Stepping down from KDE SIG
kevin.kofler at chello.at
Thu Oct 29 16:00:39 UTC 2015
in recent times, I have been increasingly unable and unwilling to adequately
(co)maintain the huge set of packages known as the KDE Software Compilation
(i.e., the union of the KDE Plasma Workspaces, KDE Frameworks and KDE
Applications release sets), which I will refer to as "KDE SC" in this mail.
* unable because:
- the set of packages keeps growing and growing (mainly due to splitting of
existing packages). The KDE SC used to be about a dozen packages. When a new
release came out, it could be updated to manually (without any scripts) by
one person in about a day. These days, we are talking about hundreds of
packages. (I am supposedly comaintaining over 300 packages at the moment.)
Those packages are impossible to deal with without scripts, and even those
scripts take hours to run (and require some amount of babysitting because
things such as rebasing patches cannot always be automated). The same
phenomenon also affects Qt, at a slightly smaller scale. Back in the day, I
felt confident about having a global view on the KDE SC packages. This is no
longer the case.
- I am having an increasingly hard time keeping up with the latest evolutions
that the KDE Project releases. Right now, I am still running Plasma 4 and am
highly unfamiliar with Plasma 5. This means I am out of touch with the
issues our users are reporting and also have only limited ability to test
- big parts of the KDE Plasma Workspaces are now written in QML, a language I
am also not very familiar with.
- for varying reasons (job, etc.), I have found myself having nowhere near
enough time to dedicate to KDE SC packaging lately. I am drowning under
bugmail, and have found myself unable to even READ all of it at times, let
alone act on it. The poor quality of KMail 2 (see below) is not helping, and
neither will the F23 release that accidentally defaults to ABRT instead of
DrKonqi (see below). I am basically the only true volunteer left in the SIG,
and the time I can invest in Fedora packaging is finite.
* unwilling because:
- the way the Fedora Project has been treating KDE since Fedora 21 (when
"Fedora.Next" was introduced) makes me feel like a second-class citizen
in the Fedora community. After years of fighting for equal treatment of KDE
in Fedora, Fedora.Next with its "Fedora is now more focused" (on GNOME)
message was a major setback and a huge disappointment. (Another symptom of
this evolution is how the PackageKit backend was rewritten with only the
exact feature set GNOME Software happens to need, leaving Apper utterly
- the evergrowing package set (see above) is making it increasingly painful
and boring to maintain KDE SC packages.
- the Akonadi/KMail stack has been a huge pain to use, due to major serious
issues, both performance and reliability issues. Upstream has been either
unwilling or unable to do anything about the problem, none of the fixes so
far really helped in any way. I am unwilling to have anything further to do
with this abomination of software.
- while I have not experienced it personally yet, the quality of KDE Plasma 5
is also reported to be very disappointing; the promise from KDE 4.0 times
that "KDE 5" (as it was referred to at the time) would be an evolutionary
rather than a revolutionary release was not followed through.
- the Fedora 23 KDE Spin (which is now final or almost final) is easily the
worst KDE Spin we have ever released:
. Firefox, a non-KDE and even non-Qt application, is the default browser. It
does not integrate into the Plasma desktop in any way.
. Due to an oversight, DrKonqi is missing, and thus ABRT (another non-KDE
application) is the default crash handler. ABRT reports all the crashes to
us downstream packagers instead of upstream where the crash reports
belong. And experience has shown that the ABRT fire&forget reporters are
unwilling to upstream the bug reports manually, if they even read our
replies at all (and that's if we even manage to reply to all the bugs to
Both of these are complete no-gos where I have said from day one that those
are not acceptable. The Firefox fiasco was a deliberate decision, the ABRT
fiasco could have been avoided if DrKonqi had not been made optional against
my recommendation, and/or if people had actually tested that the KDE Spin
uses the KDE crash handler. I do not want to be held responsible for these
decisions I did not approve of.
As a result, I AM HEREBY STEPPING DOWN FROM THE KDE SIG AND FROM
(CO)MAINTAINERSHIP OF KDE SC PACKAGES.
In particular, effective NOW:
* I hereby request to be removed from the group::kde-sig group.
* I will withdraw my comaintainership (including watchbugzilla and
watchcommits!) of Qt and of all packages in the KDE SC, EXCEPT:
- the packages for which I am upstream: kompare, libkomparediff2
- compatibility packages: qt3, kdelibs3, qt(4), kdelibs(4), kdewebdev (3),
* I may also withdraw comaintainership of, or orphan where I am the point of
contact, any packages that I am maintaining because the KDE SC depends on it,
and possibly selected other KDE-related packages.
* I am stepping down from being a voting member in the KDE SIG. I leave it to
the remaining members to nominate a replacement or reduce the member count.
* I am stepping down as a moderator of the fedora-kde mailing list. I already
cleared the moderation flag of the one person I have put on moderation. This
list moderation is an additional duty that I just do not have time for.
* I am stepping down as an operator of the #fedora-kde IRC chan. (You should
also remove the operator privileges from tigcc_bot. It does not need them, it
was only symbolic.)
* I will no longer consider it my duty to attend KDE SIG meetings, though I may
attend some meetings as an interested user. In particular, I will no longer
lead any meetings and I will not try to herd people into attending (the
necessity of which has always been annoying me).
I will KEEP (until further notice):
* my Fedora packager status,
* my provenpackager privileges in Fedora,
* my packager sponsor privileges in Fedora,
* (co)maintainership of a reasonable number of packages that I can actually
scale to (i.e., NOT the current 300-400), including:
- as mentioned above, the Kompare stack and the compatibility packages,
- KDE-related packages outside of the KDE SC (though possibly fewer than now),
- the Calamares installer,
- some non-KDE packages,
* taking care of the compatibility libraries that make existing applications
work (the Qt3/kdelibs3 stack right now, the Qt4/kdelibs4 stack if and when
needed) and of legacy applications themselves,
* developing Kannolo (what the Fedora KDE Spin SHOULD be), to which I hope being
able to devote more time than now,
* maintainership of Kompare upstream (for now), unless another maintainer with
more time can finally be found,
* using and promoting (to the current extent) Fedora with KDE (be it the KDE
Spin or Kannolo).
I am sorry, but after 8 years of volunteering and ending up with almost 400
packages to take care of, I just have to scale down. It was fun while it lasted.
I wish the remaining KDE SIG members good luck for the future.
More information about the kde