Proposed udpates policy change

[Adam Williamson]

On Mon, 2010-03-08 at 22:09 +0000, Richard W.M. Jones wrote:
> On Mon, Mar 08, 2010 at 09:59:29PM +0000, Matthew Garrett wrote:
> > We assume the following axioms:
> [..]
> > 2) It is impossible to ensure that functionality will not be reduced 
> > without sufficient testing.
> 
> Your axioms are obviously wrong.  An update which simply bumped a
> release number would have the same functionality.  Since you claim

It would be rebuilt from source, as all updates are. The build would not
be performed in the exact same circumstances as the original build. We'd
certainly hope that wouldn't change anything, but it's impossible to be
certain.

Your post is clearly nitpicking, but this is an important point; I've
certainly seen cases when I was at Mandriva where a rebuild didn't work,
whereas the initial build still did; various odd issues, such as build
environment changes, can cause this. 

Actually, it happened to me once, if I recall correctly, with Kompozer.
The initial build I'd done a few months previously for a given Mandriva
release worked in Cooker (Rawhide equivalent). If you took the exact
same SRPM and rebuilt it with the then-current Cooker environment - even
though the old binary package still worked in that environment! - the
resulting binary package did not work at all.

Which only goes to illustrate that updates are tricky and *really,
really* anything can go wrong.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net