Worthless updates

Doug Ledford dledford at redhat.com
Wed Mar 3 21:42:27 UTC 2010


On 03/03/2010 02:36 PM, Kevin Kofler wrote:
> Toshio Kuratomi wrote:
>> It is a reason but it's not the only reason. Semi-rolling releases allow
>> a subset of the entire packager community to work on an update as a set
>> and then push them when they're known to work together.  Currently rawhide
>> is not so coherent.
>>
>> We could change rawhide from a pure rolling to a semi-rolling model but
>> then would we need to have a rawerhide?
> 
> Our stable releases are already "semi-rolling" (at least in some sense), why 
> can't we just keep things the way they are?

We have no clear policy on whether updates in F(n), F(n-1), or any other
stream should be rolling updates or limited updates.  As a result, there
is a large amount of inconsistency in the distribution itself.  We have
some users complaining that rolling updates are bad, other claiming it's
a selling point.  The truth of the matter is that since we aren't
consistent, it is both and neither all at the same time.

So, at a very minimum, we need to settle on a policy, enact it, enforce
it, and make the distribution either semi rolling or stable.  A random
mix and match of the two is the worst possible way to leave things now
that the issue has been broached.  Users that want rolling updates can't
actually expect it across the board, and likewise users that want a
stable system can't expect it across the board, so we are trying to
satisfy two groups and actually satisfying none.

That's why things can't just stay the way they are.

-- 
Doug Ledford <dledford at redhat.com>
              GPG KeyID: CFBFF194
	      http://people.redhat.com/dledford

Infiniband specific RPMs available at
	      http://people.redhat.com/dledford/Infiniband

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: OpenPGP digital signature
Url : http://lists.fedoraproject.org/pipermail/devel/attachments/20100303/66283244/attachment-0001.bin 


More information about the devel mailing list