source and it'll pull that repo like you had in your command
[11:10] <rick_h_> frankban: just fyi
[11:11] <frankban> rick_h_: so  juju set juju-gui juju-gui-source=develop ?
[11:12] <rick_h_> frankban: rgr
[11:12] <frankban> rick_h_: that's cool
[11:12] <rick_h_> it's the replacement for the idea of 'trunk'
[11:12] <frankban> rick_h_: filed bug 1276532
[11:12] <rick_h_> as a shortcut
[11:12] <rick_h_> frankban: awesome thanks
[11:12] <frankban> np
[11:28] <rick_h_> frankban: heads up, going to add #1276541 to the board regarding the recent password text updates
[11:32] <rick_h_> frankban: you're supposed to just drop it on the cavas right? To deploy the local charm?
[11:32] <rick_h_> I'm getting Import from "bookie-charm-master.zip" failed. invalid charm archive: bundle file not found: metadata.yaml
[11:39] <frankban> rick_h_: are you using trunk?
[11:40] <rick_h_> frankban: yes, 2205
[11:40] <rick_h_> 2295
[11:40] <rick_h_> and the charm's version.js is trunk of the gui
[11:41] <rick_h_> well so juju --version shows 1.17.3-trusty-amd64
[11:42] <rick_h_> I went into the GOPATH/src/launchpad.net/juju-core folder to check the bzr rev
[11:42] <frankban> rick_h_: I'll try to reproduce
[11:42] <rick_h_> frankban: what charm did you deploy?
[11:42] <frankban> the ghost zip from github
[11:42] <frankban> rick_h_: where do I find yours?
[11:42] <rick_h_> I tried https://github.com/hatched/ghost-charm with the 'download zip' and https://github.com/bookieio/bookie-charm the same
[11:43] <frankban> rick_h_: uhm... weird
[11:44] <frankban> rick_h_: output from /var/log/upstart/guiserver.log?
[11:44] <rick_h_> getting 400 Bad Request from tornado
[11:44] <rick_h_> looking
[11:44] <rick_h_> [W 140205 11:43:41 web:1635] 400 POST /juju-core/charms?series=precise (41.164.30.187) 358.62ms
[11:44] <rick_h_> [W 140205 11:43:47 web:1635] 400 POST /juju-core/charms?series=precise (41.164.30.187) 350.50ms
[11:45] <rick_h_> is all it says
[11:45] <frankban> rick_h_: if you set builtin-server-logging=debug it will show the whole requests/responses
[11:46] <rick_h_> frankban: rgr, updating
[11:46] <frankban> deploying the charm now
[11:48] <rick_h_> frankban: http://pastebin.ubuntu.com/6878648/ is my updated log
[11:50] <frankban> HTTP error before end of send, stop sending
[11:50] <rick_h_> yea
[11:50] <rick_h_> 400
[11:50] <rick_h_> :/
[11:50] <frankban> interesting, so you are on ec2
[11:50] <rick_h_> right
[11:51] <rick_h_> what are you on?
[11:51] <rick_h_> just local?
[11:52]  * rick_h_ wanted to get an ec2 setup all prepped for the demo running and can leave it overnight ready to go for the morning
[11:53] <frankban> rick_h_: lxc yes
[11:53] <rick_h_> frankban: hmm, I created an lxc to build core in so didn't want to try to do lxc in lxc
[11:53]  * rick_h_ wonders if that'll work :/
[11:54] <rick_h_> frankban: ok, will add a card to investigate ec2 error. If you get time can you see if you can replicate?
[11:54] <frankban> rick_h_: sure
[11:55] <rick_h_> thanks, just to make sure I'm not doing something wrong. Will see if I can get a lxc env setup with juju trunk and try again
[11:55] <frankban> rick_h_: FWIW your bookie charm works well here in a local env
[11:55] <frankban> trying ec2
[11:55] <rick_h_> frankban: ok yea must be provider I guess
[11:56] <frankban> rick_h_: are you using quickstart to deploy the GUI?
[11:56] <rick_h_> frankban: no, I didn't
[11:56] <frankban> rick_h_: ok
[11:56] <rick_h_> I've got to setup that demo as well. When I tried to use qiuckstart in lxc I ended up with two machines still
[11:57] <rick_h_> had figured it might work on lxc since it's not sudo enabled now
[11:57] <frankban> rick_h_: asking because quickstart uses /usr/bin/juju, so it always uses stable juju
[11:57] <rick_h_> frankban: ah, no I did juju bootstrap && deploy juju-gui
[11:57] <frankban> rick_h_: yes qs + lxc installs the gui on a new machine, because the bootstrap node in that case is localhost
[11:58] <frankban> rick_h_: right, that's the same I am doing
[11:58] <rick_h_> frankban: ah that's right
[11:58] <rick_h_> frankban: ok, yea it's less impressive demo since it takes a bit to setup the second machine but maybe will do a bundle with the quickstart/gui part already running. 
[11:59] <frankban> rick_h_: to demo quickstart you might want to use juju-core 1.16 (the new version with "sudo" changes for newer revisions is not yet released
[12:00] <rick_h_> frankban: well I'll probably just 'show off' the -i and that it does bundles from a url. 
[12:00] <rick_h_> I've only got a couple of mins for that and for the local charm deploy
[12:01] <frankban> rick_h_: cool
[12:10] <frankban> juju-core trunk definitely improves debug-log readability
[12:11] <frankban> rick_h_: e.g. http://pastebin.ubuntu.com/6878732/
[12:14] <frankban> rick_h_: what's the output of your ec2 "juju status"?
[12:15] <rick_h_> frankban: ah sorry, destoryed it. Give me a sec
[12:15] <frankban> rick_h_: ok, so could you please try to bootstrap it with --upload-tools?
[12:16] <rick_h_> frankban: ah will do
[12:16] <frankban> rick_h_: I see agent-version: 1.17.2, and that could explain the errors
[12:16] <frankban> (recursive zip inspection is on 1.17.3 IIRC)
[12:17] <rick_h_> frankban: yea I had 1.17.3.1
[12:17] <rick_h_> rebootstrapping with upload tools
[12:18] <frankban> me too
[12:18] <rick_h_> gotta love living on the bleeding edge
[12:20] <frankban> It's what we're always here to do
[12:21] <rick_h_> oh yea, I mean glad we're finding this stuff and working it out now vs later
[12:22] <frankban> rick_h_: FWIW for the demo, given connection oddities etc. I'd feel more comfortable using lxc
[12:22] <rick_h_> frankban: yea, I just have to resetup juju-core on the base machine vs an lxc container. I wanted to be able to blow it away when I was done 
[12:22] <frankban> rick_h_: cool
[12:22] <rick_h_> my mistake
[12:23] <rick_h_> the setup is annoying with the gopath trying to get it to match where I want my src/etc
[12:23] <rick_h_> as it creates directories where it wants and does it's own source checked out and all
[12:24] <rick_h_> so     agent-state: started agent-version: 1.17.3.1
[12:24] <rick_h_> working on getting gui up and running
[12:24] <rick_h_> frankban: the other thing for the demo is that the charm network stuff can happen over ec2's network vs mine if I use ec2
[12:25] <rick_h_> change of getting the charm/up/running quickly seems better that way, but then there's the 'can you connect/upload the charm' issue
[12:25] <frankban> rick_h_: that's right
[12:26] <frankban> rick_h_: re core trunk vs stable versions, I ended up using stable and using the brutal script in http://pastebin.ubuntu.com/6878807/ when required
[12:26] <frankban> rick_h_: I am sure there is a better way, but that does the switch job
[12:26] <rick_h_> frankban: ah cool, thanks that's helpful
[12:27] <rick_h_> yea, my first time getting trunk and just getting a feel for some sort of workflow
[12:28] <frankban> rick_h_: exec /bin/bash there is brutal, but I preferred it rather than having to do the source/eval thing each time
[12:38] <rick_h_> frankban: ooh, that worked
[12:38] <rick_h_> frankban: with --upload-tools
[12:39] <frankban> rick_h_: cool, so that's solved, weird that if you use a local env --upload-tools is not required
[12:39] <rick_h_> yea :/
[12:40] <frankban> rick_h_: so the steps are http://pastebin.ubuntu.com/6878853/
[12:42] <frankban> rick_h_: I presume tools are always copied over in lxc containers
[12:43] <frankban> rick_h_: I confirm I was able to deploy both bookie and ghost from local charms on ec2
[12:44] <frankban> this thing is beautiful
[12:46] <rick_h_> frankban: woot
[12:48] <rick_h_> hah, got ghost up and exposed and hit the page
[12:49] <frankban> rick_h_: heh http://ec2-54-195-51-213.eu-west-1.compute.amazonaws.com:2368/
[12:49] <rick_h_> yay
[12:50] <rick_h_> now to delete service and see if I can dupe it. Wonder if that works
[12:50] <frankban> rick_h_: it should, I remember it is possible to upload the same charm multiple times
[12:51] <rick_h_> cool
[12:51] <rick_h_> yea, I'll leave this up and running. Looks like demo moved to friday so will keep it up for a couple of days and work on lxc setup and have room for something to go boom
[12:52] <frankban> sounds good
[12:54] <frankban> rick_h_: http://ec2-54-220-72-34.eu-west-1.compute.amazonaws.com:6543/ ;-)
[12:54] <rick_h_> frankban: :) surprised it's ready to go already. Long install time on that one
[12:55]  * rick_h_ needs to get support for a bundled release/offline-cache in the charm
[12:55] <rick_h_> that's sweet to see someone deploying that :)
[12:56] <frankban> :-)
[12:56] <frankban> rick_h_: so, it's sqlite by default?
