IRC logs for #baserock for Wednesday, 2015-10-28

*** gtristan has joined #baserock01:45
*** gtristan has quit IRC06:36
*** gtristan has joined #baserock06:52
lostducki sometimes wonder whether our could be just a general site with a link to the wiki, rather than a redirect to our wiki, that might make it easier to link to other interesting content/news like relevant tweets, commits to definitions etc08:39
*** toscalix__ has joined #baserock08:52
*** bashrc has joined #baserock09:04
*** bashrc has quit IRC09:09
*** bashrc has joined #baserock09:09
*** mariaderidder has joined #baserock09:10
*** bashrc has quit IRC09:12
*** bashrc has joined #baserock09:12
*** bashrc has joined #baserock09:13
*** edcragg has joined #baserock09:18
*** nowster has joined #baserock09:21
*** locallycompact has quit IRC09:29
*** jonathanmaw has joined #baserock09:30
pedroalvarezlostduck: that would be cool, but I'm afraid nobody will generate content for that09:38
pedroalvarezI'd like to see a shiny webpage, with an explanatory video about Baserock, some links, and some news, yes09:39
* SotK thinks it would be good too09:41
pedroalvarezalso a twitter account :)09:43
* CTtpollard still wonders if it would be beneficial to switch to Youtube 09:43
pedroalvarezI've been uploading videos to youtube about Baserock for ages :)09:47
*** ssam2 has joined #baserock09:47
*** ChanServ sets mode: +v ssam209:47
gtristanpedroalvarez, I am building against a patch that you wont like :)09:51
gtristanwell, who knows, maybe you will like it heh09:51
*** paulwaters_ has joined #baserock09:52
pedroalvarezgtristan: please, share with us before spending ages building09:52
pedroalvarezwhy I shouldn't like it?09:53
gtristanpedroalvarez, the thing is this - glibc is the source of default installed charsets - and actually includes a target to automatically generate the locales09:53
gtristanbecause you will build *everything* for one, and for two: It adds 102MB of mapping to every system09:53
pedroalvarez /o\09:54
gtristanmy thoughts on this though; I think that the way we stand right now, all of our systems are already thoroughly bloated09:54
gtristanand that ideally, glibc.morph itself should have a semantic for building it with all enabled (for a desktop experience you probably want that), or with only some defaults09:55
pedroalvarezwell, in your patch it looks like you are moving them to the glibc-locale artifact, which is nice because you can choose if you want it or not09:56
gtristanat least that09:56
pedroalvarezand it's not in build-essential-minimal10:20
pedroalvarezI think I don't dislike the patch :)10:20
gtristan[88 + 71/382/382]10:21
gtristanless than half way but moving along pretty well with 2 instances :)10:21
*** locallycompact has joined #baserock10:25
pedroalvarezjjardon, franred: I believe I've fixed all the issues now
pedroalvarezjjardon: no hacks in weston, and no duplication of weston10:30
pedroalvarezsolved by introducing a genivi.configure extension10:30
tiagogomes__pedroalvarez the copyright year should only be 2015. Stop copying OpenStack extensions :)10:31
pedroalvarezit was the first thing I noticed I had to change, but forgot10:35
pedroalvarezfixed, thanks!10:38
persiagtristan: re: bloating: can we split langpacks?10:43
persiaMy thought is to have selectable artifacts for the content, and only the framework by default, to allow system develops to select a subset of locales.10:45
gtristanpersia, technically: perhaps10:46
gtristanpersia, that would require that we allow the definition of a chunk have no "repo:", I think10:46
gtristanpersia, it would also require a lot of leg work afaics10:47
ssam2easy solution would be a .configure extension that deletes all of them except the ones you want10:47
gtristanssam2, no10:47
gtristanssam2, thats the thing, every language does not "add some file"10:47
gtristanevery language updates a single file which is a cache10:48
ssam2ah, balls10:48
gtristanglibc itself naturally knows how to update it all on it's own based on the charsets it installs10:49
ssam2so it has to be done at integration time... I see10:49
gtristanand perhaps based on another file which knows "what locales can be supported with those charsets"10:49
gtristannot necessarily, but that is another approach yes10:50
gtristanI like the idea of defining chunks without payloads, much more than I like the idea of allowing a stratum to run system-integration hooks10:50
gtristanthe extra legwork involved would mean that someone would have to actually maintain language packs10:52
paulsher1oodgtristan: i think ybd can cope with chunks without repos (but haven't tested it lately)10:52
gtristanas opposed to having glibc, every time we pull a new version; just "know" what it supports10:52
paulsher1oodand if it can't it should10:52
gtristanbut, I'm sure we could probably have both10:52
gtristanthats an interesting use case to consider actually; system A includes strata B and C... C asks for "all-languages" and B asks for "language-pack a, b and c"... Build should either bail out on the conflict; or assume that "all-languages" is good enough to satisfy "language-pack a, b and c"10:56
persiaThe problem is the size, really.  Hurts things like Pi or Yum10:56
gtristanpersia, so those already use build-essential-minimum anyway ?10:57
gtristanI think the quickest solution is a hard coded alternative to the full locale package10:58
gtristanbut language packs is interesting though, as you could at least configure your Pi for "only japanese"10:58
gtristanpersia, I think that requires a *lot* more thought though; because we would basically also want to filter installed translations by the available "language packs"10:59
persiaYes :)11:11
