1091 2011-05-20 04:32:41 +0000 Adding .bbappend should invalidate cache 2014-01-28 12:06:39 +0000 1 1 7 Build System, Metadata & Runtime BitBake bitbake 1.0 x86 Multiple RESOLVED FIXED Medium normal 1.1 M2 1 gary richard.purdie jessica.zhang poky.bs.watcher poky.watcher sgw tomw --- --- No (bug/feature does not impact docs) oldest_to_newest 13928 0 gary 2011-05-20 04:32:41 +0000 Here's the scenario - one builds a package, then finds that it's inadequate for the [local] requirements. A change needs to be made, but not one which warrants changing the global recipe. So, a .bbappend file is created in the local layer, affecting the change. Sadly, in this case, the .bbappend file will not be processed because the recipe has already been parsed and cached and only that data will be used. Somehow, the cache logic should take into account all parts of the recipe and if that changes, e.g. a .bbappend gets added, the cache is invalid. There might be some more elegant way to solve this, but my experience is that forcing the cache to be rebuilt solves the problem and allows the .bbappend file to operate. 14647 1 sgw 2011-07-08 15:34:52 +0000 http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=159aee22dc7d8c78ca2920dfc58d8bf9238ef720 39640 2 tomw 2013-12-31 16:53:11 +0000 This seems to be another consequence of this bug: when a new bbappend recipe is added to a working image (rootfs has been previously generated), the associated RPM is rebuilt, but the rebuild of the rootfs is not triggered (you have to do a bitbake -c clean <image> before the change is propagated from the RPM into the rootfs). Here is an example, build an image using ext3 as the output and make the system rpm based. Then, add a "<machine>/meta/recipes/recipes-core/netbase/netbase_5.1.bbappend" containing the following: FILESEXTRAPATHS_prepend := "${THISDIR}/files:" PRINC := "${@int(PRINC) + 3}" Add a hosts file into "<machine>/meta/recipes/recipes-core/netbase/files/" and build the image (e.g. bitbake core-image-*). The bbappend is processed, an RPM with the correct contents is produced, but the rootfs image is not updated. Subsequent attempts to build the image will ignore the new change until a clean is done of the rootfs image. 40059 3 richard.purdie 2014-01-28 12:06:39 +0000 I'm pretty sure you're running into http://git.yoctoproject.org/cgit.cgi/poky/commit/meta/lib/oe/sstatesig.py?id=80d522a5bc694fdb8b8a53ad4aa36410f3612c6b and this has been resolved. In future please open new bugs for new issues like this, its a different issue to the one described here which was fixed 2.5 years ago.