Lack of update information

[Mark McLoughlin]

On Mon, 2009-01-26 at 16:07 -0800, Jesse Keating wrote:

>  But Fedora /releases/ aren't your personal rawhide.  We're providing
> releases that are supposed to stay somewhat stable, not to just be a
> dumping ground for whatever upstream chooses to drop the day before.  We
> have a developmental stream for that, and it makes releases fairly
> often.  I just don't understand why we want to treat our /release/
> branches as if they were just another rawhide.

Agree completely.

Another worrying sentiment I heard lately was that it makes sense to
push a new, untested upstream release into updates-testing so that it
actually gets testing because "no-one runs rawhide".

The idea being that if you disable the bodhi-karma-auto-push thing, then
it won't hit updates until all the wrinkles have been ironed out.

To me that's just wrong - if someone enables updates-testing, they
shouldn't be immediately exposed to new upstream release. IMHO, the
process should be:

  a) push to rawhide and let stew

  b) if there's some serious bug fix or important user feature requested
     that can't be trivially backported, then push to updates-testing

  c) push to updates

Cheers,
Mark.