*** locallycompact has quit IRC11:22
*** locallycompact has joined #baserock11:23
*** paulwaters_ has quit IRC11:25
gtristanpersia, so lets keep this in mind, in the meantime; is the current situation where 'build-essential-minimum' does not include the locales... satisfactory for the smaller builds ?11:28
* gtristan just wants to be clear if we are blocking on something right now or not :)11:28
*** ssam2 has quit IRC11:41
*** ssam2 has joined #baserock11:44
*** ChanServ sets mode: +v ssam211:44
*** franred has joined #baserock11:59
*** gtristan has quit IRC12:04
*** locallycompact has quit IRC12:05
*** locallycompact has joined #baserock12:05
jjardonseems mason is still failing; pedroalvarez did you manage to increase the RAM in the end?12:16
pedroalvarezjjardon: I didn't have time, I was planning to do it over the weekind if that's oke12:17
jjardonpedroalvarez: ah, ok, thanks12:18
ssam2adding swap would be quick, i could do that now12:23
ssam2bah, can't create a swap file on btrfs12:29
ssam2ok, i can't do it now12:29
pedroalvarezoh, we might be able to create swap partitions with the partitioning support...12:31
pedroalvarezjust a thought12:31
*** brlogger has joined #baserock13:19
*** brlogger has joined #baserock13:19
persiagtristan: do not let my preference for the right solution block any actual implementations.13:55
gtristanpersia, got it13:59
* CTtpollard apologises 13:59
richard_mawCTtpollard: surprise, or irc != vim?14:00
CTtpollardrichard_maw: the latter14:00
paulsher1oodgtristan: ybd is now at 15.44 - instances now log when they're in retry mode - maybe that can help identify places to parallelise14:04
gtristanpaulsher1ood, great, I was waiting for the release tag especially for the parallel deploy bug :)14:39
paulsher1oodyup that's fixed14:42
*** mariaderidder has quit IRC16:46
*** nowster has quit IRC16:49
gtristanbefore I sleep, I should point out16:56
gtristan + = awesome16:57
* gtristan will follow up with another milestone email when he wakes up tomorrow16:58
pedroalvarezI should try this on a jetson or on a Pi16:58
gtristanyes, I'm curious how it turns out16:59
gtristanit should work fine though, so long as the mesa/dri/opengl stuff is known to run I guess17:00
gtristanit took me literally all day to compile though17:00
* gtristan spent most of that time working on his proposal17:01
pedroalvarezany exciting proposal for baserock-dev?17:03
*** franred has quit IRC17:04
gtristanit's basically what we spoke about in Seoul already, but with more thought put into the details17:05
gtristansince we've been getting into this discussion one way or another here on IRC, or in person (i.e. handling chunks differently), better to propose something solid and start from there17:06
gtristanthan random discussions which generally lead nowhere17:06
pedroalvarezI agree17:08
pedroalvarezI think a bit about it every day, and I like and dislike the idea at the same time17:09
gtristanhopefully it will be more likable after having put more thought into the edge cases17:09
gtristaninstead of just that new guy ranting away ;-P17:10
*** jonathanmaw has quit IRC17:10
gtristanlike an organic chicken !17:10
pedroalvarezwithout head17:10
pedroalvarezI keep wondering why we ended up talking about organic food17:10
pedroalvarezno need to talk about that here :)17:10
pedroalvarezgdp systems contain 30 strata..17:19
pedroalvarezruntime deps please!17:19
Kinnisonpedroalvarez: because inorganic food tastes of metal17:20
Kinnisonand stones17:21
* gtristan gets popcorn17:21
Kinnisonthat's organic17:21
pedroalvarezdid you know that organic chickens eat little stones (inorganic food)?17:22
KinnisonI should imagine they do -- birds often consume small stones to aid digestion17:22
tlsaA while ago there was some discussion of the ciat UI.  I'd have moved it more towards something like which has a bit more context17:25
toscalix__tlsa: cool17:25
*** toscalix__ is now known as toscalix17:25
pedroalvarezI like it17:26
* SotK too17:26
persiare:runtime depends: Debian has put a lot of effort into autocalculators (e.g. ${shlibs:Depends}, ${python:Depends}), which ought be borrowed to avoid needing overspecification manually17:28
paulsher1oodtlsa: it's interesting,  but so far the main feedback from the ciat discussion has been to steer towards existing solutions if they exist... for example go.cd17:29
persiaDoes GoCD support pre-merge validation?  A scan of the docs and a couple search engine queries aren't pointing me at the right docs.17:32
* paulsher1ood does not know17:34
richard_mawpersia: AIUI you can tell it to monitor a candidate branch, and it could then push a merge elsewhere after it has validated it, I don't recall whether it allowed pattern matching for candidate branches though17:34
richard_mawhm, nope, AFAICT it only accepts a single branch17:39
toscalixquestion, should we consider tlsa a non Codethink contributor? In such a case, it is the first time I see. :-) tlsa how does it feel?17:39
richard_mawand the documented API doesn't let you define which branch either17:39
richard_mawtoscalix: what? why would you consider tlsa to be a non-Codethink contributor?17:39
toscalixrichard_maw: I thought he was another person17:40
toscalixtlsa: welcome back :-)17:40
toscalixabout the layout.... the history is what we need to provide context17:41
toscalixincluding a What it is section helps17:41
toscalixso this is an implementation of what we talked about17:41
toscalixI like it17:42
toscalixit is an improvement from what we had17:42
*** ssam2 has quit IRC17:47
*** bashrc has quit IRC18:03
*** locallycompact has quit IRC18:06
*** edcragg has quit IRC19:45
*** edcragg has joined #baserock20:24
*** edcragg has quit IRC20:55
*** locallycompact has joined #baserock21:39
*** edcragg has joined #baserock21:40
*** toscalix has quit IRC21:49
*** locallycompact has quit IRC22:00
*** Chronus has joined #baserock22:48
*** Chronus has left #baserock22:50

Generated by 2.15.3 by Marius Gedminas - find it at!