[Bf-committers] Viewport FX Design

Antony Riakiotakis kalast at gmail.com
Fri Jun 1 11:11:47 CEST 2012


Hi Jason, great stuff. Actually, if you think about it, matrix
multiplications are actually not graphics state strictly, but a
CPU-side transform. So I would maybe vote to actually -substitute-
such uses with regular BLI_math functions, or simply use a BLI matrix
state machine to do such tasks, if it is hard to eliminate the state
oriented aspect completely. IMO eliminating the state machine and
explicitly specifying a transform before using it would be beneficial
because in OGL it is very hard to actually see what are the contents
of the matrix stack (and, more importantly when debugging, how they
got there) before modifying it. That's the reason you usually call
glLoadIdentity. In fact, if we ever go for a pure GL3+ core profile
this is absolutely essential as the gpuMatrixTransformCommit
(BLI_MatrixTransformCommit ?) call would need to probably calculate
and bind an inverse matrix to the uniform buffer object as well.


More information about the Bf-committers mailing list