Losing package maintainers (Re: [ACTION REQUIRED] Retiring packages for F-17)
kevin.kofler at chello.at
Sat Jan 14 17:28:15 UTC 2012
Michael Schwendt wrote:
> However, with the current features of pkgdb, each member of such a group
> would need to "subscribe to" the package in pkgdb. Not just for "commit"
> access, but also for someone to monitor bugzilla and the package-owner
> mail alias, which is convenient for team-work, too.
That's exactly why we need proper support for group ownership in pkgdb. In
particular, a new developer joining a SIG should AUTOMATICALLY get write
access to all the packages (co)maintained by that SIG.
We also need a policy that SIGs AUTOMATICALLY, and WITHOUT an option for the
primary maintainer to opt out, get comaintainership of the class of packages
they're experts for, e.g. the Perl SIG for perl-*, the KDE SIG for anything
based on the KDE Platform (or even just on Qt as long as there's no separate
Qt SIG) etc. A situation as we had recently where FESCo overturned a mass
change to give a new Perl developer at Red Hat commit access to all Perl
packages just makes no sense whatsoever. A new Perl SIG developer MUST get
write access to all Perl packages; if that's against policy, it just means
the policy is broken and needs to be fixed.
And if there's no primary owner anymore, the ownership should just go to the
SIG by default.
More information about the devel