Producing the instruction implementations -Reply

Ray Whitmer
Mon, 12 May 1997 09:51:37 -0600

** High Priority **

>>> "John Wood" <> 05/10/97 08:03am >>>
>Anyone had any thoughts on using existing compiler tools to create the 
>in-linable code for the VM?  It's only the in-lining that causes the problem,

I assume you are referring to the final (Just-In-Time) compilation activities,
and not the in-lining of non-polymorphic mechods.  The inlining seems
to me to be something the compiler can do at the meta level without
knowing the particular instructions which will be used on the particular
platform -- or even if the VM remains completely interpreted at execution
>the real thing stopping people porting at the moment is that lack of people

>who (a) know enough about the system at a system level (b) know
machine code 
>well enough and (c) have the time to write stuff in machine code
>from scratch.  Having tools to let people re-use existing code to write a
>would *really* help.

I seem to recall a fair amount of this stuff having been defined for GNU
compilers to make a portable compiler, which seems to be just what you
are looking for.  I don't know if the gnu public license is an acceptable
basis for the project, or if the goals are to make something more open that
even commercial providers could directly incorporate into commercial
offerings.  I have never found a way in the past to really evangelize or
reuse anything in GNU due to the overly-restrictive licensing, which
would probably cause commercial providers to simply stick to their own
proprietary implementations, IMHO.

Ray Whitmer