profile/ivi/mesa.git
18 years agoIn TEX instructions, use precomputed lambda only if using fragment.texcoord[k]
Brian Paul [Tue, 31 Oct 2006 14:29:01 +0000 (14:29 +0000)]
In TEX instructions, use precomputed lambda only if using fragment.texcoord[k]
to sample texture[k].  Use zero otherwise.  Not foolproof, but a good
compromise.

18 years agoFor TEX instructions use lambda=0. When sampling from texture unit K we
Brian Paul [Tue, 31 Oct 2006 01:22:38 +0000 (01:22 +0000)]
For TEX instructions use lambda=0.  When sampling from texture unit K we
were using the partial derivatives of texcoord[K] but the coordinate used
for texture sampling may be something totally different (and texcoord[K]
might not be a real texture coord at all).  Net result was a bogus LOD is
sometimes used, often resulting in using the smallest mipmap level (a
constant color).
Just use zero for now (undef LAMBDA_ZERO to override).
Plus, some additional debug code.

18 years agoFix bug #8799.
Ian Romanick [Mon, 30 Oct 2006 23:44:54 +0000 (23:44 +0000)]
Fix bug #8799.

Properly resize bith the drawable and the readable in mgaUpdateRects.
Eliminate the use of the deprecated GetBufferSize interface.  Bump
driver date.

18 years agoRefactor mgaXMesaSetFrontClipRects and mgaXMesaSetBackClipRects.
Ian Romanick [Mon, 30 Oct 2006 22:17:39 +0000 (22:17 +0000)]
Refactor mgaXMesaSetFrontClipRects and mgaXMesaSetBackClipRects.

Combine mgaXMesaSetFrontClipRects and mgaXMesaSetBackClipRects into a
single new function called mga_set_cliprects.  This enables a small
refactor in mgaDDDrawBuffer.

18 years agoGet DRI drawable from framebuffer object.
Ian Romanick [Mon, 30 Oct 2006 22:07:00 +0000 (22:07 +0000)]
Get DRI drawable from framebuffer object.

In the software rasterization fallbacks, get the __DRIdrawablePrivate
pointer from the driRenderbuffer that is passed in.  This eliminates
the need for the mesa_drawable pointer in MGA context structure.

This partially fixes bug #8799.  wincopy no longer produces a black
window for the destination window, but it does produce incorrect
results if the destination window is resized.

18 years agocode tweaks, remove old comments
Brian Paul [Mon, 30 Oct 2006 00:22:28 +0000 (00:22 +0000)]
code tweaks, remove old comments

18 years agofix _mesa_problem() text
Brian Paul [Mon, 30 Oct 2006 00:13:31 +0000 (00:13 +0000)]
fix _mesa_problem() text

18 years agoMove gl_vertex_program_machine struct out of mtypes.h and put into nvvertexec.h.
Brian Paul [Mon, 30 Oct 2006 00:12:05 +0000 (00:12 +0000)]
Move gl_vertex_program_machine struct out of mtypes.h and put into nvvertexec.h.
Massage nvvertexec.c code to work more like s_nvfragprog.c - another step
toward unifying vertex/fragment program execution.

18 years agoremove debug code
Brian Paul [Mon, 30 Oct 2006 00:07:22 +0000 (00:07 +0000)]
remove debug code

18 years agoRather than always allocating 100 instructions, put instructions into a
Brian Paul [Sun, 29 Oct 2006 21:17:18 +0000 (21:17 +0000)]
Rather than always allocating 100 instructions, put instructions into a
temporary buffer, then copy into a malloc'd array of the correct size.
According to Glean texCombine, we never need more than 21 instructions.

18 years agoNew _mesa_exec_vertex_state_program() function. Start of some re-org.
Brian Paul [Sun, 29 Oct 2006 18:33:14 +0000 (18:33 +0000)]
New _mesa_exec_vertex_state_program() function.  Start of some re-org.

18 years agofix wrong inequality
Brian Paul [Sun, 29 Oct 2006 18:14:00 +0000 (18:14 +0000)]
fix wrong inequality

18 years agoChange _mesa_init_instruction() to initialize an array of instructions.
Brian Paul [Sun, 29 Oct 2006 18:03:16 +0000 (18:03 +0000)]
Change _mesa_init_instruction() to initialize an array of instructions.

18 years agoProgram instructions are initialized before parsing so no need to set
Brian Paul [Sun, 29 Oct 2006 17:55:16 +0000 (17:55 +0000)]
Program instructions are initialized before parsing so no need to set
Abs, NegateAbs, CondMask and CondSwizzle fields to defaults here.
CondMask was getting mis-set.  Removing the incorrect assignment improves
fragment program speed by 15%.

18 years agoAdded PROGRAM_CONSTANT to get_register_pointer().
Brian Paul [Sat, 28 Oct 2006 23:10:17 +0000 (23:10 +0000)]
Added PROGRAM_CONSTANT to get_register_pointer().
More DEBUG_FRAG code.

18 years agospecial case END in _mesa_print_instruction()
Brian Paul [Sat, 28 Oct 2006 23:08:07 +0000 (23:08 +0000)]
special case END in _mesa_print_instruction()

18 years agoImplement _mesa_append_fog_code() for fragment program fog options.
Brian Paul [Sat, 28 Oct 2006 17:14:47 +0000 (17:14 +0000)]
Implement _mesa_append_fog_code() for fragment program fog options.

18 years agoIf DEBUG, check that all array indices really do fall in [start,end] in
Brian Paul [Thu, 26 Oct 2006 20:54:28 +0000 (20:54 +0000)]
If DEBUG, check that all array indices really do fall in [start,end] in
the DrawRangeElements() call.  Warn the user if that's not true.

18 years agoremove code for inserting mvp transform into position-invariant vertex progs and...
Roland Scheidegger [Thu, 26 Oct 2006 12:09:07 +0000 (12:09 +0000)]
remove code for inserting mvp transform into position-invariant vertex progs and instead use _mesa_insert_mvp_code().

18 years agoPreviously, if a fragment program referenced fragment.fogcoord but the
Brian Paul [Wed, 25 Oct 2006 23:10:14 +0000 (23:10 +0000)]
Previously, if a fragment program referenced fragment.fogcoord but the
program didn't use the ARB_fog_linear option, the fragment program's
fragment.fogcoord register wasn't loaded properly.

18 years agochange some bogus comments regarding the base e exponential function used for fog...
Roland Scheidegger [Wed, 25 Oct 2006 12:14:26 +0000 (12:14 +0000)]
change some bogus comments regarding the base e exponential function used for fog in vertex programs.

18 years agofix (per-vertex) fog when using ARB_vp by incorporating fog factor computation into...
Roland Scheidegger [Tue, 24 Oct 2006 22:37:51 +0000 (22:37 +0000)]
fix (per-vertex) fog when using ARB_vp by incorporating fog factor computation into the vertex program (not yet fixed for swtnl). Simplify (and correct) the VTX_TCL_OUTPUT_VTXFMT handling when using vertex programs, turns out it's solely driven by the needs of the past-vertex stage of the pipeline, this should fix lockups with ill-specified applications using vertex programs (for instance applications enabling fog but not writing to fog coord output will now get (conformant) undefined results instead of lockups).

18 years agoAdd missing call to glutInit.
Ian Romanick [Tue, 24 Oct 2006 20:50:08 +0000 (20:50 +0000)]
Add missing call to glutInit.

18 years agos/GLchan/GLubyte/
Brian Paul [Tue, 24 Oct 2006 13:46:39 +0000 (13:46 +0000)]
s/GLchan/GLubyte/

18 years agowhitespace clean-up
Brian Paul [Tue, 24 Oct 2006 13:45:08 +0000 (13:45 +0000)]
whitespace clean-up

18 years agoEmit cliprects in the userspace driver as required, rather than
Keith Whitwell [Mon, 23 Oct 2006 08:43:26 +0000 (08:43 +0000)]
Emit cliprects in the userspace driver as required, rather than
passing them to the kernel.  This works because all drawing commands
in the 965 driver are emitted with the lock held and the batchbuffer
is always flushed prior to releasing the lock.  This allows multiple
cliprects to be dealt with, without replaying entire batchbuffers and
redundantly re-emitting state.

