This is just a wild guess, but the avahi service listing from the sprint and Adam's and Marius's logs show that they're running in a dual stack environment. The avahi-browse dump shows the same queues being advertised on both IPv4 and IPv6 with the same mDNS service and host names.
Is it possible that cups-browsed is confusing these two advertisements, and leaving some job around that keeps re-queuing the idle function?
@Till: if you can't reproduce on your home network, is that network dual stack?
For others: does disabling IPv6 (or v4) locally and restarting cups-browsed make it behave differently?
This is just a wild guess, but the avahi service listing from the sprint and Adam's and Marius's logs show that they're running in a dual stack environment. The avahi-browse dump shows the same queues being advertised on both IPv4 and IPv6 with the same mDNS service and host names.
Is it possible that cups-browsed is confusing these two advertisements, and leaving some job around that keeps re-queuing the idle function?
@Till: if you can't reproduce on your home network, is that network dual stack?
For others: does disabling IPv6 (or v4) locally and restarting cups-browsed make it behave differently?