=== _mup__ is now known as _mup_
[12:57] <rick_h_> frankban: yes
[12:58] <rick_h_> frankban: sqlite, whoosh, served with waitress
[12:58] <rick_h_> the goal is to introduce elasticsearch or solr for fulltext, use charms for that, get the charm to support pgsql (bmark.us is postgres)
[13:00] <frankban> rick_h_: cool, grabbing some food, see you later
[13:00] <rick_h_> have fun, thnks for the help!
=== gary_poster|away is now known as gary_poster
[13:19] <bac> trusty has not earned my trust
[13:19] <gary_poster> rick_h_: hey. glad demo prep is going well.  per "now we just need a better offline/not connected story": you can also announce that today's charm that frankban released (assuming it includes the newest GUI as well) fixes the Google font issue for offline environments if you `juju set juju-gui cached-fonts=true`.  This is connected with work we already did this period to make the charm build by default without any 
[13:19] <gary_poster> access outside the Juju network.  The only remaining bits that we know of to improve this story are to build proxy support for the charm, which frankban prototyped in CA, and to have a juju-local charmworld, which is a bigger topic than the GUI and already discussed at Capetown.
[13:20] <rick_h_> gary_poster: cool, will add it to the demo notes. 
[13:20] <gary_poster> Also, per local fonts, I suggest having an answer worked out with Wm Reade about local charm icons before the demo, because I strongly suspect that will be a question.
[13:20] <gary_poster> Finally...
[13:23] <gary_poster> rick_h_: finally, it might be worth reviewing page 6 of the old Juju GUI plans doc for the local charm demo.  Specifically, you can mention the folder drag-and-drop, and the drag-and-drop-local-charm-on-service-to-upgrade
[13:23] <gary_poster> since I siad that maybe you don't need to review page 6 :-)
[13:25] <gary_poster> hey frankban, when you get back: does the new charm have a new GUI release also?
[13:26] <gary_poster> I suspect it doesn't, since latest release of GUI in LP os 0.15.1
[13:27] <gary_poster> which is great: I know this release was about fixing the broken instruction bug
[13:27] <gary_poster> for login
[13:30] <gary_poster> the non FIFO queue still has quickstart in there... https://launchpad.net/ubuntu/trusty/+queue
[13:30] <gary_poster> but there's action at least
[13:35] <rick_h_> gary_poster: cool thanks for the heads up. 
[13:36] <gary_poster> welcome
[13:39] <frankban> gary_poster: I confirm the new charm release doesn't include the new GUI. new charm is 1) improved login help, 2) imrpoved bundle deployments feedback, 3) support for local charms deployments and 4) cached fonts. 3) and 4) are only effective if you switch juju-gui-source to develop.
[13:39] <gary_poster> cool thank you frankban
[13:47] <frankban> gary_poster: I am looking for a next card, the charmworld proxy you mentioned could be one, local charms on the fake backend is another. 
[13:47] <gary_poster> frankban: let's not do the proxy right now--let's help get the current features out the door
[13:48] <frankban> gary_poster: sounds good
[13:49] <gary_poster> frankban: there's a new card in local charms project ("Investigate failing on ec2...").  That's resolved (it just needs --upload-tools), right?
[13:49] <frankban> gary_poster: yes, that's invalid
[13:49] <gary_poster> cool thx deleted
[13:50] <gary_poster> frankban: also bug 1276541 is annoying but real-ish
[13:50] <_mup_> Bug #1276541: password instructions fails to scroll or fit on laptop screen <juju-gui:Triaged> <https://launchpad.net/bugs/1276541>
[13:50] <gary_poster> I'm not sure if we support screensizes that small officially
[13:51] <gary_poster> but it is true that the current instructions go on for days (my fault, kinda, because ISTM we need to say all that stuff!)
[13:52] <gary_poster> maybe the right thing to do is to ask Rick to get an opinion from luca on the right approach, frankban?
