**** BEGIN LOGGING AT Thu Oct 12 03:00:02 2017 Oct 12 15:16:34 a BSP that overrides something in a distro.conf isn't YP compatible? Oct 12 15:20:16 yeah bsps shouldn't have distro policy in Oct 12 15:20:34 hmm. need to talk to LF training material dev :) Oct 12 15:20:39 dl9pf, ping Oct 12 15:50:04 Crofton|work: pong Oct 12 19:28:00 are there any examples of an app recipe setting kernel configs? Oct 12 19:59:29 armpit: one recipe can't alter how another recipe is built, at all, period, kernel or otherwise Oct 12 20:07:55 kergoth, that make sense. I know we have rrecommends on kernel modules but not sure if that is what I am looking for. Oct 12 20:09:47 we released a qa layer and have a automount test so I need to ensure ext2/3/4 are compiled in Oct 12 20:21:30 Does anyone see anything wrong with this configuration (dropped into local.conf) for master/rocko Oct 12 20:21:31 MULTILIB_GLOBAL_VARIANTS_append = " libn32" Oct 12 20:21:31 MULTILIBS ?= "multlib:lib32 multilib:libn32" Oct 12 20:21:31 DEFAULTTUNE_virtclass-multilib-lib32 ?= "mips" Oct 12 20:21:31 DEFAULTTUNE_virtclass-multilib-libn32 ?= "mips64-n32" Oct 12 20:21:31 require conf/multilib.conf Oct 12 20:21:39 I add that, and during parse I get: Oct 12 20:21:50 ERROR: ParseError in : Could not inherit file classes/multlib.bbclass Oct 12 20:22:48 (trying to build MACHINE = qemumips64) Oct 12 20:22:56 I get a similar error with qemuarm64 BTW Oct 12 20:25:03 Ugh.. I see the problem, I missed an 'i'.. Oct 12 20:25:07 ugh Oct 12 20:25:59 also the simple typos Oct 12 20:45:55 armpit: I proposed a way for recipes to at least check the kernel config for items but it was rejected :/ Oct 12 21:04:53 bluelightning, to bad Oct 12 21:05:28 armpit: well, you could always voice your support - I haven't given up on the idea, just haven't got back to it yet Oct 12 21:06:51 is there a patch somewhere? it seems reasonable as we do have other checks Oct 12 21:08:26 check is reason, but with the understanding the recipe could them be considered machine specific Oct 12 21:09:10 fray: for this to work the recipe would already have to depend on virtual/kernel and thus would have to be machine-specific, right? Oct 12 21:09:36 my point is not only that, but the PACKAGE_ARCH would have to be the MACHINE_ARCH (or similar) otherwies it couldc ause rebuild problems.. Oct 12 21:09:45 just another 'check' that would needed to be added. (as an example) Oct 12 21:09:55 right, I think we may have a QA check for that already Oct 12 21:09:59 let me dig out the patch Oct 12 21:12:49 k Oct 12 21:14:01 armpit: here's the discussion: http://lists.openembedded.org/pipermail/openembedded-core/2016-May/121109.html Oct 12 21:14:25 hmm, seems the branch is gone, let me re-push it Oct 12 21:15:23 ah, actually it isn't, we just changed the cgit URLs since then Oct 12 21:15:25 http://cgit.openembedded.org/openembedded-core-contrib/log/?h=paule/kernel-check Oct 12 21:16:31 fray: looks like I implemented it with a class, you could presumably add the PACKAGE_ARCH line in there Oct 12 23:24:00 bluelightning, thanks for the info. I am updating the bug with my use case **** ENDING LOGGING AT Fri Oct 13 03:00:02 2017