**** BEGIN LOGGING AT Sun Aug 24 02:59:59 2014 Aug 24 10:31:30 JaMa, ping Aug 24 11:09:25 pong Aug 24 11:11:39 good day. I want to override ANGSTROM_FEED_CONFIGS. should I put it in my image.bb or in local.conf? Aug 24 11:11:55 haven't really understood what gets parsed where Aug 24 11:14:59 hmm. nevermind. the issue was something else. changing the provider of opkg-collateral triggered something odd in my tmpdir Aug 24 11:17:51 nope, still not quite there.. overriding ANGSTROM_FEED_CONFIGS makes it build my package, but it still drags inn angstrom-feed-configs, and I have a hard time figuring out why Aug 24 11:41:56 do_rootfs says it fails installing opkg-collateral, but says nothing about why. any ideas on how to debug? Aug 24 11:52:51 now it compiled, but only because I replaced the contents of angstrom-feed-configs with the contents of the recipe I tried using instead. Aug 24 12:07:51 tasslehoff: a DISTRO setting needs local.conf or distro.conf Aug 24 12:08:03 tasslehoff: an image recipe cannot influence other recipe due to scope Aug 24 12:18:26 JaMa, hi Aug 24 12:18:49 I'm getting again bx instruction in klibc armv4 binaries :/ Aug 24 12:19:25 bx,lr ftw! Aug 24 12:21:26 yes, http://pastebin.com/eXR2Cbt1 Aug 24 12:21:55 klibc is patched since one year... Aug 24 12:22:07 I guess binutils is the culprit Aug 24 12:22:44 I have not yet investigated, just got 'Invalid instruction' heh Aug 24 12:23:05 btw, only the klibc binaries are affected Aug 24 12:23:17 armv4 rootfs boots fine Aug 24 12:24:10 sounds like a binutils issue indeed Aug 24 12:25:06 koen: I tried putting it in my machine.conf. should that have worked as well? Aug 24 12:25:46 it's the wrong place for a DISTRO setting, but yes Aug 24 12:26:26 koen: then I have other issues :) Aug 24 16:24:37 Shouldn't "bitbake -c menuconfig" trigger a new do_unpack if I have modified the defconfig Aug 24 16:51:45 kroon: yes Aug 24 16:54:06 well, unless I've also missed something... Aug 24 17:02:19 I have a defconfig in my own layer, which I use to override the defconfig in meta-fsl-arm. Aug 24 17:02:31 1. -c menuconfig first time will unpack Aug 24 17:02:51 2. remove defconfig file in my own layer, -c menuconfig will unpack Aug 24 17:03:11 3. add back defconfig in my own layer, -c menuconfig will not unpack Aug 24 17:06:36 hmm no, i cant reproduce it reliably Aug 24 17:09:52 there seems to be no dependency described that will trigger unpack when defconfig changes Aug 24 17:39:24 I don't think bitbake detects that I've removed the defconfig from my own layer :-( Aug 24 17:44:26 kroon: latest bitbake? iirc there were some caching issues Aug 24 17:44:49 JaMa, yes, latest Aug 24 17:44:57 when the dependency on some file was recorded only when it existed (but I take it back it was from "include" directives) Aug 24 17:44:58 aha Aug 24 17:45:14 oh, ok Aug 24 17:46:33 I was thinking about this commit b0bbd89a4f0b98fa1ab28b8e0526cd9ddb76fa57 Aug 24 17:46:40 but it's unrelated to your issu Aug 24 17:46:41 e Aug 24 17:50:53 Bitbake does unpack my version in my own layer, when it is there Aug 24 18:11:00 Mhm. somethings fishy with the caching I guess. bitbake wouldn't redo the unpack until I modified the defconfig in meta-fsl-arm Aug 24 18:22:03 Bitbake is checksumming the defconfig in meta-fsl-arm, but it is copying the defconfig in my own layer, since I've added a .bbappend with FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:" Aug 24 18:48:29 kroon: check this one https://bugzilla.yoctoproject.org/show_bug.cgi?id=4790 Aug 24 18:50:33 JaMa, ah thanks, this indeed looks relevant Aug 24 19:34:59 bluelightning, I'm looking at cache.py, and I don't think the file checksum cache is properly invalidated when I do the 3 steps I listed above **** ENDING LOGGING AT Mon Aug 25 02:59:58 2014