[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: A suggestion for the woody freeze

On Thu, Jan 17, 2002 at 09:46:40PM +0100, Adrian Bunk wrote:
> On Thu, 17 Jan 2002, Steve Langasek wrote:

> > There's a subtle flaw with this proposal.  Who will have to remove the
> > packages for testing?  The only people who can do this are the ftp
> > masters.  So the success of this proposal depends on the willingness of
> > the ftp masters to pick up the slack for maintainers who aren't doing
> > their job.

> Where's the flaw?
> Every freeze needs a release manager the has write access to the ftp
> archive.

> > If we need to be NMUing more aggressively to get RC bugs fixed, then
> > let's arrange to do that.  There's no reason why ftp masters need to be
> > dragged into this to make it stick.

> The ftp masters have nothing to do with NMUs.

Precisely my point.  Fixing bugs through NMUs doesn't place any 
additional burden on the ftp masters.  A policy that says "packages will 
be removed from testing/frozen after four weeks if their RC bugs aren't 
fixed" *does* make more work for ftp masters.

> > The section in the developer's reference on source NMUs has remained
> > virtually unchanged since the release of potato.  What was considered
> > best practice then does not seem to be working all that well this time
> > around.  If the presence of RC bugs is going to be allowed to delay the
> > freeze, then we as a body need to start treating these packages as if
> > they're already frozen and NMU appropriately.
> >...

> As already said in other mails in this thread:
> The rules for NMUs aren't the problem.

I disagree.  The developer's reference says "you should almost never NMU
to unstable, but NMUs to frozen are ok."  This doesn't work if packages 
never get to frozen because they're in need of an NMU.

> The problem is that there need to be people doing the NMUs and that there
> needs to be motivation for people doing NMUs in the sense that there must
> be a visible progress in the way to a newstable release.

Then perhaps some web scripts that give the community more feedback on 
the results of NMUs would be in order...

Steve Langasek
postmodern programmer

Attachment: pgpe3KVvdeTXt.pgp
Description: PGP signature

Reply to: