Hi, thanks to Ian's well-founded remarks, here is a v2 of the series where I added an explicit range parameter to the functions, and the offsets/ sizes are loaded iff that parameter is true. As a consequence, passing NULL for these parameters will crash as expected (which I do agree is better because it's less surprising), and the assertions become unnecessary.
There are no piglit regressions, but I cannot test atomic and ssbo buffers. Does this variant look good to you? Thanks, Nicolai -- src/mesa/main/bufferobj.c | 649 ++++++++++++------------------------ 1 file changed, 210 insertions(+), 439 deletions(-) _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
