Hi Catalin,

On 10/24/05, Catalin Negrila <[EMAIL PROTECTED]> wrote:
> How would you figure out the bytecode structure for
> AVM2? I'm really interested in this.

I have no idea, but I look at it with a hex editor. Simple SWFs, make
small changes, see what is different.

> I would think it's pretty close to the Java bytecode,
> but knowing the Java bytecode pretty well makes me
> think that using the Flex compiler to compile bits of
> code and then trying to figure out the binary layout
> of the compiled code wouldn't be such an easy task at
> all...

Probably not, but there are many interested, so there are many eyes
and brains that can try to figure it out bit by bit. Or byte by byte,
rather.
I believe we should have a wiki page. There's not too much going on in
the IRC, so that would be a good place to hang out and discuss aspects
of the bytecode that appear relevant and ask if somebody has an idea.

> Was there any documentation published or "leaked"
> about AVM2 byte-code structure?

I hope it doesn't "leak". As long as it doesn't nobody can doubt we
reverse engineered it in an "honest" way.

Anyway, as a little teaser (it is fun in a way) to get people interested:
It seems like it begins with a list of strings, each with its length
in front of it,e.g.:
<length: 0x21>flash.display:MovieClip$protected

this bit of code:
  str += "a";

is that bytecode (or about... not sure if it actually starts there):
  5e 0c 5d 0c 66 0c 2c 1a a0 80 07 68 0c

where 0c apparently addresses "str" in the above mentioned list, and
1a addresses "a". I think.

mark

--
http://snafoo.org/
jabber: [EMAIL PROTECTED]

_______________________________________________
osflash mailing list
[email protected]
http://osflash.org/mailman/listinfo/osflash_osflash.org

Reply via email to