Re: libthreads provenance

Roland McGrath <[email protected]> writes:

> Uncle Thomas, can you tell us a story?  I want to hear the one about where
> the cthreads code came from and what its name and personal identification
> number were in its native land of CMU.  

It was its own thing that the CMU CS department cooked up, and existed
independently of Mach.  The cproc layer, if you'll look, implements a
co-routine non-preemptively scheduled threads library--that works even
on Unix.  This was then ported and layered on top of Mach threads.

> It sure looks like hurd/libthreads started with a CMU version older than
> MK83.  I would ideally like to merge it all together by applying the GNU
> changes to the MK83 or MK83a code.  But for that I need to find the
> pristine CMU version that the hurd version forked from.  Otherwise it will
> just be purely by hand and that will make me more grumpy.

I think probably MK82 is where I got it from, but I really can't say
for sure.

Thomas