On Fri, Nov 16, 2012 at 07:59:29AM -0800, Kelly Clowers wrote: > On Fri, Nov 16, 2012 at 7:36 AM, Jon Dowland <jmtd@debian.org> wrote: > > The trouble is that testing is very variable: pretty good from freeze onwards, > > terrible immediately after freeze, etc. - not consistent. > > Neither is Sid, it does the same thing. It cleans up faster, but fair point: it does break at exactly the same times.