Re: Tachyon JS VM

2011-10-31 Thread Maxime Chevalier-Boisvert
I guess that's perhaps what people are more familiar with, but I don't think the term "meta-circular" should be limited to interpreters. I was using it in the same way as the Klein project (SELF in SELF) did. Perhaps the term "self-hosted" is more descriptive. The goal is for Tachyon to compile its

Re: Tachyon JS VM

2011-10-31 Thread Brendan Eich
On Oct 31, 2011, at 1:38 PM, Mike Samuel wrote: > 2011/10/31 Maxime Chevalier-Boisvert : >> Hello, >> I recently presented a paper about Tachyon, a research (meta-circular) >> JavaScript VM implementation at DLS 2011. Sam Tobin-Hochstadt suggested that >> I join this list and mention our effort, t

Re: Tachyon JS VM

2011-10-31 Thread Maxime Chevalier-Boisvert
We can't directly (re)use the host GC because Tachyon runs inside its own heap (memory region) and uses its own object representation, which the host VM can't directly interact with. I'm not sure what the NaCL alignment restrictions are. If it's a question of aligning machine code to specific boun

Re: Tachyon JS VM

2011-10-31 Thread Mike Samuel
2011/10/31 Maxime Chevalier-Boisvert : >> If it's not a tree-interpreter, then running inside a JVM that has >> ephemerons doesn't help you on the GC front?  Does the native code >> generated fit within the NaCL alignment restrictions? > > We're not running within a JVM. The goal is for Tachyon to

Re: Tachyon JS VM

2011-10-31 Thread Maxime Chevalier-Boisvert
> > Could you explain how this differs from other meta-circular > interpreters like Narcissus? > Is it correct to say that this is not a tree-interpreter like > Narcissus, but a JavaScript to native compiler written in JS. > Like Narcissus, it is implemented in JavaScript, but it has no interprete

Re: Tachyon JS VM

2011-10-31 Thread Mike Samuel
2011/10/31 Maxime Chevalier-Boisvert : > Hello, > I recently presented a paper about Tachyon, a research (meta-circular) > JavaScript VM implementation at DLS 2011. Sam Tobin-Hochstadt suggested that > I join this list and mention our effort, to get in touch with other JS > implementers, and perhap

Tachyon JS VM

2011-10-31 Thread Maxime Chevalier-Boisvert
Hello, I recently presented a paper about Tachyon, a research (meta-circular) JavaScript VM implementation at DLS 2011. Sam Tobin-Hochstadt suggested that I join this list and mention our effort, to get in touch with other JS implementers, and perhaps get some feedback/suggestions or contributions