Automated removal of RC buggy packages
>>>>> "Moritz" == Moritz Mühlenhoff <jmm@inutil.org> writes:
Moritz> Scott Kitterman <debian@kitterman.com> schrieb:
>> One maintainer doesn't get to block the removal of an entire
>> stack like Qt4. I think there's a reasonable point of discussion
>> about when RoQA is appropriate, but there comes a time when stuff
>> just has to go. That doesn't make it a free for all, but for
>> old, unsupported libs we should have a bias towards action.
Moritz> We should even work towards automating this further; if a
Moritz> package is RC-buggy for longer than say a year (with some
Moritz> select exceptions) it should just get auto-removed from the
Moritz> archive.
if you do this, please standardize a way for a maintainer to flag that
they don't want a version package in testing.
I had a package that was "RC buggy" for a couple of years because I
didn't have confidence in the stability of the over the wire protocol.
Similarly I've had RC buggy packages because I wasn't confident in my
ability to provide support across a stable release.
Yes, this could have been done with a block at the release team.
But I liked that I retained control as a maintainer and that when I was
ready I could just close that bug without asking anyone.
I know the RT is responsive, but those sort of little things like
letting people solve their own problems when it's possible do matter for
keeping Debian fun.
--Sam
Reply to: