d3dx9: Implement D3DXFloat16to32Array and D3DXFloat32to16Array.

[Misha Koshelev]

On Fri, Feb 11, 2011 at 3:58 PM, Stefan Dösinger <stefandoesinger at gmx.at> wrote:
> Am Freitag 11 Februar 2011, 20:48:58 schrieb Misha Koshelev:
>> Stefan et al:
>>
>> Thank you. How about this guy?
>
>> +    if (exp > 31)
>> +    {
>> +        /* too big */
>> +        ret = 0x7fff; /* INF */
>> +    }
> That's not INF, that is NaN. Or, if you don't read it as a special value the
> highest possible positive number in a float16 format that doesn't understand
> NaNs or INFs
>
> Otherwise it looks OK
>

Thank you.

Misha