Merge branch 'shader-file-reorg'

1. Move all GL entrypoint functions and files into src/mesa/main/
This includes the ARB vp/vp, NV vp/fp, ATI fragshader and GLSL bits
that were in src/mesa/shader/

2. Move src/mesa/shader/slang/ to src/mesa/slang/ to reduce the tree depth

3. Rename src/mesa/shader/ to src/mesa/program/ since all the
remaining files are concerned with GPU programs.

4. Misc code refactoring.  In particular, I got rid of most of the
GLSL-related ctx->Driver hook functions.  None of the drivers used
them.

Conflicts:
	src/mesa/drivers/dri/i965/brw_context.c

10 files changed, 7327 insertions, 0 deletions

diff --git a/src/mesa/slang/library/.gitignore b/src/mesa/slang/library/.gitignore
new file mode 100644
index 0000000000..02a89fc7df
--- /dev/null
+++ b/src/mesa/slang/library/.gitignore
@@ -0,0 +1 @@
+*_gc.h
diff --git a/src/mesa/slang/library/Makefile b/src/mesa/slang/library/Makefile
new file mode 100644
index 0000000000..5a76774208
--- /dev/null
+++ b/src/mesa/slang/library/Makefile
@@ -0,0 +1,51 @@
+# src/mesa/slang/library/Makefile
+
+TOP = ../../../..
+
+include $(TOP)/configs/current
+
+GLSL_CL = $(TOP)/src/glsl/apps/compile
+
+#
+# targets
+#
+
+.PHONY: default clean
+
+default: builtin
+
+clean:
+	-rm -f *_gc.h
+
+builtin: builtin_110 builtin_120
+
+#
+# builtin library sources
+#
+
+builtin_110: slang_common_builtin_gc.h slang_core_gc.h slang_fragment_builtin_gc.h slang_vertex_builtin_gc.h
+
+builtin_120: slang_120_core_gc.h slang_builtin_120_common_gc.h slang_builtin_120_fragment_gc.h
+
+
+slang_120_core_gc.h: slang_120_core.gc
+	$(GLSL_CL) fragment slang_120_core.gc slang_120_core_gc.h
+
+slang_builtin_120_common_gc.h: slang_builtin_120_common.gc
+	$(GLSL_CL) fragment slang_builtin_120_common.gc slang_builtin_120_common_gc.h
+
+slang_builtin_120_fragment_gc.h: slang_builtin_120_fragment.gc
+	$(GLSL_CL) fragment slang_builtin_120_fragment.gc slang_builtin_120_fragment_gc.h
+
+slang_common_builtin_gc.h: slang_common_builtin.gc
+	$(GLSL_CL) fragment slang_common_builtin.gc slang_common_builtin_gc.h
+
+slang_core_gc.h: slang_core.gc
+	$(GLSL_CL) fragment slang_core.gc slang_core_gc.h
+
+slang_fragment_builtin_gc.h: slang_fragment_builtin.gc
+	$(GLSL_CL) fragment slang_fragment_builtin.gc slang_fragment_builtin_gc.h
+
+slang_vertex_builtin_gc.h: slang_vertex_builtin.gc
+	$(GLSL_CL) vertex slang_vertex_builtin.gc slang_vertex_builtin_gc.h
+
diff --git a/src/mesa/slang/library/SConscript b/src/mesa/slang/library/SConscript
new file mode 100644
index 0000000000..792a7953d3
--- /dev/null
+++ b/src/mesa/slang/library/SConscript
@@ -0,0 +1,52 @@
+#######################################################################
+# SConscript for GLSL builtin library
+
+Import('*')
+
+env = env.Clone()
+
+# See also http://www.scons.org/wiki/UsingCodeGenerators
+
+def glsl_compile_emitter(target, source, env):
+	env.Depends(target, glsl_compile)
+	return (target, source)
+ 
+bld_frag = Builder(
+	action = Action(glsl_compile[0].abspath + ' fragment $SOURCE $TARGET', '$CODEGENCODESTR'),
+	emitter = glsl_compile_emitter,
+	suffix = '.gc',
+	src_suffix = '_gc.h')
+	
+bld_vert = Builder(
+	action = Action(glsl_compile[0].abspath + ' vertex $SOURCE $TARGET', '$CODEGENCODESTR'),
+	emitter = glsl_compile_emitter,
+	suffix = '.gc',
+	src_suffix = '_gc.h')
+
+env['BUILDERS']['bld_frag'] = bld_frag
+env['BUILDERS']['bld_vert'] = bld_vert
+
+# Generate GLSL builtin library binaries
+env.bld_frag(
+	'#src/mesa/slang/library/slang_core_gc.h',
+	'#src/mesa/slang/library/slang_core.gc')
+env.bld_frag(
+	'#src/mesa/slang/library/slang_common_builtin_gc.h',
+	'#src/mesa/slang/library/slang_common_builtin.gc')
+env.bld_frag(
+	'#src/mesa/slang/library/slang_fragment_builtin_gc.h',
+	'#src/mesa/slang/library/slang_fragment_builtin.gc')
+env.bld_vert(
+	'#src/mesa/slang/library/slang_vertex_builtin_gc.h',
+	'#src/mesa/slang/library/slang_vertex_builtin.gc')
+
+# Generate GLSL 1.20 builtin library binaries
+env.bld_frag(
+	'#src/mesa/slang/library/slang_120_core_gc.h',
+	'#src/mesa/slang/library/slang_120_core.gc')
+env.bld_frag(
+	'#src/mesa/slang/library/slang_builtin_120_common_gc.h',
+	'#src/mesa/slang/library/slang_builtin_120_common.gc')
+env.bld_frag(
+	'#src/mesa/slang/library/slang_builtin_120_fragment_gc.h',
+	'#src/mesa/slang/library/slang_builtin_120_fragment.gc')
diff --git a/src/mesa/slang/library/slang_120_core.gc b/src/mesa/slang/library/slang_120_core.gc
new file mode 100644
index 0000000000..04c5ec2ec5
--- /dev/null
+++ b/src/mesa/slang/library/slang_120_core.gc
@@ -0,0 +1,1978 @@
+/*
+ * Mesa 3-D graphics library
+ * Version:  6.5
+ *
+ * Copyright (C) 2006  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+//
+// Constructors and operators introduced in GLSL 1.20 - mostly on new
+// (non-square) types of matrices.
+//
+// One important change in the language is that when a matrix is used
+// as an argument to a matrix constructor, it must be the only argument
+// for the constructor. The compiler takes care of it by itself and
+// here we only care to re-introduce constructors for old (square)
+// types of matrices.
+//
+
+//
+// From Shader Spec, ver. 1.20, rev. 6
+//
+
+//// mat2x3: 2 columns of vec3
+
+mat2x3 __constructor(const float f00, const float f10, const float f20,
+                     const float f01, const float f11, const float f21)
+{
+   __retVal[0].x = f00;
+   __retVal[0].y = f10;
+   __retVal[0].z = f20;
+   __retVal[1].x = f01;
+   __retVal[1].y = f11;
+   __retVal[1].z = f21;
+}
+
+mat2x3 __constructor(const float f)
+{
+   __retVal = mat2x3(  f, 0.0, 0.0,
+                     0.0,   f, 0.0);
+}
+
+mat2x3 __constructor(const int  i)
+{
+   const float f = float(i);
+   __retVal = mat2x3(f);
+}
+
+mat2x3 __constructor(const bool b)
+{
+   const float f = float(b);
+   __retVal = mat2x3(f);
+}
+
+mat2x3 __constructor(const vec3 c0, const vec3 c1)
+{
+   __retVal[0] = c0;
+   __retVal[1] = c1;
+}
+
+
+
+//// mat2x4: 2 columns of vec4
+
+mat2x4 __constructor(const float f00, const float f10, const float f20, const float f30,
+                     const float f01, const float f11, const float f21, const float f31)
+{
+   __retVal[0].x = f00;
+   __retVal[0].y = f10;
+   __retVal[0].z = f20;
+   __retVal[0].w = f30;
+   __retVal[1].x = f01;
+   __retVal[1].y = f11;
+   __retVal[1].z = f21;
+   __retVal[1].w = f31;
+}
+
+mat2x4 __constructor(const float f)
+{
+   __retVal = mat2x4(  f, 0.0, 0.0, 0.0,
+                     0.0,   f, 0.0, 0.0);
+}
+
+mat2x4 __constructor(const int i)
+{
+   const float f = float(i);
+   __retVal = mat2x4(f);
+}
+
+mat2x4 __constructor(const bool b)
+{
+   const float f = float(b);
+   __retVal = mat2x4(f);
+}
+
+mat2x4 __constructor(const vec4 c0, const vec4 c1)
+{
+   __retVal[0] = c0;
+   __retVal[1] = c1;
+}
+
+
+
+//// mat3x2: 3 columns of vec2
+
+mat3x2 __constructor(const float f00, const float f10,
+                     const float f01, const float f11,
+                     const float f02, const float f12)
+{
+   __retVal[0].x = f00;
+   __retVal[0].y = f10;
+   __retVal[1].x = f01;
+   __retVal[1].y = f11;
+   __retVal[2].x = f02;
+   __retVal[2].y = f12;
+}
+
+mat3x2 __constructor(const float f)
+{
+   __retVal = mat3x2(  f, 0.0,
+                     0.0,   f,
+                     0.0, 0.0);
+}
+
+mat3x2 __constructor(const int i)
+{
+   const float f = float(i);
+   __retVal = mat3x2(f);
+}
+
+mat3x2 __constructor(const bool b)
+{
+   const float f = float(b);
+   __retVal = mat3x2(f);
+}
+
+mat3x2 __constructor(const vec2 c0, const vec2 c1, const vec2 c2)
+{
+   __retVal[0] = c0;
+   __retVal[1] = c1;
+   __retVal[2] = c2;
+}
+
+
+
+//// mat3x4: 3 columns of vec4
+
+mat3x4 __constructor(const float f00, const float f10, const float f20, const float f30,
+                     const float f01, const float f11, const float f21, const float f31,
+		     const float f02, const float f12, const float f22, const float f32)
+{
+   __retVal[0].x = f00;
+   __retVal[0].y = f10;
+   __retVal[0].z = f20;
+   __retVal[0].w = f30;
+   __retVal[1].x = f01;
+   __retVal[1].y = f11;
+   __retVal[1].z = f21;
+   __retVal[1].w = f31;
+   __retVal[2].x = f02;
+   __retVal[2].y = f12;
+   __retVal[2].z = f22;
+   __retVal[2].w = f32;
+}
+
+mat3x4 __constructor(const float f)
+{
+   __retVal = mat3x4(  f, 0.0, 0.0, 0.0,
+                     0.0,   f, 0.0, 0.0,
+                     0.0, 0.0,   f, 0.0);
+}
+
+mat3x4 __constructor(const int i)
+{
+   const float f = float(i);
+   __retVal = mat3x4(f);
+}
+
+mat3x4 __constructor(const bool b)
+{
+   const float f = float(b);
+   __retVal = mat3x4(f);
+}
+
+mat3x4 __constructor(const vec4 c0, const vec4 c1, const vec4 c2)
+{
+   __retVal[0] = c0;
+   __retVal[1] = c1;
+   __retVal[2] = c2;
+}
+
+
+
+//// mat4x2: 4 columns of vec2
+
+mat4x2 __constructor(const float f00, const float f10,
+                     const float f01, const float f11,
+		     const float f02, const float f12,
+		     const float f03, const float f13)
+{
+   __retVal[0].x = f00;
+   __retVal[0].y = f10;
+   __retVal[1].x = f01;
+   __retVal[1].y = f11;
+   __retVal[2].x = f02;
+   __retVal[2].y = f12;
+   __retVal[3].x = f03;
+   __retVal[3].y = f13;
+}
+
+mat4x2 __constructor(const float f)
+{
+   __retVal = mat4x2(  f, 0.0,
+                     0.0,   4,
+                     0.0, 0.0,
+                     0.0, 0.0);
+}
+
+mat4x2 __constructor(const int i)
+{
+   const float f = float(i);
+   __retVal = mat4x2(f);
+}
+
+mat4x2 __constructor(const bool b)
+{
+   const float f = float(b);
+   __retVal = mat4x2(f);
+}
+
+mat4x2 __constructor(const vec2 c0, const vec2 c1, const vec2 c2, const vec2 c3)
+{
+   __retVal[0] = c0;
+   __retVal[1] = c1;
+   __retVal[2] = c2;
+   __retVal[3] = c3;
+}
+
+
+
+//// mat4x3: 4 columns of vec3
+
+mat4x3 __constructor(const float f00, const float f10, const float f20,
+                     const float f01, const float f11, const float f21,
+		     const float f02, const float f12, const float f22,
+		     const float f03, const float f13, const float f23)
+{
+   __retVal[0].x = f00;
+   __retVal[0].y = f10;
+   __retVal[0].z = f20;
+   __retVal[1].x = f01;
+   __retVal[1].y = f11;
+   __retVal[1].z = f21;
+   __retVal[2].x = f02;
+   __retVal[2].y = f12;
+   __retVal[2].z = f22;
+   __retVal[3].x = f03;
+   __retVal[3].y = f13;
+   __retVal[3].z = f23;
+}
+
+mat4x3 __constructor(const float f)
+{
+   __retVal = mat4x3(  f, 0.0, 0.0,
+                     0.0,   f, 0.0,
+                     0.0, 0.0,   f,
+                     0.0, 0.0, 0.0);
+}
+
+mat4x3 __constructor(const int i)
+{
+   const float f = float(i);
+   __retVal = mat4x3(f);
+}
+
+mat4x3 __constructor(const bool b)
+{
+   const float f = float(b);
+   __retVal = mat4x3(f);
+}
+
+mat4x3 __constructor(const vec3 c0, const vec3 c1, const vec3 c2, const vec3 c3)
+{
+   __retVal[0] = c0;
+   __retVal[1] = c1;
+   __retVal[2] = c2;
+   __retVal[3] = c3;
+}
+
+
+
+//// misc assorted matrix constructors
+
+mat2 __constructor(const mat2 m)
+{
+   __retVal = m;
+}
+
+mat2 __constructor(const mat3x2 m)
+{
+   __retVal = mat2(m[0], m[1]);
+}
+
+mat2 __constructor(const mat4x2 m)
+{
+   __retVal = mat2(m[0], m[1]);
+}
+
+mat2 __constructor(const mat2x3 m)
+{
+   __retVal = mat2(m[0].xy, m[1].xy);
+}
+
+mat2 __constructor(const mat2x4 m)
+{
+   __retVal = mat2(m[0].xy, m[1].xy);
+}
+
+mat2 __constructor(const mat3 m)
+{
+   __retVal = mat2(m[0].xy, m[1].xy);
+}
+
+mat2 __constructor(const mat3x4 m)
+{
+   __retVal = mat2(m[0].xy, m[1].xy);
+}
+
+mat2 __constructor(const mat4x3 m)
+{
+   __retVal = mat2(m[0].xy, m[1].xy);
+}
+
+mat2 __constructor(const mat4 m)
+{
+   __retVal = mat2(m[0].xy, m[1].xy);
+}
+
+
+
+mat2x3 __constructor(const mat2x3 m)
+{
+   __retVal = m;
+}
+
+mat2x3 __constructor(const mat3 m)
+{
+   __retVal = mat2x3(m[0], m[1]);
+}
+
+mat2x3 __constructor(const mat4x3 m)
+{
+   __retVal = mat2x3(m[0], m[1]);
+}
+
+mat2x3 __constructor(const mat2x4 m)
+{
+   __retVal = mat2x3(m[0].xyz, m[1].xyz);
+}
+
+mat2x3 __constructor(const mat3x4 m)
+{
+   __retVal = mat2x3(m[0].xyz, m[1].xyz);
+}
+
+mat2x3 __constructor(const mat4 m)
+{
+   __retVal = mat2x3(m[0].xyz, m[1].xyz);
+}
+
+mat2x3 __constructor(const mat2 m)
+{
+   __retVal = mat2x3(m[0].x, m[0].y, 0.0,
+                     m[1].x, m[1].y, 0.0);
+}
+
+mat2x3 __constructor(const mat3x2 m)
+{
+   __retVal = mat2x3(m[0].x, m[0].y, 0.0,
+                     m[1].x, m[1].y, 0.0);
+}
+
+mat2x3 __constructor(const mat4x2 m)
+{
+   __retVal = mat2x3(m[0].x, m[0].y, 0.0,
+                     m[1].x, m[1].y, 0.0);
+}
+
+
+
+mat2x4 __constructor(const mat2x4 m)
+{
+   __retVal = m;
+}
+
+mat2x4 __constructor(const mat3x4 m)
+{
+   __retVal = mat2x4(m[0], m[1]);
+}
+
+mat2x4 __constructor(const mat4 m)
+{
+   __retVal = mat2x4(m[0], m[1]);
+}
+
+mat2x4 __constructor(const mat2x3 m)
+{
+   __retVal = mat2x4(m[0].x, m[0].y, m[0].z, 0.0,
+                     m[1].x, m[1].y, m[1].z, 0.0);
+}
+
+mat2x4 __constructor(const mat3 m)
+{
+   __retVal = mat2x4(m[0].x, m[0].y, m[0].z, 0.0,
+                     m[1].x, m[1].y, m[1].z, 0.0);
+}
+
+mat2x4 __constructor(const mat4x3 m)
+{
+   __retVal = mat2x4(m[0].x, m[0].y, m[0].z, 0.0,
+                     m[1].x, m[1].y, m[1].z, 0.0);
+}
+
+mat2x4 __constructor(const mat2 m)
+{
+   __retVal = mat2x4(m[0].x, m[1].y, 0.0, 0.0,
+                     m[1].x, m[1].y, 0.0, 0.0);
+}
+
+mat2x4 __constructor(const mat3x2 m)
+{
+   __retVal = mat2x4(m[0].x, m[0].y, 0.0, 0.0,
+                     m[1].x, m[1].y, 0.0, 0.0);
+}
+
+mat2x4 __constructor(const mat4x2 m)
+{
+   __retVal = mat2x4(m[0].x, m[0].y, 0.0, 0.0,
+                     m[1].x, m[1].y, 0.0, 0.0);
+}
+
+
+
+mat3x2 __constructor(const mat3x2 m)
+{
+   __retVal = m;
+}
+
+mat3x2 __constructor(const mat4x2 m)
+{
+   __retVal = mat3x2(m[0], m[1], m[2]);
+}
+
+mat3x2 __constructor(const mat3 m)
+{
+   __retVal = mat3x2(m[0], m[1], m[2]);
+}
+
+mat3x2 __constructor(const mat3x4 m)
+{
+   __retVal = mat3x2(m[0].x, m[0].y,
+                     m[1].x, m[1].y,
+                     m[2].x, m[2].y);
+}
+
+mat3x2 __constructor(const mat4x3 m)
+{
+   __retVal = mat3x2(m[0].x, m[0].y,
+                     m[1].x, m[1].y,
+                     m[2].x, m[2].y);
+}
+
+mat3x2 __constructor(const mat4 m)
+{
+   __retVal = mat3x2(m[0].x, m[0].y,
+                     m[1].x, m[1].y,
+                        0.0,    0.0);
+}
+
+mat3x2 __constructor(const mat2 m)
+{
+   __retVal = mat3x2(m[0], m[1], vec2(0.0));
+}
+
+mat3x2 __constructor(const mat2x3 m)
+{
+   __retVal = mat3x2(m[0].x, m[0].y,
+                     m[1].x, m[1].y,
+                        0.0,    0.0);
+}
+
+mat3x2 __constructor(const mat2x4 m)
+{
+   __retVal = mat3x2(m[0].x, m[0].y,
+                     m[1].x, m[1].y,
+                        0.0,    0.0);
+}
+
+
+
+
+mat3 __constructor(const mat3 m)
+{
+   __retVal = m;
+}
+
+mat3 __constructor(const mat4x3 m)
+{
+   __retVal = mat3 (
+        m[0],
+        m[1],
+        m[2]
+    );
+}
+
+mat3 __constructor(const mat3x4 m)
+{
+   __retVal = mat3 (
+        m[0].xyz,
+        m[1].xyz,
+        m[2].xyz
+    );
+}
+
+mat3 __constructor(const mat4 m)
+{
+   __retVal = mat3 (
+        m[0].xyz,
+        m[1].xyz,
+        m[2].xyz
+    );
+}
+
+mat3 __constructor(const mat2x3 m)
+{
+   __retVal = mat3 (
+        m[0],
+        m[1],
+        0., 0., 1.
