Re: gdiplus: Blend in the background colour.

[Vincent Povirk]

> +    return ((b     + ((BYTE)bkgnd         * (255 - alpha) + 127) / 255) |

What's the purpose of adding 127 here?
> stat = GdipBitmapLockBits(bitmap, NULL, ImageLockModeRead|ImageLockModeUserInputBuf,
>              PixelFormat32bppPARGB, &lockeddata);
>
> +        if (background & 0xffffff)
> +        {
> +            DWORD *ptr;
> +            UINT i;
> +            for (ptr = (DWORD*)bits, i = 0; i < width * height; ptr++, i++)
> +            {
> +                if ((*ptr & 0xff000000) == 0xff000000) continue;
> +                *ptr = blend_argb_no_bkgnd_alpha(*ptr, background);
> +            }
> +        }

I guess it wouldn't hurt anything, but it doesn't make much sense to
do this if LockBits fails.

gdiplus: Blend in the background colour.

[Huw Davies]

On Tue, Oct 28, 2014 at 10:58:14AM -0500, Vincent Povirk wrote:
> > +    return ((b     + ((BYTE)bkgnd         * (255 - alpha) + 127) / 255) |
>
> What's the purpose of adding 127 here?

It compensates for C's truncation on integer division.  We have exactly the
same code in the dib engine, so I borrowed it from there.
> > stat = GdipBitmapLockBits(bitmap, NULL, ImageLockModeRead|ImageLockModeUserInputBuf,
> >              PixelFormat32bppPARGB, &lockeddata);
> >
> > +        if (background & 0xffffff)
> > +        {
> > +            DWORD *ptr;
> > +            UINT i;
> > +            for (ptr = (DWORD*)bits, i = 0; i < width * height; ptr++, i++)
> > +            {
> > +                if ((*ptr & 0xff000000) == 0xff000000) continue;
> > +                *ptr = blend_argb_no_bkgnd_alpha(*ptr, background);
> > +            }
> > +        }
>
> I guess it wouldn't hurt anything, but it doesn't make much sense to
> do this if LockBits fails.

Yes, I should test that and probably move this after the Unlock.

Thanks,
Huw.