Hi all,
wanted to give you all a status update as to where we stand with pungi and
Fedora 23. the current blocker is
https://bugzilla.redhat.com/show_bug.cgi?id=1208092 there is a bug in mock
where it forcibly unmounts anything that is mounted in the chroot. I have
proposed it as a blocker for F23 Alpha.
leading up to this we hit an issue where the mounting was not happening
correctly. it turns out that when you pass os.path.join two variables the
first one is silently dropped. not sure if this is a bug in python or the way
that os.path.join is supposed to work. we have a workaround deployed in the
runroot plugin.
Right now we are working in a non optional fashion but it is one that works
with pungi.
http://koji.fedoraproject.org/koji/taskinfo?taskID=10328143 is a
failed runroot task if you look in the mock log you can see the mount being
forcibly unmounted.
we could force the use of bind mounts only in runroot and add the mounts to
the mock config, which would be an additional step to take and one that goes
against what was intended and limits the options unless we then mount them
elsewhere and bind mount. it is kind of messy.
We have branching on Tuesday and are just over two weeks from Alpha change
freeze, our window to get pungi working is rapidly shrinking. we do have other
deliverables to sort out yet also
Regards
Dennis