+    );
+}
+
+mat3 __constructor(const mat2x4 m)
+{
+   __retVal = mat3 (
+        m[0].xyz,
+        m[1].xyz,
+        0., 0., 1.
+    );
+}
+
+mat3 __constructor(const mat3x2 m)
+{
+   __retVal = mat3 (
+        m[0], 0.,
+        m[1], 0.,
+        m[2], 1.
+    );
+}
+
+mat3 __constructor(const mat4x2 m)
+{
+   __retVal = mat3 (
+        m[0], 0.,
+        m[1], 0.,
+        m[2], 1.
+    );
+}
+
+mat3 __constructor(const mat2 m)
+{
+   __retVal = mat3 (
+        m[0], 0.,
+        m[1], 0.,
+        0., 0., 1.
+    );
+}
+
+
+mat3x4 __constructor(const mat3x4 m)
+{
+   __retVal = m;
+}
+
+mat3x4 __constructor(const mat4 m)
+{
+   __retVal = mat3x4 (
+        m[0],
+        m[1],
+        m[2]
+    );
+}
+
+mat3x4 __constructor(const mat3 m)
+{
+   __retVal = mat3x4 (
+        m[0], 0.,
+        m[1], 0.,
+        m[2], 0.
+    );
+}
+
+mat3x4 __constructor(const mat4x3 m)
+{
+   __retVal = mat3x4 (
+        m[0], 0.,
+        m[1], 0.,
+        m[2], 0.
+    );
+}
+
+mat3x4 __constructor(const mat2x4 m)
+{
+   __retVal = mat3x4 (
+        m[0],
+        m[1],
+        0., 0., 1., 0.
+    );
+}
+
+mat3x4 __constructor(const mat2x3 m)
+{
+   __retVal = mat3x4 (
+        m[0], 0.,
+        m[1], 0.,
+        0., 0., 1., 0.
+    );
+}
+
+mat3x4 __constructor(const mat3x2 m)
+{
+   __retVal = mat3x4 (
+        m[0], 0., 0.,
+        m[1], 0., 0.,
+        m[2], 1., 0.
+    );
+}
+
+mat3x4 __constructor(const mat4x2 m)
+{
+   __retVal = mat3x4 (
+        m[0], 0., 0.,
+        m[1], 0., 0.,
+        m[2], 1., 0.
+    );
+}
+
+mat3x4 __constructor(const mat2 m)
+{
+   __retVal = mat3x4 (
+        m[0], 0., 0.,
+        m[1], 0., 0.,
+        0., 0., 1., 0.
+    );
+}
+
+
+mat4x2 __constructor(const mat4x2 m)
+{
+   __retVal = m;
+}
+
+mat4x2 __constructor(const mat4x3 m)
+{
+   __retVal = mat4x2 (
+        m[0].xy,
+        m[1].xy,
+        m[2].xy,
+        m[3].xy
+    );
+}
+
+mat4x2 __constructor(const mat4 m)
+{
+   __retVal = mat4x2 (
+        m[0].xy,
+        m[1].xy,
+        m[2].xy,
+        m[3].xy
+    );
+}
+
+mat4x2 __constructor(const mat3x2 m)
+{
+   __retVal = mat4x2 (
+        m[0],
+        m[1],
+        0., 0.
+    );
+}
+
+mat4x2 __constructor(const mat3 m)
+{
+   __retVal = mat4x2 (
+        m[0].xy,
+        m[1].xy,
+        m[2].xy,
+        0., 0.
+    );
+}
+
+mat4x2 __constructor(const mat3x4 m)
+{
+   __retVal = mat4x2 (
+        m[0].xy,
+        m[1].xy,
+        m[2].xy,
+        0., 0.
+    );
+}
+
+mat4x2 __constructor(const mat2 m)
+{
+   __retVal = mat4x2 (
+        m[0],
+        m[1],
+        0., 0.,
+        0., 0.
+    );
+}
+
+mat4x2 __constructor(const mat2x3 m)
+{
+   __retVal = mat4x2 (
+        m[0].xy,
+        m[1].xy,
+        0., 0.,
+        0., 0.
+    );
+}
+
+mat4x2 __constructor(const mat2x4 m)
+{
+   __retVal = mat4x2 (
+        m[0].xy,
+        m[1].xy,
+        0., 0.,
+        0., 0.
+    );
+}
+
+
+mat4x3 __constructor(const mat4x3 m)
+{
+   __retVal = m;
+}
+
+mat4x3 __constructor(const mat4 m)
+{
+   __retVal = mat4x3 (
+        m[0].xyz,
+        m[1].xyz,
+        m[2].xyz,
+        m[3].xyz
+    );
+}
+
+mat4x3 __constructor(const mat3 m)
+{
+   __retVal = mat4x3 (
+        m[0],
+        m[1],
+        m[2],
+        0., 0., 0.
+    );
+}
+
+mat4x3 __constructor(const mat3x4 m)
+{
+   __retVal = mat4x3 (
+        m[0].xyz,
+        m[1].xyz,
+        m[2].xyz,
+        0., 0., 0.
+    );
+}
+
+mat4x3 __constructor(const mat4x2 m)
+{
+   __retVal = mat4x3 (
+        m[0], 0.,
+        m[1], 0.,
+        m[2], 1.,
+        m[3], 0.
+    );
+}
+
+mat4x3 __constructor(const mat2x3 m)
+{
+   __retVal = mat4x3 (
+        m[0],
+        m[1],
+        0., 0., 1.,
+        0., 0., 0.
+    );
+}
+
+mat4x3 __constructor(const mat3x2 m)
+{
+   __retVal = mat4x3 (
+        m[0], 0.,
+        m[1], 0.,
+        m[2], 1.,
+        0., 0., 0.
+    );
+}
+
+mat4x3 __constructor(const mat2x4 m)
+{
+   __retVal = mat4x3 (
+        m[0].xyz,
+        m[1].xyz,
+        0., 0., 1.,
+        0., 0., 0.
+    );
+}
+
+mat4x3 __constructor(const mat2 m)
+{
+   __retVal = mat4x3 (
+        m[0], 0.,
+        m[1], 0.,
+        0., 0., 1.,
+        0., 0., 0.
+    );
+}
+
+
+mat4 __constructor(const mat4 m)
+{
+   __retVal = m;
+}
+
+mat4 __constructor(const mat3x4 m)
+{
+   __retVal = mat4 (
+        m[0],
+        m[1],
+        m[2],
+        0., 0., 0., 1.
+    );
+}
+
+mat4 __constructor(const mat4x3 m)
+{
+   __retVal = mat4 (
+        m[0], 0.,
+        m[1], 0.,
+        m[2], 0.,
+        m[3], 1.
+    );
+}
+
+mat4 __constructor(const mat2x4 m)
+{
+   __retVal = mat4 (
+        m[0],
+        m[1],
+        0., 0., 1., 0.,
+        0., 0., 0., 1.
+    );
+}
+
+mat4 __constructor(const mat4x2 m)
+{
+   __retVal = mat4 (
+        m[0], 0., 0.,
+        m[1], 0., 0.,
+        m[2], 1., 0.,
+        m[3], 0., 1.
+    );
+}
+
+mat4 __constructor(const mat3 m)
+{
+   __retVal = mat4 (
+        m[0], 0.,
+        m[1], 0.,
+        m[2], 0.,
+        0., 0., 0., 1.
+    );
+}
+
+mat4 __constructor(const mat2x3 m)
+{
+   __retVal = mat4 (
+        m[0], 0.,
+        m[1], 0.,
+        0., 0., 1., 0.,
+        0., 0., 0., 1.
+    );
+}
+
+mat4 __constructor(const mat3x2 m)
+{
+   __retVal = mat4 (
+        m[0], 0., 0.,
+        m[1], 0., 0.,
+        m[2], 1., 0.,
+        0., 0., 0., 1.
+    );
+}
+
+mat4 __constructor(const mat2 m)
+{
+   __retVal = mat4 (
+        m[0], 0., 0.,
+        m[1], 0., 0.,
+        0., 0., 1., 0.,
+        0., 0., 0., 1.
+    );
+}
+
+
+void __operator += (inout mat2x3 m, const mat2x3 n) {
+    m[0] += n[0];
+    m[1] += n[1];
+}
+
+void __operator += (inout mat2x4 m, const mat2x4 n) {
+    m[0] += n[0];
+    m[1] += n[1];
+}
+
+void __operator += (inout mat3x2 m, const mat3x2 n) {
+    m[0] += n[0];
+    m[1] += n[1];
+    m[2] += n[2];
+}
+
+void __operator += (inout mat3x4 m, const mat3x4 n) {
+    m[0] += n[0];
+    m[1] += n[1];
+    m[2] += n[2];
+}
+
+void __operator += (inout mat4x2 m, const mat4x2 n) {
+    m[0] += n[0];
+    m[1] += n[1];
+    m[2] += n[2];
+    m[3] += n[3];
+}
+
+void __operator += (inout mat4x3 m, const mat4x3 n) {
+    m[0] += n[0];
+    m[1] += n[1];
+    m[2] += n[2];
+    m[3] += n[3];
+}
+
+
+void __operator -= (inout mat2x3 m, const mat2x3 n) {
+    m[0] -= n[0];
+    m[1] -= n[1];
+}
+
+void __operator -= (inout mat2x4 m, const mat2x4 n) {
+    m[0] -= n[0];
+    m[1] -= n[1];
+}
+
+void __operator -= (inout mat3x2 m, const mat3x2 n) {
+    m[0] -= n[0];
+    m[1] -= n[1];
+    m[2] -= n[2];
+}
+
+void __operator -= (inout mat3x4 m, const mat3x4 n) {
+    m[0] -= n[0];
+    m[1] -= n[1];
+    m[2] -= n[2];
+}
+
+void __operator -= (inout mat4x2 m, const mat4x2 n) {
+    m[0] -= n[0];
+    m[1] -= n[1];
+    m[2] -= n[2];
+    m[3] -= n[3];
+}
+
+void __operator -= (inout mat4x3 m, const mat4x3 n) {
+    m[0] -= n[0];
+    m[1] -= n[1];
+    m[2] -= n[2];
+    m[3] -= n[3];
+}
+
+
+void __operator /= (inout mat2x3 m, const mat2x3 n) {
+    m[0] /= n[0];
+    m[1] /= n[1];
+}
+
+void __operator /= (inout mat2x4 m, const mat2x4 n) {
+    m[0] /= n[0];
+    m[1] /= n[1];
+}
+
+void __operator /= (inout mat3x2 m, const mat3x2 n) {
+    m[0] /= n[0];
+    m[1] /= n[1];
+    m[2] /= n[2];
+}
+
+void __operator /= (inout mat3x4 m, const mat3x4 n) {
+    m[0] /= n[0];
+    m[1] /= n[1];
+    m[2] /= n[2];
+}
+
+void __operator /= (inout mat4x2 m, const mat4x2 n) {
+    m[0] /= n[0];
+    m[1] /= n[1];
+    m[2] /= n[2];
+    m[3] /= n[3];
+}
+
+void __operator /= (inout mat4x3 m, const mat4x3 n) {
+    m[0] /= n[0];
+    m[1] /= n[1];
+    m[2] /= n[2];
+    m[3] /= n[3];
+}
+
+
+vec3 __operator * (const mat2x3 m, const vec2 v)
+{
+   __retVal.x = v.x * m[0].x + v.y * m[1].x;
+   __retVal.y = v.x * m[0].y + v.y * m[1].y;
+   __retVal.z = v.x * m[0].z + v.y * m[1].z;
+}
+
+vec4 __operator * (const mat2x4 m, const vec2 v)
+{
+   __retVal.x = v.x * m[0].x + v.y * m[1].x;
+   __retVal.y = v.x * m[0].y + v.y * m[1].y;
+   __retVal.z = v.x * m[0].z + v.y * m[1].z;
+   __retVal.w = v.x * m[0].w + v.y * m[1].w;
+}
+
+vec2 __operator * (const mat3x2 m, const vec3 v)
+{
+   __retVal.x = v.x * m[0].x + v.y * m[1].x + v.z * m[2].x;
+   __retVal.y = v.x * m[0].y + v.y * m[1].y + v.z * m[2].y;
+}
+
+vec4 __operator * (const mat3x4 m, const vec3 v)
+{
+   __retVal.x = v.x * m[0].x + v.y * m[1].x + v.z * m[2].x;
+   __retVal.y = v.x * m[0].y + v.y * m[1].y + v.z * m[2].y;
+   __retVal.z = v.x * m[0].z + v.y * m[1].z + v.z * m[2].z;
+   __retVal.w = v.x * m[0].w + v.y * m[1].w + v.z * m[2].w;
+}
+
+vec2 __operator * (const mat4x2 m, const vec4 v)
+{
+   __retVal.x = v.x * m[0].x + v.y * m[1].x + v.z * m[2].x + v.w * m[3].x;
+   __retVal.y = v.x * m[0].y + v.y * m[1].y + v.z * m[2].y + v.w * m[3].y;
+}
+
+vec3 __operator * (const mat4x3 m, const vec4 v)
+{
+   __retVal.x = v.x * m[0].x + v.y * m[1].x + v.z * m[2].x + v.w * m[3].x;
+   __retVal.y = v.x * m[0].y + v.y * m[1].y + v.z * m[2].y + v.w * m[3].y;
+   __retVal.z = v.x * m[0].z + v.y * m[1].z + v.z * m[2].z + v.w * m[3].z;
+}
+
+
+mat3x2 __operator * (const mat2 m, const mat3x2 n)
+{
+   //return mat3x2 (m * n[0], m * n[1], m * n[2]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+}
+
+mat4x2 __operator * (const mat2 m, const mat4x2 n)
+{
+   //return mat4x2 (m * n[0], m * n[1], m * n[2], m * n[3]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+   __retVal[3] = m * n[3];
+}
+
+
+mat2x3 __operator * (const mat2x3 m, const mat2 n)
+{
+   //return mat2x3 (m * n[0], m * n[1]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+}
+
+mat3 __operator * (const mat2x3 m, const mat3x2 n)
+{
+   //return mat3 (m * n[0], m * n[1], m * n[2]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+}
+
+mat4x3 __operator * (const mat2x3 m, const mat4x2 n)
+{
+   //return mat4x3 (m * n[0], m * n[1], m * n[2], m * n[3]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+   __retVal[3] = m * n[3];
+}
+
+
+mat2x4 __operator * (const mat2x4 m, const mat2 n)
+{
+   //return mat2x4 (m * n[0], m * n[1]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+}
+
+mat3x4 __operator * (const mat2x4 m, const mat3x2 n)
+{
+   //return mat3x4 (m * n[0], m * n[1], m * n[2]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+}
+
+mat4 __operator * (const mat2x4 m, const mat4x2 n)
+{
+   //return mat4 (m * n[0], m * n[1], m * n[2], m * n[3]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+   __retVal[3] = m * n[3];
+}
+
+
+mat2 __operator * (const mat3x2 m, const mat2x3 n)
+{
+   //return mat2 (m * n[0], m * n[1]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+}
+
+mat3x2 __operator * (const mat3x2 m, const mat3 n)
+{
+   //return mat3x2 (m * n[0], m * n[1], m * n[2]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+}
+
+mat4x2 __operator * (const mat3x2 m, const mat4x3 n)
+{
+   //return mat4x2 (m * n[0], m * n[1], m * n[2], m * n[3]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+   __retVal[3] = m * n[3];
+}
+
+
+mat2x3 __operator * (const mat3 m, const mat2x3 n)
+{
+   //return mat2x3 (m * n[0], m * n[1]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+}
+
+mat4x3 __operator * (const mat3 m, const mat4x3 n)
+{
+   //return mat4x3 (m * n[0], m * n[1], m * n[2], m * n[3]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+   __retVal[3] = m * n[3];
+}
+
+
+mat2x4 __operator * (const mat3x4 m, const mat2x3 n)
+{
