EGL_MESA_screen_surface version 4

[Ian Romanick]

Jon Smirl wrote:

> I thought the whole point of FB objects was to get rid of the
> distinction between window/pixmap/pbuffer. Could the base object
> always be an FBO and window/pixmap/pbuffer are then derived off from
> the FBO? Then if the base scan out object is also an FBO everything
> can be scanned out if it is compatible with the mode.

Not at all.  The point of EXT_framebuffer_object is to replace pbuffers 
and WGL_ARB_render_texture with something window-system independent. 
Before FBOs, apps needed multiple code paths to support off-screen 
rendering on Windows, X, and Mac OS.  This alone was pretty painful.

Another goal of EXT_fbo was to make it more efficient to switch between 
off-screen rendering targets.  With pbuffers, you have to do a 
glXMakeContextCurrent (or it's moral equivalent) to swtich rendering 
targets.  Yuck!  The mechanisms in EXT_fbo are much lighter weight and 
are window-system independent.  Yay!  :)

The thing to take home here is the FBOs live in the GL impelementation, 
but windows, pixmaps, and pbuffers are part of the window-system (e.g., 
WGL, GLX, EGL, etc.).  Because of that, they can't really be managed in 
the same way.