Re: [Kawa-commonlisp-dev] tagbodies

[Charles Turner]

On 26 July 2012 20:13, Helmut Eller <address@hidden> wrote:
> State machines can be written with LABELS and (local) tail-calls.  Kawa
> should recognize those and implement them with gotos.  E.g.

Ah yes, forgot about tail-calls :-)

> I think implementing TAGBODY now is a bit of a distraction and messaging
> the source into LABELS isn't that hard.
>> How hard would it be to implement TAGBODY?  Seems like I could just
>> use BlockExp to do it.
>
> In general you need to do a modicum of escape analysis before you can
> turn it into BlockExp.

I was just curious about it's implementation. Thanks for the pointers!

> So what still on the TODO stack?

When I started with destructuring I immediately pushed lambdas
parameters on my stack, then I ran into the whole

((lambda (&optional x &rest y) ...)

unleashing the hounds if and only if &optional is not proceed by an
ordinary var. Still need to fix this one, but mercifully SBCL doesn't
seem to use this in it's implementation :-) so I postponed it. But now
I've been told several times to just get on with the workarounds
(sorry for the slowness!) several items popped of the stack.

> You've already done a lot to make &optional/&key work for normal lambdas.
> Do you have some test cases for that?  Should it be reviewed?

The code for that is in OrdinaryLambda, which is (for now, sorry
future people) online at
http://chturner.com/kawa/trunk/kawa/lang/OrdinaryLambda.java if you
don't have my repository locally. It's still brittle (for instance, it
currently only supports 10 &optional/&key parameters because I haven't
added a dynamic array yet). I'm afraid I haven't written any tests for
it yet.