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

Re: [E-devel] Buffer canvas and transparent background



> On Sun, 2006-04-02 at 15:18 +0000, jose_ogp@juno.com wrote:
> >     In your example, you have an output size of (640, 480), and
> > a viewport at (0,0) of size (800, 600).
> >     You then create a rectangle object of *canvas* coordinate
> > size (700, 50) and move it to the *canvas* position (0, 300).
> 
> I understand that.  My test case wasn't meant to demonstrate best
> practices. :)
> 
> >     This means that in *output* coordinates, your rectangle will
> > be of *output* size (560, 40) and its top-left at *output*
> position (0, 240).
> >     At the output position of (0, 300) there is NO rectangle since
> > your rectangle begins at output y-pos of y = 240, and ends at 
> > output y-pos of y = (240 + 40).
> 
> That is why I sampled in the buffer at y=250.  The code was correct.  > It wasn't pretty, but it was correct, and demonstrated my point, 
> I hope. :)
> 
> ....
> ....
> 
> Pixel values still wrong, but at least now it's zeroing the alpha
> at the old location after object move.
> 
> So there seems to be two separate problems with BGRA32 here.
> 
> Cheers!
> Jason.

    Ummm... you did indeed.  There is indeed a problem here with the
BGRA32 format - likely there was a change in the way the 'convert'
functions deal with this format.

   The internal data format for both evas and imlib2 is ARGB32,
but there are places where it's not clear if the byte order of some
data is assumed one way or another.. especially in imlib2.

   Jose.