Netbench on HEAD code

[Jeremy Allison]

Can the person who was complaining about HEAD being only
%50 of the speed of the 2.0.x code branch please do an
update to their HEAD CVS tree and try again please.

I spent the afternoon in the SGI lab with Herb, and we
fixed a couple of nasty performance glitches that had
crept into the code (vfs port had re-introduced some
fsync calls, and the tdb tables were slow due to the default
hash size being non-prime).

On IRIX we're back within random variance levels of 2.0.x
performance using the HEAD code (more testing being done)
so I'd appreciate independent NetBench confirmation.

Cheers,

	Jeremy Allison,
	Samba Team.

-- 
--------------------------------------------------------
Buying an operating system without source is like buying
a self-assembly Space Shuttle with no instructions.
--------------------------------------------------------