**** BEGIN LOGGING AT Mon Mar 09 02:59:58 2015 Mar 09 07:30:07 morning Mar 09 08:03:55 good night... Mar 09 08:07:59 good morning Mar 09 08:08:19 morning mckoan, evening nerdboy Mar 09 09:16:08 morning woglinde, so i tried renaming the java -initial and -bootstrap recipes into -initial-native & -bootstrap-native and now it seems to run without warnings. i will probably also submit a report to yocto bugzilla to document the issue with autotools.bbclass, but i don't really see it getting fixed anytime soon. Before i submit this patch to meta-java, should i run any particular tests? ive already tried Mar 09 09:16:10 building jamvm and openjdk-7-jre + openjdk-6-jre, and no sstate-control-related warnings. Openjdk-7-jre seem to emit some qa-issues (openjdk-7-jre quires libjli.so but no providers in its RDEPENDS), do you know if this is a known issue? Mar 09 10:17:00 gm mago Mar 09 10:17:25 libjli.so comes with openjdk but it is a plugin Mar 09 10:17:40 so we should adjust the QA stuff Mar 09 10:17:59 mago did you some runtime tests? Mar 09 11:53:38 anyone here know about a channel like this for kernel development. Mar 09 12:04:08 pompomJuice: there's some info here: http://kernelnewbies.org/IRC Mar 09 12:21:57 hello channel Mar 09 12:22:13 anyone else having trouble fetching all the stuff from meta-qt5? Mar 09 12:22:53 I have the feeling that qt.gitorious.org has disabled the git protocol Mar 09 12:23:25 http://blog.gitorious.org/2014/12/04/announcing-gitorious-3-2/ Mar 09 12:23:25 I've been seeing a bunch of people have issues with gitorious over the weekend Mar 09 12:23:38 So I need to show a webcam working via an ssh connection Mar 09 12:23:48 what are good packages to install to do this Mar 09 12:25:01 I am patching my tree with protocol=https for most cases, but in a few months a new patch will be needed because gitorious is being eaten by gitlab Mar 09 12:25:01 Crofton, very open question :) Mar 09 12:26:17 if BW and manual configuration is not a big issue, any video chat package can work over SSH if you tunnel the correct ports Mar 09 12:27:09 just having the image on something ssh with X forwarding will et marketing off my back :) Mar 09 12:28:27 my feeling is that video over SSH with X forwarding would give very slow results, even in a local connection Mar 09 12:29:16 at least try that setup on desktop distributions and see if it's acceptable before putting it on OE :) Mar 09 12:29:40 for your actual question, I have no idea of OE packages for video capture or chat, sorry Mar 09 12:29:40 that is fine Mar 09 12:29:46 I ahve set their expectatins low Mar 09 12:29:57 mainly need to show the board can get video Mar 09 12:33:25 or both Mar 09 12:33:28 get video from the network or capture video from HW? Mar 09 12:35:19 mainly show the webcam works Mar 09 12:35:45 ulitimately he wants to take mages from the camera and transmit them via an sdr radio running on the hw Mar 09 12:35:58 and fly the thin on a drone Mar 09 12:36:10 I think this was an excuse to buy the drone Mar 09 12:37:04 challinan, gm I'll have some comments for you at some point :) Mar 09 12:37:55 I don't know much about network video Mar 09 12:39:18 at least get started on video capture (based on your BSP) and see where you can take that output (v4l2? gstreamer?). Then find a SW that can take that and put it over the network Mar 09 12:39:29 gotta go, good luck with that Mar 09 12:42:02 thanks Mar 09 12:42:25 yes Mar 09 12:42:27 gstreamer Mar 09 12:42:29 also what's a good resolution for the preferred provides for lib jpeg turbo? Mar 09 12:42:35 has all the love Mar 09 12:42:49 hm wait Mar 09 12:42:58 what does the cam produces in hw? Mar 09 12:43:04 /dev/video0 Mar 09 12:43:06 or the v4l2 driver Mar 09 12:43:12 uvc webcam Mar 09 12:43:21 no I meant which streams it support Mar 09 12:43:53 the device can spit out a lot of stuff Mar 09 12:43:58 yeah, not sure Mar 09 12:44:05 on rpi it can even do h264 Mar 09 12:44:47 first thing to find out Mar 09 12:52:29 crofton haha you could also try to forward the /dev/video device with socat and ssh Mar 09 12:53:06 for pppd and at-modem that worked for zs Mar 09 12:53:08 for us Mar 09 12:56:49 woglinde, that may be tricky with /dev/video, it's full of ioctl's and I am not sure socat manages that very well Mar 09 13:05:49 fredcadete hehe yes, but would be worth a try Mar 09 13:06:34 yesm if it works