Am Dienstag, den 25.06.2019, 17:52 -0400 schrieb Paul Wouters:
This was a mistake on my end. I thought I was the owner of the
package, but I think I was only the owner of it back in el6. I assume
systemd then wasn't depending on it. I saw a PR the other day, assumed
it was to me as package owner, and saw no reason to not upgrade since
it was long over due. I didn't realise this would break systemd.
Note, it is a bit worrying that systemd depends on this package, which
wasn't updated in 5 years, and for which the upstream changelog mostly
states "bug fixes".
nirik untagged the package for me. I pinged Zbyszek to coordinate
things. I've reverted the commits for f29 and f30 (no updates were
issued for these branches yet)
Ironically, this seems to not be the first time this happened either.
I see someone else made the exact same mistake in January 2018 and
reverted their changes to the package. So let's see if we can fix this
now in rawhide so it does not happen again in another year.
Paul
I've fixed the package in a way so-name bumps can be detected easily.
If bootstrapping for such a bump (the next one) is needed one can do
that now by simply editing some lines of the spec-file.
During that fix, I've rebuilt systemd and the other consumers of
qrencode against the new so-name.
Cheers
Björn