Re: wined3d: Add surface conversion function from WINED3DFMT_YUY2 to WINED3DFMT_B8G8R8X8_UNORM

[Stefan Dösinger]

Am 09.02.2010 um 18:11 schrieb Sandijs Ribaks:
> This is a patch for http://bugs.winehq.org/show_bug.cgi?id=21629 .

It seems ok in concept, just a few points:
> +static void convert_yuy2_x8r8g8b8(const BYTE *src, BYTE *dst,
> -
>    DWORD pitch_in, DWORD pitch_out, unsigned int w, unsigned int h)
>
> ...
> -
>    const BYTE *src_line = (const BYTE *)(src + y * pitch_in);

The cast here is redundant
> -
>        if (!(x & 1)) {       /* for every even pixel, read new U and V */

The placement of the { is inconsistent with the rest of the patch.
> -
>        dst_line[x] = 0xff000000
> -
>            | cliptobyte((c2 + r2) >> 8) << 16    /* red   */
> -
>            | cliptobyte((c2 + g2) >> 8) << 8     /* green */
> -
>            | cliptobyte((c2 + b2) >> 8);         /* blue  */

I recommend to add a comment explaining that (x >> 8) << 8 is not redundant but in used to mask out lower bits.

Just to make sure I understand this correctly:
> -
>            r2 = 409 * e + 128;
> -
>            g2 = - 100 * d - 208 * e + 128;
> -
>            b2 = 516 * d + 128;

The +128 here is for proper rounding in the right shift step afterwards?