I wrote:
> But FWIW, when it comes to KDE in particular, the whole thing is moot or
> soon to be moot anyway because parts of KDE are now being redefined as
> "critical path", resulting in even more annoying update policies, even
> though there was clear consensus in KDE SIG that such policies are neither
> necessary for nor of any benefit to KDE. FESCo just asked us to come up
> with a list of critical KDE packages and shut up. So we did. (My proposal
> to submit an empty list was voted down in KDE SIG on the grounds of being
> against the spirit of what FESCo asked of us, even though it did get some
> support due to our objections to the critical path process as a whole.) We
> (KDE SIG) have been more or less forced to participate in a process most
> of us (and me in particular) do not agree with and consider outright
> harmful.
PS:
1. The critical path update rules (and thus also the clause in the general
update rules which references them) were initially defined as requiring only
1 proventester to approve. (This was left somewhat vague in the actual
policy, but 1 proventester was what was mentioned in all the discussion
inside FESCo.) This was modified to 1 proventester + 1 other tester to match
existing practice for freezes (the Critical Path Policy implemented as part
of No Frozen Rawhide). FESCo never actually voted to approve that change, it
was single-handedly made in the wiki by one person. This makes this policy
much more of a PITA than it could have been. It also shows that we aren't
even trusting PROVEN testers to reliably test a package! This is really
ridiculous!
2. FESCo also rejected an amendment I suggested to make sure that the
proventesters group should include at least one member of each of the main 4
desktops' SIGs. And in fact, no KDE SIG member was included in the initial
proventesters seed, despite Rex Dieter:
(i) having applied WEEKS before the proventesters group was seeded and
(ii) having YEARS of experience with approving freeze overrides, as he had
been processing freeze override requests all over the years in the old rel-
eng-Trac-ticket-based process.
This really hurts the abilities of SIGs to self-organize, instead promoting
a kind of centralized power distribution which just does not scale to our
evergrowing distribution. If you want KDE to be considered critical path,
you also have to allow KDE people to approve critical path packages. (In
fact, I think we actually need much more than one KDE proventester in the
long run.) And likewise for XFCE and LXDE.
IMHO, FESCo should be abolished, Fedora needs to be ruled by the SIGs!
Kevin Kofler
--
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/devel