18 years agoassorted clean-ups
Brian Paul [Sun, 22 Oct 2006 23:21:45 +0000 (23:21 +0000)]
assorted clean-ups

18 years agodon't save color/z if fragment is killed
Brian Paul [Sun, 22 Oct 2006 22:58:42 +0000 (22:58 +0000)]
don't save color/z if fragment is killed

18 years agoadd minstall script to MAIN_FILES (bug 8720)
Brian Paul [Sun, 22 Oct 2006 22:37:30 +0000 (22:37 +0000)]
add minstall script to MAIN_FILES (bug 8720)

18 years agominor tweaks good for about 12% faster execution
Brian Paul [Sun, 22 Oct 2006 22:36:24 +0000 (22:36 +0000)]
minor tweaks good for about 12% faster execution

18 years agosilence used var warnings
Brian Paul [Sun, 22 Oct 2006 22:25:46 +0000 (22:25 +0000)]
silence used var warnings

18 years agoColor clamping fixes.
Brian Paul [Sun, 22 Oct 2006 17:18:50 +0000 (17:18 +0000)]
Color clamping fixes.

18 years agomove/improve comments, clean up code in a few places
Brian Paul [Sat, 21 Oct 2006 23:21:35 +0000 (23:21 +0000)]
move/improve comments, clean up code in a few places

18 years agofix another logic error in computing deferredTexture
Brian Paul [Sat, 21 Oct 2006 22:56:21 +0000 (22:56 +0000)]
fix another logic error in computing deferredTexture

18 years agoFor deferredTexture, need to check if the fragment program/shader writes
Brian Paul [Sat, 21 Oct 2006 22:43:15 +0000 (22:43 +0000)]
For deferredTexture, need to check if the fragment program/shader writes
the fragment Z.

18 years agomove interpolate_varying() call
Brian Paul [Sat, 21 Oct 2006 21:14:20 +0000 (21:14 +0000)]
move interpolate_varying() call

18 years agodon't use params passed to Clear()
Brian Paul [Sat, 21 Oct 2006 21:13:40 +0000 (21:13 +0000)]
don't use params passed to Clear()

18 years agoMove interpolate_texcoords() and interpolate_varying() calls into shade_texture_span().
Brian Paul [Sat, 21 Oct 2006 00:20:16 +0000 (00:20 +0000)]
Move interpolate_texcoords() and interpolate_varying() calls into shade_texture_span().
Added some INLINE hints.

18 years agoConsolidate some code into shade_texture_span().
Brian Paul [Thu, 19 Oct 2006 22:27:44 +0000 (22:27 +0000)]
Consolidate some code into shade_texture_span().
Clean up logic for deferred texturing/shading.

18 years agofixed Fixed/Int bug
Brian Paul [Thu, 19 Oct 2006 22:26:29 +0000 (22:26 +0000)]
fixed Fixed/Int bug

18 years agoremove temp code
Brian Paul [Thu, 19 Oct 2006 22:13:31 +0000 (22:13 +0000)]
remove temp code

18 years agoremove some debug/test changes
Brian Paul [Thu, 19 Oct 2006 20:11:15 +0000 (20:11 +0000)]
remove some debug/test changes

18 years agoNew bin/minstall script - a minimal replacement for 'install'.
Brian Paul [Thu, 19 Oct 2006 20:09:05 +0000 (20:09 +0000)]
New bin/minstall script - a minimal replacement for 'install'.
Correctly handles symlinks so we can get rid of the COPY_LIBS stuff.

18 years agofix problem when CHAN_BITS == 32 (bug 8704), remove unneeded assertion
Brian Paul [Thu, 19 Oct 2006 14:02:42 +0000 (14:02 +0000)]
fix problem when CHAN_BITS == 32 (bug 8704), remove unneeded assertion

18 years agoFix nonportable pointer+unsigned arithmetic. Use pointer+int.
Michal Krol [Thu, 19 Oct 2006 08:07:00 +0000 (08:07 +0000)]
Fix nonportable pointer+unsigned arithmetic. Use pointer+int.

