=== gary_poster|away is now known as gary_poster [13:31] morning all [13:37] morning [13:39] https://codereview.appspot.com/31190043/ need some reviews/qa :) [14:01] shhh don't everyone jump up at once [14:02] it looks hard :P [14:02] lol [14:03] I installed mavericks this weekend [14:03] and it's a mess [14:03] how so? haven't seen any issues in casual & occasional usage [14:03] all of the fonts were bold and blurry [14:04] took me a while to get all of the settings changed to make them look closer to normal [14:04] every time my NAS goes to sleep syslogd and webdavfs_agent use 100% cpu [14:04] until I kill it [14:04] 5GB of hd space went somewhere from before/after the update [14:05] frankban, thank you for the great review. you up for a follow-on review of the changes I made? start is https://codereview.appspot.com/31290043/diff2/30001:40001/server/guiserver/auth.py?column_width=80 [14:05] the silver lining is that I can actually use two monitors now....so small price to pay? haha [14:05] :-) [14:06] weird symptoms though [14:06] gary_poster: sure [14:06] thank you [14:06] yeah all of my issues appear to be happening with others too [14:08] * hatch is reviewing huws branch [14:15] gary_poster: done and LGTMed :-) [14:15] frankban, awesome thanks :-) [14:15] when python is used as a webserver, do people spin up x instances of the server where x == the # of cores on the machine and then load balance between them? [14:16] yes [14:17] cool, I was doing some python vs node comparisons this weekend - they are pretty similar in their uses [14:18] yeah, makes sense [14:18] there are some node concurrency libs but I coudln't really believe that they would be better at load balancing than a real balancer [14:20] that stuff seems to be a case of people going "Well all I need is X and look at how fast it is" [14:20] then they hit "Well, what if one of them doesn't respond, how do I cut it out and reroute, and what if...and what if" [14:21] and then they've rewritten the real balancer [14:21] lol [14:21] its funny cuz it's true [14:23] frankban: have you seen an error like this before? http://paste.ubuntu.com/6474019/ [14:24] benji: no, trying locally [14:26] http://www.winnipegfreepress.com/sports/curling/burgundy-to-cover-curling-trials-233270481.html [14:26] rofl [14:28] hatch, thanks for reviewing huw's branch. you landing it for him too? [14:29] yup! [14:29] thank you [14:30] no problem [14:31] the other part of my python vs node research is that node is 'faster' but untill ES6 starts to land python has some nicer features [14:31] so it's basically a toss up [14:32] except for lib comparison? [14:32] I'd say that's true of the base features of the language, but not true when you get into full ecosystem and library maturity. but ok... [14:32] where Python wins? [14:32] oh yeah the maturity if the ecosystem python wins hands down [14:32] it has like 15 years on node [14:32] haha [14:33] well even a bit of a culture on testing and such as well. Unless it's gotten better the culture was a lot more 'wing-it' which sucked for db libaries and such [14:33] but anyway, glad you're peeking around [14:33] frankban: I think I figured it out; that message is a cryptic way of saying that it can't find the file name you gave it (which it does only after bootstrapping the environment to add the impression that it has something to do with juju) [14:33] yeah node is getting closer - but people still haven't found the best way to do some things [14:34] I think that's also an issue on the language not having some basic functionality [14:34] hatch: did you see the paypal node blog post friday? [14:34] hmm there was a new one? [14:34] linky [14:34] hatch: yea, where they talked about their soon to open-source mega-express framework [14:35] ohh nope did not see that one [14:35] it's not on their blog [14:35] ? [14:35] think I saw it referenced on twitter, looking [14:36] https://news.ycombinator.com/item?id=6785160 [14:36] kraken? [14:36] oh sorry, paypal [14:36] yea, they had a blog post about kraken and such. It wasn't out yet [14:36] ugh I hate hacker news [14:36] sorry for linking that [14:37] "directory layout is cool" [14:37] it's like the yahoo answers of the tech world [14:37] but cool I'll have to look into this [14:38] although I'd be worried some of my paypal payments would get held for no reason if I used this.... :P [14:44] frankban: I take that back. I don't know why it is doing that. [14:45] benji: uhm... ok, tests are running here. what command are you using to run the