On the stability of current 2.5

I am very pleased. The current 2.5 branch appears to me very stable.
I've had a setup running for about 2 weeks now non-stop.
So far it handled over 41 million HTTP hits, over 5 million ICP reqs, and
despite some noticeable heap fragmentation (about 25% free heap space
as reported by mallinfo()) apparently it leaked no memory, running
NTLM auth via NTLMSSP without helper-fail-open (almost 6 million
authentication
operations).

All in all, I'd consider this a pretty good real-life test environment.
The state of aufs is pretty unclear (when I tested it I noticed 100% CPU
usage
in the main thread under very moderate loads), but diskd is fine.

IIRC today was slated as release date for STABLE1, but all is silent.
What is the latest news on that?

-- 
	/kinkie