IRC logs for #baserock for Friday, 2014-08-01

*** fay [] has joined #baserock06:48
*** fay [] has quit [Client Quit]06:48
*** fay [] has joined #baserock06:51
*** fay [] has quit [Client Quit]06:51
*** fay [] has joined #baserock07:33
fay is now known as Guest4142207:34
*** Guest41422 [] has quit [Client Quit]07:34
paulsherwoodpersia: CI *may* be best practice. or it may just be a way to screw up a project. depends on how well the solution is implemented, and how effective the process is around it07:34
* paulsherwood has seen some very large CI implementations which were disastrous07:35
*** sambishop [] has joined #baserock08:05
*** tiagogomes [] has joined #baserock08:11
*** jonathanmaw [] has joined #baserock08:17
*** ssam2 [] has joined #baserock08:22
persiapaulsherwood: I assert that *some* form of CI is best practice: pushing untested stuff into production is bad.  There's lots of ways to do CI that make everything much, much worse than trusting developers to have done local tests before pushing.08:54
robtaylorpersia: persia: i think you guys may be violently agreeing ;)08:57
persiaTechnically, probably.  Semantically, not quite :)08:58
paulsherwoodthanks, rjek :)09:22
persiaOh shiny!09:23
* persia wonders if essentially copying rjeks work would get a working mips32 port09:23
KinnisonGiven the state of the toolchain, you'd need to be super-careful09:25
persiaIs that video single-frame?09:25
persiaKinnison: How careful?09:25
Kinnisonpersia: I think it took rjek and I almost an entire day of poking at things, attempting to move patches between gcc versions etc, and we still ended up stuck on gcc 4.609:26
* persia has Buffalo BHR-4GRV and would like to be able to set more than 32 port-filtering rules, which seems to require a replacement OS09:26
rjekpersia: is the original file.09:26
rjek(I find that Vimeo videos sometimes do not play)09:27
Kinnisonthe vimeo plays for me if I disable flash blockers entirely09:27
KinnisonI think vimeo has multiple flash elements09:27
radiofreei thought vimeo was html5?09:28
persiarjek: Thanks.  By the way, main(void) is accepted if you aren't parsing arguments ;p09:28
rjekNo, that's never acceptable.09:28
persiaAccepted by SUS API.  Perhaps not by some humans.09:29
rjekIt may be /permitted/09:29
ssam2radiofree: video streaming services giving you html5 video seems to be purely the stuff of fantasy09:29
Kinnisonradiofree: it might depend on your browser09:29
ssam2on youtube you had to opt-in, so I did, and I still get exclusively flash videos09:29
ssam2on systems without flash installed sometimes videos play as just the audio, so I guess something is trying to work, sometimes09:29
radiofreefrom firefox 33 youtube will use the html5 player09:30
ssam2do we have that in writing?09:32
radiofreemake sure media.gstreamer.enabled is set to true, and media.mediasource.enable if you want all that mediasource/webm crpa09:32
ssam2thanks, seems they are enabled09:33
radiofreeevery update to firefox seems to break the html5 video playback in some way09:33
persiaYou could try another browser...09:34
radiofreelike.. epiphany?09:34
* persia likes browers where upstream does not feel they have authority to deploy directly to my machine09:35
radiofreei think the main issue with youtube is mpeg-dash09:35
radiofreethere's an extension you can get for FF that disables that (youtube centre or something) and everything works fine09:35
*** locallycompact [] has joined #baserock09:36
ssam2ah, OK, I'll try that09:36
* petefoth giggles at rjek10:19
* rjek looks at persia 10:21
rjekAnd petefoth.10:21
* rjek looks at you all.10:21
persiaWhat did I do now?10:21
rjekI'm sure you did something :)10:21
* petefoth was giggling at '10:28 < rjek> No, that's never acceptable' and '10:29 < rjek> It may be /permitted/'10:23
* persia triple-checks the tape over the cameras, the state of the blinds, the fit of the aluminium hat, and other items on the checklist redacted for security reasons10:23
petefothand read it in a 'Lady Bracknell voice'10:23
*** locallycompact [] has quit [Ping timeout: 255 seconds]10:28
ssam2seems my system artifact's split rules are failing to match the chef-runtime artifact, so it's not getting included in the system10:35
ssam2given that the system morph contains : {'morph': 'chef/chef.morph', 'artifacts': ['chef-runtime']} that is a little confusing10:35
richard_mawssam2: add name: chef10:39
richard_mawfor backwards compatibility it falls back to using "morph" as the name, which means you'd need chef/chef.morph-runtime as your 'artifacts' match10:39
richard_mawit's simpler to set the name10:39
ssam2richard_maw: ah, OKL10:39
ssam2we need to add a warning to make this clearer10:40
richard_mawand once everything has been moved, rip out the bass-ackwards compatibility10:40
*** jonathanmaw [] has quit [Read error: No route to host]10:41
*** jonathanmaw [] has joined #baserock10:43
*** locallycompact [] has joined #baserock10:44
*** locallycompact [] has quit [Read error: No route to host]10:46
*** locallycompact [] has joined #baserock11:05
* petefoth apologises for his slowness in coming to terms with the 'Upgrading a Baserock installation' workflow. has gone a long way towards rmedying that11:19
