**** BEGIN LOGGING AT Thu Jul 14 02:59:57 2011 Jul 14 09:24:02 hi bluelightning Jul 14 09:24:14 hi ant_work Jul 14 09:25:02 I think there is an issue in oe-core wrt undeleted files in DEPLOY_DIR after do_clean and do_cleansstate Jul 14 09:25:29 i.e. the initramfs cpio's ar notremoved Jul 14 09:26:12 that's why linux-kexecboot built even with broken kernel.bbclass Jul 14 09:27:00 I copied form somewhere (JaMa?) the 'inherit deploy' stuff for replacing package_stagefile_shell Jul 14 09:29:37 the files deploying directly there, as updater.sh, can be removed by do_clean iirc Jul 14 09:50:31 bluelightning: ah Jul 14 09:50:38 seems bitbake standard behavior Jul 14 09:50:43 clean Jul 14 09:50:45 cleans the package (tmp/work). Does not touch deploy dir (this has to be done manually). Jul 14 09:51:47 hmm, I wonder if it's worth looking at changing however Jul 14 09:51:54 seems inconsistent Jul 14 09:52:10 i.e. bitbake -c clean virtual/kernel did remove the zImages in the past Jul 14 09:52:21 from DEPLOY_DIR_IMAGE Jul 14 09:52:24 I'm sure Jul 14 09:53:09 maybe a fix after the Poky fork? Jul 14 09:53:32 I'll check again in oe-dev if you mind, but I'm pretty sure Jul 14 09:53:57 my workflow was always the same in oe-dev Jul 14 09:54:20 bitbake -c clean kexecboot-klibc initramfs-kexecboot-klibc-image linux-kexecboot Jul 14 09:54:28 bitbake linux-kexecboot Jul 14 09:54:43 this assured a clean build (like from scratch) Jul 14 09:56:57 hmm, well if oe-dev does it then we should definitely bring across the patch that ensures it does it Jul 14 10:05:28 the task should be managed by bitbake code, isn't? Jul 14 10:06:48 I'm not entirely sure Jul 14 10:27:10 offhand I see one difference in kernel.bbclass Jul 14 10:27:46 do_deploy[dirs] = "${DEPLOYDIR} ${B}" vs do_deploy[dirs] = "${S}" but the files are indeed different Jul 14 11:10:49 image.bbclass differs too **** ENDING LOGGING AT Fri Jul 15 02:59:57 2011