18 years agoadded clamp_colors(), updated assertions
Brian Paul [Wed, 18 Oct 2006 23:14:02 +0000 (23:14 +0000)]
added clamp_colors(), updated assertions

18 years agominor re-org, doxygen comments
Brian Paul [Wed, 18 Oct 2006 22:38:13 +0000 (22:38 +0000)]
minor re-org, doxygen comments

18 years agoChanged install filenames to libGL* and libOSMesa* so things work on Darwin
Brian Paul [Wed, 18 Oct 2006 22:05:13 +0000 (22:05 +0000)]
Changed install filenames to libGL* and libOSMesa* so things work on Darwin
where GL_LIB_NAME = libGL.dylib but the other library name is libGL.1.dylib
Unlike other platforms, the version number is in the middle of the name,
not at the end.
See bug 8671

18 years agoEnable GLX_SGI_make_current_read for radeon.
Ian Romanick [Wed, 18 Oct 2006 20:27:11 +0000 (20:27 +0000)]
Enable GLX_SGI_make_current_read for radeon.

Added code to track the drawable bound to the context for reading.  In
addition, when a drawable is initially bound (for reading or drawing)
or when the size of the drawable changes, update the size of the
framebuffer object that back the drawable (for software fallbacks).

Deprecate the old GetBufferSize interface.

Bump the driver date.

These changes were tested with wincopy on both direct rendering and
accelerated indirect rendering (AIGLX).

18 years agoDon't use x/y/w/h params given to Clear().
Brian Paul [Wed, 18 Oct 2006 20:03:20 +0000 (20:03 +0000)]
Don't use x/y/w/h params given to Clear().

18 years agoWant to stop passing x/y/width/height to Clear() function.
Brian Paul [Wed, 18 Oct 2006 20:02:42 +0000 (20:02 +0000)]
Want to stop passing x/y/width/height to Clear() function.
The coordinates need to be computed after we've got the hw lock.
Code updated to:
1. Ignore all/x/y/width/height/ params passed to Clear func.
2. Pass 0,0,0,0,0 to _swrast_Clear() until they're totally removed.

18 years agoDon't pass x/y/width/height to ctx->Driver.Accum().
Brian Paul [Wed, 18 Oct 2006 18:35:09 +0000 (18:35 +0000)]
Don't pass x/y/width/height to ctx->Driver.Accum().
Compute the region after we've locked (and possibly updated the buffer's size).
Same thing is needed for ctx->Driver.Clear().

18 years agoglDrawPixels, line clipping bug fixes
Brian Paul [Wed, 18 Oct 2006 18:03:54 +0000 (18:03 +0000)]
glDrawPixels, line clipping bug fixes

18 years agoFix broken line clipping.
Brian Paul [Wed, 18 Oct 2006 18:02:05 +0000 (18:02 +0000)]
Fix broken line clipping.
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.

18 years agoFix a dangerous use of ASSERT in an else-clause not enclosed in braces.
Brian Paul [Wed, 18 Oct 2006 17:36:57 +0000 (17:36 +0000)]
Fix a dangerous use of ASSERT in an else-clause not enclosed in braces.
We've been lucky if this hasn't been causing line rendering bugs.

18 years agoAlways convert colors to float before running a fragment shader/program.
Brian Paul [Tue, 17 Oct 2006 22:23:32 +0000 (22:23 +0000)]
Always convert colors to float before running a fragment shader/program.

18 years agofix fog color bug
Brian Paul [Tue, 17 Oct 2006 22:22:42 +0000 (22:22 +0000)]
fix fog color bug

18 years agoMove struct atifs_machine into s_atifragshader.c
Brian Paul [Tue, 17 Oct 2006 21:12:32 +0000 (21:12 +0000)]
Move struct atifs_machine into s_atifragshader.c
Add support for runtime colorbuffer types.

18 years agoSeperate interpolate_colors() and interpolate_specular().
Brian Paul [Tue, 17 Oct 2006 20:40:22 +0000 (20:40 +0000)]
Seperate interpolate_colors() and interpolate_specular().
Fix bugs in apply_aa_coverage().

