Qish Re: fuzzyness....
Sat, 22 Jan 2000 22:52:45 +0100 (MET)
>>>>> "Tom" == Tom Novelli <firstname.lastname@example.org> writes:
Tom> Rather than something as dubious as Java bytecode, you could
Tom> use syntax trees or tokenized "Forth" code. Either of those
Tom> could be interpreted sorta quickly, or compiled to machine
Tom> code on demand.
Compiling a bytecode to x86 machine code is reasonably easy (and
probably doable in less than 5KLOC of source). I'am trying to do this
(using Ian Piumarta's Ccg for machine code generation and Vladimir
Makarov' Cocom/Nona for code selection).
My current program (doing x86 code generation) is Qish, in alpha
stage. See my home page, or fetch it from
"http://perso.wanadoo.fr/starynkevitch/basile/qish.tar.gz" If you
fetch it, *please* do send positive or negative feedback (to me and/or
the tunes lists).
I am slowly working on Qish and the tar ball is updated every other
day (I am working on Qish on my spare time only).
The longterm aim of Qish is to provide a runtime for an easy
reflective system. So compiling (a specially designed bytecode) to
machine code is a must. The LLL (ie the compiled bytecode) is designed
in the details while I am coding Qish itself (but I have rather
precise ideas about it)
Basile STARYNKEVITCH - 8 rue de la Faiencerie, 92340 BOURG LA REINE (France)
tel 01.46.65.45.53. email = basile point starynkevitch at wanadoo point fr