you get extra points for hackiness Mar 09 13:24:11 ANy suggestions for setting a preferrefprovider for jpeg/jpeg-native? Mar 09 13:27:43 crfoton no Mar 09 13:27:49 heh Mar 09 13:27:55 just take one Mar 09 13:28:49 we shoudl amke a section ion the wiki for cases where there are multiple providers a you add layers and try and explain benfits of each path Mar 09 13:31:24 we need to deal with this page :) Mar 09 13:31:25 http://openembedded.org/wiki/Documentation Mar 09 13:32:35 Crofton|work: Thanks. built wxPython over the weekend, hope to find time to test it this week Mar 09 13:32:47 masochist Mar 09 13:32:55 still hitting that 4GB issue, need to find a workaround Mar 09 13:33:09 what is that issue Mar 09 13:33:15 I want to beat wx out of gnuradio :) Mar 09 13:33:24 supporting two gui packages is madness Mar 09 13:33:34 using "live" image type, one is restricted to 4GB image size, and I'm up to 6! ;) Mar 09 13:34:01 but some of those qui sinks are very cool. I'd love to see them in pure Qt. Mar 09 13:34:14 * challinan agrees about it being madness Mar 09 13:34:31 there is a lot of madness due to extremely conservative devs Mar 09 13:34:43 as usual ;) Mar 09 13:34:46 and shortage of people to do gui work Mar 09 13:35:18 I do not like wx also Mar 09 13:39:22 no OE recipes is a bad sign for embedded platforms Mar 09 13:42:48 arg Mar 09 13:43:08 vlc want libjpeg-turbo, something else want to build jsut jpeg Mar 09 15:34:22 hmm I inherit from externalsrc and it still tries to patch Mar 09 16:34:09 does github have a way to commen ton a pull request inline? Mar 09 16:34:33 what do you mean? Mar 09 16:35:17 I hav epull request and I would like to make some inline comments on the patch Mar 09 16:35:37 clock on the line Mar 09 16:35:40 click* Mar 09 16:35:52 that will add a comment to the commit, which will be visible in the PR Mar 09 16:36:20 yeah, click on the view of the contents/files and click on a line, nothing to it Mar 09 16:36:23 this is to obvious Mar 09 16:36:24 do it every day Mar 09 16:37:57 we inherit pkgconfig if we create a pc file or use on? Mar 09 16:38:00 or both Mar 09 16:38:13 * Crofton|work always forgets Mar 09 16:38:23 in the past, it was both. today, it's the latter Mar 09 16:38:57 afaik all it does is adds pkgconfig-native to the depends nowadays. still the preferred method, but it just maeks sure you can call the native pkg-config Mar 09 16:38:58 yep, it is down to adding the DEPEND on pkgconfig-native Mar 09 16:45:04 notably pkgconfig-native is not pulled in by autotools Mar 09 16:45:40 so some packags which used it indirectly will start failing unless you inherit pkgconfig explicitly Mar 09 16:45:47 which is an improvement Mar 09 16:50:16 I'm being long winded so heopefully I remember Mar 09 16:50:35 I suspcet many recipes that need pkhconfig-native don't have the inherit :) Mar 09 16:50:51 this is sort of like herd immunity :) Mar 09 19:41:33 If I have a recipe that need commercial licenses whitelisted, only builds using that recipe will fail? Mar 09 21:58:32 khem`: gcc-cross-canadian does fail; we are doing same change there. Mar 09 23:42:06 I have troubles getting USE_DEVFS = "0" to work properly Mar 09 23:42:22 makdevs isn't called within pseudo Mar 09 23:42:27 and no devs are created Mar 09 23:44:44 abelloni: you need devmmpfs in kernel Mar 09 23:46:10 that won't work for an initramfs Mar 09 23:46:22 you need at least /dev/console Mar 09 23:47:47 abelloni: devtmpfs works; Mar 09 23:50:02 it doesn't for me... Mar 09 23:55:52 abelloni: are you mounting it byhand or using the mounting in defconfig? Mar 09 23:58:19 mounting by hand is too late Mar 09 23:59:18 and CONFIG_DEVTMPFS_MOUNT doesn't have any effects with an initramfs Mar 09 23:59:50 basically, /dev/console has to exist before init is run Mar 10 00:07:20 I'm not sure how USE_DEVFS is supposed to work... Mar 10 00:07:51 check the initramfs recipes Mar 10 00:28:42 otavio: basically there are no USE_DEVFS = "0" users Mar 10 00:28:54 and I believe it doesn't work Mar 10 00:29:19 it was previously using FAKEROOT Mar 10 00:29:32 which is deprecated Mar 10 01:03:40 ok, fixed Mar 10 01:10:13 ? Mar 10 01:12:56 I'm using the makedevs from buildroot, running it from a ROOTFS_POSTPROCESS_COMMAND and using FAKEROOTENV/FAKEROOTCMD Mar 10 01:13:07 so it should be possible to make that work **** ENDING LOGGING AT Tue Mar 10 02:59:58 2015