18 years agoapply_aa_coverage() for ubyte/ushort/float
Brian Paul [Tue, 17 Oct 2006 20:22:53 +0000 (20:22 +0000)]
apply_aa_coverage() for ubyte/ushort/float

18 years agonew assertions - one of which is currently failing with arbfslight.c
Brian Paul [Tue, 17 Oct 2006 19:57:25 +0000 (19:57 +0000)]
new assertions - one of which is currently failing with arbfslight.c

18 years agoremove obsolete intelBufferSize(), bump driver date
Brian Paul [Tue, 17 Oct 2006 18:54:19 +0000 (18:54 +0000)]
remove obsolete intelBufferSize(), bump driver date

18 years agoRemove get_buffer_size()
Brian Paul [Tue, 17 Oct 2006 17:51:39 +0000 (17:51 +0000)]
Remove get_buffer_size()
Always check window size in XMesaMakeCurrent() in case glViewport is not
called.

18 years agoLots of fixes/changes to select color depth at runtime.
Brian Paul [Tue, 17 Oct 2006 17:44:43 +0000 (17:44 +0000)]
Lots of fixes/changes to select color depth at runtime.

18 years agocheck if ctx->Driver.GetBufferSize==NULL and no-op
Brian Paul [Tue, 17 Oct 2006 17:43:47 +0000 (17:43 +0000)]
check if ctx->Driver.GetBufferSize==NULL and no-op

18 years agoinclude x86-64.h to silence warning
Brian Paul [Tue, 17 Oct 2006 17:03:21 +0000 (17:03 +0000)]
include x86-64.h to silence warning

18 years agoFix CHAN_BITS=32 issues.
Brian Paul [Tue, 17 Oct 2006 17:03:00 +0000 (17:03 +0000)]
Fix CHAN_BITS=32 issues.
Save/restore array->ChanType in case colors are converted.

18 years agoAdded -clip option to test glCopyPixels beyond window bounds.
Brian Paul [Mon, 16 Oct 2006 23:22:44 +0000 (23:22 +0000)]
Added -clip option to test glCopyPixels beyond window bounds.
Clear dest window to black before copying.
Use glWindowPos2iARB().

18 years agoInit _UseTexEnvProgram to fix failed assertion in i915_render_start() - this may...
Brian Paul [Mon, 16 Oct 2006 22:02:33 +0000 (22:02 +0000)]
Init _UseTexEnvProgram to fix failed assertion in i915_render_start() - this may be temporary

18 years agoUpdates to intelWindowMoved():
Brian Paul [Mon, 16 Oct 2006 21:39:34 +0000 (21:39 +0000)]
Updates to intelWindowMoved():
Only need to call _mesa_resize_framebuffer() when we've detected a window
size change.
Set the drawFb->Initalized flag to GL_TRUE to avoid obsolete
Driver.GetBufferSize/ResizeBuffer calls in the Mesa code.

18 years agoTrack the currently bound read drawable and make sure its info is up
Ian Romanick [Mon, 16 Oct 2006 20:59:53 +0000 (20:59 +0000)]
Track the currently bound read drawable and make sure its info is up
to date.

18 years agoFix build when DEBUG is defined but DEBUG_MATH is not.
Ian Romanick [Mon, 16 Oct 2006 20:57:09 +0000 (20:57 +0000)]
Fix build when DEBUG is defined but DEBUG_MATH is not.

A couple places attempted to call _math_test_all_*_functions when
DEBUG was defined.  These routines only exist when DEBUG_MATH is
defined.

18 years agos/X86_SOURCES/ASM_SOURCES/ in ALL_SOURCES to try to fix make depend warnings (bug...
Brian Paul [Mon, 16 Oct 2006 20:51:50 +0000 (20:51 +0000)]
s/X86_SOURCES/ASM_SOURCES/ in ALL_SOURCES to try to fix make depend warnings (bug 8669)

18 years agoFix bug #8563.
Ian Romanick [Mon, 16 Oct 2006 17:50:19 +0000 (17:50 +0000)]
Fix bug #8563.

Fixes a GLX protocol problem when binding an indirect rendering context
after a direct rendering context.  In this case, the oldContetTag sent to
the server should be None, but the currectContextTag stored in the direct
rendering context (i.e., -1) was sent instead.

