**** BEGIN LOGGING AT Fri Aug 15 03:00:00 2014 Aug 15 08:19:18 what Aug 15 08:19:42 I update the srcrev of a u-boot git recipe, and now OE wants to rebuild chromium Aug 15 08:20:06 is there a way to get information why OE decides to rebuild a recipe? Aug 15 08:52:45 morning all Aug 15 09:15:08 hi bl Aug 15 09:16:53 hi woglinde Aug 15 09:42:48 monring Aug 15 10:26:18 I see the x11-server-nodm init script is still broken Aug 15 10:26:32 root cause is this line: xinit /home/root/moo -- /usr/bin/Xorg Aug 15 10:26:50 err, copypaste error :) I meant xinit /etc/X11/Xsession -- `which $XSERVER` $ARGS >/var/log/Xsession.log Aug 15 10:28:19 with this line, sometimes X starts with 1024x768 (on a FullHD screen), sometimes not at all. replacing it with: xinit /etc/X11/custom-session -- /usr/bin/Xorg fixes it, where /etc/X11/custom-session contains just two lines: "/usr/bin/matchbox-window-manager &" and "/usr/bin/matchbox-terminal" Aug 15 10:28:43 and looking into that Xsession it seems it tries to set a resolution based on ... machine names? seems like a fundamentally broken approach Aug 15 10:29:15 dv__: which one is this though? the one from OE-Core or the one from meta-oe? Aug 15 10:29:30 how can I find it out? I built with meta-oe added to bblayers.conf Aug 15 10:30:16 if it's the 2.0 version it's the one from meta-oe Aug 15 10:30:37 and that's probably the one you'll have been using if you're building with meta-oe Aug 15 10:31:21 xserver-nodm-init-2.0-r22.all.rpm Aug 15 10:33:23 so what's up with this? oe-core vs meta-oe xserver-nodm-init ? Aug 15 10:33:34 and hm, perhaps I should just mask the meta-oe one Aug 15 10:35:27 it's an annoying duplication that nobody has had time to sort out as yet... Aug 15 10:35:53 to be honest there is a bug filed to sort it out and it's assigned to me... Aug 15 10:38:23 why not just delete the one in meta-oe ? Aug 15 10:38:54 i wonder if there is some added value/functionality in it that I am missing Aug 15 10:40:23 either way, I'll mask the meta-oe for now. thanks for the info! Aug 15 10:41:02 further details in the bug: https://bugzilla.yoctoproject.org/show_bug.cgi?id=5546 Aug 15 10:44:38 oh, I see. tough one. Aug 15 11:50:34 anyone having experience adding support for freescale flexcan (driver and device) on imx6 (wandboard-solo? Aug 15 14:28:41 Has anyone played with meta-amd? I have a gizmo board and am trying to build an image for the steppe eagle machine. This succeeds and i get a hddimg which I dd to a USB stick. It starts to boot but the kernel panics with an invalid instruction message. Any thoughts? Aug 15 14:32:50 martinmeba: wrong architecture or toolchain issue probably Aug 15 14:34:31 tizen-meta? Aug 15 14:34:46 Yeah - I thought about that - I thought steppe eagle was the code name for the G Series APU which is on the Gizmo. I will keep digging. Aug 15 16:39:50 what is the difference between python classes distutils and setuptools Aug 15 16:40:03 * Crofton|work need to add python-zmq to meta-python Aug 15 16:40:14 and is trying to figure out what recipe to copy :) Aug 15 16:41:47 whats zmq? Aug 15 16:43:49 zero message queue Aug 15 16:43:52 some ipc thing Aug 15 16:44:02 someone added support for it to gnuradio Aug 15 16:48:11 ah okay Aug 15 16:54:42 martinmeba: check default tune files someone might have copied the intel ones verbatim Aug 15 18:00:22 Help - How do I see the output from a raise command? I'm trying to use the USERADD_ERROR_DYNAMIC feature, but while the feature is working, I see no errors about skipped packages, only overall failures due to a skipped package. Aug 15 18:05:21 skipped packages don't generate errors unless you try to build or depend on what's being skipped Aug 15 18:05:37 bitbake the recipe being skipped and you'll see it Aug 15 18:06:25 Yes exactly and that's basically the problem. The package is "silently" skipped and then the whole build fails for "unknown" reasons ... see what I'm getting at? Aug 15 18:07:08 oh, it's a package, not a recipe, i see Aug 15 18:07:12 There is a raise bb.build.FuncFailed("Useful error message"); line, but I don't know where that message is printed, stored, etc. Aug 15 18:07:22 if a funcfailed was raised, the build would halt Aug 15 18:07:29 they result in task failure Aug 15 18:07:38 so it sounds like it's not hitting that at all Aug 15 18:08:38 Hmm, but if I put a bb.warn("Same useful error message"); one line before the raise command, I get the bb.warn output ... just not the raise output. I'm just wanting to make sure I'm not reinventing the wheel (and it sorta feels like I am). Aug 15 18:09:41 it'd be better to use bb.fatal() rather than that raise, from a correctness standpoint Aug 15 18:09:56 FuncFailed is a bitbake internal class used to identify task/function execution failure Aug 15 18:10:02 recipes/classes shouldn't be using it directly Aug 15 18:10:43 great, maybe that would be better and actually stop the build like I want :) Thanks for the input. Aug 15 18:11:56 worth a shot anyway :) Aug 15 18:22:58 kergoth: Thanks, bb.fatal gets me the behavior I wanted. Aug 15 18:23:07 cool, np **** ENDING LOGGING AT Sat Aug 16 02:59:58 2014