IRC logs for #automotive for Wednesday, 2015-09-16

*** ToxicGumbo1 has joined #automotive00:17
*** jjardon_ has joined #automotive00:21
*** jjardon_ has quit IRC00:21
*** jjardon_ has joined #automotive00:21
*** ToxicGumbo has quit IRC00:21
*** jjardon has quit IRC00:21
*** ToxicGumbo1 has quit IRC00:28
*** ToxicGumbo has joined #automotive00:28
*** persia_ has joined #automotive00:29
*** dzen_ has joined #automotive00:32
*** dzen has quit IRC00:35
*** jjardon_ has quit IRC00:35
*** persia has quit IRC00:36
*** jjardon_ has joined #automotive00:42
*** jjardon_ has joined #automotive00:42
*** riazantsev has quit IRC00:48
*** fernandod has quit IRC01:47
*** fernandod has joined #automotive01:47
*** jjardon_ has quit IRC02:28
*** jjardon_ has joined #automotive02:33
*** riazantsev has joined #automotive02:42
*** riazantsev has quit IRC04:53
*** KlausUhl has joined #automotive07:06
*** jeremiah has joined #automotive07:12
*** KlausUhl has quit IRC07:19
*** jjardon_ has quit IRC07:19
*** toscalix__ has joined #automotive07:24
*** KlausUhl has joined #automotive07:24
*** jjardon_ has joined #automotive07:24
*** dzen_ is now known as dzem07:31
*** dzem is now known as dzen07:31
*** jacobo has joined #automotive07:34
*** apinheiro has joined #automotive07:38
*** kooltux_ has joined #automotive07:59
*** kooltux__ has joined #automotive08:00
*** kooltux_ has quit IRC08:01
*** kooltux__ has quit IRC08:01
*** kooltux_ has joined #automotive08:02
*** CTtpollard has quit IRC08:02
*** Tarnyko has joined #automotive08:03
*** CTtpollard has joined #automotive08:16
*** Luming has quit IRC08:23
*** Luming has joined #automotive08:24
*** jlrmagnus has joined #automotive08:26
*** fredcadete has quit IRC08:31
Figurejeremiah: parellization doesn't always mean faster. Initializing cores can take quite a much time. from my experience when tinkering with imx6 I was able to get ~3sec boot time from cold boot to qt5 qml ui application running from user space08:34
Figurejeremiah: I'm not saying that systemd is bad I'm just saying that when creating agl the architecture should be designed from ground up to start with screen on as soon as possible and only after that start initializing rest of the services08:35
Figuree.g. genivi is really really slow to start and has huge root image footprint08:35
Figurein general so does yocto and if you try to build minimal image it tends to draw in extra libraries that are not really required08:36
Figurejeremiah: also what you mean by everything in 1 sec?08:36
*** jlrmagnus has quit IRC08:37
Figurejeremiah: and sure I've seen the camera demos also using framebuffer but that's still long way from actual application ui it's just camera video nothing more. With open gl running you can already create your hmi or instrument cluster etc08:38
Figurebtw any of you guys coming to Qt World Summit 2015 in Berlin?08:40
*** CTtpollard has quit IRC08:50
*** Luming has quit IRC08:53
*** CTtpollard has joined #automotive08:53
*** Luming has joined #automotive08:54
*** fredcadete has joined #automotive09:32
*** CTtpollard has quit IRC11:19
*** CTtpollard has joined #automotive11:26
*** jeremiah has quit IRC11:30
fredcadeteI have noticed that the GDP is installing libgbm-dev into the target11:47
fredcadeteand I imagine it's because wayland will not start up correctly if it's not there11:47
fredcadeteI have tracked the root cause to libgbm and proposed a patch to the maintainer:
fredcadetewith this patch in libgbm it should not be necessary anymore to install libgbm-dev11:48
CTtpollardfredcadete: cool :)11:59
*** fredcadete has quit IRC12:24
*** persia_ has quit IRC12:48
*** persia has joined #automotive12:50
*** jacobo has quit IRC13:23
*** jacobo has joined #automotive13:23
*** riazantsev has joined #automotive13:24
*** KlausUhl` has joined #automotive13:28
*** jjardon__ has joined #automotive13:29
*** KlausUhl has quit IRC13:29
*** jjardon_ has quit IRC13:29
*** Tanikawa has joined #automotive13:56
*** stevel has joined #automotive14:00
*** stevel is now known as Guest3710814:01
*** yoshi_ito has joined #automotive14:01
Guest37108nick steve_l14:02
*** Guest37108 is now known as steve_l14:03
CTtpollardhi steve_l14:03
*** waltminer has joined #automotive14:08
paulsherwood== AGL Layer Design Meeting started ==14:08
paulsherwoodJenkins is now running14:08
steve_lwoohoo :)14:08
paulsherwoodQt patch has been rejected because it's human instructions, not recipe14:08
paulsherwoodwe recommend Tanikawa post his Qt5 patch direct to gerrit14:09
paulsherwoodNed has joined the meeting14:09
paulsherwoodothers are welcome to join too
paulsherwoodwe suggest Tanikawa mentions Jens contribution to the patch in the commit message14:12
paulsherwooddiscussion about package group design based on the discussion at AMM documented in this thread...
