[parenscript-devel] Returning (values)

[Vladimir Sedach]

Hi Robert,

(values) looks weird when you return it, because it gets picked up by
the multiple-value return functionality. If you do (ps (lambda
()(side-effect-call-1)(side-effect-call-2)(side-effect-call-3) nil))

the output will look like:

"(function () { sideEffectCall1(); sideEffectCall2();
sideEffectCall3(); return null; });"

There are two things I can see that Parenscript can do to handle
situations like yours in a more graceful way:

(return (values)) => return null;

(lambda () (foobar) undefined)) => (function () { foobar; })

If nobody else does this in the meantime, I'll make these
optimizations when I get around to working on the backlog of
Parenscript tasks I have.

Happy hacking,
Vladimir

On Thu, Apr 4, 2013 at 1:00 AM, Robert Monfera <monfera.robert at gmail.com> wrote:
> Hi Guys,
>
> (ps (lambda ()(side-effect-call-1)(side-effect-call-2)(side-effect-call-3)))
> "(function () {
>     sideEffectCall1();
>     sideEffectCall2();
>     return sideEffectCall3();
> });"
>
> (ps (lambda ()(side-effect-call-1)(side-effect-call-2)(side-effect-call-3)(values)))
> "(function () {
>     sideEffectCall1();
>     sideEffectCall2();
>     sideEffectCall3();
>     __PS_MV_REG = {};
>     return null;
> });"
>
> What I naively expected is this:
>
> "(function () {
>     sideEffectCall1();
>     sideEffectCall2();
>     sideEffectCall3();
> });"
>
> The value of the last subexpression of the lambda expression gets
> returned by the return statement, which is in line with the intent of
> the lisp expression. However, the translated version of functions
> which exist purely for their side effect look weird. I tried to use
> (values) as the last subexpression, but then it will explicitly return
> null and set a PS flag in the translated code. A quick search in the
> doc or the net didn't reveal a good solution. Isn't it desirable to be
> able to generate lambda expressions that just don't return anything
> such that it avoids the clutter of additional lines, or am I missing
> something
> obvious? Maybe it's to do with the lisp semantics of returning multiple
> values (including no values), causing (values) not map well to the
> intuitive version?
>
> Thank you,
>
> Robert
>
> _______________________________________________
> parenscript-devel mailing list
> parenscript-devel at common-lisp.net
> http://lists.common-lisp.net/cgi-bin/mailman/listinfo/parenscript-devel