Paul Wain wrote:

> Im sure that at one point both Pete and I sent our similar but slightly
> different pthread fixes to Mark for integration (there is a bug in my
> version in a macro that as it transpires is never used but just so people
> are aware and all that!). If Mark doesnt still have them I can sent mine to
> him on Monday morning PST.

The 'bugged macro' isn't used internally by pthreads, but could be used
by external programs, eg it would break Postgres when built with the
POSIX threads option. At atomic test-and-set that isn't atomic
*wouldn't* be a nice bug to track down ;)

> Note however that I cant send the whole of NCIs pthreads tree to people
> since it contains enhancements but the arm32 machine dependant files can be
> made avaliable. These are actually fairly trival files to write!

That would be really useful. In return for the generosity I can send
back a fixed tas :-)

	Cheers and thanks,

	Neil

-- 
Neil A. Carson