paulsherwoodaction on Tanikawa to document based on Rudi's comments14:17
paulsherwoodat AMM there was discussion about staying close to poky versions14:18
*** ToxicGumbo-work has joined #automotive14:18
paulsherwoodTanikawa suggests it may be best to learn from GENIVI how best to organise to track poky14:21
paulsherwoodNed suggests he can discuss this with new WR meta-ivi maintainer and comment by end of this week14:22
paulsherwoodquestion is whether AGL explicitly specifies versions, or adopts versions from poky14:23
paulsherwoodNed suggests best-practice approach depends on situation. for example some components move to gplv3, maybe we need to stay with a preferred gplv2 version14:24
paulsherwoodTanikawa notes we currently have two gstreamer versions14:25
paulsherwoodeven with poky 1.7 we use gstreamer 1.2314:25
paulsherwoodsteve_l: suggests best practice is not to use versions in bb.appends14:27
paulsherwoodpaulsherwood asks if we are expecting to snapshot poky at release times, or take ongoing changes14:29
paulsherwoodwaltminer says we shoudl choose a release, until decision to move to next version14:29
paulsherwoodsome discussion about GENIVI's approach on this. steve_l mentions that GENIVI9 will be 1.8, for example - Ned says normal approach is to consider at the time (every six months), pick current stable version14:31
paulsherwoodNed asks what is the proposal for AGL cadence14:33
paulsherwoodwaltminer thinks case-by-case basis, depending on silicon vendor recommendations14:34
paulsherwoodand other ecosystem considerations14:34
paulsherwoodsteve_l: it would be hard for silicon partners to track every yocto release14:35
radiofreewhat Qt patch would that be?14:38
paulsherwoodpaulsherwood raises Fulup ar Foll's question from last week about governance for decisions14:39
paulsherwood(eg how is cadence decision to be taken)14:39
paulsherwoodwaltminer needs to consider this further14:39
paulsherwoodpaulsherwood raises question of what happened to minutes for SAT meetings (as raised by ToxicGumbo-work yesterday)14:40
paulsherwoodwaltminer says he'll reply to Eugene's question on list14:41
paulsherwood16:48 < ToxicGumbo-work> Anyone know why the AGL SAT meeting minutes stop in February?  No more minute taking or no meetings?14:41
paulsherwood16:48 < ToxicGumbo-work>
paulsherwoodTanikawa raises systemd question - he has made a systemd patch - when should ths go to gerrit?14:44
paulsherwoodwe agreed to seek formal feedback on this question on the list. hence above email. systemd haters do your worst! :)14:47
Figurepaulsherwood: definately systemd over systemv14:48
paulsherwoodFigure: on list please :)14:48
Figurebut I would say that maybe busybox could be optional or something like that14:48
paulsherwoodsome further discussion about repositories14:50
paulsherwood- where does code for demo applications go?14:52
paulsherwood(one project or multiple projects)14:52
waltmineruse AGL/DemoApps/AppName14:53
Figurewaltminer: +114:53
* paulsherwood has a hard cut-off on the hour, fwiw14:53
paulsherwoodwaltminer is a good timekeeper14:54
steve_lhmm lost audio in the gotomeeting. As it's almost time I'm dropping out. Interesting to hear how things are going on AGL side.14:55
* waltminer thanks paulsherwood for keeping minutes14:58
waltminer=== AGL Layer Design Meeting Ended ===14:58
steve_lRe the minute "it would be hard for silicon partners to track every yocto release" I would say it would be hard to trace every Yocto release if *all* upgrades in the associated poky meta would also need to be supported15:01
steve_lat least those that touch h/w.15:01
radiofreewhat was the patch for qt?15:05
*** steve_l has quit IRC15:08
*** jjardon__ is now known as jjardon15:10
*** toscalix__ has quit IRC15:16
*** waltminer has quit IRC15:21
*** yoshi_ito has left #automotive15:28
*** waltminer has joined #automotive15:30
*** KlausUhl` has quit IRC15:32
*** Tanikawa has quit IRC15:54
paulsherwoodradiofree: instructions for qt integration into agl15:59
paulsherwoodit was provided as instructions in the readme, rather than an actual recipe16:00
*** fredcadete has joined #automotive16:40
*** waltminer has quit IRC16:45
*** jacobo has quit IRC17:20
*** jacobo has joined #automotive17:21
*** jacobo has quit IRC17:45
*** jacobo has joined #automotive18:14
*** Tarnyko has quit IRC18:38
*** jacobo has quit IRC18:57
*** waltminer has joined #automotive18:59
*** circ-user-pHY1t has joined #automotive19:07
*** circ-user-pHY1t has left #automotive19:07
*** DanCauchy has joined #automotive19:10
*** apinheiro has quit IRC19:52
*** DanCauchy has quit IRC21:34
*** ToxicGumbo-work has quit IRC21:41
*** waltminer has quit IRC22:16
*** waltminer has joined #automotive23:22
*** waltminer has joined #automotive23:22

Generated by 2.14.0 by Marius Gedminas - find it at!