18 years agoSet _NEW_BUFFERS in _mesa_resize_framebuffer(). This makes sure the swrast
Brian Paul [Mon, 16 Oct 2006 17:26:30 +0000 (17:26 +0000)]
Set _NEW_BUFFERS in _mesa_resize_framebuffer().  This makes sure the swrast
module updates its state (CLIP_BIT) upon resizes.  The driver could do this
instead, but doing it here is safer.

18 years agoremove some old code that could cause infinite recursion
Brian Paul [Mon, 16 Oct 2006 16:39:54 +0000 (16:39 +0000)]
remove some old code that could cause infinite recursion

18 years agoallow >10 mipmap levels (patch by Benno Schulenberg) (bug 3130)
Brian Paul [Sun, 15 Oct 2006 22:30:35 +0000 (22:30 +0000)]
allow >10 mipmap levels (patch by Benno Schulenberg) (bug 3130)

18 years agoregenerated to get DRI_CONF_EXCESS_MIPMAP()
Brian Paul [Sun, 15 Oct 2006 22:29:16 +0000 (22:29 +0000)]
regenerated to get DRI_CONF_EXCESS_MIPMAP()

18 years agofix handling of textures with a base internal format that does not have all four...
Roland Scheidegger [Sun, 15 Oct 2006 21:47:56 +0000 (21:47 +0000)]
fix handling of textures with a base internal format that does not have all four rgba values set for radeon and r200 (discovered with a modified glean pixelFormats test, noone ever noticed in over 2 years). For radeon, use hw format I8 as previously, and change tex env to make the correct default values appear for both GL_ALPHA and GL_LUMINANCE textures. For r200, which supports GL_LUMINANCE just fine,  use the AL88 hw format for GL_ALPHA textures, since it seems like it's probably not worth the effort to fix up the texture environment (certainly complicated in case of ATI_fragment_shader programs).

18 years agoAdded #if / #else / #endif around code related to framebuffer initialization
Brian Paul [Sun, 15 Oct 2006 19:26:43 +0000 (19:26 +0000)]
Added #if / #else / #endif around code related to framebuffer initialization
to describe what should be changed in the drivers.

18 years agoupdated comments, mark GetBufferSize() and ResizeBuffers() as obsolete
Brian Paul [Sun, 15 Oct 2006 19:20:29 +0000 (19:20 +0000)]
updated comments, mark GetBufferSize() and ResizeBuffers() as obsolete

18 years agominor clean-up: s/intel->driDrawable/dPriv/
Brian Paul [Sun, 15 Oct 2006 19:08:11 +0000 (19:08 +0000)]
minor clean-up: s/intel->driDrawable/dPriv/

18 years agoRemove calls to _mesa_ResizeBuffersMESA() - generally replace with code to
Brian Paul [Sun, 15 Oct 2006 18:51:22 +0000 (18:51 +0000)]
Remove calls to _mesa_ResizeBuffersMESA() - generally replace with code to
check the current window size, then call _mesa_resize_framebuffer().

18 years agoRemove a bunch of "ctx->Driver.function = _swrast_Function" lines since
Brian Paul [Sun, 15 Oct 2006 18:31:29 +0000 (18:31 +0000)]
Remove a bunch of "ctx->Driver.function = _swrast_Function" lines since
default/fallback functions are already plugged in by the call to
_mesa_init_driver_functions().

18 years agoUse endianness test function provided by mesa.
Jerome Glisse [Sun, 15 Oct 2006 18:22:28 +0000 (18:22 +0000)]
Use endianness test function provided by mesa.

Use the _mesa_little_endian helper function insted
of the MESA_BIG_ENDIAN flag. There is still place
using this flag but we need to find how to submit
16bits elements on big endian computer.

18 years agoRemove Driver.ResizeBuffers = _mesa_resize_framebuffer lines.
Brian Paul [Sun, 15 Oct 2006 18:18:39 +0000 (18:18 +0000)]
Remove Driver.ResizeBuffers = _mesa_resize_framebuffer lines.
This is already done by the preceeding call to _mesa_init_driver_functions()
which plugs in default functions like that.

