| Age | Commit message (Collapse) | Author |
|---|
|
processing"
This reverts commit f058b25881e08c9d89a33345e5c84e1357396932.
This change is completely wrong in so many ways. When clip distances
are generated as part of vertex processing, they must be interpolated
to perform clipping. Geometric clipping goes right out the window.
|
|
Once the clipping distance is calculated and stored per vertex, the
distances can be re-used when clipping is actually performed. This
doesn't have any immediate benefit, but it paves the way for
implementing gl_ClipDistance in vertex shaders and result.clip[] in
vertex programs.
This has not produces any oglconform regressions on my G31 system
which uses software TNL.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
|
|
|
|
|
|
When both ends of the line were clipped, we were using the new v0 instead
of the original v0 when computing the location of the second vertex. Thus,
the second vertex's position was incorrect.
Thanks to Heath Feather for finding a test case.
|
|
We've been lucky if this hasn't been causing line rendering bugs.
|
|
|
|
This fix is to discard the line if both dot products are negative.
|
|
- no need to update ClipMask on generated vertices
- remove the VB->LastClipped value
Line clipping algorithm changed and simplified somewhat. The old one
was based on the triangle routine and probably wouldn't have recognized
lines that were clipped down to nothing (ie culled, but not by a single
plane).
|
|
|
|
|
|
|
|
Moved type conversion and interpolation macros into macros.h
Updated all the files that used to include mmath.h
|
|
|
|
|
|
|
|
clear that these are owned by t_vb_render.c.
Make swrast_setup opaque - it now hooks itself directly into
tnl->Driver.Render.*. Add a _swsetup_Wakeup() call that does this.
Update X11 (tested), osmesa and FX drivers for this change.
FX compiles but is probably broken as the changes there are large. It was the
only remaining driver that used the internal _swsetup_ functions for
interp and copy_pv. This usage has been replaced with code from the DRI
tdfx driver.
|
|
|
|
|
|
|
|
|
|
functions from core mesa -- if drivers need these fallbacks they
must now call them themselves.
Introduced hooks for clip-vertex-interpolation and the rendering
of clipped lines and polygons. Allows drivers to interpolate
their hardware-format vertices directly. Used in dri drivers to
replace fastpath code.
Slight optimizations to pipeline build/run routines.
|
|
|
|
Allow drivers to perform the perspective divide themselves. Assembly
to do cliptesting without perspective divide for size-4 vectors.
|
|
is always the last vertex parameter.
Modify clipping to preserve pv colors.
Modify swrast and X11 driver to expect the pv in the last vertex
(was looking in the first vertex previously).
Remove all handling of flatshading from swrast_setup.
Allow drivers to override the unclipped render tabs in tnl_render_stage
directly. (Like in 3.4). Removed fxsimplerender stage.
Modified t_vb_rendertmp.h to remove the need for 'parity' arguments
in RENDER_TRI macros.
|
|
Bump MAX_TEXTURE_UNITS to 8
Fix mem. leak in destroy_lists
Fix crash in q3 (cva generally)
|
|
|
|
Fixes for compiling without debug.
Fix line clipping
Fix unfilled polygon clipping (should be correct now).
|
|
New array_cache module
Support 8 texture units in core mesa (now support 8 everywhere)
Rework core mesa statechange operations to avoid flushing on many
noop statechanges.
|
