Re: Thread deficiencies

[Alan Cox <alan@lxorguk.ukuu.org.uk>]

> I think George meant that clone() ought to be somehow specced
> (allowed)---even if only a subset---not that it ought to be disallowed.

Ok. He seemed to be saying we shouldn't specify clone this time, and his
previous message talked about "temporarily" specifying it.

If he means we shouldnt specify what happens if you mix clone with pthreads
then dead right. 

For LSB 1.0 we should be looking at

	clone		linux specific		specified
	posix.4		portable		specified subset

and folks who want portability should be given as close to .4 as we can get
them without harming other stuff (eg cluttering the kernel up too much).

Alan