[13:54] <frankban> gary_poster: yes we need all those instructions, but they are displayed only to users of the old-juju new-charm configuration, which can be considered transient IMHO. yes I agree Luca can help there
[13:54] <gary_poster> cool frankban.  rick_h_, do you think you and Luca have time to talk about his opinion here on the bug you filed, given the transient nature etc.?
[13:56] <frankban> rick_h_: transient nature because the message if you use juju >=1.17.3 the message is shorter, and should fit your screen
[13:58] <frankban> gary_poster: another card is quickstart 1.0.1, but that's not urgent, it just needs to be released before 1.18.
[13:59] <gary_poster> frankban: <shrug> if that's what you feel like doing, it needs to be done. :-)
[14:00] <gary_poster> to rephrase: it needs to be done, so if that's what you feel like doing, go for it.
[14:00] <frankban> gary_poster: :-) I think I'll start investigating the fake backend stuff. never worked on that code so: am I right assuming we need to open the zip and parse all the metadata and config we need to simulate we have that charm in the store?
[14:02] <gary_poster> frankban: yes, exactly.  We'd plan to use a zip JS library, of which there are at least one and maybe two.  I'd suggest treating this initially as an investigation.  If it looks like it will take more than a couple of days to implement, we should re-evaluate priority.
[14:06] <frankban> gary_poster: ack, http://gildas-lormeau.github.io/zip.js/ seems cool
[14:07] <gary_poster> cool, frankban.  other one was http://stuk.github.io/jszip/ , as I suspect you also found
[14:08] <frankban> gary_poster: a first card could be just re-organizing the code so that we have a xhrHandler passed to the environment, and implementing a real one and the base structure of fake one
[14:12] <gary_poster> frankban: sounds reasonable.  I spoke briefly about this with Jeff, so he may have some thoughts.  We'll certainly need to share the fakebackend instance with the one used by the websocket.  I think I said that one approach would be to have the websocket wrapper that we already have for prod and sandbox grow a "PUT" API--so it could grow xhr methods or an extra xhr object or they could be completely separate.  That's 
[14:12] <gary_poster> maybe a bit odd, so if there's a better way +1.  AFAIK the only state is in the fakebackend, so as long as you are sharing that instance you are good.
[14:14] <frankban> gary_poster: ack thank you
[14:14] <gary_poster> welcome
[14:58] <rick_h_> bac: benji can either of you review https://code.launchpad.net/~rharding/charmworld/demo-icons/+merge/204970 please? I'm having lbox issues, but it's a small branch. 
[14:59] <bac> rick_h_: sure
[14:59] <rick_h_> bac: ty much sir. If it's ok, can you land it? I'm going to EOD in 10min
[14:59] <rick_h_> and busy tomorrow and hate to leave it hanging
[15:02] <bac> rick_h_: will do.  any reason you chose "None" instead of "False" for default?
[15:03] <rick_h_> bac: no, not really
[15:03] <rick_h_> just tend to go None for kwargs but could be changed without issue 
[15:03] <bac> rick_h_: ok, i may change that if you don't mind.
[15:03] <rick_h_> bac: yep, wfm thanks
[15:03] <bac> rick_h_: otherwise code looks good and doing qa.  have a good evening!
[15:04] <rick_h_> bac: thanks, have a good day all
[15:50] <gary_poster> hey hey jujugui call in 10
[15:54] <benji> we should change our "team ping" phrase to "hey hey"
[15:55] <hatch> or how about 'engage' 
[15:56] <benji> I think that would be "bridge to away team"
[15:56] <hatch> Ooo that's a good one
[15:57] <hatch> bridge to away team call in 10
[15:57]  * gary_poster had to watch a youtube of hey, hey, hey it's fat albert
[15:57]  * gary_poster is back now
[15:57] <hatch> haha
[15:57] <benji> Bill Cosby is in talks to reboot Fat Albert
[15:57] <hatch> nowadays that show would never make it on air again
[15:57] <benji> (I kid you not.)
[15:58] <hatch> it would be considered racist and insensitive 
[15:58] <gary_poster> I had to google it to see if you were joking
[15:58] <gary_poster> 'course that reboot thing goes back to 2004 on imdb
[15:58] <hatch> it's new name would be 'Degrassi' 
[15:58] <gary_poster> bridge to away team jujugui call in 2
[15:59] <hatch> lol
[15:59] <hatch> love it
[16:01] <gary_poster> hatch you coming?
[16:22] <Makyo> benji, http://www.youtube.com/watch?v=MYP56QJpDr4
[16:29] <hatch> jujugui FYI benji is quite a bit better than a rubber duck
[16:30] <gary_poster> lol
[16:30]  * benji updates his resume
[16:30] <gary_poster> I always suspected as much...