persiaJust be careful: I think there's still a bug that if you delete the "factory" system when you're cleaning up afterwards you end up making the system impossible-to-upgrade.11:22
petefothpersia: I'm not going to *do* it - I just need to learn *how* it is to be done :)11:24
* persia recommends doing it once or twice a day: running old versions of Baserock just means running buggy code11:24
* petefoth should probably install a Baserock devel system so that he can try upgrading it11:25
petefothbut that will have to wait11:25
persiaYou can just download an image, and boot in your hypervisor of choice: should be fairly easy.11:28
persiaWe're *close* to making it even easier, but the necessary magic bits to fit into popular solutions haven't quite been finished yet.11:29
radiofreeit works on a jetson now11:29
paulsherwoodpersia: fairly easy modulo all of the manual frbbing with git config, setting up src etc11:29
paulsherwoodradiofree: video? :)11:29
persiaYour patch series got merged?  Hurrah!11:29
radiofreepersia: waiting for it to be merged11:30
persiapaulsherwood: Yes, "fairly" still applies, but ought be considered a bug :)11:30
* persia is reminded about 9p-virtio, and suspects that all the precursors have landed since it was last discussed11:32
radiofreepaulsherwood: no video, no time atm11:36
paulsherwoodradiofree: no problem. must be someone else i can 'persuade' to make one... :)11:38
*** tiagogomes [] has quit [Ping timeout: 250 seconds]12:16
*** tiagogomes [] has joined #baserock12:20
*** tiagogomes_ [] has joined #baserock12:31
*** tiagogomes [] has quit [Ping timeout: 264 seconds]12:32
ssam2/usr/lib/ruby/2.0.0/rubygems/dependency.rb:296:in `to_specs': Could not find 'ohai' (~> 7.2) - did find: [ohai-7.2.0.rc.1] (Gem::LoadError)13:00
ssam2how confusing.13:00
ssam2ah, I think I understand. I'm getting the dependencies from the Gemfile in chef.git, but then actually using a premade Chef Gem from rubygems.org13:21
ssam2I guess I need to build the Chef Gem from source, instead13:21
ssam2or ensure that the Git repo is at the tag matching the Gem that's used, but I doubt there's a reliable way to do that13:22
paulsherwoodssam2: might be connected to the tag containing 'rc'14:17
* radiofree hangs head in shame14:39
radiofreei figured out my distbuild problem14:39
radiofree(other than .morph.morph)14:40
radiofree"time /root/morph/morph"14:40
radiofreedisbuild-controller was treating the output from time as an error message14:40
* radiofree will updated to say "DON'T USE THIS IF YOU WANT TO USE IT AS A DISTBUILD CONTROLLER"14:41
ssam2oh, yeah, obvious and yet completely unobvious at the same time!14:41
richard_mawI want to get rid of the time part of that14:42
persiaSomehow I think the instructions on using latest morph should be a little harder to find.14:42
richard_mawpaulsherwood: since you're the one that put the time command in, what do you think?14:42
paulsherwoodrichard_maw: fine by me14:58
* paulsherwood will continue to use time though, until morph actually reports how long it's taken14:59
radiofreemake sure it doesn't report it on stderr though15:00
*** tiagogomes_ [] has quit [Ping timeout: 255 seconds]16:37
*** ssam2 [] has quit [Quit: Leaving]17:11
*** jonathanmaw [] has quit [Quit: Leaving]17:15
*** franred [] has quit [Quit: Leaving]17:20
*** tiagogomes [] has joined #baserock17:26
richard_mawhi straycat_17:47
straycat_we seem to have lorried the "wrong" git-fat17:47
richard_mawas in we're not using the canonical upstream repository?17:48
straycat_We're lorrying from git:// not git://
straycat_Would you agree that we should be lorrying from git:// ?17:54
straycat_radiofree, I don't think that you should suggest that, master *should* always work.17:54
persiathe cyaninc repo seems derivative of the jedbrown repo, and the jedbrown repo has more recent commits.  Why is one better than the other?17:57
richard_mawAccording to GitHub's own metrics, the cyaninc one is a fork17:57
straycat_According to locallycompact jed brown's is out of date even though it has more recent commits.17:58
locallycompactCan confirm after a weekend of playing with it on a personal project that is more featureful and less wonky.17:59
persiajedbrown certainly has fewer commits than cyaninc18:00
locallycompactand also, it's the one that gets installed when you run easy_install git-fat18:00
persialocallycompact: And pypi is fed from the cyaninc repo?18:00
* persia prefers `pip install git-fat` but would expect the same result doing that18:00
persiaWait: 0.3.1?18:01
persiacyaninc only seems to have
persiaHow did 0.3.1 come to be?18:01
richard_mawhm, well, I'm finishing for the week, can you send a mail to the list so we don't forget it, and can look at it on Monday?18:01
*** locallycompact [] has quit [Ping timeout: 264 seconds]18:19
*** tiagogomes [] has quit [Quit: Leaving]18:39

Generated by 2.14.0 by Marius Gedminas - find it at!