Slate do use bytecode

[Brian Rice]

On Oct 16, 2004, at 8:20 AM, Paul Dufresne wrote:

> Pupeno asked a 'stupid question' on IRC.
>
> I like stupid question, because not only am I
> able to understand the answer, but actually I am
> able to give the answer.
>
> The question he asked: 'Does Slate have some kind
> of Slate bytecodes?'.
>
> The answer is yes.
>
> The file src/mobius/vm/base/vm.slate
> list them at the beginning, and the interpret
> method at the end, is the one in charge of
> 'executing' them.
>
> What is not clear to me, is the role of
> src/mobius/vm/interp directory, more
> precisely the compiler.slate file in it.

That area is for all image-side code that needs to know details of the 
VM representation, and also for VM simulation (by inference).

> Note that there is a bytecode.slate
> file describing the bytecodes too there.

It's available in the image as the namespace VM ByteCode.

> I guess, this is a compiler from compiled
> Slate methode, to bytecode. Probably used
> to build an initial Slate image.
> Am I right?

Compiling a Slate method is what produces bytecode (or native code), 
not consumes it. Of course, it still is involved in bootstrapping to 
produce the initial standard library code. This is also the compiler 
that drives the REPL; it's all one mechanism.

--
Brian T. Rice
LOGOS Research and Development
http://tunes.org/~water/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 1452 bytes
Desc: not available
Url : /archives/slate/attachments/20041017/71253520/attachment.bin