+   //return mat2x4 (m * n[0], m * n[1]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+}
+
+mat3x4 __operator * (const mat3x4 m, const mat3 n)
+{
+   //return mat3x4 (m * n[0], m * n[1], m * n[2]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+}
+
+mat4 __operator * (const mat3x4 m, const mat4x3 n)
+{
+   //return mat4 (m * n[0], m * n[1], m * n[2], m * n[3]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+   __retVal[3] = m * n[3];
+}
+
+
+mat2 __operator * (const mat4x2 m, const mat2x4 n)
+{
+   //return = mat2 (m * n[0], m * n[1]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+}
+
+mat3x2 __operator * (const mat4x2 m, const mat3x4 n)
+{
+   //return mat3x2 (m * n[0], m * n[1], m * n[2]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+}
+
+mat4x2 __operator * (const mat4x2 m, const mat4 n)
+{
+   //return mat4x2 (m * n[0], m * n[1], m * n[2], m * n[3]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+   __retVal[3] = m * n[3];
+}
+
+
+mat2x3 __operator * (const mat4x3 m, const mat2x4 n)
+{
+   //return mat2x3 (m * n[0], m * n[1]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+}
+
+mat3 __operator * (const mat4x3 m, const mat3x4 n)
+{
+   //return mat3 (m * n[0], m * n[1], m * n[2]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+}
+
+mat4x3 __operator * (const mat4x3 m, const mat4 n)
+{
+   //return mat4x3 (m * n[0], m * n[1], m * n[2], m * n[3]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+   __retVal[3] = m * n[3];
+}
+
+
+mat2x4 __operator * (const mat4 m, const mat2x4 n)
+{
+   //return mat2x4 (m * n[0], m * n[1]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+}
+
+mat3x4 __operator * (const mat4 m, const mat3x4 n)
+{
+   //return mat3x4 (m * n[0], m * n[1], m * n[2]);
+   __retVal[0] = m * n[0];
+   __retVal[1] = m * n[1];
+   __retVal[2] = m * n[2];
+}
+
+
+void __operator *= (inout mat2x3 m, const mat2 n) {
+    m = m * n;
+}
+
+void __operator *= (inout mat2x4 m, const mat2 n) {
+    m = m * n;
+}
+
+void __operator *= (inout mat3x2 m, const mat3 n) {
+    m = m * n;
+}
+
+void __operator *= (inout mat3x4 m, const mat3 n) {
+    m = m * n;
+}
+
+void __operator *= (inout mat4x2 m, const mat4 n) {
+    m = m * n;
+}
+
+void __operator *= (inout mat4x3 m, const mat4 n) {
+    m = m * n;
+}
+
+
+vec3 __operator * (const vec2 v, const mat3x2 m)
+{
+   __retVal.x = dot(v, m[0]);
+   __retVal.y = dot(v, m[1]);
+   __retVal.z = dot(v, m[2]);
+}
+
+vec4 __operator * (const vec2 v, const mat4x2 m)
+{
+   __retVal.x = dot(v, m[0]);
+   __retVal.y = dot(v, m[1]);
+   __retVal.z = dot(v, m[2]);
+   __retVal.w = dot(v, m[3]);
+}
+
+vec2 __operator * (const vec3 v, const mat2x3 m)
+{
+   __retVal.x = dot(v, m[0]);
+   __retVal.y = dot(v, m[1]);
+}
+
+vec4 __operator * (const vec3 v, const mat4x3 m)
+{
+   __retVal.x = dot(v, m[0]);
+   __retVal.y = dot(v, m[1]);
+   __retVal.z = dot(v, m[2]);
+   __retVal.w = dot(v, m[3]);
+}
+
+vec2 __operator * (const vec4 v, const mat2x4 m)
+{
+   __retVal.x = dot(v, m[0]);
+   __retVal.y = dot(v, m[1]);
+}
+
+vec3 __operator * (const vec4 v, const mat3x4 m)
+{
+   __retVal.x = dot(v, m[0]);
+   __retVal.y = dot(v, m[1]);
+   __retVal.z = dot(v, m[2]);
+}
+
+
+void __operator += (inout mat2x3 m, const float a) {
+    m[0] += a;
+    m[1] += a;
+}
+
+void __operator += (inout mat2x4 m, const float a) {
+    m[0] += a;
+    m[1] += a;
+}
+
+void __operator += (inout mat3x2 m, const float a) {
+    m[0] += a;
+    m[1] += a;
+    m[2] += a;
+}
+
+void __operator += (inout mat3x4 m, const float a) {
+    m[0] += a;
+    m[1] += a;
+    m[2] += a;
+}
+
+void __operator += (inout mat4x2 m, const float a) {
+    m[0] += a;
+    m[1] += a;
+    m[2] += a;
+    m[3] += a;
+}
+
+void __operator += (inout mat4x3 m, const float a) {
+    m[0] += a;
+    m[1] += a;
+    m[2] += a;
+    m[3] += a;
+}
+
+
+void __operator -= (inout mat2x3 m, const float a) {
+    m[0] -= a;
+    m[1] -= a;
+}
+
+void __operator -= (inout mat2x4 m, const float a) {
+    m[0] -= a;
+    m[1] -= a;
+}
+
+void __operator -= (inout mat3x2 m, const float a) {
+    m[0] -= a;
+    m[1] -= a;
+    m[2] -= a;
+}
+
+void __operator -= (inout mat3x4 m, const float a) {
+    m[0] -= a;
+    m[1] -= a;
+    m[2] -= a;
+}
+
+void __operator -= (inout mat4x2 m, const float a) {
+    m[0] -= a;
+    m[1] -= a;
+    m[2] -= a;
+    m[3] -= a;
+}
+
+void __operator -= (inout mat4x3 m, const float a) {
+    m[0] -= a;
+    m[1] -= a;
+    m[2] -= a;
+    m[3] -= a;
+}
+
+
+void __operator *= (inout mat2x3 m, const float a) {
+    m[0] *= a;
+    m[1] *= a;
+}
+
+void __operator *= (inout mat2x4 m, const float a) {
+    m[0] *= a;
+    m[1] *= a;
+}
+
+void __operator *= (inout mat3x2 m, const float a) {
+    m[0] *= a;
+    m[1] *= a;
+    m[2] *= a;
+}
+
+void __operator *= (inout mat3x4 m, const float a) {
+    m[0] *= a;
+    m[1] *= a;
+    m[2] *= a;
+}
+
+void __operator *= (inout mat4x2 m, const float a) {
+    m[0] *= a;
+    m[1] *= a;
+    m[2] *= a;
+    m[3] *= a;
+}
+
+void __operator *= (inout mat4x3 m, const float a) {
+    m[0] *= a;
+    m[1] *= a;
+    m[2] *= a;
+    m[3] *= a;
+}
+
+
+void __operator /= (inout mat2x3 m, const float a) {
+    m[0] /= a;
+    m[1] /= a;
+}
+
+void __operator /= (inout mat2x4 m, const float a) {
+    m[0] /= a;
+    m[1] /= a;
+}
+
+void __operator /= (inout mat3x2 m, const float a) {
+    m[0] /= a;
+    m[1] /= a;
+    m[2] /= a;
+}
+
+void __operator /= (inout mat3x4 m, const float a) {
+    m[0] /= a;
+    m[1] /= a;
+    m[2] /= a;
+}
+
+void __operator /= (inout mat4x2 m, const float a) {
+    m[0] /= a;
+    m[1] /= a;
+    m[2] /= a;
+    m[3] /= a;
+}
+
+void __operator /= (inout mat4x3 m, const float a) {
+    m[0] /= a;
+    m[1] /= a;
+    m[2] /= a;
+    m[3] /= a;
+}
+
+
+mat2x3 __operator + (const mat2x3 m, const mat2x3 n) {
+    return mat2x3 (m[0] + n[0], m[1] + n[1]);
+}
+
+mat2x4 __operator + (const mat2x4 m, const mat2x4 n) {
+    return mat2x4 (m[0] + n[0], m[1] + n[1]);
+}
+
+mat3x2 __operator + (const mat3x2 m, const mat3x2 n) {
+    return mat3x2 (m[0] + n[0], m[1] + n[1], m[2] + n[2]);
+}
+
+mat3x4 __operator + (const mat3x4 m, const mat3x4 n) {
+    return mat3x4 (m[0] + n[0], m[1] + n[1], m[2] + n[2]);
+}
+
+mat4x2 __operator + (const mat4x2 m, const mat4x2 n) {
+    return mat4x2 (m[0] + n[0], m[1] + n[1], m[2] + n[2], m[3] + n[3]);
+}
+
+mat4x3 __operator + (const mat4x3 m, const mat4x3 n) {
+    return mat4x3 (m[0] + n[0], m[1] + n[1], m[2] + n[2], m[3] + n[3]);
+}
+
+
+mat2x3 __operator - (const mat2x3 m, const mat2x3 n) {
+    return mat2x3 (m[0] - n[0], m[1] - n[1]);
+}
+
+mat2x4 __operator - (const mat2x4 m, const mat2x4 n) {
+    return mat2x4 (m[0] - n[0], m[1] - n[1]);
+}
+
+mat3x2 __operator - (const mat3x2 m, const mat3x2 n) {
+    return mat3x2 (m[0] - n[0], m[1] - n[1], m[2] - n[2]);
+}
+
+mat3x4 __operator - (const mat3x4 m, const mat3x4 n) {
+    return mat3x4 (m[0] - n[0], m[1] - n[1], m[2] - n[2]);
+}
+
+mat4x2 __operator - (const mat4x2 m, const mat4x2 n) {
+    return mat4x2 (m[0] - n[0], m[1] - n[1], m[2] - n[2], m[3] - n[3]);
+}
+
+mat4x3 __operator - (const mat4x3 m, const mat4x3 n) {
+    return mat4x3 (m[0] - n[0], m[1] - n[1], m[2] - n[2], m[3] - n[3]);
+}
+
+
+mat2x3 __operator / (const mat2x3 m, const mat2x3 n) {
+    return mat2x3 (m[0] / n[0], m[1] / n[1]);
+}
+
+mat2x4 __operator / (const mat2x4 m, const mat2x4 n) {
+    return mat2x4 (m[0] / n[0], m[1] / n[1]);
+}
+
+mat3x2 __operator / (const mat3x2 m, const mat3x2 n) {
+    return mat3x2 (m[0] / n[0], m[1] / n[1], m[2] / n[2]);
+}
+
+mat3x4 __operator / (const mat3x4 m, const mat3x4 n) {
+    return mat3x4 (m[0] / n[0], m[1] / n[1], m[2] / n[2]);
+}
+
+mat4x2 __operator / (const mat4x2 m, const mat4x2 n) {
+    return mat4x2 (m[0] / n[0], m[1] / n[1], m[2] / n[2], m[3] / n[3]);
+}
+
+mat4x3 __operator / (const mat4x3 m, const mat4x3 n) {
+    return mat4x3 (m[0] / n[0], m[1] / n[1], m[2] / n[2], m[3] / n[3]);
+}
+
+
+mat2x3 __operator + (const float a, const mat2x3 n) {
+    return mat2x3 (a + n[0], a + n[1]);
+}
+
+mat2x3 __operator + (const mat2x3 m, const float b) {
+    return mat2x3 (m[0] + b, m[1] + b);
+}
+
+mat2x4 __operator + (const float a, const mat2x4 n) {
+    return mat2x4 (a + n[0], a + n[1]);
+}
+
+mat2x4 __operator + (const mat2x4 m, const float b) {
+    return mat2x4 (m[0] + b, m[1] + b);
+}
+
+mat3x2 __operator + (const float a, const mat3x2 n) {
+    return mat3x2 (a + n[0], a + n[1], a + n[2]);
+}
+
+mat3x2 __operator + (const mat3x2 m, const float b) {
+    return mat3x2 (m[0] + b, m[1] + b, m[2] + b);
+}
+
+mat3x4 __operator + (const float a, const mat3x4 n) {
+    return mat3x4 (a + n[0], a + n[1], a + n[2]);
+}
+
+mat3x4 __operator + (const mat3x4 m, const float b) {
+    return mat3x4 (m[0] + b, m[1] + b, m[2] + b);
+}
+
+mat4x2 __operator + (const mat4x2 m, const float b) {
+    return mat4x2 (m[0] + b, m[1] + b, m[2] + b, m[3] + b);
+}
+
+mat4x2 __operator + (const float a, const mat4x2 n) {
+    return mat4x2 (a + n[0], a + n[1], a + n[2], a + n[3]);
+}
+
+mat4x3 __operator + (const mat4x3 m, const float b) {
+    return mat4x3 (m[0] + b, m[1] + b, m[2] + b, m[3] + b);
+}
+
+mat4x3 __operator + (const float a, const mat4x3 n) {
+    return mat4x3 (a + n[0], a + n[1], a + n[2], a + n[3]);
+}
+
+
+mat2x3 __operator - (const float a, const mat2x3 n) {
+    return mat2x3 (a - n[0], a - n[1]);
+}
+
+mat2x3 __operator - (const mat2x3 m, const float b) {
+    return mat2x3 (m[0] - b, m[1] - b);
+}
+
+mat2x4 __operator - (const float a, const mat2x4 n) {
+    return mat2x4 (a - n[0], a - n[1]);
+}
+
+mat2x4 __operator - (const mat2x4 m, const float b) {
+    return mat2x4 (m[0] - b, m[1] - b);
+}
+
+mat3x2 __operator - (const float a, const mat3x2 n) {
+    return mat3x2 (a - n[0], a - n[1], a - n[2]);
+}
+
+mat3x2 __operator - (const mat3x2 m, const float b) {
+    return mat3x2 (m[0] - b, m[1] - b, m[2] - b);
+}
+
+mat3x4 __operator - (const float a, const mat3x4 n) {
+    return mat3x4 (a - n[0], a - n[1], a - n[2]);
+}
+
+mat3x4 __operator - (const mat3x4 m, const float b) {
+    return mat3x4 (m[0] - b, m[1] - b, m[2] - b);
+}
+
+mat4x2 __operator - (const mat4x2 m, const float b) {
+    return mat4x2 (m[0] - b, m[1] - b, m[2] - b, m[3] - b);
+}
+
+mat4x2 __operator - (const float a, const mat4x2 n) {
+    return mat4x2 (a - n[0], a - n[1], a - n[2], a - n[3]);
+}
+
+mat4x3 __operator - (const mat4x3 m, const float b) {
+    return mat4x3 (m[0] - b, m[1] - b, m[2] - b, m[3] - b);
+}
+
+mat4x3 __operator - (const float a, const mat4x3 n) {
+    return mat4x3 (a - n[0], a - n[1], a - n[2], a - n[3]);
+}
+
+
+mat2x3 __operator * (const float a, const mat2x3 n)
+{
+   //return mat2x3 (a * n[0], a * n[1]);
+   __retVal[0] = a * n[0];
+   __retVal[1] = a * n[1];
+}
+
+mat2x3 __operator * (const mat2x3 m, const float b)
+{
+   //return mat2x3 (m[0] * b, m[1] * b);
+   __retVal[0] = m[0] * b;
+   __retVal[1] = m[1] * b;
+}
+
+mat2x4 __operator * (const float a, const mat2x4 n)
+{
+   //return mat2x4 (a * n[0], a * n[1]);
+   __retVal[0] = a * n[0];
+   __retVal[1] = a * n[1];
+}
+
+mat2x4 __operator * (const mat2x4 m, const float b)
+{
+   //return mat2x4 (m[0] * b, m[1] * b);
+   __retVal[0] = m[0] * b;
+   __retVal[1] = m[1] * b;
+}
+
+mat3x2 __operator * (const float a, const mat3x2 n)
+{
+   //return mat3x2 (a * n[0], a * n[1], a * n[2]);
+   __retVal[0] = a * n[0];
+   __retVal[1] = a * n[1];
