> 
> Note that you never really *need* sigsetjmp/siglongjmp, since you
> can always replace with setjmp/longjmp or _setjmp/_longjmp according
> to the 'save signal mask' argument to sigsetjmp.  However, since we
> just had this on the PMAX port I've done a first cut of a VAX
> implementation.  Since I don't know anything about VAX assembly
> language, someone should check these out, but they *do* pass the
> regression test, for whatever that's worth.
> 
> I'm quite unsure about how the stack frame handling works -  my guess
> is that sigsetjmp can do the maximum save from setjmp /_setjmp
> (max[R6,0] = R6), then jump directly into the routines.
> There is probably a nicer way to express this direct jump than my
> 	jmp	_setjmp+2
> hack, too...
> 
There is nothing wrong with your code implementation, and it works
really as expected. I have added it to the source tree.

-- Ragge