"gabriele renzi" <surrender_it / remove.yahoo.it> schrieb im Newsbeitrag
news:6tou30dgnpok7ohji94nj06jniasjuhhna / 4ax.com...
> il Fri, 27 Feb 2004 17:59:07 +0900, Mark Hubbart <discord / mac.com> ha
> scritto::
>
>
> >
> >Which leads to a question that has been bouncing around in my mind for
> >several weeks now: What obstacles are there against being able to
> >compile Ruby? I would guess that eval() would be one, but other than
> >that, would it be feasible to create a compilable subset of Ruby?

The question is, how attractive would this subset be?  I mean, there's a
lot of code around that relys on beeing able to use eval and friends.

> IIRC lisp has eval() too, and there are plenty of compiler out ther ;)

But not all of them compile to native.  GNU CLISP compiles to byte code
AFAIK.  Dunno what the native compiling compilers support, basically
you'll have to include the compiler in the runtime.  And looking at the
high efforts done at Sun and MS to enhance their VM's, that shows that
it's not too easy to be able to come up with such a system that does work
correctly *and* is fast.

> I bet you could write a compiler from ruby to at&t asm and let GAS do
> the rest :)

Hm...

    robert