FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

Patrice Dumas pertusus at free.fr
Fri Feb 26 19:10:58 UTC 2010

On Fri, Feb 26, 2010 at 12:14:41PM -0500, Bill Nottingham wrote:
> > 
> > Bringinig down productivity of good packagers for a few bad ones, is,
> > in my opinion, not a good move. 
> Fedora doesn't exist for the productivity of packagers. It exists for
> the productivity of our users. 

Both are related (except of course for people paid to work on

> To put it a different way, a large regression for our users far outweighs
> tha cost of any number (heck, even hundreds) of bugfixes having to wait
> a day, or two, or even a week.

I fully agree with that. But pushing to stable rapidly may help 
correct rapidly regerssions, too.

> "No update is pushed to users without verification and testing from entities
> other than the packager."

That works for known packages with a lot of users, but for specialized
packages, it just doesn't make sense. Leaving it to the packager in that
case is better.

Maybe something that would work would be to flag especially packages
that are important, especially when other packages depend on them,
and have sufficient users/co-maintainers such that having verification
isn't a waste of time. Something like 'core' packages, but with a 
definition based on their coreness, not their maintainer. For example
the definition could be everything in Core + Base + Basic X + base gnome
and KDE + ... + their dependencies.


More information about the devel mailing list