who is Petr Pisar from redhat ?

[Adam Williamson]

On Fri, 2010-07-02 at 21:58 +1000, Dave Airlie wrote:
> On Fri, 2010-07-02 at 13:55 +0200, Léon Keijser wrote:
> > On Fri, 2010-07-02 at 11:34 +0200, Michael Schwendt wrote: 
> > > On Fri, 02 Jul 2010 13:07:45 +0530, Rahul wrote:
> > > 
> > > > IMO,  there is absolutely nothing wrong with anyone
> > > > with commit access updating packages in Rawhide 
> > > 
> > > Of course there is. There ought to be prior communication about such plans
> > > to upgrade a package. The primary package maintainer may have good reasons
> > > for not upgrading the package. Just ask!
> > 
> > I think you are spot-on with this comment. There probably wouldn't be
> > such a discussion if the (proven)packager simply stated his intentions
> > to the package maintainer.
> > 
> 
> That doesn't really scale, on a single package basis yeah maybe, if I
> have to bump 10 or 15 packages after a mass rebuild it get ugly quick.
> 
> You get mails from CVS, its all in version control, if you disagree with
> what they did, back it out, state in the spec  what they did wrong.

Or you can act pre-emptively. Mandriva's opensync packages have (still
had, actually) a paragraph-long comment at the start explicitly
explaining why no-one should 'helpfully' bump them to the 'newer'
version available upstream, as this newer version in fact is completely
useless. I added this after a similar situation happened (someone else
'helpfully' bumped the packages, which I was intentionally leaving at
the older version).

We can argue over who should have the burden to notify whom and so
forth, but practically speaking, whether you think you 'ought' to or
not, if you don't want other people touching 'your' packages, there's
ways you can help this happen.

This is better than trying to fix the change retrospectively; it's quite
hard to revert a package after someone bumps it, and tends to involve
the use of an Epoch (urgh).
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net