Wrapping a body in without-interrupts is useless when multiprocessing. Casual use of CLX with implementations other than SBCL (which used a lock) proved to exercise a race condition where the operation threaded-nconc looped over the buffer forever, because circular list was created.
Also we use CAS operation on implementations which implement it.
Wrapping a body in without-interrupts is useless when multiprocessing. Casual use of CLX with implementations other than SBCL (which used a lock) proved to exercise a race condition where the operation threaded-nconc looped over the buffer forever, because circular list was created.
Also we use CAS operation on implementations which implement it.