From: Ruben <ryu@mundivia.es>
  To  : ggi-develop@eskimo.com
  Date: Sat, 24 Jul 1999 23:29:46 +0200 (CEST)

Re: GII problems and transparent blitter

On Sat, 24 Jul 1999, Andreas Beck wrote:

> > Also, on a somewhat unrelated topic, I've heard that a transparent blitter
> > for GGI is currently nonexistent. Is anyone working on one, have plans to
> > work on one, or are we on our own and must implement one ourselves?
> I have one and I'll woork with Marcus on integrating all of it into a
> Sprite/Overlay and a Blit extension as soon as he is back.

	I'm the Sprites GGL developer, I will explain a bit on how we do
blitting (ourselves of course, because GGI doesn't support it yet): 
	We have a bitmap and a mask, the bitmap is saved as is, but the
mask is compressed in a RLE-like way. Our RLE algorithm works in each line
independently (don't compress between lines) and it stores alternatively
number of transparents/number of non transparents, starting each line with
number of transparents, as they are the more probable at begin of lines.

	It enhaces very much the drawing of sprites because we can do a
memcpy for each sequence of non-transparents and skip a group of
transparents in one instruction. But it also enhaces our collision
detection, we don't need to test all pixels, only the groups of
transparents/non-transparents.

	We have also the algorithm of creation of the mask from a bitmap
(saying the transparent color, of course).

	I't the fastest way I know of handling sprites, but surely it
isn't the best...

	Well, the most important thing of this email: I need to know the
way transparency blitting will be implemented in GGI, because I will need
to change GGL to incorporate it, and probabily I will need to change the
object model. Please, if you decide to use your own implementation, tell
me the way it runs, I need to know it to make new collision detection.
	
Regards.
--
  _
 /_) \/ / /  email: mailto:ryu@mundivia.es
/ \  / (_/   www  : http://programacion.mundivia.es/ryu
[GGL developer] [IRCore developer]