On Sun, 17 Dec 2006, Basile STARYNKEVITCH wrote:

>  use some library capable of generating machine code in memory; in
>  particular libjit http://www.southern-storm.com.au/libjit.html
>  http://dotgnu.org/pipermail/libjit-developers/ or
>  GNU lightning http://www.gnu.org/software/lightning/lightning.html
>  https://savannah.gnu.org/projects/lightning/ . Ok, some of them are
>  lacking
>  important features like tail recursion or good support of a
>  garbaged-collected runtime.

I am unimpressed with the performance given by virtual machine 
implementations, even those with signifigant developer backing (such as 
the JVM).

If I didn't give a fig about performance, I'd probably implement my own 
VM.

>
>  implementing the functional language within a
>  language/implementation providing enough hooks for that, I am
>  thinking of MetaOcaml http://metaocaml.org/ (somehow nearly dead
>  unfortunately), Parrot http://parrotcode.org/ (which does have some
>  tail recursion support and JIT compiling) or even SBCL common lisp
>  implementation (provided you use the internals of it, which means
>  making a big dive!)
>
Using another language runs afoul of many of the same problems using the 
JVM or CLR does.

>>
>> I think the three new things I'd like to see out of C-- are (in rough
>> order of priority):
>> 1) x86-64 support
>
> I definitely agree that this is of topmost priority for C-- to survive!

But I don't think it'd be hard to implement.  The x86-64 instruction set 
is very similiar to the x86-32 instruction set, so large amounts of the 
back end could be cribbed.

Brian

_______________________________________________
Cminusminus mailing list
[email protected]
https://cminusminus.org/mailman/listinfo/cminusminus

Reply via email to