Dan --
> While I'm not fond of segfaults myself, the place to check isn't in the
> interpreter loop. It's not unwarranted in assuming that it's told to go OK
> places. If we want to check the better place is inside the jump ops.
I'm in agreement, although we still need to watch out for falling
At 03:53 PM 9/17/2001 +0100, Simon Cozens wrote:
>On Mon, Sep 17, 2001 at 11:01:32AM -0400, Dan Sugalski wrote:
> > While I'm not fond of segfaults myself, the place to check isn't in the
> > interpreter loop.
>
>My view is that if you screw up writing assembly code, you should be
>thankful that y
At 11:02 AM 9/17/2001 -0400, Gregor N. Purdy wrote:
>Dan --
> > Also, we're trying to keep the stuff in the loop to a minimum, so for this
> > I'd rather have a separate runops function, as well as having the actual
> > funky stuff in the body separated out. (I'd really like it abstracted out
> >
On Mon, Sep 17, 2001 at 11:01:32AM -0400, Dan Sugalski wrote:
> While I'm not fond of segfaults myself, the place to check isn't in the
> interpreter loop.
My view is that if you screw up writing assembly code, you should be
thankful that you get the protection of a segfault.
Same happens if y
Dan --
> The big one is you shouldn't assume that we are only going to have a single
> chunk of bytecode--we may well have several loaded from disk, and more
> created on the fly by eval/do/dynamic recompilation.
Yeah. It seems we should have a parrot_bytecode structure that keeps the
pointer
At 10:58 AM 9/17/2001 -0400, Gregor N. Purdy wrote:
>Nor do I. Note that the checks themselves are about as cheap as they
>*can* be, though. And I think that not checking is dangerous. I *did*
>get some segfaults when I was working on the jump_i example. Just try
>running blamo.pasm (attached) on
Simon --
> On Mon, Sep 17, 2001 at 10:33:35AM -0400, Gregor N. Purdy wrote:
> > as *word* counts not *byte* counts -- D'Oh!)
>
> Isn't assembly programming fun? :)
Woo-hoo!
> > */
> >
> > static int
> > -check_magic(void** program_code) {
> > +check_magic(void** program_code, long* program
On Mon, Sep 17, 2001 at 10:33:35AM -0400, Gregor N. Purdy wrote:
> as *word* counts not *byte* counts -- D'Oh!)
Isn't assembly programming fun? :)
> */
>
> static int
> -check_magic(void** program_code) {
> +check_magic(void** program_code, long* program_size) {
> +program_size -= sizeof
At 10:33 AM 9/17/2001 -0400, Gregor N. Purdy wrote:
>Any thoughts?
The big one is you shouldn't assume that we are only going to have a single
chunk of bytecode--we may well have several loaded from disk, and more
created on the fly by eval/do/dynamic recompilation.
Also, we're trying to keep
All --
While I was working on understanding jump_i, I was wishing I could
see a trace of where the PC was going and what it was executing.
(BTW, the jump.pasm test works now that I hard-coded the offsets
as *word* counts not *byte* counts -- D'Oh!)
I was also concerned that even though I was pro
10 matches
Mail list logo
