While building qemux86-64 and couple of armv7a and armv4t machines in the same tmpdir I got strange error in do_package: NOTE: recipe packagegroup-core-x11-1.0-r40: task do_package_write_ipk: Started ERROR: Error executing a python function in /OE/shr-core/openembedded-core/meta/recipes-graphics/packagegroups/packagegroup-core-x11.bb: ExpansionError: Failure expanding variable SUMMARY, expression was ${SUMMARY} - Debugging files which triggered exception Exception: variable SUMMARY references itself! ERROR: The stack trace of python calls that resulted in this exception/failure was: ERROR: File "do_package_ipk", line 193, in <module> ERROR: ERROR: File "do_package_ipk", line 96, in do_package_ipk ERROR: ERROR: File "/OE/shr-core/bitbake/lib/bb/data_smart.py", line 293, in getVar ERROR: return self.expand(value, var) ERROR: ERROR: File "/OE/shr-core/bitbake/lib/bb/data_smart.py", line 157, in expand ERROR: return self.expandWithRefs(s, varname).value ERROR: ERROR: File "/OE/shr-core/bitbake/lib/bb/data_smart.py", line 147, in expandWithRefs ERROR: raise ExpansionError(varname, s, exc) ERROR: ERROR: The code that was being executed was: ERROR: 0189: bb.utils.unlockfile(lf) ERROR: 0190: ERROR: 0191: ERROR: 0192: ERROR: *** 0193:do_package_ipk(d) ERROR: 0194: ERROR: [From file: 'do_package_ipk', lineno: 193, function: <module>] ERROR: 0092: if localdata.getVar(f) is None: ERROR: 0093: raise KeyError(f) ERROR: 0094: # Special behavior for description... ERROR: 0095: if 'DESCRIPTION' in fs: ERROR: *** 0096: summary = localdata.getVar('SUMMARY', True) or localdata.getVar('DESCRIPTION', True) or "." ERROR: 0097: description = localdata.getVar('DESCRIPTION', True) or "." ERROR: 0098: description = textwrap.dedent(description).strip() ERROR: 0099: ctrlfile.write('Description: %s\n' % summary) ERROR: 0100: ctrlfile.write('%s\n' % textwrap.fill(description, width=74, initial_indent=' ', subsequent_indent=' ')) ERROR: [From file: 'do_package_ipk', lineno: 96, function: do_package_ipk] ERROR: Function failed: do_package_ipk ERROR: Logfile of failure stored in: /OE/shr-core/tmp-eglibc/work/all-oe-linux-gnueabi/packagegroup-core-x11-1.0-r40/temp/log.do_package_write_ipk.24073 NOTE: recipe packagegroup-core-x11-1.0-r40: task do_package_write_ipk: Failed ERROR: Task 2773 (/OE/shr-core/openembedded-core/meta/recipes-graphics/packagegroups/packagegroup-core-x11.bb, do_package_write_ipk) failed with exit code '1' While looking in stamps directory I've noticed that qemux86-64 was using different subdirectory then arm builds. packagegroup.bbclass does not inherit allarch.bbclass, only sets PACKAGE_ARCH = "all" should it set also TARGET_OS? OE nokia900@shr ~/shr-core/tmp-eglibc/stamps $ ls */packagegroup-core-x11-1* all-oe-linux-gnueabi/packagegroup-core-x11-1.0-r40.do_build all-oe-linux-gnueabi/packagegroup-core-x11-1.0-r40.do_package_setscene all-oe-linux-gnueabi/packagegroup-core-x11-1.0-r40.do_package_write all-oe-linux-gnueabi/packagegroup-core-x11-1.0-r40.do_populate_lic_setscene all-oe-linux-gnueabi/packagegroup-core-x11-1.0-r40.do_populate_sysroot_setscene.nokia900 all-oe-linux-gnueabi/packagegroup-core-x11-1.0-r40.do_populate_sysroot_setscene.om-gta02 all-oe-linux-gnueabi/packagegroup-core-x11-1.0-r40.do_populate_sysroot_setscene.om-gta04 all-oe-linux-gnueabi/packagegroup-core-x11-1.0-r40.do_populate_sysroot_setscene.tuna all-oe-linux/packagegroup-core-x11-1.0-r40.do_build all-oe-linux/packagegroup-core-x11-1.0-r40.do_compile all-oe-linux/packagegroup-core-x11-1.0-r40.do_configure all-oe-linux/packagegroup-core-x11-1.0-r40.do_fetch all-oe-linux/packagegroup-core-x11-1.0-r40.do_install all-oe-linux/packagegroup-core-x11-1.0-r40.do_package all-oe-linux/packagegroup-core-x11-1.0-r40.do_package.sigdata.63dc7a62d8e9740c34d18728830aed44 all-oe-linux/packagegroup-core-x11-1.0-r40.do_package_write all-oe-linux/packagegroup-core-x11-1.0-r40.do_package_write_ipk_setscene all-oe-linux/packagegroup-core-x11-1.0-r40.do_patch all-oe-linux/packagegroup-core-x11-1.0-r40.do_populate_lic all-oe-linux/packagegroup-core-x11-1.0-r40.do_populate_lic.sigdata.8451162b66b25e639ed440c09fbbf766 all-oe-linux/packagegroup-core-x11-1.0-r40.do_populate_sysroot_setscene.qemux86-64 all-oe-linux/packagegroup-core-x11-1.0-r40.do_unpack OE nokia900@shr ~/shr-core/tmp-eglibc/stamps $ ls */packagegroup-core-ssh-openssh* all-oe-linux-gnueabi/packagegroup-core-ssh-openssh-1.0-r1.do_build all-oe-linux-gnueabi/packagegroup-core-ssh-openssh-1.0-r1.do_package_setscene all-oe-linux-gnueabi/packagegroup-core-ssh-openssh-1.0-r1.do_package_write all-oe-linux-gnueabi/packagegroup-core-ssh-openssh-1.0-r1.do_populate_lic_setscene all-oe-linux-gnueabi/packagegroup-core-ssh-openssh-1.0-r1.do_populate_sysroot_setscene.nokia900 all-oe-linux-gnueabi/packagegroup-core-ssh-openssh-1.0-r1.do_populate_sysroot_setscene.om-gta02 all-oe-linux-gnueabi/packagegroup-core-ssh-openssh-1.0-r1.do_populate_sysroot_setscene.om-gta04 all-oe-linux-gnueabi/packagegroup-core-ssh-openssh-1.0-r1.do_populate_sysroot_setscene.tuna all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_build all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_compile all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_configure all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_fetch all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_install all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_package all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_package.sigdata.c398f583265815c38b5295be14baf99e all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_package_write all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_package_write_ipk_setscene all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_patch all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_populate_lic all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_populate_lic.sigdata.adbcefb359671185ec720b654b496707 all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_populate_sysroot_setscene.qemux86-64 all-oe-linux/packagegroup-core-ssh-openssh-1.0-r1.do_unpack Only those 2 recipes seem to have this problem (from what I'm building here)
packagegroup inherits allarch now (as of a few weeks after this bug was filed), and allarch sets TARGET_OS, so this issue should now be fixed.