Pixmaps are bitmaps in main memory, Textures are areas of video ram. They are connected via a bus over which the data must travel. OpenGL lives in video ram and everything it has access to has to be accessible to the graphics card GPU, thus has to be in video ram. You have to get the data from main memory to video ram by using e.g. GlTexFromPixmap. That copies the pixmap to video ram as a texture. The Graphics card then can access the texture data. It cannot access the pixmap.
Once you've generated a texture from a pixmap (i think that function returns a texture handle integer?), the pixmap still exists, and you could easily just save the pixmap. If you've thrown the pixmap away, you'll need to draw the texture image to the backbuffer and then either use GrabPixmap or GLReadPixels to transfer it to main memory, then SavePixmapPNG().
You cannot use GLTexCoord2f with pixmaps, you can only use it with textures in video ram. The whole reason there are two distinct entities - textures and pixmaps, is because the GPU graphics card can only access the special video ram that is `on board` the graphics card itself. It cannot access main memory - unless you have integrated graphics which is pretty crappy. So because the GPU has to work with video ram and know where to look for stuff, and for performance reasons, it doesn't access main memory. This creates the need for a) a storage system in main memory for bitmap data (pixmaps), and b) a bus to transfer between them. So you're pretty much stuck with two separate storage spaces.
|
