A lot of obsolete rpms in FE currently

[Jeff Spaleta]

On 10/18/05, Greg DeKoenigsberg <gdk at redhat.com> wrote:
> Would this constitute a sensible policy?

Sure exactly one old version. I'd even be happy if it were "exactly
one old version that is at most X number of days old."  X could be a
week or two for my concerns. I basically want a fallback for
forehead-slapping bad bugs that slipped through maintainer testing. If
something isn't reported in the first week of an update
release..well..maybe its just a subtle bug and we just learn to live
with it.

Any case that would require an old package several revisions old would
indicate that a maintainer felt the need to publish several buggy
revisions of a package in short order. And if we start seeing behavior
like that..then it would be better to have a scratch area like an
extras-testing for maintainers to publish.

-jef