+   __retVal[2] = a * n[2];
+}
+
+mat3x2 __operator * (const mat3x2 m, const float b)
+{
+   //return mat3x2 (m[0] * b, m[1] * b, m[2] * b);
+   __retVal[0] = m[0] * b;
+   __retVal[1] = m[1] * b;
+   __retVal[2] = m[2] * b;
+}
+
+mat3x4 __operator * (const float a, const mat3x4 n)
+{
+   //return mat3x4 (a * n[0], a * n[1], a * n[2]);
+   __retVal[0] = a * n[0];
+   __retVal[1] = a * n[1];
+   __retVal[2] = a * n[2];
+}
+
+mat3x4 __operator * (const mat3x4 m, const float b)
+{
+   //return mat3x4 (m[0] * b, m[1] * b, m[2] * b);
+   __retVal[0] = m[0] * b;
+   __retVal[1] = m[1] * b;
+   __retVal[2] = m[2] * b;
+}
+
+mat4x2 __operator * (const mat4x2 m, const float b)
+{
+   //return mat4x2 (m[0] * b, m[1] * b, m[2] * b, m[3] * b);
+   __retVal[0] = m[0] * b;
+   __retVal[1] = m[1] * b;
+   __retVal[2] = m[2] * b;
+   __retVal[3] = m[3] * b;
+}
+
+mat4x2 __operator * (const float a, const mat4x2 n)
+{
+   //return mat4x2 (a * n[0], a * n[1], a * n[2], a * n[3]);
+   __retVal[0] = a * n[0];
+   __retVal[1] = a * n[1];
+   __retVal[2] = a * n[2];
+   __retVal[3] = a * n[3];
+}
+
+mat4x3 __operator * (const mat4x3 m, const float b)
+{
+   //return mat4x3 (m[0] * b, m[1] * b, m[2] * b, m[3] * b);
+   __retVal[0] = m[0] * b;
+   __retVal[1] = m[1] * b;
+   __retVal[2] = m[2] * b;
+   __retVal[3] = m[3] * b;
+}
+
+mat4x3 __operator * (const float a, const mat4x3 n)
+{
+   //return mat4x3 (a * n[0], a * n[1], a * n[2], a * n[3]);
+   __retVal[0] = a * n[0];
+   __retVal[1] = a * n[1];
+   __retVal[2] = a * n[2];
+   __retVal[3] = a * n[3];
+}
+
+
+mat2x3 __operator / (const float a, const mat2x3 n)
+{
+   //return mat2x3 (a / n[0], a / n[1]);
+   const float inv = 1.0 / a;
+   __retVal[0] = inv * n[0];
+   __retVal[1] = inv * n[1];
+}
+
+mat2x3 __operator / (const mat2x3 m, const float b)
+{
+   //return mat2x3 (m[0] / b, m[1] / b);
+   const float inv = 1.0 / b;
+   __retVal[0] = m[0] * inv;
+   __retVal[1] = m[1] * inv;
+}
+
+mat2x4 __operator / (const float a, const mat2x4 n)
+{
+   //return mat2x4 (a / n[0], a / n[1]);
+   const float inv = 1.0 / a;
+   __retVal[0] = inv * n[0];
+   __retVal[1] = inv * n[1];
+}
+
+mat2x4 __operator / (const mat2x4 m, const float b)
+{
+   //return mat2x4 (m[0] / b, m[1] / b);
+   const float inv = 1.0 / b;
+   __retVal[0] = m[0] * inv;
+   __retVal[1] = m[1] * inv;
+}
+
+mat3x2 __operator / (const float a, const mat3x2 n)
+{
+   //return mat3x2 (a / n[0], a / n[1], a / n[2]);
+   const float inv = 1.0 / a;
+   __retVal[0] = inv * n[0];
+   __retVal[1] = inv * n[1];
+   __retVal[2] = inv * n[2];
+}
+
+mat3x2 __operator / (const mat3x2 m, const float b)
+{
+   //return mat3x2 (m[0] / b, m[1] / b, m[2] / b);
+   const float inv = 1.0 / b;
+   __retVal[0] = m[0] * inv;
+   __retVal[1] = m[1] * inv;
+   __retVal[2] = m[2] * inv;
+}
+
+mat3x4 __operator / (const float a, const mat3x4 n)
+{
+   //return mat3x4 (a / n[0], a / n[1], a / n[2]);
+   const float inv = 1.0 / a;
+   __retVal[0] = inv * n[0];
+   __retVal[1] = inv * n[1];
+   __retVal[2] = inv * n[2];
+}
+
+mat3x4 __operator / (const mat3x4 m, const float b)
+{
+   //return mat3x4 (m[0] / b, m[1] / b, m[2] / b);
+   const float inv = 1.0 / b;
+   __retVal[0] = m[0] * inv;
+   __retVal[1] = m[1] * inv;
+   __retVal[2] = m[2] * inv;
+}
+
+mat4x2 __operator / (const mat4x2 m, const float b)
+{
+   //return mat4x2 (m[0] / b, m[1] / b, m[2] / b, m[3] / b);
+   const float inv = 1.0 / b;
+   __retVal[0] = m[0] * inv;
+   __retVal[1] = m[1] * inv;
+   __retVal[2] = m[2] * inv;
+   __retVal[3] = m[3] * inv;
+}
+
+mat4x2 __operator / (const float a, const mat4x2 n)
+{
+   //return mat4x2 (a / n[0], a / n[1], a / n[2], a / n[3]);
+   const float inv = 1.0 / a;
+   __retVal[0] = inv * n[0];
+   __retVal[1] = inv * n[1];
+   __retVal[2] = inv * n[2];
+   __retVal[3] = inv * n[3];
+}
+
+mat4x3 __operator / (const mat4x3 m, const float b)
+{
+   //return mat4x3 (m[0] / b, m[1] / b, m[2] / b, m[3] / b);
+   const float inv = 1.0 / b;
+   __retVal[0] = m[0] * inv;
+   __retVal[1] = m[1] * inv;
+   __retVal[2] = m[2] * inv;
+   __retVal[3] = m[3] * inv;
+}
+
+mat4x3 __operator / (const float a, const mat4x3 n)
+{
+   //return mat4x3 (a / n[0], a / n[1], a / n[2], a / n[3]);
+   const float inv = 1.0 / a;
+   __retVal[0] = inv * n[0];
+   __retVal[1] = inv * n[1];
+   __retVal[2] = inv * n[2];
+   __retVal[3] = inv * n[3];
+}
+
+
+mat2x3 __operator - (const mat2x3 m) {
+    return mat2x3 (-m[0], -m[1]);
+}
+
+mat2x4 __operator - (const mat2x4 m) {
+    return mat2x4 (-m[0], -m[1]);
+}
+
+mat3x2 __operator - (const mat3x2 m) {
+    return mat3x2 (-m[0], -m[1], -m[2]);
+}
+
+mat3x4 __operator - (const mat3x4 m) {
+    return mat3x4 (-m[0], -m[1], -m[2]);
+}
+
+mat4x2 __operator - (const mat4x2 m) {
+    return mat4x2 (-m[0], -m[1], -m[2], -m[3]);
+}
+
+mat4x3 __operator - (const mat4x3 m) {
+    return mat4x3 (-m[0], -m[1], -m[2], -m[3]);
+}
+
+
+void __operator -- (inout mat2x3 m) {
+    --m[0];
+    --m[1];
+}
+
+void __operator -- (inout mat2x4 m) {
+    --m[0];
+    --m[1];
+}
+
+void __operator -- (inout mat3x2 m) {
+    --m[0];
+    --m[1];
+    --m[2];
+}
+
+void __operator -- (inout mat3x4 m) {
+    --m[0];
+    --m[1];
+    --m[2];
+}
+
+void __operator -- (inout mat4x2 m) {
+    --m[0];
+    --m[1];
+    --m[2];
+    --m[3];
+}
+
+void __operator -- (inout mat4x3 m) {
+    --m[0];
+    --m[1];
+    --m[2];
+    --m[3];
+}
+
+
+void __operator ++ (inout mat2x3 m) {
+    ++m[0];
+    ++m[1];
+}
+
+void __operator ++ (inout mat2x4 m) {
+    ++m[0];
+    ++m[1];
+}
+
+void __operator ++ (inout mat3x2 m) {
+    ++m[0];
+    ++m[1];
+    ++m[2];
+}
+
+void __operator ++ (inout mat3x4 m) {
+    ++m[0];
+    ++m[1];
+    ++m[2];
+}
+
+void __operator ++ (inout mat4x2 m) {
+    ++m[0];
+    ++m[1];
+    ++m[2];
+    ++m[3];
+}
+
+void __operator ++ (inout mat4x3 m) {
+    ++m[0];
+    ++m[1];
+    ++m[2];
+    ++m[3];
+}
+
diff --git a/src/mesa/slang/library/slang_builtin_120_common.gc b/src/mesa/slang/library/slang_builtin_120_common.gc
new file mode 100644
index 0000000000..c6264c3b47
--- /dev/null
+++ b/src/mesa/slang/library/slang_builtin_120_common.gc
@@ -0,0 +1,200 @@
+/*
+ * Mesa 3-D graphics library
+ * Version:  6.6
+ *
+ * Copyright (C) 2006  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+//
+// From Shader Spec, ver. 1.20, rev. 6
+//
+
+//
+// 8.5 Matrix Functions
+//
+
+mat2x3 matrixCompMult (mat2x3 m, mat2x3 n) {
+    return mat2x3 (m[0] * n[0], m[1] * n[1]);
+}
+
+mat2x4 matrixCompMult (mat2x4 m, mat2x4 n) {
+    return mat2x4 (m[0] * n[0], m[1] * n[1]);
+}
+
+mat3x2 matrixCompMult (mat3x2 m, mat3x2 n) {
+    return mat3x2 (m[0] * n[0], m[1] * n[1], m[2] * n[2]);
+}
+
+mat3x4 matrixCompMult (mat3x4 m, mat3x4 n) {
+    return mat3x4 (m[0] * n[0], m[1] * n[1], m[2] * n[2]);
+}
+
+mat4x2 matrixCompMult (mat4x2 m, mat4x2 n) {
+    return mat4x2 (m[0] * n[0], m[1] * n[1], m[2] * n[2], m[3] * n[3]);
+}
+
+mat4x3 matrixCompMult (mat4x3 m, mat4x3 n) {
+    return mat4x3 (m[0] * n[0], m[1] * n[1], m[2] * n[2], m[3] * n[3]);
+}
+
+mat2 outerProduct (vec2 c, vec2 r) {
+    return mat2 (
+        c.x * r.x, c.y * r.x,
+        c.x * r.y, c.y * r.y
+    );
+}
+
+mat3 outerProduct (vec3 c, vec3 r) {
+    return mat3 (
+        c.x * r.x, c.y * r.x, c.z * r.x,
+        c.x * r.y, c.y * r.y, c.z * r.y,
+        c.x * r.z, c.y * r.z, c.z * r.z
+    );
+}
+
+mat4 outerProduct (vec4 c, vec4 r) {
+    return mat4 (
+        c.x * r.x, c.y * r.x, c.z * r.x, c.w * r.x,
+        c.x * r.y, c.y * r.y, c.z * r.y, c.w * r.y,
+        c.x * r.z, c.y * r.z, c.z * r.z, c.w * r.z,
+        c.x * r.w, c.y * r.w, c.z * r.w, c.w * r.w
+    );
+}
+
+mat2x3 outerProduct (vec3 c, vec2 r) {
+    return mat2x3 (
+        c.x * r.x, c.y * r.x, c.z * r.x,
+        c.x * r.y, c.y * r.y, c.z * r.y
+    );
+}
+
+mat3x2 outerProduct (vec2 c, vec3 r) {
+    return mat3x2 (
+        c.x * r.x, c.y * r.x,
+        c.x * r.y, c.y * r.y,
+        c.x * r.z, c.y * r.z
+    );
+}
+
+mat2x4 outerProduct (vec4 c, vec2 r) {
+    return mat2x4 (
+        c.x * r.x, c.y * r.x, c.z * r.x, c.w * r.x,
+        c.x * r.y, c.y * r.y, c.z * r.y, c.w * r.y
+    );
+}
+
+mat4x2 outerProduct (vec2 c, vec4 r) {
+    return mat4x2 (
+        c.x * r.x, c.y * r.x,
+        c.x * r.y, c.y * r.y,
+        c.x * r.z, c.y * r.z,
+        c.x * r.w, c.y * r.w
+    );
+}
+
+mat3x4 outerProduct (vec4 c, vec3 r) {
+    return mat3x4 (
+        c.x * r.x, c.y * r.x, c.z * r.x, c.w * r.x,
+        c.x * r.y, c.y * r.y, c.z * r.y, c.w * r.y,
+        c.x * r.z, c.y * r.z, c.z * r.z, c.w * r.z
+    );
+}
+
+mat4x3 outerProduct (vec3 c, vec4 r) {
+    return mat4x3 (
+        c.x * r.x, c.y * r.x, c.z * r.x,
+        c.x * r.y, c.y * r.y, c.z * r.y,
+        c.x * r.z, c.y * r.z, c.z * r.z,
+        c.x * r.w, c.y * r.w, c.z * r.w
+    );
+}
+
+mat2 transpose (mat2 m) {
+    return mat2 (
+        m[0].x, m[1].x,
+        m[0].y, m[1].y
+    );
+}
+
+mat3 transpose (mat3 m) {
+    return mat3 (
+        m[0].x, m[1].x, m[2].x,
+        m[0].y, m[1].y, m[2].y,
+        m[0].z, m[1].z, m[2].z
+    );
+}
+
+mat4 transpose (mat4 m) {
+    return mat4 (
+        m[0].x, m[1].x, m[2].x, m[3].x,
+        m[0].y, m[1].y, m[2].y, m[3].y,
+        m[0].z, m[1].z, m[2].z, m[3].z,
+        m[0].w, m[1].w, m[2].w, m[3].w
+    );
+}
+
+mat2x3 transpose (mat3x2 m) {
+    return mat2x3 (
+        m[0].x, m[1].x, m[2].x,
+        m[0].y, m[1].y, m[2].y
+    );
+}
+
+mat3x2 transpose (mat2x3 m) {
+    return mat3x2 (
+        m[0].x, m[1].x,
+        m[0].y, m[1].y,
+        m[0].z, m[1].z
+    );
+}
+
+mat2x4 transpose (mat4x2 m) {
+    return mat2x4 (
+        m[0].x, m[1].x, m[2].x, m[3].x,
+        m[0].y, m[1].y, m[2].y, m[3].y
+    );
+}
+
+mat4x2 transpose (mat2x4 m) {
+    return mat4x2 (
+        m[0].x, m[1].x,
+        m[0].y, m[1].y,
+        m[0].z, m[1].z,
+        m[0].w, m[1].w
+    );
+}
+
+mat3x4 transpose (mat4x3 m) {
+    return mat3x4 (
+        m[0].x, m[1].x, m[2].x, m[3].x,
+        m[0].y, m[1].y, m[2].y, m[3].y,
+        m[0].z, m[1].z, m[2].z, m[3].z
+    );
+}
+
+mat4x3 transpose (mat3x4 m) {
+    return mat4x3 (
+        m[0].x, m[1].x, m[2].x,
+        m[0].y, m[1].y, m[2].y,
+        m[0].z, m[1].z, m[2].z,
+        m[0].w, m[1].w, m[2].w
+    );
+}
+
diff --git a/src/mesa/slang/library/slang_builtin_120_fragment.gc b/src/mesa/slang/library/slang_builtin_120_fragment.gc
new file mode 100644
index 0000000000..7d516046a1
--- /dev/null
+++ b/src/mesa/slang/library/slang_builtin_120_fragment.gc
@@ -0,0 +1,30 @@
+/*
+ * Mesa 3-D graphics library
+ * Version:  6.5
+ *
+ * Copyright (C) 2006  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+//
+// From Shader Spec, ver. 1.20, rev. 6
+//
+
+varying vec2 gl_PointCoord;
+
diff --git a/src/mesa/slang/library/slang_common_builtin.gc b/src/mesa/slang/library/slang_common_builtin.gc
new file mode 100644
index 0000000000..d75354deff
--- /dev/null
+++ b/src/mesa/slang/library/slang_common_builtin.gc
@@ -0,0 +1,1887 @@
+/*
+ * Mesa 3-D graphics library
+ * Version:  7.3
+ *
+ * Copyright (C) 2006  Brian Paul   All Rights Reserved.
+ * Copyright (C) 2008  VMware, Inc.  All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+//
+// From Shader Spec, ver. 1.10, rev. 59
+//
+
+// Note: the values assigned to these constants here aren't actually used.
+// They're set by the compiler according to the GL context limits.
