Re: [Gnash-commit] gnash/testsuite/actionscript.all ops.as

[zou lunkai]

I don't care what are the internal conversions that Ming has made.

x = 0xffffffff;
x = x << 16;
trace(x); // should output -1, no mater on a 32bit or 64bit system.
shouldn't it?

So if Ming(cvs version) give us a different result, it must be a Ming
compiler error.

BTW,

x = 4.29497e+09;
x = x << 16;
trace(x); //of course the output won't be -1.  it is 177209344 as I tested.


> This reduces to the double 4.29497e+09 being converted to -1 by the pp
> but NOT by gnash.

Em..., What do you mean?


On 8/7/07, strk <address@hidden> wrote:
> I tracked the difference between the two movies down to this:
>
> My local version:
>     [000]   Constant: 75 "x"
>     [001]   Double: 4.29497e+09
>
> The version on gnashdev:
>     [000]   Constant: 75 "x"
>     [001]   Integer: -1
>
> Basically, a=0xffffffff becomes:
>
>     [001]   Integer: -1
>
> or
>
>     [001]   Double: 4.29497e+09
>
> Depending on whether automatic conversion to double is implemented in Ming we 
> get
> two different pushes. Still, it seems the properietary player returns -65536
> when shifting both numbers 16 bits left.
>
> This reduces to the double 4.29497e+09 being converted to -1 by the pp
> but NOT by gnash.
>
> --strk;
>