**** BEGIN LOGGING AT Wed Nov 20 02:59:58 2013 Nov 20 09:02:27 after an update I get link errors on libgles-omap3: http://pastebin.com/LKSTGnAF Nov 20 09:02:39 meta-ti hasn't changed, so it must be Something Else Nov 20 09:19:21 nobody else with this issue? I see some old reports on it, but nothing recent. Nov 20 10:17:07 morning all Nov 20 10:37:40 JaMa: there is a bug in that bitbake patch Nov 20 10:37:48 JaMa: will send out a revised version Nov 20 10:38:34 RP: thanks for info, will update my local branch when you do Nov 20 11:09:49 tasselhoff ping Nov 20 11:10:07 tasselhoff you are using now diffrent compile option than ti-gles uses Nov 20 11:10:09 d Nov 20 11:10:15 for vfp registerts Nov 20 11:20:36 morning all Nov 20 11:37:19 morning pb_ Nov 20 11:37:56 hi bluelightning Nov 20 12:07:27 hi all, bluelightning , pb_ , woglinde Nov 20 12:07:53 hi silvio_ Nov 20 12:25:56 hi pb Nov 20 12:48:53 RP: Nov 20 12:49:05 I'll try that again... Nov 20 12:49:16 RP: http://git.yoctoproject.org/cgit.cgi/poky-contrib/patch/?id=7b49d336c4672f3dfa78e1c3f1f5c7384264a118 seems to work for me. Nov 20 12:51:15 RP: in as far as the build no longer fails. It seems to negate the benefit of the sstate cache for me though because the bundle_initramfs task causes the kernel to be compiled anyway. :( Nov 20 13:00:36 quboid: right, it fixes the bitbake bug and lets builds happen, we now need to fix bundle_initramfs to be covered by sstate Nov 20 13:00:42 quboid: two different bugs Nov 20 13:01:52 RP: OK, great. Nov 20 13:02:00 pb_: The v2 of the patch appears to fix the issue and not break anything else Nov 20 13:02:35 RP: Please can you point me at v2 of the patch? Nov 20 13:03:09 quboid: its on bitbake -devel or in master-next Nov 20 13:03:23 quboid: there was a small logic bug in the first version Nov 20 13:08:16 RP: v2 (f21910157d873c030b149c4cdc5b57c5062ab5a6) seems to solve the populate_sysroot problem for me. I'll set some more thorough tests going. Thanks! Nov 20 13:09:15 quboid: I'm also running the autobuilder against this patch at the moment fwiw Nov 20 13:09:38 quboid: we hit issues for the v1 but I can see what the problem was there Nov 20 13:20:52 where we having a conversation about changing the default armv7 tune to hard-float abi? Nov 20 13:42:48 *** 0003: oe.path.copyhardlinktree(d.expand("/home/balister/src/oe-repo-test/build/tmp-eglibc/work/zedboard_zynq7-oe-linux-gnueabi/linux-xlnx/3.8-xilinx+gitf4ff79d44a966ebea6229213816d17eb472b303e-r1/image/usr/src/kernel"), d.expand("/home/balister/src/oe-repo-test/build/tmp-eglibc/work/zedboard_zynq7-oe-linux-gnueabi/linux-xlnx/3.8-xilinx+gitf4ff79d44a966ebea6229213816d17eb472b303e-r1/sysroot-destdir//usr/src/kernel")) Nov 20 13:45:12 RP: cool, thanks Nov 20 13:49:34 do_rootfs takes too long Nov 20 14:48:08 hi all Nov 20 14:48:35 I was looking a layer which can provide html5 Nov 20 14:48:49 can anybody point me at it Nov 20 14:51:20 denix, did you have some recipes for meta-sdr? Nov 20 14:53:08 hi. can someone confirm if my statements are correct: 1) when using require/include files are searched in BBPATH 'order', not priority. And BBPATH is contructed by appending all layers 'root' based on the order used in BBLAYERS. 2) Layer priority is used when we need to sort recipes. e.g. when sorting .bbappend, or when multiple recipe with same version exist. Nov 20 14:53:29 bluelightning, did I submit meta-sdr already? Nov 20 14:53:54 Crofton|work: yep, and I approved it - http://layers.openembedded.org/layerindex/branch/master/layer/meta-sdr/ Nov 20 14:54:20 I think I tweaked the description a little bit (since the short desc doesn't appear on the layer page, the long desc has to stand on its own) Nov 20 14:54:37 oh and the subdir was wrong, you had put a URL in it Nov 20 14:54:40 and they do not sort alpha ... Nov 20 14:54:52 it's sorted by type then alpha Nov 20 14:55:24 why doesn't it find gnuradio ... Nov 20 14:55:40 dog Nov 20 14:55:43 doh Nov 20 14:55:47 defaults to layers Nov 20 14:55:49 not recipes Nov 20 16:22:29 re Nov 20 16:22:58 hm if I have 2 linux-foo.bb with diffrent versions how can I select the version Nov 20 16:23:08 looks like the highestnumber wins Nov 20 16:23:38 PREFERRED_VERSION? Nov 20 16:23:39 PFERRED_VERSION-linux-foo seems not work Nov 20 16:23:39 unless you set PREFERRED_VERSION_pn-linux-foo Nov 20 16:24:49 rburton what is pn in this case Nov 20 16:25:11 looks like its a bit more compilcated in case of kernel stuff Nov 20 16:25:34 "-pn-" is literal text Nov 20 16:25:35 ${PN} isn't in external overrides anymore Nov 20 16:25:40 pn-${PN} is Nov 20 16:25:48 args Nov 20 16:25:54 damn underscore Nov 20 16:26:14 hm no Nov 20 