http://autobuilder.yoctoproject.org:8010/builders/cedartrail/builds/96 ------------ snip ------------- | PATH=.:$PATH /srv/home/pokybuild/yocto-autobuilder/yocto-slave/cedartrail/build/scripts/help2man --section=1 -i ./docs/man/grub-mkrescue.h2m -o grub-mkrescue.1 grub-mkrescue ... | make[2]: *** [grub-mkrescue.1] Error 1 ------------ snip ------------- help2man is being invoked directly, and our fake help2man script returns an error (1) which is intended to force a fallback to proper manpages, but for grub-efi there is apparently no such fallback, and thus the build fails. Incidentally it might be worth making our fake help2man script actually print something so we know exactly what's gone wrong.
Error with the grub-efi-native_1.99.bb of openembedded core, the description is in this link http://comments.gmane.org/gmane.linux.embedded.yocto.general/7102
If grub-efi +DEPENDS = "help2man-native" Then why does our script come into play at all? Shouldn't it find the help2man-native we built?
That DEPENDS entry should not be there. Our help2man script comes out of the box, no need to build anything.
There is a workaround in place in master: 70067f082af279b58247467ec29283b310a1575e I don't know if this is the proper fix or not.
From the description it looks like an ideal solution to me. I'm guessing though that we should probably remove help2man-native from DEPENDS.
cedartrail build ok now. removed help2man dependency as suggested, patch pending review.
Please do not mark a bug as RESOLVED until the patch is pulled into master (or the approprirate stable branch). If someone finds a flaw with your patch, this could still sit in development for a while. When the patch is pulled in, then we change to RESOLVED/FIXED and copy a link to the commit (on git.yoctoproject.org) here in bugzilla.
(In reply to comment #7) > Please do not mark a bug as RESOLVED until the patch is pulled into master > (or the approprirate stable branch). If someone finds a flaw with your > patch, this could still sit in development for a while. When the patch is > pulled in, then we change to RESOLVED/FIXED and copy a link to the commit > (on git.yoctoproject.org) here in bugzilla. I moved the bug to RESOLVED only after the patch was merged, but did not paste a link in here. The patch in master is 24340ed5d1f153610a0452e48babcadd4fc50b63
(In reply to comment #8) > (In reply to comment #7) > > Please do not mark a bug as RESOLVED until the patch is pulled into master > > (or the approprirate stable branch). If someone finds a flaw with your > > patch, this could still sit in development for a while. When the patch is > > pulled in, then we change to RESOLVED/FIXED and copy a link to the commit > > (on git.yoctoproject.org) here in bugzilla. > > I moved the bug to RESOLVED only after the patch was merged, but did not > paste a link in here. > The patch in master is 24340ed5d1f153610a0452e48babcadd4fc50b63 OK, your comment above read "patch pending review", which I took to mean it had not been merged yet. Great that it's in and this is done, Thanks!
On autobuildert this is not reproduced anymore. Setting this bug to verified for revision 8b8748c8f963900b83dc0fdd7757556f917fe4fd