**** BEGIN LOGGING AT Thu Jul 02 02:59:59 2015 Jul 02 08:07:05 so i am making my baby steps to adopt to poky coming from long years of oe-classic / early oe-core .. Jul 02 08:07:46 i am used to use bitbake -b ... but it seems current poky will not even pull in pkg-config if i do it like this .. Jul 02 10:46:21 can you use relative paths in bblayers, or is there a script which can add/remove paths for you Jul 02 11:06:49 ah, reading about templateconf now, seems to be what I'm looking for Jul 02 11:49:25 rob_w, what did bitbake -b do? Jul 02 11:49:34 gm jackmitchell Jul 02 11:49:49 Crofton|work: morning Jul 02 11:51:53 Crofton|work: in my old chains or in the uptodate chain ? Jul 02 11:52:09 old Jul 02 11:52:57 bitbake -b did actually work in that sense that i didnt fail on configure, where it seemed to use the chains pkg-config Jul 02 11:53:21 in the new chain it fails for me as configure even doesnt know about PKG_CHECK_MODULES() stuff Jul 02 11:54:15 as i am a total agnostic in terms of autotools, i am not of any help i guess Jul 02 11:54:57 maybe pastebing the entire equence Jul 02 11:58:21 http://pastebin.com/2SfegSmV my recipes might well be outofdate too here Jul 02 11:58:28 but this is the failed sequence Jul 02 11:58:45 if i do a bitbake e2manager -c cleansstate && bitbake e2manager .. all is fine Jul 02 12:03:35 I'd fix the tabs in CFLAGS warning Jul 02 12:03:52 done already Jul 02 12:04:12 also try bitbake pkgconfig-native Jul 02 12:04:19 or do you think that this might cause my issue Jul 02 12:04:25 ah ok tring that Jul 02 12:04:48 PKG_PROG_PKG_CONFIG: command not found Jul 02 12:04:58 ah yeah that was a try Jul 02 12:04:59 I'd look at the configure script line that makes that message Jul 02 12:05:07 ignore that message Jul 02 12:05:35 ok let me redo the thing and pastebin again maybe Jul 02 12:07:41 also I wouldn't run with -b Jul 02 12:09:15 http://pastebin.com/fUa2eHbZ http://pastebin.com/Bhu4dLbc Jul 02 12:09:38 yeah -b was the old way i did it for many years, of course assuming all dependcys are done Jul 02 12:10:09 i am fine with doing it without -b ... and therefore having the chain do its magic Jul 02 12:10:25 i just wondered why and what changed in this repsect Jul 02 12:10:49 pls dont laugh at my recipe file ;-) Jul 02 12:24:13 | /data/mibtec-bsp/fido-am437x/build/tmp/work/cortexa9hf-vfp-neon-poky-linux-gnueabi/e2manager/0.1-r0/e2manager/configure: line 5483: syntax error near unexpected token `PACKAGE,' Jul 02 12:24:15 weird Jul 02 12:24:49 k inherit pkgconfig build pkg-config-native btw Jul 02 12:25:33 double append to CFLAGS, not sure what that will result in :) Jul 02 12:26:32 I'd review the logs carefully and look at the autofoo in your source Jul 02 12:30:43 CFLAGS_append += does work .. althoug its a ugly thing to get a date inject into the compilat Jul 02 12:30:58 anyway thx for the time .. Jul 02 12:31:03 sure Jul 02 12:31:16 I'd drop either the append or the += Jul 02 12:31:27 you'll need to check whouch one adds a space Jul 02 12:31:46 and try to understand the syntax error Jul 02 12:32:03 I didn't see any gross issues Jul 02 12:35:17 http://pastebin.com/8rahdS2D check this out .. the 2 log_configure ,., old is from -b and new is without Jul 02 12:35:39 the new one does know to expand into the chains pkg-config Jul 02 12:36:22 the old using -b .. doenst even try to, hence the lines end up in the configure script and spit the syntax error Jul 02 12:44:08 http://pastebin.com/8kCJAFeA the actuall configure scripts generate by the 2 ways Jul 02 12:44:30 the new does set the PKG_X stuff which -b seems to skip Jul 02 12:49:24 so it might bail down to my old configure.ac which misses some stuff .. which would -b make to work Jul 02 12:50:15 at least i now understand it a bit better , which is all i want Jul 02 12:56:17 I'd avoid -b unless you have a really good reason to use it Jul 02 12:58:06 rob_w, and thanks for updating to oe-core :) Jul 02 13:10:46 Crofton|work: will do .. yeah oe-core rocks, but i will be with my old chains for some more time .. until my stupid hw gets obsolete Jul 02 13:11:17 *g* Jul 02 13:18:46 too hot ,, i am off Jul 02 17:10:12 joshuagl: I have to destroy your autobuilder build. :/ I need to bring it down in about an hour to fix some stuff Jul 02 17:18:48 :q Jul 02 19:06:54 oeclassic, if I have multiple overlays, and there's a foo_1.0.bbappend at multiple levels, do they stack (both get used) or does one override the other? Jul 02 19:07:14 it's cumulative Jul 02 19:07:17 both in classic and current Jul 02 19:07:23 thx kergoth Jul 02 19:45:31 oeclassic, is pkg_postinst_${MACHINE} allowed? Jul 02 19:49:00 not sure what you mean by allowed. Jul 02 19:49:29 first, pkg_postinst has to be pkg suffixed, as the packaging code doesn't use 'pkg_postinst', but 'pkg_postinst_', at least as far as i know Jul 02 19:50:09 well, it might work, but you don't want it to, or it'll end up affecting -dev, -dbg, etc Jul 02 19:50:10 i *think* pkg_postinst__${MACHINE} should work.. i know it does now, don't recall precisely how well old bitbake handled multiple-overrides Jul 02 19:51:24 * kergoth thinks the packaging code really doesn't need to use overrides the way it does, could explicitly grab what it needs instead, but i can see why we did it this way, easier Jul 02 20:53:31 kergoth ping? Jul 02 20:53:36 woglinde: pong Jul 02 20:53:55 hm I asked bluelightning some time ago Jul 02 20:54:22 how can trigger external module rebuilding when I upgrade the git revision of some kernel Jul 02 20:55:16 hmmmm.. not sure what the best approach is there offhand Jul 02 20:56:11 and it must work in dora and master Jul 02 20:56:39 if do not retrigger the build the modules will not match the kernel **** ENDING LOGGING AT Fri Jul 03 02:59:58 2015