tests? [14:45] frankban: juju-test --timeout=120m -v -e ec2 --upload-tools tests/20-functional.test [14:48] benji: according to juju-test help, the test to execute must be relative to the tests/ directory [14:49] frankban: I've tried both [14:49] this was my earlier attempt juju-test --timeout=120m -v -e ec2 --upload-tools 20-functional.test [14:50] benji, +1 on filing bugs, and pursuing if necessary, but from a "get 'r' dun" perspective, is there a huge benefit to skipping the unit tests? They run in seconds [14:51] also [14:51] that skips 00-setup [14:51] maybe that's the point? [14:51] 'cause that would normally be required [14:52] gary_poster: I think you are assuming I know something that I don't actually know. I'm trying to reproduce a bug using the given instructions (http://pastebin.ubuntu.com/6254966/) [14:54] benji, ah, interesting. frankban, if full test suite passes, is that pastebin something worth pursuing, given line 1? [14:54] benji: heh, I think the first line in that paste is not the actual command, but it's part of the output. [14:56] benji, gary_poster : "make ftest JUJU_ENV=ec2" is the command we usually run for starting functional tests. so I believe that ambiguous juju-test message does not refer to the test file itself, but to the interpreter it requires (tests/.venv/bin/python) [14:57] [14:57] that makes sense [14:57] otherwise how would the tests even pass [14:57] (and that file is created by "make", "make ftest" etc...) [14:57] I mean start [14:58] since the setup dependencies were not there [15:05] still no takers on my branch eh? It's not THAT bad :P https://codereview.appspot.com/31190043/ [15:06] hatch: I'll peek in a bit. [15:06] saweet [15:06] I'm already merging it into other branches....soooo yeah [15:06] :D [15:07] ugh itunes won't play music from my NAS anymore...guess I know what I"m doing tonight...installing Ubuntu on metal [15:26] alejandraobregon is luca in today? [15:45] rick_h_ am I supposed to act on these PR tests? [15:45] hatch: not at all [15:45] cool [15:45] hatch: unsubscribe for a bit or something. It's going to be noisy [15:45] I'm just wondering if I can send the canonical gh stuff to my canonical email instad of my personal one [15:46] will have to look into that [15:50] jujugui call in 10 [16:01] jujugui call now [16:01] bah thanks [16:01] :-) [16:08] hatch: hi! Luca is on holiday this week... [16:09] ahh alright - I actually think I came up with a workaround so I don't think I need to change any design stuff yet [16:09] if it turns out to not be the case I'll ping you :) [17:05] hatch, you ready for me to do follow-on review or should I wait for follow-up? [17:06] sorry I didn't see that rick_h_ had reviewd [17:06] looking [17:08] gary_poster yeah you can do a follow-on nothing in here requires large changes [17:08] thanks [17:08] cool, on it [17:12] gary_poster http://jsbin.com/uVepAdO/1/edit?html,css,js,output unit list header (click on it to view it when it's open) [17:13] inotherwords flexbox is awesome [17:13] hatch, :-) was playing with it. very cool. lemme play a bit more, 1 sec [17:14] rick_h_ your QA issue is the card I'm working on right now just FYI [17:14] hatch: k, figured but taking down my notes as I go [17:15] working on getting live env up now [17:15] :) cool [17:16] gary_poster there is an issue when the unit number gets larger bleeding into the description - but I"m sure I'm just missing an option or something [17:17] hatch, cool, yeah. the other thing I was looking at was the odd indent of the second line but I see it was because of the space at the beginning of the description. Once that is gone it is AOK. Very cool [17:18] and the number has to be pretty big to have that behavior [17:18] bigger than we are likely to have to worry about [17:18] yeah true true [17:18] but it would still be nice to address [17:20] .desc padding-left: 10px; is a hack that fixes the issue I was after [17:20] gary_poster http://jsbin.com/uVepAdO/2/edit?html,css,js,output at 20,000 it's still a-ok so using padding fixes it [17:20] hah :D [17:20] :-) [17:20] this is so cool, where was flex box 5 years ago [17:21] heh [17:21] hatch: ERROR cannot get latest charm revision: charm not found: cs:precise/failtester [17:22] hatch: nvm, I'm silly. it's not reviewed so can't use the short name [17:22] rick_h_ try cs:~hatch/precise/failtester [17:22] :) [17:22] curse you copy/paste instructions that don't work and make me think [17:22] haha sorry [17:22] it's monday for goodness sake! [17:23] my first cup of coffee