[R6RS] record-constructor suggestion

[dyb at cs.indiana.edu]

> > I think I prefer to leave the door open for this change in the
> > future by requiring the rtd argument for record-constructor, but I'm
> > willing to go either way.
>
> OK, but if there's still leeway even after the change, is this change
> really worth it?  Few programmers are going to type calls to these
> procedures anyway.

Yes.  I think it will be easier to change the lesser-used
make-record-constructor-descriptor procedure than the record-constructor
descriptor if we decide to do so down the road.

But in keeping with the use of #f to signify a default
constructor-descriptor in other contexts, I think we should use that
convention for record-constructor as well, so record-constructor takes two
arguments:  rtd and record-constructor, where record-constructor may be
#f.  I'll go ahead and make this change unless I hear an objection.

Kent