+// See slang_simplify.c
+const int gl_MaxLights = 8;
+const int gl_MaxClipPlanes = 6;
+const int gl_MaxTextureUnits = 8;
+const int gl_MaxTextureCoords = 8;
+const int gl_MaxVertexAttribs = 16;
+const int gl_MaxVertexUniformComponents = 512;
+const int gl_MaxVaryingFloats = 32;
+const int gl_MaxVertexTextureImageUnits = 0;
+const int gl_MaxCombinedTextureImageUnits = 2;
+const int gl_MaxTextureImageUnits = 2;
+const int gl_MaxFragmentUniformComponents = 64;
+const int gl_MaxDrawBuffers = 1;
+
+uniform mat4 gl_ModelViewMatrix;
+uniform mat4 gl_ProjectionMatrix;
+uniform mat4 gl_ModelViewProjectionMatrix;
+uniform mat4 gl_TextureMatrix[gl_MaxTextureCoords];
+
+uniform mat3 gl_NormalMatrix;
+
+uniform mat4 gl_ModelViewMatrixInverse;
+uniform mat4 gl_ProjectionMatrixInverse;
+uniform mat4 gl_ModelViewProjectionMatrixInverse;
+uniform mat4 gl_TextureMatrixInverse[gl_MaxTextureCoords];
+
+uniform mat4 gl_ModelViewMatrixTranspose;
+uniform mat4 gl_ProjectionMatrixTranspose;
+uniform mat4 gl_ModelViewProjectionMatrixTranspose;
+uniform mat4 gl_TextureMatrixTranspose[gl_MaxTextureCoords];
+
+uniform mat4 gl_ModelViewMatrixInverseTranspose;
+uniform mat4 gl_ProjectionMatrixInverseTranspose;
+uniform mat4 gl_ModelViewProjectionMatrixInverseTranspose;
+uniform mat4 gl_TextureMatrixInverseTranspose[gl_MaxTextureCoords];
+
+uniform float gl_NormalScale;
+
+struct gl_DepthRangeParameters {
+    float near;
+    float far;
+    float diff;
+};
+
+uniform gl_DepthRangeParameters gl_DepthRange;
+
+uniform vec4 gl_ClipPlane[gl_MaxClipPlanes];
+
+struct gl_PointParameters {
+    float size;
+    float sizeMin;
+    float sizeMax;
+    float fadeThresholdSize;
+    float distanceConstantAttenuation;
+    float distanceLinearAttenuation;
+    float distanceQuadraticAttenuation;
+};
+
+uniform gl_PointParameters gl_Point;
+
+struct gl_MaterialParameters {
+    vec4 emission;
+    vec4 ambient;
+    vec4 diffuse;
+    vec4 specular;
+    float shininess;
+};
+
+uniform gl_MaterialParameters gl_FrontMaterial;
+uniform gl_MaterialParameters gl_BackMaterial;
+
+/* NOTE: the order of these fields is significant!
+ * See the definition of the lighting state vars such as STATE_SPOT_DIRECTION.
+ */
+struct gl_LightSourceParameters {
+    vec4 ambient;
+    vec4 diffuse;
+    vec4 specular;
+    vec4 position;
+    vec4 halfVector;
+    vec3 spotDirection;
+    float spotCosCutoff;
+
+    float constantAttenuation;
+    float linearAttenuation;
+    float quadraticAttenuation;
+    float spotExponent;
+
+    float spotCutoff;
+};
+
+uniform gl_LightSourceParameters gl_LightSource[gl_MaxLights];
+
+struct gl_LightModelParameters {
+    vec4 ambient;
+};
+
+uniform gl_LightModelParameters gl_LightModel;
+
+struct gl_LightModelProducts {
+    vec4 sceneColor;
+};
+
+uniform gl_LightModelProducts gl_FrontLightModelProduct;
+uniform gl_LightModelProducts gl_BackLightModelProduct;
+
+struct gl_LightProducts {
+    vec4 ambient;
+    vec4 diffuse;
+    vec4 specular;
+};
+
+uniform gl_LightProducts gl_FrontLightProduct[gl_MaxLights];
+uniform gl_LightProducts gl_BackLightProduct[gl_MaxLights];
+
+uniform vec4 gl_TextureEnvColor[gl_MaxTextureImageUnits];
+uniform vec4 gl_EyePlaneS[gl_MaxTextureCoords];
+uniform vec4 gl_EyePlaneT[gl_MaxTextureCoords];
+uniform vec4 gl_EyePlaneR[gl_MaxTextureCoords];
+uniform vec4 gl_EyePlaneQ[gl_MaxTextureCoords];
+uniform vec4 gl_ObjectPlaneS[gl_MaxTextureCoords];
+uniform vec4 gl_ObjectPlaneT[gl_MaxTextureCoords];
+uniform vec4 gl_ObjectPlaneR[gl_MaxTextureCoords];
+uniform vec4 gl_ObjectPlaneQ[gl_MaxTextureCoords];
+
+struct gl_FogParameters {
+    vec4 color;
+    float density;
+    float start;
+    float end;
+    float scale;
+};
+
+uniform gl_FogParameters gl_Fog;
+
+
+
+
+
+//
+// 8.1 Angle and Trigonometry Functions
+//
+
+//// radians
+
+float radians(const float deg)
+{
+   const float c = 3.1415926 / 180.0;
+   __asm vec4_multiply __retVal, deg, c;
+}
+
+vec2 radians(const vec2 deg)
+{
+   const float c = 3.1415926 / 180.0;
+   __asm vec4_multiply __retVal.xy, deg.xy, c.xx;
+}
+
+vec3 radians(const vec3 deg)
+{
+   const float c = 3.1415926 / 180.0;
+   __asm vec4_multiply __retVal.xyz, deg.xyz, c.xxx;
+}
+
+vec4 radians(const vec4 deg)
+{
+   const float c = 3.1415926 / 180.0;
+   __asm vec4_multiply __retVal, deg, c.xxxx;
+}
+
+
+//// degrees
+
+float degrees(const float rad)
+{
+   const float c = 180.0 / 3.1415926;
+   __asm vec4_multiply __retVal, rad, c;
+}
+
+vec2 degrees(const vec2 rad)
+{
+   const float c = 180.0 / 3.1415926;
+   __asm vec4_multiply __retVal.xy, rad.xy, c.xx;
+}
+
+vec3 degrees(const vec3 rad)
+{
+   const float c = 180.0 / 3.1415926;
+   __asm vec4_multiply __retVal.xyz, rad.xyz, c.xxx;
+}
+
+vec4 degrees(const vec4 rad)
+{
+   const float c = 180.0 / 3.1415926;
+   __asm vec4_multiply __retVal, rad, c.xxxx;
+}
+
+
+//// sin
+
+float sin(const float radians)
+{
+   __asm float_sine __retVal, radians;
+}
+
+vec2 sin(const vec2 radians)
+{
+   __asm float_sine __retVal.x, radians.x;
+   __asm float_sine __retVal.y, radians.y;
+}
+
+vec3 sin(const vec3 radians)
+{
+   __asm float_sine __retVal.x, radians.x;
+   __asm float_sine __retVal.y, radians.y;
+   __asm float_sine __retVal.z, radians.z;
+}
+
+vec4 sin(const vec4 radians)
+{
+   __asm float_sine __retVal.x, radians.x;
+   __asm float_sine __retVal.y, radians.y;
+   __asm float_sine __retVal.z, radians.z;
+   __asm float_sine __retVal.w, radians.w;
+}
+
+
+//// cos
+
+float cos(const float radians)
+{
+   __asm float_cosine __retVal, radians;
+}
+
+vec2 cos(const vec2 radians)
+{
+   __asm float_cosine __retVal.x, radians.x;
+   __asm float_cosine __retVal.y, radians.y;
+}
+
+vec3 cos(const vec3 radians)
+{
+   __asm float_cosine __retVal.x, radians.x;
+   __asm float_cosine __retVal.y, radians.y;
+   __asm float_cosine __retVal.z, radians.z;
+}
+
+vec4 cos(const vec4 radians)
+{
+   __asm float_cosine __retVal.x, radians.x;
+   __asm float_cosine __retVal.y, radians.y;
+   __asm float_cosine __retVal.z, radians.z;
+   __asm float_cosine __retVal.w, radians.w;
+}
+
+
+
+//// tan
+
+float tan(const float angle)
+{
+   const float s = sin(angle);
+   const float c = cos(angle);
+   return s / c;
+}
+
+vec2 tan(const vec2 angle)
+{
+   const vec2 s = sin(angle);
+   const vec2 c = cos(angle);
+   return s / c;
+}
+
+vec3 tan(const vec3 angle)
+{
+   const vec3 s = sin(angle);
+   const vec3 c = cos(angle);
+   return s / c;
+}
+
+vec4 tan(const vec4 angle)
+{
+   const vec4 s = sin(angle);
+   const vec4 c = cos(angle);
+   return s / c;
+}
+
+
+
+float asin(const float x)
+{
+   const float a0 = 1.5707288;  // PI/2?
+   const float a1 = -0.2121144;
+   const float a2 = 0.0742610;
+   //const float a3 = -0.0187293;
+   const float halfPi = 3.1415926 * 0.5;
+   const float y = abs(x);
+   // three terms seem to be enough:
+   __retVal = (halfPi - sqrt(1.0 - y) * (a0 + y * (a1 + a2 * y))) * sign(x);
+   // otherwise, try four:
+   //__retVal = (halfPi - sqrt(1.0 - y) * (a0 + y * (a1 + y * (a2 + y * a3)))) * sign(x);
+}
+
+vec2 asin(const vec2 v)
+{
+   __retVal.x = asin(v.x);
+   __retVal.y = asin(v.y);
+}
+
+vec3 asin(const vec3 v)
+{
+   __retVal.x = asin(v.x);
+   __retVal.y = asin(v.y);
+   __retVal.z = asin(v.z);
+}
+
+vec4 asin(const vec4 v)
+{
+   __retVal.x = asin(v.x);
+   __retVal.y = asin(v.y);
+   __retVal.z = asin(v.z);
+   __retVal.w = asin(v.w);
+}
+
+float acos(const float x)
+{
+   const float halfPi = 3.1415926 * 0.5;
+   __retVal = halfPi - asin(x);
+}
+
+vec2 acos(const vec2 v)
+{
+   __retVal.x = acos(v.x);
+   __retVal.y = acos(v.y);
+}
+
+vec3 acos(const vec3 v)
+{
+   __retVal.x = acos(v.x);
+   __retVal.y = acos(v.y);
+   __retVal.z = acos(v.z);
+}
+
+vec4 acos(const vec4 v)
+{
+   __retVal.x = acos(v.x);
+   __retVal.y = acos(v.y);
+   __retVal.z = acos(v.z);
+   __retVal.w = acos(v.w);
+}
+
+float atan(const float x)
+{
+   __retVal = asin(x * inversesqrt(x * x + 1.0));
+}
+
+vec2 atan(const vec2 y_over_x)
+{
+   __retVal.x = atan(y_over_x.x);
+   __retVal.y = atan(y_over_x.y);
+}
+
+vec3 atan(const vec3 y_over_x)
+{
+   __retVal.x = atan(y_over_x.x);
+   __retVal.y = atan(y_over_x.y);
+   __retVal.z = atan(y_over_x.z);
+}
+
+vec4 atan(const vec4 y_over_x)
+{
+   __retVal.x = atan(y_over_x.x);
+   __retVal.y = atan(y_over_x.y);
+   __retVal.z = atan(y_over_x.z);
+   __retVal.w = atan(y_over_x.w);
+}
+
+float atan(const float y, const float x)
+{
+   float r;
+   if (abs(x) > 1.0e-4) {
+      r = atan(y / x);
+      if (x < 0.0) {
+         r = r + sign(y) * 3.141593;
+      }
+   }
+   else {
+      r = sign(y) * 1.5707965;  // pi/2
+   }
+   return r;
+}
+
+vec2 atan(const vec2 u, const vec2 v)
+{
+   __retVal.x = atan(u.x, v.x);
+   __retVal.y = atan(u.y, v.y);
+}
+
+vec3 atan(const vec3 u, const vec3 v)
+{
+   __retVal.x = atan(u.x, v.x);
+   __retVal.y = atan(u.y, v.y);
+   __retVal.z = atan(u.z, v.z);
+}
+
+vec4 atan(const vec4 u, const vec4 v)
+{
+   __retVal.x = atan(u.x, v.x);
+   __retVal.y = atan(u.y, v.y);
+   __retVal.z = atan(u.z, v.z);
+   __retVal.w = atan(u.w, v.w);
+}
+
+
+//
+// 8.2 Exponential Functions
+//
+
+//// pow
+
+float pow(const float a, const float b)
+{
+   __asm float_power __retVal, a, b;
+}
+
+vec2 pow(const vec2 a, const vec2 b)
+{
+   __asm float_power __retVal.x, a.x, b.x;
+   __asm float_power __retVal.y, a.y, b.y;
+}
+
+vec3 pow(const vec3 a, const vec3 b)
+{
+   __asm float_power __retVal.x, a.x, b.x;
+   __asm float_power __retVal.y, a.y, b.y;
+   __asm float_power __retVal.z, a.z, b.z;
+}
+
+vec4 pow(const vec4 a, const vec4 b)
+{
+   __asm float_power __retVal.x, a.x, b.x;
+   __asm float_power __retVal.y, a.y, b.y;
+   __asm float_power __retVal.z, a.z, b.z;
+   __asm float_power __retVal.w, a.w, b.w;
+}
+
+
+//// exp
+
+float exp(const float a)
+{
+   // NOTE: log2(e) = 1.44269502
+   float t = a * 1.44269502;
+   __asm float_exp2 __retVal, t;
+}
+
+vec2 exp(const vec2 a)
+{
+   vec2 t = a * 1.44269502;
+   __asm float_exp2 __retVal.x, t.x;
+   __asm float_exp2 __retVal.y, t.y;
+}
+
+vec3 exp(const vec3 a)
+{
+   vec3 t = a * 1.44269502;
+   __asm float_exp2 __retVal.x, t.x;
+   __asm float_exp2 __retVal.y, t.y;
+   __asm float_exp2 __retVal.z, t.z;
+}
+
+vec4 exp(const vec4 a)
+{
+   vec4 t = a * 1.44269502;
+   __asm float_exp2 __retVal.x, t.x;
+   __asm float_exp2 __retVal.y, t.y;
+   __asm float_exp2 __retVal.z, t.z;
+   __asm float_exp2 __retVal.w, t.w;
+}
+
+
+
+//// log2
+
+float log2(const float x)
+{
+   __asm float_log2 __retVal, x;
+}
+
+vec2 log2(const vec2 v)
+{
+   __asm float_log2 __retVal.x, v.x;
+   __asm float_log2 __retVal.y, v.y;
+}
+
+vec3 log2(const vec3 v)
+{
+   __asm float_log2 __retVal.x, v.x;
+   __asm float_log2 __retVal.y, v.y;
+   __asm float_log2 __retVal.z, v.z;
+}
+
+vec4 log2(const vec4 v)
+{
+   __asm float_log2 __retVal.x, v.x;
+   __asm float_log2 __retVal.y, v.y;
+   __asm float_log2 __retVal.z, v.z;
+   __asm float_log2 __retVal.w, v.w;
+}
+
+
+//// log  (natural log)
+
+float log(const float x)
+{
+   // note:  logBaseB(x) = logBaseN(x) / logBaseN(B)
+   // compute log(x) = log2(x) / log2(e)
+   // c = 1.0 / log2(e) = 0.693147181
+   const float c = 0.693147181;
+   return log2(x) * c;
+}
+
+vec2 log(const vec2 v)
+{
+   const float c = 0.693147181;
+   return log2(v) * c;
+}
+
+vec3 log(const vec3 v)
+{
+   const float c = 0.693147181;
+   return log2(v) * c;
+}
+
+vec4 log(const vec4 v)
+{
+   const float c = 0.693147181;
+   return log2(v) * c;
+}
+
+
+//// exp2
+
+float exp2(const float a)
+{
+   __asm float_exp2 __retVal, a;
+}
+
+vec2 exp2(const vec2 a)
+{
+   __asm float_exp2 __retVal.x, a.x;
+   __asm float_exp2 __retVal.y, a.y;
+}
+
+vec3 exp2(const vec3 a)
+{
+   __asm float_exp2 __retVal.x, a.x;
+   __asm float_exp2 __retVal.y, a.y;
+   __asm float_exp2 __retVal.z, a.z;
+}
+
+vec4 exp2(const vec4 a)
+{
+   __asm float_exp2 __retVal.x, a.x;
+   __asm float_exp2 __retVal.y, a.y;
+   __asm float_exp2 __retVal.z, a.z;
+   __asm float_exp2 __retVal.w, a.w;
+}
+
+
+//// sqrt
+
+float sqrt(const float x)
+{
+   const float nx = -x;
+   float r;
+   __asm float_rsq r, x;
+   r = r * x;
+   __asm vec4_cmp __retVal, nx, r, 0.0;
+}
+
+vec2 sqrt(const vec2 x)
+{
+   const vec2 nx = -x, zero = vec2(0.0);
+   vec2 r;
+   __asm float_rsq r.x, x.x;
+   __asm float_rsq r.y, x.y;
+   r = r * x;
+   __asm vec4_cmp __retVal, nx, r, zero;
+}
+
+vec3 sqrt(const vec3 x)
+{
+   const vec3 nx = -x, zero = vec3(0.0);
+   vec3 r;
+   __asm float_rsq r.x, x.x;
+   __asm float_rsq r.y, x.y;
+   __asm float_rsq r.z, x.z;
+   r = r * x;
+   __asm vec4_cmp __retVal, nx, r, zero;
+}
+
+vec4 sqrt(const vec4 x)
+{
+   const vec4 nx = -x, zero = vec4(0.0);
+   vec4 r;
+   __asm float_rsq r.x, x.x;
+   __asm float_rsq r.y, x.y;
+   __asm float_rsq r.z, x.z;
+   __asm float_rsq r.w, x.w;
+   r = r * x;
+   __asm vec4_cmp __retVal, nx, r, zero;
+}
+
+
+//// inversesqrt
+
+float inversesqrt(const float x)
+{
+   __asm float_rsq __retVal.x, x;
+}
+
+vec2 inversesqrt(const vec2 v)
+{
+   __asm float_rsq __retVal.x, v.x;
+   __asm float_rsq __retVal.y, v.y;
+}
+
+vec3 inversesqrt(const vec3 v)
+{
+   __asm float_rsq __retVal.x, v.x;
+   __asm float_rsq __retVal.y, v.y;
+   __asm float_rsq __retVal.z, v.z;
+}
+
+vec4 inversesqrt(const vec4 v)
+{
+   __asm float_rsq __retVal.x, v.x;
+   __asm float_rsq __retVal.y, v.y;
+   __asm float_rsq __retVal.z, v.z;
+   __asm float_rsq __retVal.w, v.w;
+}
+
+
+//// normalize
+
+float normalize(const float x)
+{
+   __retVal = 1.0;
+}
+
+vec2 normalize(const vec2 v)
+{
+   const float s = inversesqrt(dot(v, v));
+   __asm vec4_multiply __retVal.xy, v, s;
+}
+
+vec3 normalize(const vec3 v)
+{
+//   const float s = inversesqrt(dot(v, v));
+//   __retVal = v * s;
+// XXX note, we _could_ use __retVal.w instead of tmp and save a
+// register, but that's actually a compilation error because v is a vec3
+// and the .w suffix is illegal.  Oh well.
