jon at alumni.caltech.edu
Mon Jul 9 13:49:45 PDT 2007
On Mon, Jul 09, 2007 at 12:59:47PM -0700, Wichmann, Mats D wrote:
> >I just noticed that libGLU is not part of the LSB. One would
> >think that almost everyone using GL also uses GLU, I think
> >the library should be part of 3.2
> I agree. I've agreed for several years now, but to date
> that's not been enough to push it over the hump...
> I'm not enough of an expert in the area to know how
> difficult it would be to implement.
GLU, or at least the SGI reference implementation of it that I think
everyone including Mesa uses these days, is C++ client code (with a
C-callable API) that calls into libGL. Nothing weird about it, other
than the C++ runtime issue. The polygon tesselator is complicated and
poorly documented, but hasn't required attention in years.
Along the same lines, we'd had some discussion a while back about
updating the Linux OpenGL ABI (which is, or at least was, one of the LSB
reference documents), to mandate a higher minimum version of OpenGL
support at link and runtime. That stalled out for lack of interest, but
it shouldn't actually be *hard* to converge on this, if people want. We
would need to determine what OpenGL versions Mesa, Intel, NVIDIA, and
AMD support on which cards, and correlate that against which graphics
cards LSB wishes to support.
More information about the lsb-discuss