At 4:57 AM +0530 11/12/02, Gopal V wrote:
Oh, sure. But whether the metadata is trustable is an interesting question, as is whether the JIT can generate code that's safe to execute from an unsafe base. It's distinctly possible that when running in safe mode you don't get the JIT.If memory serves me right, Dan Sugalski wrote: > that case, why bother verifying?Hmm.... wouldn't the JIT benifit from a pre knowledge of basic blocks and types or some information ? ... (I seem to think so ...).
More like I'm not expecting to use the JIT for untrusted code. I'm not sure we'll be able to reasonably use the CG core, though I expect we probably will.> at runtime anyway. With a full scan of the bytecode, of course, and > you'd need to figure where each and every instruction starts, which > can be costly. Can't that be added onto the JIT'ing process ? ... viz during conversion ,check for jump targets ?.. I still have this assumption that JITs need to maintain some sort of basic block identification for peephole optimisations ?.. Or is that totally irrelvant for register VMs ? ... (this is the first register VM I have encountered...)
The JIT can likely use basic block info in normal circumstances. I leave that up to the JIT folks--if there's useful metadata they can do things with, we can see about getting it into the bytecode.
Oh, absolutely not. Some benchmarks are too poor to consider. :)So, Parrot is more secure than perl is ? (that being your benchmark).
VMS is my benchmark system. I want the safe interpreters to be as safe as a locked down VMS system. Whether we get there or not's an open question, but it's where we're trying.
--
Dan
--------------------------------------"it's like this"-------------------
Dan Sugalski even samurai
[EMAIL PROTECTED] have teddy bears and even
teddy bears get drunk