+   float tmp;
+   __asm vec3_dot tmp, v, v;
+   __asm float_rsq tmp, tmp;
+   __asm vec4_multiply __retVal.xyz, v, tmp;
+}
+
+vec4 normalize(const vec4 v)
+{
+   float tmp;
+   __asm vec4_dot tmp, v, v;
+   __asm float_rsq tmp, tmp;
+   __asm vec4_multiply __retVal.xyz, v, tmp;
+}
+
+
+
+//
+// 8.3 Common Functions
+//
+
+
+//// abs
+
+float abs(const float a)
+{
+   __asm vec4_abs __retVal, a;
+}
+
+vec2 abs(const vec2 a)
+{
+   __asm vec4_abs __retVal.xy, a;
+}
+
+vec3 abs(const vec3 a)
+{
+   __asm vec4_abs __retVal.xyz, a;
+}
+
+vec4 abs(const vec4 a)
+{
+   __asm vec4_abs __retVal, a;
+}
+
+
+//// sign
+
+float sign(const float x)
+{
+   float p, n;
+   __asm vec4_sgt p, x, 0.0;            // p = (x > 0)
+   __asm vec4_sgt n, 0.0, x;            // n = (x < 0)
+   __asm vec4_subtract __retVal, p, n;  // sign = p - n
+}
+
+vec2 sign(const vec2 v)
+{
+   vec2 p, n;
+   __asm vec4_sgt p.xy, v, 0.0;
+   __asm vec4_sgt n.xy, 0.0, v;
+   __asm vec4_subtract __retVal.xy, p, n;
+}
+
+vec3 sign(const vec3 v)
+{
+   vec3 p, n;
+   __asm vec4_sgt p.xyz, v, 0.0;
+   __asm vec4_sgt n.xyz, 0.0, v;
+   __asm vec4_subtract __retVal.xyz, p, n;
+}
+
+vec4 sign(const vec4 v)
+{
+   vec4 p, n;
+   __asm vec4_sgt p, v, 0.0;
+   __asm vec4_sgt n, 0.0, v;
+   __asm vec4_subtract __retVal, p, n;
+}
+
+
+//// floor
+
+float floor(const float a)
+{
+   __asm vec4_floor __retVal, a;
+}
+
+vec2 floor(const vec2 a)
+{
+   __asm vec4_floor __retVal.xy, a;
+}
+
+vec3 floor(const vec3 a)
+{
+   __asm vec4_floor __retVal.xyz, a;
+}
+
+vec4 floor(const vec4 a)
+{
+   __asm vec4_floor __retVal, a;
+}
+
+
+//// ceil
+
+float ceil(const float a)
+{
+   // XXX this could be improved
+   float b = -a;
+   __asm vec4_floor b, b;
+   __retVal = -b;
+}
+
+vec2 ceil(const vec2 a)
+{
+   vec2 b = -a;
+   __asm vec4_floor b, b;
+   __retVal.xy = -b;
+}
+
+vec3 ceil(const vec3 a)
+{
+   vec3 b = -a;
+   __asm vec4_floor b, b;
+   __retVal.xyz = -b;
+}
+
+vec4 ceil(const vec4 a)
+{
+   vec4 b = -a;
+   __asm vec4_floor b, b;
+   __retVal = -b;
+}
+
+
+//// fract
+
+float fract(const float a)
+{
+   __asm vec4_frac __retVal, a;
+}
+
+vec2 fract(const vec2 a)
+{
+   __asm vec4_frac __retVal.xy, a;
+}
+
+vec3 fract(const vec3 a)
+{
+   __asm vec4_frac __retVal.xyz, a;
+}
+
+vec4 fract(const vec4 a)
+{
+   __asm vec4_frac __retVal, a;
+}
+
+
+//// mod  (very untested!)
+
+float mod(const float a, const float b)
+{
+    float oneOverB;
+    __asm float_rcp oneOverB, b;
+    __retVal = a - b * floor(a * oneOverB);
+}
+
+vec2 mod(const vec2 a, const float b)
+{
+    float oneOverB;
+    __asm float_rcp oneOverB, b;
+    __retVal.xy = a - b * floor(a * oneOverB);
+}
+
+vec3 mod(const vec3 a, const float b)
+{
+    float oneOverB;
+    __asm float_rcp oneOverB, b;
+    __retVal.xyz = a - b * floor(a * oneOverB);
+}
+
+vec4 mod(const vec4 a, const float b)
+{
+    float oneOverB;
+    __asm float_rcp oneOverB, b;
+    __retVal = a - b * floor(a * oneOverB);
+}
+
+vec2 mod(const vec2 a, const vec2 b)
+{
+    vec2 oneOverB;
+    __asm float_rcp oneOverB.x, b.x;
+    __asm float_rcp oneOverB.y, b.y;
+    __retVal = a - b * floor(a * oneOverB);
+}
+
+vec3 mod(const vec3 a, const vec3 b)
+{
+    vec3 oneOverB;
+    __asm float_rcp oneOverB.x, b.x;
+    __asm float_rcp oneOverB.y, b.y;
+    __asm float_rcp oneOverB.z, b.z;
+    __retVal = a - b * floor(a * oneOverB);
+}
+
+vec4 mod(const vec4 a, const vec4 b)
+{
+    vec4 oneOverB;
+    __asm float_rcp oneOverB.x, b.x;
+    __asm float_rcp oneOverB.y, b.y;
+    __asm float_rcp oneOverB.z, b.z;
+    __asm float_rcp oneOverB.w, b.w;
+    __retVal = a - b * floor(a * oneOverB);
+}
+
+
+//// min
+
+float min(const float a, const float b)
+{
+   __asm vec4_min __retVal, a, b;
+}
+
+vec2 min(const vec2 a, const vec2 b)
+{
+   __asm vec4_min __retVal.xy, a.xy, b.xy;
+}
+
+vec3 min(const vec3 a, const vec3 b)
+{
+   __asm vec4_min __retVal.xyz, a.xyz, b.xyz;
+}
+
+vec4 min(const vec4 a, const vec4 b)
+{
+   __asm vec4_min __retVal, a, b;
+}
+
+vec2 min(const vec2 a, const float b)
+{
+   __asm vec4_min __retVal, a.xy, b;
+}
+
+vec3 min(const vec3 a, const float b)
+{
+   __asm vec4_min __retVal, a.xyz, b;
+}
+
+vec4 min(const vec4 a, const float b)
+{
+   __asm vec4_min __retVal, a, b;
+}
+
+
+//// max
+
+float max(const float a, const float b)
+{
+   __asm vec4_max __retVal, a, b;
+}
+
+vec2 max(const vec2 a, const vec2 b)
+{
+   __asm vec4_max __retVal.xy, a.xy, b.xy;
+}
+
+vec3 max(const vec3 a, const vec3 b)
+{
+   __asm vec4_max __retVal.xyz, a.xyz, b.xyz;
+}
+
+vec4 max(const vec4 a, const vec4 b)
+{
+   __asm vec4_max __retVal, a, b;
+}
+
+vec2 max(const vec2 a, const float b)
+{
+   __asm vec4_max __retVal, a.xy, b;
+}
+
+vec3 max(const vec3 a, const float b)
+{
+   __asm vec4_max __retVal, a.xyz, b;
+}
+
+vec4 max(const vec4 a, const float b)
+{
+   __asm vec4_max __retVal, a, b;
+}
+
+
+//// clamp
+
+float clamp(const float val, const float minVal, const float maxVal)
+{
+   __asm vec4_clamp __retVal, val, minVal, maxVal;
+}
+
+vec2 clamp(const vec2 val, const float minVal, const float maxVal)
+{
+   __asm vec4_clamp __retVal, val, minVal, maxVal;
+}
+
+vec3 clamp(const vec3 val, const float minVal, const float maxVal)
+{
+   __asm vec4_clamp __retVal, val, minVal, maxVal;
+}
+
+vec4 clamp(const vec4 val, const float minVal, const float maxVal)
+{
+   __asm vec4_clamp __retVal, val, minVal, maxVal;
+}
+
+vec2 clamp(const vec2 val, const vec2 minVal, const vec2 maxVal)
+{
+   __asm vec4_clamp __retVal, val, minVal, maxVal;
+}
+
+vec3 clamp(const vec3 val, const vec3 minVal, const vec3 maxVal)
+{
+   __asm vec4_clamp __retVal, val, minVal, maxVal;
+}
+
+vec4 clamp(const vec4 val, const vec4 minVal, const vec4 maxVal)
+{
+   __asm vec4_clamp __retVal, val, minVal, maxVal;
+}
+
+
+//// mix
+
+float mix(const float x, const float y, const float a)
+{
+   __asm vec4_lrp __retVal, a, y, x;
+}
+
+vec2 mix(const vec2 x, const vec2 y, const float a)
+{
+   __asm vec4_lrp __retVal, a, y, x;
+}
+
+vec3 mix(const vec3 x, const vec3 y, const float a)
+{
+   __asm vec4_lrp __retVal, a, y, x;
+}
+
+vec4 mix(const vec4 x, const vec4 y, const float a)
+{
+   __asm vec4_lrp __retVal, a, y, x;
+}
+
+vec2 mix(const vec2 x, const vec2 y, const vec2 a)
+{
+   __asm vec4_lrp __retVal, a, y, x;
+}
+
+vec3 mix(const vec3 x, const vec3 y, const vec3 a)
+{
+   __asm vec4_lrp __retVal, a, y, x;
+}
+
+vec4 mix(const vec4 x, const vec4 y, const vec4 a)
+{
+   __asm vec4_lrp __retVal, a, y, x;
+}
+
+
+//// step
+
+float step(const float edge, const float x)
+{
+   __asm vec4_sge __retVal, x, edge;
+}
+
+vec2 step(const vec2 edge, const vec2 x)
+{
+   __asm vec4_sge __retVal.xy, x, edge;
+}
+
+vec3 step(const vec3 edge, const vec3 x)
+{
+   __asm vec4_sge __retVal.xyz, x, edge;
+}
+
+vec4 step(const vec4 edge, const vec4 x)
+{
+   __asm vec4_sge __retVal, x, edge;
+}
+
+vec2 step(const float edge, const vec2 v)
+{
+   __asm vec4_sge __retVal.xy, v, edge;
+}
+
+vec3 step(const float edge, const vec3 v)
+{
+   __asm vec4_sge __retVal.xyz, v, edge;
+}
+
+vec4 step(const float edge, const vec4 v)
+{
+   __asm vec4_sge __retVal, v, edge;
+}
+
+
+//// smoothstep
+
+float smoothstep(const float edge0, const float edge1, const float x)
+{
+    float t = clamp((x - edge0) / (edge1 - edge0), 0.0, 1.0);
+    return t * t * (3.0 - 2.0 * t);
+}
+
+vec2 smoothstep(const vec2 edge0, const vec2 edge1, const vec2 v)
+{
+   vec2 t = clamp((v - edge0) / (edge1 - edge0), 0.0, 1.0);
+   return t * t * (3.0 - 2.0 * t);
+}
+
+vec3 smoothstep(const vec3 edge0, const vec3 edge1, const vec3 v)
+{
+   vec3 t = clamp((v - edge0) / (edge1 - edge0), 0.0, 1.0);
+   return t * t * (3.0 - 2.0 * t);
+}
+
+vec4 smoothstep(const vec4 edge0, const vec4 edge1, const vec4 v)
+{
+   vec4 t = clamp((v - edge0) / (edge1 - edge0), 0.0, 1.0);
+   return t * t * (3.0 - 2.0 * t);
+}
+
+vec2 smoothstep(const float edge0, const float edge1, const vec2 v)
+{
+   vec2 t = clamp((v - edge0) / (edge1 - edge0), 0.0, 1.0);
+   return t * t * (3.0 - 2.0 * t);
+}
+
+vec3 smoothstep(const float edge0, const float edge1, const vec3 v)
+{
+   vec3 t = clamp((v - edge0) / (edge1 - edge0), 0.0, 1.0);
+   return t * t * (3.0 - 2.0 * t);
+}
+
+vec4 smoothstep(const float edge0, const float edge1, const vec4 v)
+{
+   vec4 t = clamp((v - edge0) / (edge1 - edge0), 0.0, 1.0);
+   return t * t * (3.0 - 2.0 * t);
+}
+
+
+
+//
+// 8.4 Geometric Functions
+//
+
+
+//// length
+
+float length(const float x)
+{
+   return abs(x);
+}
+
+float length(const vec2 v)
+{
+   float r;
+   const float p = dot(v, v);      // p = v.x * v.x + v.y * v.y
+   __asm float_rsq r, p;           // r = 1 / sqrt(p)
+   __retVal = p * r;               // p * r = sqrt(p);
+}
+
+float length(const vec3 v)
+{
+   float r;
+   const float p = dot(v, v);      // p = v.x * v.x + v.y * v.y + v.z * v.z
+   __asm float_rsq r, p;           // r = 1 / sqrt(p)
+   __retVal = p * r;               // p * r = sqrt(p);
+}
+
+float length(const vec4 v)
+{
+   float r;
+   const float p = dot(v, v);      // p = v.x * v.x + v.y * v.y + ...
