new version of systemd seems to break autopkgtests of policykit-1, kwin and python-dbusmock

Michael Biebl biebl at
Sat Jun 23 12:44:28 BST 2018

Hi Paul,

you are quick!

Am 23.06.2018 um 13:30 schrieb Paul Gevers:
> Dear systemd maintainers, all,
> With the upload of version 239-1 of systemd the autopkgtests of the
> policykit-1, kwin and python-dbusmock started to failĀ¹.
> It appear that policykit-1 fails to install since the 239-1 version of
> systemd with (in the unstable test run, in the testing test run as well
> as in the python-dbusmock test run):
> Setting up policykit-1 (0.105-20) ...
> --runtime cannot be used with unmask
> dpkg: error processing package policykit-1 (--configure):
>  installed policykit-1 package post-installation script subprocess
> returned error exit status 1
> dpkg: dependency problems prevent configuration of autopkgtest-satdep:
>  autopkgtest-satdep depends on policykit-1; however:
>   Package policykit-1 is not configured yet.
> This looks very serious to me, can you, together with the maintainers of
> the failing packages (which I CC'ed), please investigate if the failure
> is really caused by systemd or if policykit-1 needs updating. In the
> latter case, I think a "Breaks" in systemd is very much justified.
> (Please let me know when this happens, I may have to help autopkgtest a
> bit due to bug #896023).

I already tracked down the reason for the policykit-1 failure and it is
indeed a regression/behaviour change in systemd. Another affected
package is avahi-daemon (which also uses systemctl unmask --runtime),
unfortunately it doesn't have an autopkgtest yet.

Yay for autopkgtests!

I've filed this upstream for now

I'll see if upstream agrees to change back to the old behaviour. If not,
we'll indeed have to update avahi-daemon and policykit-1

and add appropriate Breaks to systemd.

We already have another, unrelated RC bug in systemd which will prevent
testing migration for now.

> In the kwin case, it is currently failing in unstable for some time
> already and detected regressions in quite some packages in the KDE
> stack, but the error in this test is different:
> FAIL!  : KeyboardLayoutTest::testWindowPolicy() Compared values are not
> the same
>    Actual   (((xkb->layoutName())))

The breakage in kwin looks unrelated to me (maybe caused by the libinput
update, dunno).

Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <>

More information about the Pkg-systemd-maintainers mailing list