16:36:58 rburton: pn- isn't needed for PREFERRED_* Nov 20 16:38:06 woglinde: did you try PREFERRED_VERSION_linux-foo or PREFERRED_VERSION-linux-foo (the latter would never work) Nov 20 16:38:41 rburton: in fact I suspect pn- won't work at all for PREFERRED_* Nov 20 16:42:29 bluelightning PREFERRED_VERSION_linux-foo Nov 20 16:43:32 woglinde: ok; if that doesn't work you should check that that has not already been set somewhere else Nov 20 16:43:40 bitbake -e should help there Nov 20 16:45:57 bluelightning I have Nov 20 16:46:20 infect -DDD tells me there is only one provider Nov 20 16:49:32 woglinde: are you sure the build system is actually picking up the other recipe? Nov 20 16:49:44 yes Nov 20 16:49:57 er, then I can't tell what could be going on... Nov 20 16:50:04 I see both .bb getting parsed Nov 20 16:50:14 otherwise I would not have been asked Nov 20 17:01:28 woglinde: something like that happened to me once.... one .bb file had DEFAULT_PREFERENCE="-1" which caused it to be silently ignored... Nov 20 17:02:55 ndec right I thought I can trick with it but does not work either Nov 20 17:04:34 have the 2 .bb the same name? Nov 20 17:04:39 filename? Nov 20 17:05:21 no Nov 20 17:05:31 maybee I have a typo Nov 20 17:24:20 hm yes typo Nov 20 17:24:22 hrms Nov 20 17:24:25 sorry for noise Nov 20 17:25:39 it's too quiet anyway... Nov 20 17:25:50 but then, i don't make pilicy Nov 20 17:25:57 *policy evenb Nov 20 17:26:19 * mr_science smacks fingers in keyboard drawer Nov 20 17:31:58 lol Nov 20 22:10:12 pb_: what's the best way to restore e.g. FREESMARTPHONE_GIT used in meta-oe and E_MIRROR, E_SVN in meta-efl? Adding them to layer.conf works, but I wonder if there is better place Nov 20 22:10:51 heh even one HANDHELD_CVS was found :) Nov 20 22:10:53 The URL: '${HANDHELDS_CVS};module=apps/atd;tag=ATD-0_70' is invalid and cannot be interpreted Nov 20 22:15:50 i been sent here from #angstrom to ask why this is happening: Nov 20 22:15:52 Missing or unbuildable dependency chain was: ['udev-hostap-cs-vcc-workaround'] Nov 20 23:02:42 ZetaNeta, and be warned we are morning people mostly Nov 20 23:02:59 and we need more info about the setup and what you are tyingg to do Nov 20 23:15:13 Crofton|work: someone reported incorrect checksums for uhd in oe-core did you notice? Nov 20 23:31:34 Crofton|work: https://lists.yoctoproject.org/pipermail/yocto/2013-November/017294.html Nov 20 23:31:52 Crofton|work: https://lists.yoctoproject.org/pipermail/yocto/2013-November/017295.html Nov 20 23:35:05 I saw that Nov 20 23:35:19 have you seen it? Nov 20 23:35:26 I need to force a re-download Nov 20 23:35:51 what is odd is a few poeple should have tripped it in the past few days I think Nov 20 23:37:16 I have seen it, but I don't care about uhd anymore :) Nov 20 23:37:23 I've even replied just now Nov 20 23:37:44 this is odd Nov 20 23:38:13 bluelightning, let me know also, due to the confusing subject line Nov 20 23:38:28 but I've talked a couple of guys thourhg builds recently and they didn;t notice Nov 20 23:38:46 Crofton|work: btw there is another patch for gnuradio which I wont probably finish now, see http://patchwork.openembedded.org/patch/56967/ Nov 20 23:39:21 yeah Nov 20 23:39:27 I may have added those already Nov 20 23:40:18 I've been testing cross building gnuradio with output of populate_sdk today Nov 20 23:40:21 so far so good Nov 20 23:43:22 IIRC I haven't merged this one in meta-oe, because test-dependencies.sh was still showing the same errors in gnuradio or some more and I wanted to fix them all in one go Nov 20 23:44:52 yeah it was showing one more Nov 20 23:44:53 WARN: packages/armv5te-oe-linux-gnueabi/gnuradio/gnuradio/latest lost dependency on libpython3 Nov 20 23:44:57 http://lists.openembedded.org/pipermail/openembedded-core/2013-October/084905.html Nov 20 23:51:02 python3? Nov 21 00:10:41 Crofton|work: yes, from khem's patches Nov 21 00:38:07 heh PID 1 now ported to libsystemd-bus, so long with my issues to get some MACHINEs to at least 3.0 kernel to support latest systemd from oe-core :) Nov 21 02:51:40 i just noticed, earlier today, that OVERRIDES, by default, doesn't (seem to) include FILESOVERRIDES Nov 21 02:52:01 am i investigating OVERRIDES correctly? Nov 21 02:54:11 is FILESOVERRIDES not part of the OVERRIDES facility? Nov 21 02:55:34 FILESOVERRIDES is a subset of OVERRIDES for use by FILESPATH, as minimizing the FILESPATH is important for performance Nov 21 02:55:45 afaik, that is Nov 21 02:56:26 ah okay, i see it being used in base_set_filespath() Nov 21 02:56:52 kergoth: thanks for the clarification :-) Nov 21 02:56:56 np **** ENDING LOGGING AT Thu Nov 21 02:59:59 2013