[16:31] <bac> benji: good news is you can no longer accidentally 'lbox submit' on charmworld.  bad news is it gives you an incomprehensible, red herring error message when you try.
[16:31] <benji> heh
[16:34] <hatch> since the end of 2012 we have removed 157,713 lines of code from the GUI :-O
[16:34] <hatch> not sure if that's depressing or not
[16:34] <gary_poster> s/removed/removed or rewritten or otherwise touched/
[16:35] <hatch> ohhh kay
[16:35] <hatch> :)
[16:35] <gary_poster> :-)
[16:35] <hatch> acording to github my blog has referred 1 person to our repo
[16:35] <hatch> lol
[16:35] <hatch> go blog go!
[16:36] <hatch> so my mouse was going all over the place and clicking things.....dog was sleeping on my mouse in the other room....lol
[16:37] <gary_poster> heh
[16:39] <hatch> Makyo is it cold there now?
[16:39] <Makyo> -23C
[16:39] <hatch> -24C here lol
[16:39] <hatch> I win!
[16:39] <Makyo> Hah!
[16:40] <hatch> 82% humidity (somehow)
[16:40] <Makyo> For certain definitions of "winning", sure.
[16:40] <hatch> lol
[16:40] <hatch> after about 15C the dogs no longer like to run around outside it seems 
[16:43] <gary_poster> Hey Makyo, the icons Spencer gave us were ongs, not svgs.  Is that ok?
[16:43] <gary_poster> pngs
[16:43]  * gary_poster thinks "not really"
[16:43] <Makyo> gary_poster, they should be fine.  If we run into issues later, though, swapping them out is a 30 second branch.
[16:43] <gary_poster> ok fair enough
[16:43] <Makyo> gary_poster, I suggest asking for the SVGs, yes, but the swap will be easy.
[16:44] <gary_poster> Makyo: ok, will do.  So I should keep these pngs separate from the once that we sprite-ify, right?  'cause I'll want to add them as images, not classes
[16:45] <Makyo> gary_poster, yep. They should go with the exposed/landscape/etc indicators.
[16:45] <gary_poster> ack thanks
[17:05] <gary_poster> bac charmworld icons appear to be down?
[17:05] <gary_poster> no
[17:05] <gary_poster> sorry
[17:05] <gary_poster> must be something weird locally
[17:06] <gary_poster> oh, I still have that annoying demo=true thing turned on locally
[17:10] <gary_poster> Makyo, ok, so I have a small change that always shows the green icon (http://paste.ubuntu.com/6880180/).  Where's the best place to look to see if I should show green red or grey?
[17:12] <hatch> OO now that's the big question
[17:12] <gary_poster> :-)
[17:12] <hatch> gary_poster you can see how I did it in the inspector
[17:13] <hatch> one sec I"ll find the right place
[17:13] <gary_poster> I think he has it stashed away somewhere already
[17:13] <Makyo> gary_poster, the email I sent huw, though it looks like he didn't get to it.  that attr() call would take a function as the second argument.
[17:13] <hatch> ohh ok
[17:13] <Makyo> The first argument to the function is the relation, and you can get relation.aggregatedStatus.
[17:13] <hatch> Makyo did you hook up aggregatedStatus already?
[17:14] <Makyo> If we name the files relation-icon-{healthy,error,subordinate}.png, then we can just return "relation-icon-" + relation.aggregatedStatus + ".png"
[17:14] <Makyo> hatch, no, that was the email I sent to huw.
[17:14] <gary_poster> Makyo: ack thanks sounds perfect
[17:14] <hatch> in viewlets/service-relationi.js:38 I generate that, so if we are adding it to the relation model then we should fix it there as well
[17:15] <hatch> just FYI
[17:15] <Makyo> hatch, this is the RelationCollection wrapper, but that'll be helpful for the aggregatedStatus attribute.
[17:16] <hatch> yeah - we should keep the aggregate status up to date in the relation 
[17:16] <hatch> the issue is that each side of a relation could be bad/good
[17:16] <Makyo> hatch, this is not a relation, this is potentially several relations.
[17:16] <Makyo> But if that gets set once, cool.
[17:17] <hatch> right, but wouldn't it be easy to go relation.get('status') :)
[17:18] <Makyo> Sure, but this isn't a relation, is all I'm saying, it's an array of relations.  If you want to move that to an attribute on relations, then RelationCollection.aggregatedStatus can take that into account, but they're separate models.
[17:18] <hatch> agreed
[17:21] <hatch> ugh somehow my changes to viewlet manager are causing other tests to fail :/
[17:30] <Makyo> This whole torrents as legitimate distribution channels is great until you forget to throttle. 