+   __asm float_rsq r, p;           // r = 1 / sqrt(p)
+   __retVal = p * r;               // p * r = sqrt(p);
+}
+
+
+//// distance
+
+float distance(const float x, const float y)
+{
+   const float d = x - y;
+   __retVal = length(d);
+}
+
+float distance(const vec2 v, const vec2 u)
+{
+   const vec2 d2 = v - u;
+   __retVal = length(d2);
+}
+
+float distance(const vec3 v, const vec3 u)
+{
+   const vec3 d3 = v - u;
+   __retVal = length(d3);
+}
+
+float distance(const vec4 v, const vec4 u)
+{
+   const vec4 d4 = v - u;
+   __retVal = length(d4);
+}
+
+
+//// cross
+
+vec3 cross(const vec3 v, const vec3 u)
+{
+   __asm vec3_cross __retVal.xyz, v, u;
+}
+
+
+//// faceforward
+
+float faceforward(const float N, const float I, const float Nref)
+{
+    // this could probably be done better
+    const float d = dot(Nref, I);
+    float s;
+    __asm vec4_sgt s, 0.0, d;  // s = (0.0 > d) ? 1 : 0
+    return mix(-N, N, s);
+}
+
+vec2 faceforward(const vec2 N, const vec2 I, const vec2 Nref)
+{
+    // this could probably be done better
+    const float d = dot(Nref, I);
+    float s;
+    __asm vec4_sgt s, 0.0, d;  // s = (0.0 > d) ? 1 : 0
+    return mix(-N, N, s);
+}
+
+vec3 faceforward(const vec3 N, const vec3 I, const vec3 Nref)
+{
+    // this could probably be done better
+    const float d = dot(Nref, I);
+    float s;
+    __asm vec4_sgt s, 0.0, d;  // s = (0.0 > d) ? 1 : 0
+    return mix(-N, N, s);
+}
+
+vec4 faceforward(const vec4 N, const vec4 I, const vec4 Nref)
+{
+    // this could probably be done better
+    const float d = dot(Nref, I);
+    float s;
+    __asm vec4_sgt s, 0.0, d;  // s = (0.0 > d) ? 1 : 0
+    return mix(-N, N, s);
+}
+
+
+//// reflect
+
+float reflect(const float I, const float N)
+{
+   return I - 2.0 * dot(N, I) * N;
+}
+
+vec2 reflect(const vec2 I, const vec2 N)
+{
+   return I - 2.0 * dot(N, I) * N;
+}
+
+vec3 reflect(const vec3 I, const vec3 N)
+{
+   return I - 2.0 * dot(N, I) * N;
+}
+
+vec4 reflect(const vec4 I, const vec4 N)
+{
+   return I - 2.0 * dot(N, I) * N;
+}
+
+//// refract
+
+float refract(const float I, const float N, const float eta)
+{
+   float n_dot_i = dot(N, I);
+   float k = 1.0 - eta * eta * (1.0 - n_dot_i * n_dot_i);
+   float retval;
+   if (k < 0.0)
+      retval = 0.0;
+   else
+      retval = eta * I - (eta * n_dot_i + sqrt(k)) * N;
+   return retval;
+}
+
+vec2 refract(const vec2 I, const vec2 N, const float eta)
+{
+   float n_dot_i = dot(N, I);
+   float k = 1.0 - eta * eta * (1.0 - n_dot_i * n_dot_i);
+   vec2 retval;
+   if (k < 0.0)
+      retval = vec2(0.0);
+   else
+      retval = eta * I - (eta * n_dot_i + sqrt(k)) * N;
+   return retval;
+}
+
+vec3 refract(const vec3 I, const vec3 N, const float eta)
+{
+   float n_dot_i = dot(N, I);
+   float k = 1.0 - eta * eta * (1.0 - n_dot_i * n_dot_i);
+   vec3 retval;
+   if (k < 0.0)
+      retval = vec3(0.0);
+   else
+      retval = eta * I - (eta * n_dot_i + sqrt(k)) * N;
+   return retval;
+}
+
+vec4 refract(const vec4 I, const vec4 N, const float eta)
+{
+   float n_dot_i = dot(N, I);
+   float k = 1.0 - eta * eta * (1.0 - n_dot_i * n_dot_i);
+   vec4 retval;
+   if (k < 0.0)
+      retval = vec4(0.0);
+   else
+      retval = eta * I - (eta * n_dot_i + sqrt(k)) * N;
+   return retval;
+}
+
+
+
+
+//
+// 8.5 Matrix Functions
+//
+
+mat2 matrixCompMult (mat2 m, mat2 n) {
+    return mat2 (m[0] * n[0], m[1] * n[1]);
+}
+
+mat3 matrixCompMult (mat3 m, mat3 n) {
+    return mat3 (m[0] * n[0], m[1] * n[1], m[2] * n[2]);
+}
+
+mat4 matrixCompMult (mat4 m, mat4 n) {
+    return mat4 (m[0] * n[0], m[1] * n[1], m[2] * n[2], m[3] * n[3]);
+}
+
+
+
+
+//
+// 8.6 Vector Relational Functions
+//
+
+//// lessThan
+
+bvec2 lessThan(const vec2 u, const vec2 v)
+{
+   __asm vec4_slt __retVal.xy, u, v;
+}
+
+bvec3 lessThan(const vec3 u, const vec3 v)
+{
+   __asm vec4_slt __retVal.xyz, u, v;
+}
+
+bvec4 lessThan(const vec4 u, const vec4 v)
+{
+   __asm vec4_slt __retVal, u, v;
+}
+
+bvec2 lessThan(const ivec2 u, const ivec2 v)
+{
+   __asm vec4_slt __retVal.xy, u, v;
+}
+
+bvec3 lessThan(const ivec3 u, const ivec3 v)
+{
+   __asm vec4_slt __retVal.xyz, u, v;
+}
+
+bvec4 lessThan(const ivec4 u, const ivec4 v)
+{
+   __asm vec4_slt __retVal, u, v;
+}
+
+
+//// lessThanEqual
+
+bvec2 lessThanEqual(const vec2 u, const vec2 v)
+{
+   __asm vec4_sle __retVal.xy, u, v;
+}
+
+bvec3 lessThanEqual(const vec3 u, const vec3 v)
+{
+   __asm vec4_sle __retVal.xyz, u, v;
+}
+
+bvec4 lessThanEqual(const vec4 u, const vec4 v)
+{
+   __asm vec4_sle __retVal, u, v;
+}
+
+bvec2 lessThanEqual(const ivec2 u, const ivec2 v)
+{
+   __asm vec4_sle __retVal.xy, u, v;
+}
+
+bvec3 lessThanEqual(const ivec3 u, const ivec3 v)
+{
+   __asm vec4_sle __retVal.xyz, u, v;
+}
+
+bvec4 lessThanEqual(const ivec4 u, const ivec4 v)
+{
+   __asm vec4_sle __retVal, u, v;
+}
+
+
+//// greaterThan
+
+bvec2 greaterThan(const vec2 u, const vec2 v)
+{
+   __asm vec4_sgt __retVal.xy, u, v;
+}
+
+bvec3 greaterThan(const vec3 u, const vec3 v)
+{
+   __asm vec4_sgt __retVal.xyz, u, v;
+}
+
+bvec4 greaterThan(const vec4 u, const vec4 v)
+{
+   __asm vec4_sgt __retVal, u, v;
+}
+
+bvec2 greaterThan(const ivec2 u, const ivec2 v)
+{
+   __asm vec4_sgt __retVal.xy, u.xy, v.xy;
+}
+
+bvec3 greaterThan(const ivec3 u, const ivec3 v)
+{
+   __asm vec4_sgt __retVal.xyz, u, v;
+}
+
+bvec4 greaterThan(const ivec4 u, const ivec4 v)
+{
+   __asm vec4_sgt __retVal, u, v;
+}
+
+
+//// greaterThanEqual
+
+bvec2 greaterThanEqual(const vec2 u, const vec2 v)
+{
+   __asm vec4_sge __retVal.xy, u, v;
+}
+
+bvec3 greaterThanEqual(const vec3 u, const vec3 v)
+{
+   __asm vec4_sge __retVal.xyz, u, v;
+}
+
+bvec4 greaterThanEqual(const vec4 u, const vec4 v)
+{
+   __asm vec4_sge __retVal, u, v;
+}
+
+bvec2 greaterThanEqual(const ivec2 u, const ivec2 v)
+{
+   __asm vec4_sge __retVal.xy, u, v;
+}
+
+bvec3 greaterThanEqual(const ivec3 u, const ivec3 v)
+{
+   __asm vec4_sge __retVal.xyz, u, v;
+}
+
+bvec4 greaterThanEqual(const ivec4 u, const ivec4 v)
+{
+   __asm vec4_sge __retVal, u, v;
+}
+
+
+//// equal
+
+bvec2 equal(const vec2 u, const vec2 v)
+{
+   __asm vec4_seq __retVal.xy, u, v;
+}
+
+bvec3 equal(const vec3 u, const vec3 v)
+{
+   __asm vec4_seq __retVal.xyz, u, v;
+}
+
+bvec4 equal(const vec4 u, const vec4 v)
+{
+   __asm vec4_seq __retVal, u, v;
+}
+
+bvec2 equal(const ivec2 u, const ivec2 v)
+{
+   __asm vec4_seq __retVal.xy, u, v;
+}
+
+bvec3 equal(const ivec3 u, const ivec3 v)
+{
+   __asm vec4_seq __retVal.xyz, u, v;
+}
+
+bvec4 equal(const ivec4 u, const ivec4 v)
+{
+   __asm vec4_seq __retVal, u, v;
+}
+
+bvec2 equal(const bvec2 u, const bvec2 v)
+{
+   __asm vec4_seq __retVal.xy, u, v;
+}
+
+bvec3 equal(const bvec3 u, const bvec3 v)
+{
+   __asm vec4_seq __retVal.xyz, u, v;
+}
+
+bvec4 equal(const bvec4 u, const bvec4 v)
+{
+   __asm vec4_seq __retVal, u, v;
+}
+
+
+
+
+//// notEqual
+
+bvec2 notEqual(const vec2 u, const vec2 v)
+{
+   __asm vec4_sne __retVal.xy, u, v;
+}
+
+bvec3 notEqual(const vec3 u, const vec3 v)
+{
+   __asm vec4_sne __retVal.xyz, u, v;
+}
+
+bvec4 notEqual(const vec4 u, const vec4 v)
+{
+   __asm vec4_sne __retVal, u, v;
+}
+
+bvec2 notEqual(const ivec2 u, const ivec2 v)
+{
+   __asm vec4_sne __retVal.xy, u, v;
+}
+
+bvec3 notEqual(const ivec3 u, const ivec3 v)
+{
+   __asm vec4_sne __retVal.xyz, u, v;
+}
+
+bvec4 notEqual(const ivec4 u, const ivec4 v)
+{
+   __asm vec4_sne __retVal, u, v;
+}
+
+bvec2 notEqual(const bvec2 u, const bvec2 v)
+{
+   __asm vec4_sne __retVal.xy, u, v;
+}
+
+bvec3 notEqual(const bvec3 u, const bvec3 v)
+{
+   __asm vec4_sne __retVal.xyz, u, v;
+}
+
+bvec4 notEqual(const bvec4 u, const bvec4 v)
+{
+   __asm vec4_sne __retVal, u, v;
+}
+
+
+
+//// any
+
+bool any(const bvec2 v)
+{
+   float sum;
+   __asm vec4_add sum.x, v.x, v.y;
+   __asm vec4_sne __retVal.x, sum.x, 0.0;
+}
+
+bool any(const bvec3 v)
+{
+   float sum;
+   __asm vec4_add sum.x, v.x, v.y;
+   __asm vec4_add sum.x, sum.x, v.z;
+   __asm vec4_sne __retVal.x, sum.x, 0.0;
+}
+
+bool any(const bvec4 v)
+{
+   float sum;
+   __asm vec4_add sum.x, v.x, v.y;
+   __asm vec4_add sum.x, sum.x, v.z;
+   __asm vec4_add sum.x, sum.x, v.w;
+   __asm vec4_sne __retVal.x, sum.x, 0.0;
+}
+
+
+//// all
+
+bool all (const bvec2 v)
+{
+   float prod;
+   __asm vec4_multiply prod, v.x, v.y;
+   __asm vec4_sne __retVal, prod, 0.0;
+}
+
+bool all (const bvec3 v)
+{
+   float prod;
+   __asm vec4_multiply prod, v.x, v.y;
+   __asm vec4_multiply prod, prod, v.z;
+   __asm vec4_sne __retVal, prod, 0.0;
+}
+
+bool all (const bvec4 v)
+{
+   float prod;
+   __asm vec4_multiply prod, v.x, v.y;
+   __asm vec4_multiply prod, prod, v.z;
+   __asm vec4_multiply prod, prod, v.w;
+   __asm vec4_sne __retVal, prod, 0.0;
+}
+
+
+
+//// not
+
+bvec2 not (const bvec2 v)
+{
+   __asm vec4_seq __retVal.xy, v, 0.0;
+}
+
+bvec3 not (const bvec3 v)
+{
+   __asm vec4_seq __retVal.xyz, v, 0.0;
+}
+
+bvec4 not (const bvec4 v)
+{
+   __asm vec4_seq __retVal, v, 0.0;
+}
+
+
+
+//// Texture Lookup Functions  (for both fragment and vertex shaders)
+
+vec4 texture1D(const sampler1D sampler, const float coord)
+{
+   __asm vec4_tex_1d __retVal, sampler, coord;
+}
+
+vec4 texture1DProj(const sampler1D sampler, const vec2 coord)
+{
+   // need to swizzle .y into .w
+   __asm vec4_tex_1d_proj __retVal, sampler, coord.xyyy;
+}
+
+vec4 texture1DProj(const sampler1D sampler, const vec4 coord)
+{
+   __asm vec4_tex_1d_proj __retVal, sampler, coord;
+}
+
+
+vec4 texture2D(const sampler2D sampler, const vec2 coord)
+{
+   __asm vec4_tex_2d __retVal, sampler, coord;
+}
+
+vec4 texture2DProj(const sampler2D sampler, const vec3 coord)
+{
+   // need to swizzle 'z' into 'w'.
+   __asm vec4_tex_2d_proj __retVal, sampler, coord.xyzz;
+}
+
+vec4 texture2DProj(const sampler2D sampler, const vec4 coord)
+{
+   __asm vec4_tex_2d_proj __retVal, sampler, coord;
+}
+
+
+vec4 texture3D(const sampler3D sampler, const vec3 coord)
+{
+   __asm vec4_tex_3d __retVal, sampler, coord;
+}
+
+vec4 texture3DProj(const sampler3D sampler, const vec4 coord)
+{
+   __asm vec4_tex_3d_proj __retVal, sampler, coord;
+}
+
+
+vec4 textureCube(const samplerCube sampler, const vec3 coord)
+{
+   __asm vec4_tex_cube __retVal, sampler, coord;
+}
+
+
+
+vec4 shadow1D(const sampler1DShadow sampler, const vec3 coord)
+{
+   __asm vec4_tex_1d_shadow __retVal, sampler, coord;
+}
+
+vec4 shadow1DProj(const sampler1DShadow sampler, const vec4 coord)
+{
+   // .s and .p will be divided by .q
+   __asm vec4_tex_1d_proj_shadow __retVal, sampler, coord;
+}
+
+vec4 shadow2D(const sampler2DShadow sampler, const vec3 coord)
+{
+   __asm vec4_tex_2d_shadow __retVal, sampler, coord;
+}
+
+vec4 shadow2DProj(const sampler2DShadow sampler, const vec4 coord)
+{
+   // .s, .t and .p will be divided by .q
+   __asm vec4_tex_2d_proj_shadow __retVal, sampler, coord;
+}
+
+
+//// GL_ARB_texture_rectangle:
+vec4 texture2DRect(const sampler2DRect sampler, const vec2 coord)
+{
+   __asm vec4_tex_rect __retVal, sampler, coord;
+}
+
+vec4 texture2DRectProj(const sampler2DRect sampler, const vec3 coord)
+{
+   // need to swizzle .y into .w
+   __asm vec4_tex_rect_proj __retVal, sampler, coord.xyzz;
+}
+
+vec4 texture2DRectProj(const sampler2DRect sampler, const vec4 coord)
+{
+   __asm vec4_tex_rect_proj __retVal, sampler, ccoord;
+}
+
+vec4 shadow2DRect(const sampler2DRectShadow sampler, const vec3 coord)
+{
+   __asm vec4_tex_rect_shadow __retVal, sampler, coord;
+}
+
+vec4 shadow2DRectProj(const sampler2DRectShadow sampler, const vec4 coord)
+{
+   __asm vec4_tex_rect_proj_shadow __retVal, sampler, coord;
+}
+
+
+
+//// GL_EXT_texture_array
+vec4 texture1DArray(const sampler1DArray sampler, const vec2 coord)
+{
+   __asm vec4_tex_1d_array __retVal, sampler, coord;
+}
+
+vec4 texture2DArray(const sampler2DArray sampler, const vec3 coord)
+{
+   __asm vec4_tex_2d_array __retVal, sampler, coord;
+}
+
+
+//
+// 8.9 Noise Functions
+//
+// AUTHOR: Stefan Gustavson (stegu@itn.liu.se), Nov 26, 2005
+//
+
+float noise1(const float x)
+{
+   __asm float_noise1 __retVal, x;
+}
+
+
+float noise1(const vec2 x)
+{
+    __asm float_noise2 __retVal, x;
+}
+
+float noise1(const vec3 x)
+{
+    __asm float_noise3 __retVal, x;
+}
+
+float noise1(const vec4 x)
+{
+    __asm float_noise4 __retVal, x;
+}
+
+vec2 noise2(const float x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + 19.34);
+}
+
+vec2 noise2(const vec2 x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + vec2(19.34, 7.66));
+}
+
+vec2 noise2(const vec3 x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + vec3(19.34, 7.66, 3.23));
+}
+
+vec2 noise2(const vec4 x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + vec4(19.34, 7.66, 3.23, 2.77));
+}
+
+vec3 noise3(const float x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + 19.34);
+   __retVal.z = noise1(x + 5.47);
+}
+
+vec3 noise3(const vec2 x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + vec2(19.34, 7.66));
+   __retVal.z = noise1(x + vec2(5.47, 17.85));
+}
+
+vec3 noise3(const vec3 x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + vec3(19.34, 7.66, 3.23));
+   __retVal.z = noise1(x + vec3(5.47, 17.85, 11.04));
+}
+
+vec3 noise3(const vec4 x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + vec4(19.34, 7.66, 3.23, 2.77));
+   __retVal.z = noise1(x + vec4(5.47, 17.85, 11.04, 13.19));
+}
+
+vec4 noise4(const float x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + 19.34);
+   __retVal.z = noise1(x + 5.47);
+   __retVal.w = noise1(x + 23.54);
+}
+
+vec4 noise4(const vec2 x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + vec2 (19.34, 7.66));
+   __retVal.z = noise1(x + vec2 (5.47, 17.85));
+   __retVal.w = noise1(x + vec2 (23.54, 29.11));
+}
+
+vec4 noise4(const vec3 x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + vec3(19.34, 7.66, 3.23));
+   __retVal.z = noise1(x + vec3(5.47, 17.85, 11.04));
+   __retVal.w = noise1(x + vec3(23.54, 29.11, 31.91));
+}
+
+vec4 noise4(const vec4 x)
+{
+   __retVal.x = noise1(x);
+   __retVal.y = noise1(x + vec4(19.34, 7.66, 3.23, 2.77));
+   __retVal.z = noise1(x + vec4(5.47, 17.85, 11.04, 13.19));
+   __retVal.w = noise1(x + vec4(23.54, 29.11, 31.91, 37.48));
+}
diff --git a/src/mesa/slang/library/slang_core.gc b/src/mesa/slang/library/slang_core.gc
new file mode 100644
index 0000000000..0a0d15903b
--- /dev/null
+++ b/src/mesa/slang/library/slang_core.gc
@@ -0,0 +1,2619 @@
+/*
+ * Mesa 3-D graphics library
+ * Version:  6.5
+ *
+ * Copyright (C) 2006  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+//
+// This file defines nearly all constructors and operators for built-in data
+// types, using extended language syntax. In general, compiler treats
+// constructors and operators as ordinary functions with some exceptions.
+// For example, the language does not allow functions to be called in
+// constant expressions - here the exception is made to allow it.
+//
+// Each implementation provides its own version of this file. Each
+// implementation can define the required set of operators and constructors
+// in its own fashion.
+//
+// The extended language syntax is only present when compiling this file.
+// It is implicitly included at the very beginning of the compiled shader,
+// so no built-in functions can be used.
