Rethinking proventester and critpath

[Nils Philippsen]

On Tue, 2011-11-22 at 07:42 -0700, Ken Dreyer wrote:
> On Tue, Nov 22, 2011 at 6:06 AM, Nils Philippsen <nils at redhat.com> wrote:
> > Hm. The list of (implicitly labeled) critpath packages seems to have
> > proliferated recently
> 
> Why not impose a self-restriction on the number of critpath packages?
> Make a rule like "The ratio of proventesters to critpath packages must
> be x or higher". When the critpath numbers start creeping up, drop a
> few other packages from critpath. This will help focus testing
> efforts, and allow it to expand along with the growth of the
> community.

But that wouldn't really help, would it? I assume packages are
explicitly labeled critpath for a reason: if they break, the system
might not boot up, or users might not be able to log in (or several
other reasons).

I think the current system is a bit too simplistic when it comes to the
packages upon which critpath components depend, however. For instance
(picking this example because it affects me), I don't think that logging
in should fail because of problems the scanner library may have. Or
because (drawing examples from a hat) the desktop background can't be
set, or there's an issue with the camera in your laptop. Yet, by way of
simply following package dependencies, we do as if that were the case.

Nils
-- 
Nils Philippsen      "Those who would give up Essential Liberty to purchase 
Red Hat               a little Temporary Safety, deserve neither Liberty
nils at redhat.com       nor Safety."  --  Benjamin Franklin, 1759
PGP fingerprint:      C4A8 9474 5C4C ADE3 2B8F  656D 47D8 9B65 6951 3011