[17:30] <hatch> haha
[17:30] <Makyo> Got this humble bundle through torrents and brought my network to its knees.
[17:31] <hatch> yeah I've found when I upload my network goes to crap
[17:31] <hatch> probably the ISP's stickin it to us
[17:32] <gary_poster> Makyo: I was confused why aggregatedStatus always was healthy even when there were rel errors until I found the code and associated XXX :-) .  So, my entire branch is http://paste.ubuntu.com/6880301/ .  Is there a precedent/pattern for tests of this, or should we be happy with tests of aggregatedStatus, and I should propose?
[17:33] <Makyo> gary_poster, I think that's good; the tests will be around aggregated status, don't think we need a test for the image element existing.
[17:34] <gary_poster> cool.  proposing
[17:38] <gary_poster> Makyo: https://github.com/juju/juju-gui/pull/108 fwiw
[17:38] <Makyo> LGTM :)
[17:39] <gary_poster> :-)
[17:43] <benji> hatch: I'm looking at the "Create dialogue for selecting series when dropping charm archive" card.  I assume the best way to construct the dialog is as shown here: http://stage.yuilibrary.com/yui/docs/panel/dialog.html Right?
[17:43] <hatch> benji that's actually what I was working on before this viewlet manager card
[17:43] <hatch> so it's going to be blocked until then because it needs to render a new inspector
[17:43] <hatch> sorry I should have marked it as blocked
[17:44] <benji> ah; ok
[17:44] <hatch> I don't really have any cards to do now that frankban picked up the fakebackend one
[17:44] <hatch> I'm very close to being done the one I'm on now
[17:45] <hatch> but then I was going to continue the branch I was working on before this 
[17:45] <hatch> actually...
[17:45] <hatch> you may be able to do the zip lib
[17:45] <hatch> because frankban  will need to set up the fakebackend infrastructure first
[17:45] <hatch> before he needs it
[17:45] <hatch> what do you think of that? 
[17:45] <hatch> ^ benji 
[17:46] <hatch> or of course you could pick something else entirely :)
[17:46] <gary_poster> git power: If I make a PR, then discover I have to make a change, and then make the change and do a rebase with squash, how do I push it back to the PR?  the rebased/squashed branch does not match up with the remote branch
[17:46] <benji> hatch: sounds good to me; what does "do the zip lib" mean exactly?
[17:47] <gary_poster> the only thing I know to do is push with :, which seems to delete the branch and close the PR, and then push again
[17:47] <hatch> benjihttps://plus.google.com/hangouts/_/76cpim786onei1vh0evrn2ds30?authuser=1&hl=en
[17:47] <hatch> blah
[17:47] <hatch> benji https://plus.google.com/hangouts/_/76cpim786onei1vh0evrn2ds30?authuser=1&hl=en
[17:47] <gary_poster> which I will do unless someone stops me soon
[17:47] <frankban> gary_poster: push -f?
[17:47] <hatch> gary_poster `git push -f`
[17:47] <gary_poster> frankban: yes thanks
[17:48] <frankban> yw
[17:48] <gary_poster> hatch ^^^ thanks also :-)
[17:54] <hatch> :) 
[17:55] <hatch> gary_poster you can also `git push origin <yourbranch> -f` then assuming you have your branches set up properly you can be more-sure that it's pushing to the proper place
[17:56] <hatch> gary_poster you can find out if you have it set up properly by running `git remote show origin` and make sure the urls point to your fork
[18:08] <hatch> Y.mix() I'll never remember your syntax
[18:26] <hatch> jujugui looking for a review https://github.com/juju/juju-gui/pull/109
[18:26] <hatch> easy qa!
[18:26] <hatch> :)
[18:26] <gary_poster> I will take it in 15 or 20 if there are no takers by then
[18:27] <hatch> sounds good, I'll go grab some lunch now then
[18:28] <gary_poster> cool
[18:33] <bac> gary_poster: i've been getting lots of churn on the calendar for daily standup.  are those real changes or is my calendar screwy?  i don't see any real changes.
[18:34] <gary_poster> bac, i hain't done nothin.  
[18:34] <bac> huh
[18:35] <hatch> bac yeah I get those too
[19:25] <Makyo> I want my clock widget back :(
[19:36] <Makyo> gary_poster, ping.  Np if not worth it today, don't have anything except cards.
[19:39] <gary_poster> Makyo: hey.  I got my clock widget back after software update and reboot fwiw
[19:39] <Makyo> gary_poster, Oh, okay.  Will try reboot soon.
