[DGD]varargs void foo(object bar)

[Felix A. Croes]

Gregory D Lewis <glewis at maths.adelaide.edu.au> wrote:

> Dworkin wrote:
>
> > Note that 3) and 4) are always identical, from the called function's
> > point of view.  Unlike MudOS, DGD has no way to distinguish a
> > destructed object value from a zero value.
>
> I believe at one point you suggested a "null" value may be introduced in
> a future version (2.0?) of DGD?

Yes.  Note that my drive to radically clean up the LPC language and
release DGD 2.0 is not that large, considering that it would break
backward compatibility (having all those muds running DGD is a
great asset even just for testing out my code).  For 1.2 I plan
to add multiprocessor support + full LPC documentation, and I can
vaguely foresee a 1.3 with bugfixes and new platform ports, so
2.0 would be at least several years in the future.  I'd much rather
concentrate on writing a good mudlib.

Note that a non-mud project has priority now (yes, I really have
been working full time on DGD for a while.  Being self-employed is
fun!) and that I only do bugfixes at the moment.  I don't expect to
get back to honest DGD work until February.


> > Both LPC and Java have their strengths and weaknesses.  The greatest
> > advantage that LPC has over Java may be multiple inheritance.
>
> Some would argue:
> 1) Multiple inheritance isn't an advantage, its an unnecessary complication.

I know of only one good argument against multiple inheritance.


> 2) Java (at least partially) solves this issue with interfaces and inner
>    classes.

You have been listening to Sun propaganda. :)  The idea that
interfaces are a substitute for multiple inheritance is ludicrous.
Interfaces are useful in their own right, however.  The same
goes for inner classes.


> Not that I'm necessarily one of these "some" :).  I actually like DGD's multiple
> inheritance model, its very clean and implemented much better than that in
> the other LP drivers (that I know of).
>
> As for LPC displacing Java, folks are aware of Pike?  If not then I suggest
> a look at http://pike.idonex.se/ for a lnaguage that looks "strangely" like
> LPC (ok, well not strangely once you know who wrote it :).

Not so strangely once you take a look at the source.  Though the
ghost of Lars has been laid to rest, there are still some LPC4-isms
in there.

I don't think Pike has much of a chance against Java.  Actually,
I think it will lose even from Perl.


> It would be nice (imho) to see LPC getting byte compiled into byte code that 
> could be run on a Java VM.  Maybe an extra niceness would be the ability to 
> interact with Java objects in this situation a la JPython.

A one-to-one mapping of LPC to Java isn't possible, precisely
because Java lacks multiple inheritance.  There are other problems,
such as the ability to turn off typechecking in LPC with the
mixed type.  Obviously it can be done, but not as nicely as one
might hope.

Regards,
Dworkin