18 years agoincreate MaxTextureLevels to 11 (bug 3130)
Brian Paul [Sun, 15 Oct 2006 17:57:53 +0000 (17:57 +0000)]
increate MaxTextureLevels to 11 (bug 3130)

18 years agocheck if files exist before installing (Thierry Reding)
Brian Paul [Sun, 15 Oct 2006 16:38:09 +0000 (16:38 +0000)]
check if files exist before installing (Thierry Reding)

18 years agoallow specifying image size on cmd line
Brian Paul [Sat, 14 Oct 2006 23:24:45 +0000 (23:24 +0000)]
allow specifying image size on cmd line

18 years agoSome code movement so that the gl_framebuffer is fully initialized before
Brian Paul [Sat, 14 Oct 2006 23:02:26 +0000 (23:02 +0000)]
Some code movement so that the gl_framebuffer is fully initialized before
calling _mesa_make_current() to avoid hitting the initialize_framebuffer_size()
stuff in context.c

18 years agoBig re-org of buffer size management.
Brian Paul [Sat, 14 Oct 2006 22:52:17 +0000 (22:52 +0000)]
Big re-org of buffer size management.
All buffer resizes now handled by xmesa_check_and_update_buffer_size() which
uses the _mesa_resize_framebuffer() function.
Moved all low-level XImage/Pixmap resizing into xm_buffers.c file.
Also, update lots of comments for Doxygen.

18 years agocut and paste error: s/Width/Height/
Brian Paul [Sat, 14 Oct 2006 21:33:06 +0000 (21:33 +0000)]
cut and paste error: s/Width/Height/

18 years agobuild fix: s/Table/TableUB/
Brian Paul [Sat, 14 Oct 2006 20:28:36 +0000 (20:28 +0000)]
build fix: s/Table/TableUB/

18 years agoassorted clean-ups
Brian Paul [Sat, 14 Oct 2006 15:22:18 +0000 (15:22 +0000)]
assorted clean-ups

18 years agoimplement ARB_point_parameters and ARB_point_sprite on r200. The code is nearly the...
Roland Scheidegger [Fri, 13 Oct 2006 22:10:05 +0000 (22:10 +0000)]
implement ARB_point_parameters and ARB_point_sprite on r200. The code is nearly the same as outlined in bug #4707, except it disables perspective correction for point sprites to make them actually work. And, separate the state atom into two as the tcl parameters would overwrite vertex program parameters when active. Also implement the GL_VERTEX_PROGRAM_POINT_SIZE_ARB option to make vertex programs outputting a point size work correctly (untested). Smooth points will still always be size 1. While here, enable gouraud shading for fog when using fog coord.

18 years agoremove needless condition
Brian Paul [Fri, 13 Oct 2006 17:45:08 +0000 (17:45 +0000)]
remove needless condition

18 years agoConsolidate, move, fix code related to color index and stencil image transfer
Brian Paul [Fri, 13 Oct 2006 17:41:23 +0000 (17:41 +0000)]
Consolidate, move, fix code related to color index and stencil image transfer
operations (shift, offset, table lookup, etc).

18 years agoFix dumb build errors.
Ian Romanick [Fri, 13 Oct 2006 17:34:01 +0000 (17:34 +0000)]
Fix dumb build errors.

I always build with -DGLX_USE_TLS, so I never hit these paths.  glapi.h is
required in some places because _glapi_Dispatch is declared there, but
_glapi_tls_Dispatch is declared in glthread.h.

18 years agoIn _mesa_pack_rgba_span_float() we don't need to make a temporary copy of
Brian Paul [Fri, 13 Oct 2006 16:34:25 +0000 (16:34 +0000)]
In _mesa_pack_rgba_span_float() we don't need to make a temporary copy of
incoming colors when applying pixel transfer ops.  In all cases, the
caller either indicates there's no pixel transfer ops, or the incoming
colors are coming from temporary storage already and can be safely
modified.

18 years agos/GLuint/GLbitfield/
Brian Paul [Fri, 13 Oct 2006 15:11:47 +0000 (15:11 +0000)]
s/GLuint/GLbitfield/