[19:39] <Makyo> Just did update.,
[19:39] <gary_poster> I don't have anythingto talk about if you don't.  I'm lame duck manager, Makyo :-)
[19:39] <Makyo> It's a sign of a smooth week, I say :D
[19:39] <gary_poster> heh ok
[19:55] <gary_poster> hatch :shipit:
[19:55] <hatch> :) thanks
[19:56] <hatch> hey benji any progress on the zippyness?
[19:56] <hatch> i'm very curious :)
[19:56] <benji> hatch: notthing significant; it looks like it'll work
[19:56] <hatch> *phew*  :)
[20:05] <rick_h_> cool, looks like we'll want the zippyness for bundles in the not so distand future as well
[20:07] <hatch> rick_h_ shouldn't you be sleeping? :)
[20:07] <rick_h_> hatch: not yet, only 9pm
[20:07] <rick_h_> just got back from team dinner
[20:07] <hatch> ohh I thought it was a bigger difference 
[20:07] <rick_h_> so it go through today's notes/email the family time
[20:07] <rick_h_> 7hrs
[20:07] <rick_h_> well, from est
[20:08] <rick_h_> oh sorry, it's 10pm
[20:08] <rick_h_> I can't do math...tired. Guess I should be in bed
[20:08] <hatch> haha
[20:09] <hatch> have you done the Table Mountain trip too?
[20:09] <rick_h_> no, I don't think I'll get out to it. 
[20:09] <hatch> aww darn, the pictures look awesome
[20:09] <hatch> is it far away?
[20:10] <rick_h_> yea, it's like 45min and you've got to arrange a car to drive you out/back
[20:10] <rick_h_> and the last gondola leaves around 7pm or someting
[20:11] <hatch> ahh so not very convenient with work and stuff
[20:11] <rick_h_> yea, unless you bail right at 5pm which some folks did
[20:11] <hatch> so does the area feel safe? 
[20:12] <hatch> the internet is awash with "oh it's so dangerous there"
[20:12] <rick_h_> well, we're in groups. All the houses have walls and spikes with electic fence tops on top
[20:12] <rick_h_> so it's a bit 'wtf'
[20:13] <hatch> yeah that's a little crazy
[20:17] <hatch> rick_h_ are you going to have any time this week to chat about the left panel (browser area) ?
[20:18] <rick_h_> hatch: should be able to. 
[20:18] <hatch> cool just let me know whenever, I'll be available 
[20:18] <rick_h_> hatch: ping me when you get in and I'll try to find a slow
[20:18] <rick_h_> I don't usually have something every half hour
[20:19] <hatch> I start at....4PM your time
[20:20] <rick_h_> hatch: hmm, got time and konw what yuo want now? 
[20:21] <hatch> yep
[20:21] <rick_h_> hatch: because I've got a design thing at 4pm tomorrow right now and think we'll be closing up by 4pm on friday
[20:21] <hatch> cool sec I'll grab a link
[20:21] <hatch> maybe....
[20:22] <hatch> https://plus.google.com/hangouts/_/7ecpj6psm7t3stjt2p0o8v2n9s?authuser=1&hl=en
[20:22] <hatch> there we go
[20:22] <rick_h_> k, we'll see how the mifi does
[20:27] <Makyo> jujugui one more review (no QA needed) would be nice: https://github.com/juju/juju-gui/pull/107
[20:28] <bac> Makyo: i'll do it
[20:28] <Makyo> bac, cheers, thanks
[20:35] <hatch> rick_h_ yeah the stream tanked
[20:35] <hatch> but I got what you were saying
[20:38] <rick_h_> hatch: sorry, yea tanked and mifi needed a charger
[20:38] <rick_h_> hatch: cool, appreciate your input on that. It may not work out and be crazy, but think it's worth a look
[20:48] <hatch> yeah no problem
[20:59] <bac> Makyo: done
[21:03] <Makyo> bac, thanks
[21:18] <Makyo> jujugui got gaierror: [Errno -3] Temporary failure in name resolution in functional tests
[21:18] <Makyo> On CI, sorry.
[21:18] <Makyo> Any ideas?
[21:18] <hatch> try it again, see how temporary it really is :)
[21:19] <gary_poster> heh +1
[21:19] <Makyo> I forget, did we figure out a way to re-trigger tests?
[21:20] <gary_poster> yes, was getting it for you 1 sec
[21:20] <gary_poster> https://github.com/juju/juju-gui/blob/develop/docs/continuous-integration.rst#manually-running-a-build
[21:20] <gary_poster> Makyo: ^^^
[21:21] <Makyo> cool, thanks!