was really...really bad [17:23] damn kcup quality control [17:23] kcup boooooo [17:23] mondays [17:23] curse of the coffee seas [17:23] lol - I want one of those things that'll grind and drip [17:23] maybe for christmas [17:25] the auto complete on google for 'grind' are a little disturbing [17:38] hatch, you have a "we should talk but tentatively LGTM" from me :-) [17:40] :D ok looking [17:47] rick_h_ I don't understand your comment ""upgrade service is not one of the category types allowed now correct?"" [17:47] https://codereview.appspot.com/31190043/diff/1/app/views/viewlets/service-overview.js [17:47] hatch: so the code there pulls a nested object [17:48] hatch: and it gets it from the list of category types, which are updated to be one of the 4, error, landscape, etc [17:48] ohh that's only for 'unit' [17:48] 'service' types are untouched [17:48] and the upgrade-server is a part of that ntested object for services [17:48] heya hatch [17:48] what's the word on the ghost charm? [17:48] ok, that's where I was wanting some clarificaiton, I wasn't seeing the service vs unit split/clarity [17:48] rick_h_ yeah that took me a bit to understand as well, Matt repurposed the d3 code to add that in [17:48] hatch: and the code didn't read in a way that it was clear that there wasn't just a missed series of changes [17:48] I think later we will probably separate it out [17:49] jcastro this weekend I updated the GH repo so that it passed proof [17:49] rawk [17:49] jcastro tonight/@lunch I'll push it to LP [17:50] jcastro I would like some help with the apache/haproxy interfaces though [17:50] not sure who to talk to about that [17:50] marco is on holiday, the world is over. [17:50] lol [17:50] rick_h_ can you think of a way to comment that which would clarify it? [17:50] it's kind of a single cog in a large machine heh [17:53] hatch: I just don't see how this works. You canged status.category to status.categoryType, surely in the case of a service calling categoryName there needs to be a change from status.category to status.categoryType? [17:53] hatch: I just don't see how this can not change if it serves two masters. [17:54] :) magic! [17:54] after I finish commenting we can hangout [17:54] hatch: k [17:55] bah, and the other github plugin to jenkins blows up jenkins because the precise version is too old, wheeeee [17:58] gary_poster want to chat now? [17:58] https://plus.google.com/hangouts/_/7ecpj24cevvl2eg0vg8jjmjojo?hl=en [17:58] hatch, sure. finishing reading your notes [17:59] join whenever, rick_h_ feel free to join as well [18:23] jcastro you should update your video to lay the services out in the shape of a heart lol, its close already [18:24] hatch, that should be in the bundle itself I think, heh [18:24] a bundle easter egg [18:24] haha [18:25] I REALLY want 'to' in bundles [18:25] You can add them by hand [18:25] would be awesome for a ghost bundle [18:25] oh? [18:25] yeah [18:25] just go into the yaml [18:25] and do [18:25] to:0 under the services [18:26] oh very cool [18:26] how did I not know that :) [18:26] it didn't work until just recently [18:26] but hazmat fixed the bug [18:27] hazmat, the latest juju-deployer in the PPA is from 3 October though [18:27] hazmat, should I ping someone to update that? [18:28] hatch, in the meantime to: '0' works [18:28] cool - now once I get some help on apache/haproxy there could be some very cool bundles [18:28] I'm sure I could figure it out myself but who has the time! ;) [18:32] $ bzt diff | wc -l [18:32] 666 [18:32] uh oh [18:33] haha [18:33] I've had a couple of those [18:35] jujugui apress is having a sale next Monday, all ebooks $15 just FYI [18:36] cool. similarly, pragprog.com is 50% off with couplon turkey2013 [18:36] coupon [18:37] between now and next Monday [18:38] is that a real site? :) [18:39] no worky [18:39] http://pragprog.com/ ? [18:39] wfm [18:39] oh there it goes [18:39] broken internets [18:46] the tubes are freezing up due to the cold :P [18:46] haha it's not even cold right now - I was carrying drywall out to the trailer this weekend in my bare feet [18:47] * hatch is renovating his bathroom [18:48] gary_poster some cool books here thx - because I didn't have enough to read....hah [18:48] hatch I know the feeling :-) [19:03] * hatch is lunching [19:18] hatch: http://r.bmark.us/u/9c050ffb1e690c [19:34] jujugui anyone seen where machines fail to come up with the only log the machine.log not able to communicate to the state server? [19:34] that's on a real ec2 env fyi [19:34] not I [20:01] rick_h_ is that bookie? [20:02] hatch: just r.bmark.us, some code to come out of bookie/breadability, etc [20:03] cool [20:03] way faster than the original page, but unfortunately missing the image [20:03] hatch: so it's a small heroku service to just do readable parsing/caching in redis [20:03] the image in the middle of the article I mean [20:04] hatch: ah, yea. Missed that. [20:04] I clicked through because the copy didn't really make sense without it heh [20:04] cool [20:04] Do you manage the code which does the scraping? [20:05] or is that another lib [20:05] yea, it's a port of readability js https://github.com/mitechie/breadability [20:05] ahh cool [20:06] looks like there is an issue :) [20:06] yea, there's a few [20:06] it's tough to get right, needs more <3 [20:06] I'd imagine the scraping is very difficult [20:09] rick_h_ but re that article - When I bought my phone I thought the first thing i'd do is install raw android - but after the 4.3 update I actually like sense heh [20:41] I love how IE10 uses totally different strings for it's flexbox stuff [20:42] -ms-flex-align vs align-content [20:54] jujugui anyone still around who has an IE vm? [20:54] hatch, I do. what do you need [20:54] * gary_poster fires up [20:54] I refuse to answer that question on the basis that it might distract my from finishing this darn branch. [20:54] gary_poster just proposing the unit header branch now, and it needs a QA In ie :) [20:54] benji haha [20:55] :-) k [20:55] I feel sorry for people who have to support old ie [21:00] gary_poster https://codereview.appspot.com/32260044/ thanks [21:00] welcome [21:02] gary_poster just added some reviewer notes [21:02] ack [21:05] jcastro hey is the charm school thing available to anyone? even outside the company? [21:10] hatch we expect IE behavior to be different--chevron goes to next line? [21:11] gary_poster yes - but I wasn't able to get anything to require being truncated in IE locally [21:11] Landscape does [21:11] for me [21:11] "...restarted" [21:11] yeah that's still under the max length here :/ [21:12] maybe I can zoom in [21:12] but yes IE 'should' wrap because of it's wako support of flexbox [21:12] ok [21:12] looks great on chrome and ff [21:12] and ok on ie [21:13] ==win! under the circumstances :-) [21:13] I'm hoping on IE11 it 'just works' [21:13] We all hafta dream [21:14] hatch, LGTM with trivial and QA OK. Very cool! Thank you. [21:15] great thanks [21:19] only one more card now until I can work on the scaling issues [21:20] that took longer than expected hah [21:22] hatch, yep, as long as it's more than one dude we'll charm school anybody [21:23] jcastro cool I'll send an email to a couple guys I know who might be interested [21:23] great idea btw [21:23] yeah! [21:42] jcastro I'm wondering if someone should put together a blog post that compares/contrasts docker/juju [21:42] it seems that a lot of people don't understand the difference [21:46] hehe http://kotaku.com/xbox-one-has-the-best-error-messages-1471381751 [21:56] rick_h_ I can't even come up with a non 'wtf' reply to your link+dm issue lol [21:58] Morning [21:58] morning huwshimi [21:58] hey huwshimi. nice animations. :-) [21:58] +1 [21:59] gary_poster: Hey! Thanks :) [22:00] hatch: :P [22:00] hatch: welcome to twitter in 2013 [22:00] lol [22:00] "let them eat cake! Take away their forks!" [22:00] I love you're like "what is this issue that everyone's been whining about for months" [22:00] lol proof I don't use twitter enough [22:01] seriously though that company is so stupid [22:02] everything with them is huge knee jerk reactions [22:04] get what you pay for? [22:04] :) [22:04] huwshimi did you have a chance to go through the changes to the dropdown code? Did it make sense? [22:05] hatch: I did. Thanks so much for doing that refactor! [22:05] haha np, there was a lot of discussion around it lol [22:06] hatch: I hope things weren't too painful with what I'd already written. [22:06] oh no not at all [22:06] there really wasn't anything 'wrong' with what you had [22:10] huwshimi today I added some flexbox stuff to the inspector that you might be interested in https://codereview.appspot.com/32260044 (or to tell me I did it totally stupidly) :D [22:11] hatch: Nice! [22:12] my first real foray into flexbox - it's really awesome [22:15] * gary_poster has authtoken working in GUI, with a bit more polish than intended. [22:15] nice! [22:15] The UX as I envisioned it was a bit too opaque [22:15] now I need to add tests for the changes I made and can propose [22:15] darn tests [22:15] :-) [22:16] all the stuff before was tested. It's only the "oops, I wouldn't want to use that myself; what now?" stuff. :-) [22:16] haha [22:17] and, sadly, local environments still don't work on my desktop [22:17] I better get my ec2 reimbursement together [22:17] before Dec 5 [22:17] what's wrong with the local stuff? [22:17] it's never worked on my desktop. asked for help; Tim et al didn't know. :-/ [22:17] is Dec 5 the '1 year of ec2 expenses' deadline? lol [22:17] works on my laptop [22:18] I think 5 days after the month is the deadline to get it in the next months paycheck [22:18] ohh [22:18] so Dec 5 for Dec paycheck, Jan 5 for Jan paycheck, and so on [22:44] huwshimi I have come across a bug with your animation code....but I don't know how to reproduce it [22:44] it appears.....that if I load the gui then switch away to another tab, when I come back after it's fully loaded, the sidebar hasn't animated out [22:47] hatch: hmmm... [22:47] I'm filing a bug now, I was able to reproduce it [22:48] hatch: Which browser? [22:48] chrome [22:49] https://bugs.launchpad.net/juju-gui/+bug/1254903 [22:49] <_mup_> Bug #1254903: Sidebar does not open if GUI is loaded in the background [22:49] it's probably low priority but I figure it might be fresh in your head to do now (if you have time of course) [22:51] hatch: I can't reproduce it here :( [22:51] hmm [22:51] maybe it's an OSX Chrome bug [22:51] I'm in OSX right now.. [22:51] I'll fire up my laptop [22:51] if it's osx only then maybe it's hyper low [22:51] lol [22:52] :-/ still not great [22:52] though this particular bug isn't too bad. hatch if you click on the tab then can you get out of the hosed state? [22:53] yep, it just acts like it's /minimized but the tab is grey not orange [22:55] I can reproduce on Ubuntu 12.10 Chrome 30.0 [22:56] just added that to the bug... [22:56] oh wait [22:57] maybe back later [22:57] have a good night === gary_poster is now known as gary_poster|away [22:57] you too [22:57] huwshimi I just updated the bug [22:57] it doesn't actually act normal [22:57] it 'thinks' it's open [22:57] so clicking it 'closes it' heh [22:58] hatch: Oh, so is it a state bug that is being made obvious by the animation? [22:58] well no the url shows that it should be open [22:59] then clicking it changes the url to /minimized [22:59] which changes the tab to orange but stays closed [22:59] then clicking it again triggers the animation [23:00] huwshimi it open happens in chrome [23:00] so I'm guessing it's a chrome bug [23:00] hatch: Right, but it might not be updating a class or something... [23:01] hatch: which bug is this? [23:01] or maybe clicking it changes the state so that it starts working again [23:01] https://bugs.launchpad.net/juju-gui/+bug/1254903 [23:01] <_mup_> Bug #1254903: Sidebar does not open if GUI is loaded in the background [23:01] hatch: oh hmm, hey did we make sure we can skip animations for tests or anything? [23:02] I know it's a bit late in the game now, but I'm a bit nervous that there's some awesome side effects of this that we'll 'discover' [23:02] rick_h_: I'd be interested to see if you can reproduce that bug... I can't here. [23:02] checking [23:03] rick_h_: It's just change positions of elements. I'm not sure if it could break a test [23:03] *changing [23:03] huwshimi what class should I be looking for when it's minimized? [23:04] hatch: The tab should have a parent of #subapp-browser-min [23:04] hatch: Otherwise the parent should be #subapp-browser [23:04] or grandparent or something up the chain [23:04] ok well both are in the DOM [23:05] min shows hidden="true" [23:05] and subapp-browser doesn't have that [23:05] when I click again, the hidden="true" moves to the subapp-browser [23:05] I can demo it in a hangout if you like [23:06] hatch: No that makes sense. So it's doing the correct thing, it's probably the animation then. [23:06] I wonder how hard it would be to create a repro to see if this is indeed a chrome bug [23:06] hatch: So it's starting on the first keyframe but not animating to the second by the sounds of things... [23:06] I can demo if you like [23:07] hatch: huwshimi can't dupe here. [23:07] hmm, and I could dupe on two computers [23:07] rick_h_: Could it be something like this: http://blog.wercker.com/2013/10/02/Chrome-css-animation.html [23:07] oops, hatch: ^ [23:08] rick_h_: Yeah, I think we should just blame hatch [23:08] lol incoming call [23:08] https://plus.google.com/hangouts/_/lite/72cpjbdnhbd1meck125vvcnibk?hl=en [23:09]