In May last year, we made
good progress with the bug backlog. This was before the systemd-230
release. In November, we were preparing for systemd-232, a we made a
nice dip too. After four more months of development, systemd-233 was
released. I was a bit disappointed to see that this time we didn't
make such visible progress:
We did close a lot of issues, but we also opened many new ones.
If you look closely at the right edge of the graph, the slope of
the total number of issues is fairly low in Janary 2017 (many
people on vacation or recovering from the x-mas break), but suddenly
accelerates in February 2017 (we are opening bugs, but closing them
at a maching rate).
My hypothesis is that this is caused by an increase in testing: CI on
Ubuntu Zesty (autopkg tests, thanks to Martin Pitt), in addition to
Ubuntu Xenial (autopkg tests), Ubuntu Trusty (Semaphore), CentOS
(Jenkins, thanks to Daniel Mack), and an increase of architectures
that get regular use: ppc64-be, ppc64-le, arm64 (now primary
architectures in Fedora), s390x (Ubuntu autopkg).
Another aspect is the accumulation of RFE
issues. Not bugs — rather a sign that people use systemd and wish
it could do more.
Pull request status
Finally, pull requests. We keep going burning through them at at a
steady clip. Somewhere before the end of 2015 github stopped losing
comments on patches which have been force-pushed over, and we could
stop opening duplicate PRs. The rate is approximately constant since
then, with a slight uptick in the last month:
There are comments.