Re: interesting timings for socket streams

[Sam S. <sd...@gn...>]

Don Cohen wrote:
> 
> (POSIX:BOGOMIPS) = 400 (that's odd, it's a pentium 200 MHz)

bogomips is highly unreliable.
I suspect that at least part of the problem is that the bogomips code
is compiled with -O2 and it should be compiled with -O0.
I suspect that there is #pragma to change that, but I am not sure how
portable it is...

> Of course, if I could read the number of bytes available
> or the number of places available for writing then this overhead could
> be avoided.  For now I don't expect that to be a big problem.

IIUC, there is no way to get this information.
Again, as I said, prove me wrong (by supplying a system call which does
this)
and we will add this functionality to CLISP.

I hope Bruno will be more helpful than me.