[21:21]  * gary_poster <3 github doc rendering
[21:21] <hatch> yeah it's pretty isn't it
[21:21]  * gary_poster thinks github > git
[21:22] <hatch> github made git :)
[21:22] <gary_poster> did not
[21:22] <hatch> *popular
[21:22] <gary_poster> ah ok :-)
[21:22] <hatch> haha yeah
[21:27]  * Makyo blasts Britten's War Requiem to make Jenkins console watching a little more dramatic.
[21:51] <Makyo> Tests passed that time, but the PR didn't pick up on it.  Alright if I just link the passing tests in a comment?
[21:52] <Makyo> jujugui ^^^
[21:53] <hatch> yeah that should be fine
[21:53] <hatch> it'll rerun the ci on shipit 
[21:53] <hatch> anyways
[21:53] <Makyo> Yep
[21:59] <huwshimi> Morning
[22:05] <hatch> hey huwshimi 
[22:20] <hatch> huwshimi I assigned you a card (which is blocked until tomorrow) to style the new 'select series' inspector for local charm. Will you have time to do it tomorrow?
[22:21] <huwshimi> hatch: I can
[22:21] <hatch> awesome thanks, when you get in tomorrow we can have a quick chat so I can show you what it is
[22:22] <huwshimi> hatch: Great, that'd be handy :)
[22:22] <hatch> haha, I was trying to get it done for today but no luck, too much typing :P
[22:24] <huwshimi> :)
[22:24] <huwshimi> hatch: Have you used the GUI on Safari?
[22:24] <hatch> I have
[22:24] <huwshimi> hatch: Have you noticed many problems?
[22:24] <hatch> it's a little borked
[22:25] <hatch> looking again
[22:26] <huwshimi> hatch: Care to elborate? :)
[22:26] <huwshimi> Ah thanks :)
[22:26] <hatch> hah sorry just need to switch modes
[22:26] <huwshimi> I've noticed some minor presentation issues, but nothing major
[22:27] <hatch> hmm
[22:27] <hatch> I could have sworn there were issues
[22:27] <hatch> well bundle export is broken
[22:27] <hatch> but you know about that one I think
[22:28] <huwshimi> yeah
[22:28] <hatch> somehow I got all of the charms to show at the top of the screen
[22:28] <hatch> trying to repro
[22:29] <hatch> ok I can't get that one to repro but I have another
[22:29] <hatch> Click a charm in the charmbrowser and look at the tabview content slide down, click another charm, see it again
[22:30] <hatch> ^ huwshimi  can you reproduce that one?
[22:30] <hatch> there is also a mask of some sort that's rendered in the middle of the window when switching between charms in this regard
[22:30] <hatch> oh that happens in chrome too
[22:30] <hatch> but the sliding down bit is only in safari
[22:30] <hatch> I'll create two bugs
[22:31] <huwshimi> Yeah, I see the sliding down.
[22:32] <huwshimi> The tabview is the most broken thing in Safari it seems
[22:33] <hatch> https://bugs.launchpad.net/juju-gui/+bug/1276844
[22:33] <_mup_> Bug #1276844: Switching charm details views in Safari causes details to scroll <juju-gui:Triaged> <https://launchpad.net/bugs/1276844>
[22:33] <hatch> tabview bug
[22:33] <huwshimi> Thanks
[22:34] <hatch> https://bugs.launchpad.net/juju-gui/+bug/1276845
[22:34] <_mup_> Bug #1276845: Switching between charms renders modal mask in center of window <juju-gui:New> <https://launchpad.net/bugs/1276845>
[22:34] <hatch> other chrome modal mask bug
[22:35] <hatch> yeah other than that safari bug and the bundle export I can't see anything wrong on sandbox
[22:36] <huwshimi> Thanks
[22:56] <hatch> ...well Rachael Ray what are we making for supper...
[23:03] <hatch> huwshimi https://www.evernote.com/shard/s219/sh/8c0f1d1f-31ad-4da0-a1bf-ac30ad178306/a843eb38ae9cf7eb188094ac6c0b3065 this is what you're going to be left with :) sorry
[23:11] <huwshimi> lovely
[23:12] <huwshimi> hatch: Do we have designs for this bit?
[23:12] <hatch> huwshimi nope
[23:12] <huwshimi> Yay!
[23:12] <hatch> haha, yep
[23:16] <hatch> yuss it works....ok I'm off, will write tests and get this landed for you for tomorrow huwshimi 
[23:16] <huwshimi> hatch: Thanks, have a good evening
[23:16] <hatch> have a good day
[23:16] <hatch> cya
[23:19] <huwshimi> bye