[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnash-dev] Re: no AVM2 for a while?

From: Benjamin Wolsey
Subject: Re: [Gnash-dev] Re: no AVM2 for a while?
Date: Sun, 16 Aug 2009 13:01:39 +0200

> Speaking of which, as3compile has its own testsuite (only for 
> syntax/bytecode, not for API), that generates all kinds of obscure
> (but working!) bytecode combinations. I wonder whether you can use 
> that for testing Gnash's AVM2 implementation, too:

This may be very useful, thanks! I've used swfmill a bit for generating
weird bytecode, but as its AS3 support isn't in a released version,
practically no one else runs those tests.

Does as3compile have a way to produce particular bytecode - like ming's
asm instruction, for instance? That would be very useful for testing
Gnash's machine.

> Well you can use the built-in typeof(), but that's about it. as3compile,
> unlike Haxe, doesn't have it's own API. 
> Maybe whatever Haxe does for "Type.typeof" can be ported to vanilla
> ActionScript.

Indeed, it's rather important for us work out what on earth typeof()
means for haxe, instead of relying on whatever black magic (or more
likely just plain typeof or instanceof opcodes) haxe uses. Inheritance
is one of the things Gnash is poor at, in both AVM1 and AVM2, because of
the way objects are implemented using C++ subclasses.

It's also the central reason why not much apart from tracing works in
AS3: because the main timeline is a subclass of MovieClip or Sprite,
which Gnash simply can't instantiate with the present code. I hope to
get this sorted out in the process of fixing prototypes and object
instantiation in the next couple of weeks.


The current release of Gnash is 0.8.5

Benjamin Wolsey, Software Developer -

Attachment: signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil

reply via email to

[Prev in Thread] Current Thread [Next in Thread]