On Fri, Jul 08, 2005 at 03:58:50PM +0200, Leopold Toetsch wrote:
> Chip Salzenberg wrote:
> >Thus, this is probably better:
> >
> > .sub "foo"
> > .param int beta
> > .param string gamma :optional(have_gamma)
> > .param string delta :optional(have_delta)
>
> This needs more parsing rules in imcc.y so it's not (yet) implemented.
>
> But I've now implemented :opt_count, with an extension to the original
> proposal: the opt_count is reset after :opt_count is seen, so that
> multiple :opt_count flags work.
I'll note that with Leo's resetting version of :opt_count, we might
not really need the parameterized form of :optional, since one
can simply do:
.sub "foo"
.param int beta
.param string gamma :optional
.param int have_gamma :opt_count
.param string delta :optional
.param int have_delta :opt_count
and each of have_gamma and have_delta is 0 or 1, depending on the parameter
before it. And if a subroutine only needs the total count of
optional parameters, there's an easy way to do that.
OTOH, the :optional(reg) syntax has its uses as well.
At any rate, having both available seems to give a nice level of
flexibility to subroutine authors (I haven't thought too deeply yet
about potential conflicts/confusions).
Pm
