Christophe Rhodes <csr21@...> writes:
> The attached patch fixes the problem, which seems to have been a relic
> of the old calling convention (changing the return address instead of
> using the carry flag to indicate multiple values). On the other hand,
> it removes the (presumed) optimization provided by jmp-short. Can
> x86-64 experts please comment?
I've convinced myself that jmp-short was an artifact of the previous
calling convention on x86 and x86-64, and so I've deleted it and
merged the patch I sent as sbcl-18.104.22.168.