**** BEGIN LOGGING AT Thu Aug 10 03:00:01 2017 Aug 10 06:20:06 hi Aug 10 06:20:26 I proposed a patch for meta-qt5 to correct an issue in gstreamer when it comes to building its qt element Aug 10 06:20:38 not sure now if it really is a gstreamer issue though Aug 10 06:20:54 the Qt5Core.pc file created by meta-qt5 contains this: host_bins=/usr/bin/qt5 Aug 10 06:21:31 shouldn't it contain ${RECIPE_SYSROOT_NATIVE}/usr/bin/qt5 ? Aug 10 08:53:23 can anybody easily see why http://cgit.openembedded.org/meta-openembedded/tree/meta-oe/recipes-graphics/libvncserver/libvncserver_0.9.11.bb has a DEPENDS on libsdl? according to its own documentation, this is just for included example code Aug 10 12:41:24 RP, in your never ending quest for speed, a friend is making bizarre claims hat setting cmkae/autotools to run on one cpu speeds things up Aug 10 12:44:44 Crofton: Its a full circle. You know I was the one who discovered this? Aug 10 12:45:23 Crofton: http://linux-kernel.2935.n7.nabble.com/Performance-of-forky-workloads-td1255622.html (Dec 2015) Aug 10 12:45:57 Crofton: Its all about cache reuse I suspect Aug 10 12:50:36 Crofton: I even have a patch which does do_configure cpu pinning: http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/wip-rss2&id=aa692cdc13f1efa132d4786d4c467ac916b5e708 Aug 10 12:54:55 interesting Aug 10 12:56:23 I wonder if turbo boost could be having an impact as well. Doesn't it only boost a single core's speed if the others are idle? Aug 10 12:57:48 If so that would be a pretty small percentage of the issue but... Aug 10 14:03:00 RP, thanks, I'll pass those to the guy rediscoverin this Aug 10 14:05:20 RP, a pox on stubborn people, he is now whining about nabble Aug 10 14:05:38 of course, he has had some support email threads abdly derailed by nabble users Aug 10 18:24:10 lol, He found an official lnk and read it **** BEGIN LOGGING AT Thu Aug 10 19:21:31 2017 Aug 10 20:16:58 Does it rings any bells? .../cgo.a(_all.o): unknown relocation type 30 Aug 10 20:27:56 otavio: if I update the gst recipes to 0.12.2 , what to do about the qt5 gst plugins? would it make sense to add the qt5 packageconfig to the gstreamer1.0-plugins-base.inc file even though it is not part of oe-core or meta-openembedded ? Aug 10 20:28:21 "it" = qt5 support Aug 10 20:32:31 I'd say no Aug 10 20:32:37 I'd add to the specific one Aug 10 20:33:07 but I don't know for sure if base.inc has other items supported by few only Aug 10 20:57:49 behanw: glad to hear you have some automation in that 3000 line script! Aug 10 20:59:11 moto-timo: 4392 src/ready-for.sh Aug 10 20:59:11 2869 check-requirements/check-requirements.sh Aug 10 20:59:32 moto-timo: "som automation"? I'm insulted! ;) Aug 10 20:59:40 That is "some automation" Aug 10 21:00:25 The whole thing automates what I otherwise have to do manually. It *is* automation. ;) Aug 10 21:00:25 :) Aug 10 21:00:38 I automated myself. Aug 10 21:00:51 I'm far too lazy to do things by hand... Aug 10 21:00:59 behanw: that's usually what drives me to write anything. I hate repetition. Aug 10 21:01:02 moto-timo: So pink was okay? Aug 10 21:01:48 behanw: fine with me Aug 10 21:01:52 :) Aug 10 21:02:10 It's all I could find. Aug 10 21:30:55 otavio: the problem is that the --with-moc/uic/rcc options got introduced in 0.12.2 Aug 10 21:31:36 or ... hmm well I guess if it gets added to master it could be ok, assuming oe-core master would also have 0.12.2 Aug 10 21:31:43 yes Aug 10 21:32:37 otavio: something else: kernel config fragments + linux-imx kernels, what do you think about that? Aug 10 21:33:03 dv_: possible to add; for linux-fslc and like Aug 10 21:33:09 yes Aug 10 21:33:32 well, if it were moved from linux-yocto to the base linux bbclass, it would be automatically added.. Aug 10 21:41:49 dv_: there's an open yocto issue in bugzilla about adding fragments support to non-linux-yocto kernels in a more generic way Aug 10 21:42:11 it's slightly non-trivial, because there's no standard way in which defconfig is pulled in and written to .config, it varies from one kernel recipe to the next Aug 10 21:43:20 sounds like a good case for a dedicated task Aug 10 21:43:28 prepare_defconfig() or something like that Aug 10 21:58:19 indeed. do_preconfigure has been done in some recipes, iirc. no real preference, but yeah, something should be done to clearly deliate setting up .config from actually calling make oldconfig & whatnot Aug 10 21:58:40 delineate, that is **** ENDING LOGGING AT Fri Aug 11 03:00:01 2017