+//
+// To communicate with the implementation, a special extended "__asm" keyword
+// is used, followed by an instruction name (any valid identifier), a
+// destination variable identifier and a list of zero or more source
+// variable identifiers.
+//
+// A variable identifier is a variable name declared earlier in the code
+// (as a function parameter, local or global variable).
+//
+// An instruction name designates an instruction that must be exported
+// by the implementation.  Each instruction receives data from source
+// variable identifiers and returns data in the destination variable
+// identifier.
+//
+// It is up to the implementation how to define a particular operator
+// or constructor. If it is expected to being used rarely, it can be
+// defined in terms of other operators and constructors,
+// for example:
+//
+// ivec2 __operator + (const ivec2 x, const ivec2 y) {
+//    return ivec2 (x[0] + y[0], x[1] + y[1]);
+// }
+//
+// If a particular operator or constructor is expected to be used very
+// often or is an atomic operation (that is, an operation that cannot be
+// expressed in terms of other operations or would create a dependency
+// cycle) it must be defined using one or more __asm constructs.
+//
+// Each implementation must define constructors for all scalar types
+// (bool, float, int).  There are 9 scalar-to-scalar constructors
+// (including identity constructors). However, since the language
+// introduces special constructors (like matrix constructor with a single
+// scalar value), implementations must also implement these cases.
+// The compiler provides the following algorithm when resolving a constructor:
+// - try to find a constructor with a prototype matching ours,
+// - if no constructor is found and this is a scalar-to-scalar constructor,
+//   raise an error,
+// - if a constructor is found, execute it and return,
+// - count the size of the constructor parameter list - if it is less than
+//   the size of our constructor's type, raise an error,
+// - for each parameter in the list do a recursive constructor matching for
+//   appropriate scalar fields in the constructed variable,
+//
+// Each implementation must also define a set of operators that deal with
+// built-in data types.
+// There are four kinds of operators:
+// 1) Operators that are implemented only by the compiler: "()" (function
+//    call), "," (sequence) and "?:" (selection).
+// 2) Operators that are implemented by the compiler by expressing it in
+//    terms of other operators:
+//    - "." (field selection) - translated to subscript access,
+//    - "&&" (logical and) - translated to "<left_expr> ? <right_expr> :
+//      false",
+//    - "||" (logical or) - translated to "<left_expr> ? true : <right_expr>",
+// 3) Operators that can be defined by the implementation and if the required
+//    prototype is not found, standard behaviour is used:
+//    - "==", "!=", "=" (equality, assignment) - compare or assign
+//      matching fields one-by-one;
+//      note that at least operators for scalar data types must be defined
+//      by the implementation to get it work,
+// 4) All other operators not mentioned above. If no required prototype is
+//    found, an error is raised. An implementation must follow the language
+//    specification to provide all valid operator prototypes.
+//
+
+
+
+//// Basic, scalar constructors/casts
+
+int __constructor(const float f)
+{
+   __asm vec4_to_ivec4 __retVal, f;
+}
+
+int __constructor(const bool b)
+{
+   __retVal = b;
+}
+
+int __constructor(const int i)
+{
+   __retVal = i;
+}
+
+bool __constructor(const int i)
+{
+   __asm vec4_sne __retVal, i, 0.0;
+}
+
+bool __constructor(const float f)
+{
+   __asm vec4_sne __retVal, f, 0.0;
+}
+
+bool __constructor(const bool b)
+{
+   __retVal = b;
+}
+
+float __constructor(const int i)
+{
+    __asm ivec4_to_vec4 __retVal, i;
+}
+
+float __constructor(const bool b)
+{
+    __asm ivec4_to_vec4 __retVal, b;
+}
+
+float __constructor(const float f)
+{
+   __retVal = f;
+}
+
+
+//// vec2 constructors
+
+vec2 __constructor(const float x, const float y)
+{
+   __retVal.x = x;
+   __retVal.y = y;
+}
+
+vec2 __constructor(const float f)
+{
+   __asm vec4_move __retVal.xy, f;
+}
+
+vec2 __constructor(const int i)
+{
+   __asm ivec4_to_vec4 __retVal.xy, i;
+}
+
+vec2 __constructor(const bool b)
+{
+   __asm ivec4_to_vec4 __retVal.xy, b;
+}
+
+vec2 __constructor(const bvec2 b)
+{
+//   __retVal = b;
+   __asm ivec4_to_vec4 __retVal.xy, b;
+}
+
+vec2 __constructor(const vec3 v)
+{
+   __asm vec4_move __retVal.xy, v.xy;
+}
+
+vec2 __constructor(const vec4 v)
+{
+   __asm vec4_move __retVal.xy, v.xy;
+}
+
+
+//// vec3 constructors
+
+vec3 __constructor(const float x, const float y, const float z)
+{
+   __retVal.x = x;
+   __retVal.y = y;
+   __retVal.z = z;
+}
+
+vec3 __constructor(const float f)
+{
+   // Note: this could be "__retVal.xyz = f" but that's an illegal assignment
+   __asm vec4_move __retVal.xyz, f;
+}
+
+vec3 __constructor(const int i)
+{
+   __asm ivec4_to_vec4 __retVal.xyz, i;
+}
+
+vec3 __constructor(const bool b)
+{
+   __asm ivec4_to_vec4 __retVal.xyz, b;
+}
+
+vec3 __constructor(const bvec3 b)
+{
+   __asm ivec4_to_vec4 __retVal.xyz, b;
+}
+
+vec3 __constructor(const vec4 v)
+{
+   __asm vec4_move __retVal.xyz, v;
+}
+
+
+//// vec4 constructors
+
+vec4 __constructor(const float x, const float y, const float z, const float w)
+{
+   __retVal.x = x;
+   __retVal.y = y;
+   __retVal.z = z;
+   __retVal.w = w;
+}
+
+vec4 __constructor(const float f)
+{
+   // Note: this could be "__retVal = f" but that's an illegal assignment
+   __asm vec4_move __retVal, f;
+}
+
+vec4 __constructor(const int i)
+{
+   __asm ivec4_to_vec4 __retVal, i;
+}
+
+vec4 __constructor(const bool b)
+{
+   __asm ivec4_to_vec4 __retVal, b;
+}
+
+vec4 __constructor(const bvec4 b)
+{
+   __asm ivec4_to_vec4 __retVal, b;
+}
+
+vec4 __constructor(const ivec4 i)
+{
+   __asm ivec4_to_vec4 __retVal, i;
+}
+
+vec4 __constructor(const vec3 v3, const float f)
+{
+   // XXX this constructor shouldn't be needed anymore
+   __retVal.xyz = v3;
+   __retVal.w = f;
+}
+
+vec4 __constructor(const vec2 v2, const float f1, const float f2)
+{
+   // XXX this constructor shouldn't be needed anymore
+   __retVal.xy = v2;
+   __retVal.z = f1;
+   __retVal.w = f2;
+}
+
+
+//// ivec2 constructors
+
+ivec2 __constructor(const int i, const int j)
+{
+   __retVal.x = i;
+   __retVal.y = j;
+}
+
+ivec2 __constructor(const int i)
+{
+   __asm vec4_move __retVal.xy, i;
+}
+
+ivec2 __constructor(const float f)
+{
+   __asm vec4_to_ivec4 __retVal.xy, f;
+}
+
+ivec2 __constructor(const bool b)
+{
+   __asm vec4_to_ivec4 __retVal.xy, b;
+}
+
+
+//// ivec3 constructors
+
+ivec3 __constructor(const int i, const int j, const int k)
+{
+   __retVal.x = i;
+   __retVal.y = j;
+   __retVal.z = k;
+}
+
+ivec3 __constructor(const int i)
+{
+   __asm vec4_move __retVal.xyz, i;
+}
+
+ivec3 __constructor(const float f)
+{
+   __asm vec4_to_ivec4 __retVal.xyz, f;
+}
+
+ivec3 __constructor(const bool b)
+{
+   __asm vec4_move __retVal.xyz, b;
+}
+
+
+//// ivec4 constructors
+
+ivec4 __constructor(const int x, const int y, const int z, const int w)
+{
+   __retVal.x = x;
+   __retVal.y = y;
+   __retVal.z = z;
+   __retVal.w = w;
+}
+
+ivec4 __constructor(const int i)
+{
+   __asm vec4_move __retVal, i;
+}
+
+ivec4 __constructor(const float f)
+{
+   __asm vec4_to_ivec4 __retVal, f;
+}
+
+ivec4 __constructor(const bool b)
+{
+   __asm vec4_to_ivec4 __retVal, b;
+}
+
+
+//// bvec2 constructors
+
+bvec2 __constructor(const bool b1, const bool b2)
+{
+   __retVal.x = b1;
+   __retVal.y = b2;
+}
+
+bvec2 __constructor(const int i1, const int i2)
+{
+   __asm vec4_sne __retVal.x, i1, 0.0;
+   __asm vec4_sne __retVal.y, i2, 0.0;
+}
+
+
+bvec2 __constructor(const bool b)
+{
+   __asm vec4_move __retVal.xy, b;
+}
+
+bvec2 __constructor(const float f)
+{
+   __asm vec4_sne __retVal.xy, f, 0.0;
+}
+
+bvec2 __constructor(const int i)
+{
+   __asm vec4_sne __retVal.xy, i, 0.0;
+}
+
+bvec2 __constructor(const vec2 v)
+{
+   __asm vec4_sne __retVal.xy, v, 0.0;
+}
+
+bvec2 __constructor(const ivec2 v)
+{
+   __asm vec4_sne __retVal.xy, v, 0.0;
+}
+
+
+
+//// bvec3 constructors
+
+bvec3 __constructor(const bool b1, const bool b2, const bool b3)
+{
+   __retVal.x = b1;
+   __retVal.y = b2;
+   __retVal.z = b3;
+}
+
+bvec3 __constructor(const float f1, const float f2, const float f3)
+{
+   __asm vec4_sne __retVal.x, f1, 0.0;
+   __asm vec4_sne __retVal.y, f2, 0.0;
+   __asm vec4_sne __retVal.z, f3, 0.0;
+}
+
+bvec3 __constructor(const bool b)
+{
+   __asm vec4_move __retVal.xyz, b;
+}
+
+bvec3 __constructor(const float f)
+{
+   __asm vec4_sne __retVal.xyz, f, 0.0;
+}
+
+bvec3 __constructor(const int i)
+{
+   __asm vec4_sne __retVal.xyz, i, 0.0;
+}
+
+bvec3 __constructor(const vec3 v)
+{
+   __asm vec4_sne __retVal.xyz, v, 0.0;
+}
+
+bvec3 __constructor(const ivec3 v)
+{
+   __asm vec4_sne __retVal.xyz, v, 0.0;
+}
+
+
+
+//// bvec4 constructors
+
+bvec4 __constructor(const bool b1, const bool b2, const bool b3, const bool b4)
+{
+   __retVal.x = b1;
+   __retVal.y = b2;
+   __retVal.z = b3;
+   __retVal.w = b4;
+}
+
+bvec4 __constructor(const float f1, const float f2, const float f3, const float f4)
+{
+   const float zero = 0.0;
+   __asm vec4_sne __retVal.x, f1, zero;   
+   __asm vec4_sne __retVal.y, f2, zero;   
+   __asm vec4_sne __retVal.z, f3, zero;   
+   __asm vec4_sne __retVal.w, f4, zero;   
+}
+
+bvec4 __constructor(const bool b)
+{
+   __asm vec4_move __retVal.xyzw, b;
+}
+
+bvec4 __constructor(const float f)
+{
+   __asm vec4_sne __retVal.xyzw, f, 0.0;
+}
+
+bvec4 __constructor(const int i)
+{
+   __asm vec4_sne __retVal.xyzw, i, 0.0;
+}
+
+bvec4 __constructor(const vec4 v)
+{
+   __asm vec4_sne __retVal.xyzw, v, 0.0;
+}
+
+bvec4 __constructor(const ivec4 v)
+{
+   __asm vec4_sne __retVal.xyzw, v, 0.0;
+}
+
+
+
+//// mat2 constructors
+
+mat2 __constructor(const float m00, const float m10,
+                   const float m01, const float m11)
+{
+   __retVal[0].x = m00;
+   __retVal[0].y = m10;
+   __retVal[1].x = m01;
+   __retVal[1].y = m11;
+}
+
+mat2 __constructor(const float f)
+{
+   __retVal[0].x = f;
+   __retVal[0].y = 0.0;
+   __retVal[1].x = 0.0;
+   __retVal[1].y = f;
+}
+
+mat2 __constructor(const int i)
+{
+   return mat2(float(i));
+}
+
+mat2 __constructor(const bool b)
+{
+   return mat2(float(b));
+}
+
+mat2 __constructor(const vec2 c0, const vec2 c1)
+{
+   __retVal[0] = c0;
+   __retVal[1] = c1;
+}
+
+
+//// mat3 constructors
+
+mat3 __constructor(const float m00, const float m10, const float m20,
+                   const float m01, const float m11, const float m21,
+                   const float m02, const float m12, const float m22)
+{
+   __retVal[0].x = m00;
+   __retVal[0].y = m10;
+   __retVal[0].z = m20;
+   __retVal[1].x = m01;
+   __retVal[1].y = m11;
+   __retVal[1].z = m21;
+   __retVal[2].x = m02;
+   __retVal[2].y = m12;
+   __retVal[2].z = m22;
+}
+
+mat3 __constructor(const float f)
+{
+   vec2 v = vec2(f, 0.0);
+   __retVal[0] = v.xyy;
+   __retVal[1] = v.yxy;
+   __retVal[2] = v.yyx;
+}
+
+mat3 __constructor(const int i)
+{
+   return mat3(float(i));
+}
+
+mat3 __constructor(const bool b)
+{
+   return mat3(float(b));
+}
+
+mat3 __constructor(const vec3 c0, const vec3 c1, const vec3 c2)
+{
+   __retVal[0] = c0;
+   __retVal[1] = c1;
+   __retVal[2] = c2;
+}
+
+
+//// mat4 constructors
+
+mat4 __constructor(const float m00, const float m10, const float m20, const float m30,
+                   const float m01, const float m11, const float m21, const float m31,
+                   const float m02, const float m12, const float m22, const float m32,
+                   const float m03, const float m13, const float m23, const float m33)
+{
+   __retVal[0].x = m00;
+   __retVal[0].y = m10;
+   __retVal[0].z = m20;
+   __retVal[0].w = m30;
+   __retVal[1].x = m01;
+   __retVal[1].y = m11;
+   __retVal[1].z = m21;
+   __retVal[1].w = m31;
+   __retVal[2].x = m02;
+   __retVal[2].y = m12;
+   __retVal[2].z = m22;
+   __retVal[2].w = m32;
+   __retVal[3].x = m03;
+   __retVal[3].y = m13;
+   __retVal[3].z = m23;
+   __retVal[3].w = m33;
+}
+
+
+mat4 __constructor(const float f)
+{
+   vec2 v = vec2(f, 0.0);
+   __retVal[0] = v.xyyy;
+   __retVal[1] = v.yxyy;
+   __retVal[2] = v.yyxy;
+   __retVal[3] = v.yyyx;
+}
+
+mat4 __constructor(const int i)
+{
+   return mat4(float(i));
+}
+
+mat4 __constructor(const bool b)
+{
+   return mat4(float(b));
+}
+
+mat4 __constructor(const vec4 c0, const vec4 c1, const vec4 c2, const vec4 c3)
+{
+   __retVal[0] = c0;
+   __retVal[1] = c1;
+   __retVal[2] = c2;
+   __retVal[3] = c3;
+}
+
+
+
+//// Basic int operators
+
+int __operator + (const int a, const int b)
+{
+   __asm vec4_add __retVal, a, b;
+}
+
+int __operator - (const int a, const int b)
+{
+   __asm vec4_subtract __retVal, a, b;
+}
+
+int __operator * (const int a, const int b)
+{
+   __asm vec4_multiply __retVal, a, b;
+}
+
+int __operator / (const int a, const int b)
+{
+   float bInv, x;
+   __asm float_rcp bInv, b;
+   __asm vec4_multiply x, a, bInv;
+   __asm vec4_to_ivec4 __retVal, x;
+}
+
+
+//// Basic ivec2 operators
+
+ivec2 __operator + (const ivec2 a, const ivec2 b)
+{
+   __asm vec4_add __retVal, a, b;
+}
+
+ivec2 __operator - (const ivec2 a, const ivec2 b)
+{
+   __asm vec4_subtract __retVal, a, b;
+}
+
+ivec2 __operator * (const ivec2 a, const ivec2 b)
+{
+   __asm vec4_multiply __retVal, a, b;
+}
+
+ivec2 __operator / (const ivec2 a, const ivec2 b)
+{
+   vec2 bInv, x;
+   __asm float_rcp bInv.x, b.x;
+   __asm float_rcp bInv.y, b.y;
+   __asm vec4_multiply x, a, bInv;
+   __asm vec4_to_ivec4 __retVal, x;
+}
+
+
+//// Basic ivec3 operators
+
+ivec3 __operator + (const ivec3 a, const ivec3 b)
+{
+   __asm vec4_add __retVal, a, b;
+}
+
+ivec3 __operator - (const ivec3 a, const ivec3 b)
+{
+   __asm vec4_subtract __retVal, a, b;
+}
+
+ivec3 __operator * (const ivec3 a, const ivec3 b)
+{
+   __asm vec4_multiply __retVal, a, b;
+}
+
+ivec3 __operator / (const ivec3 a, const ivec3 b)
+{
+   vec3 bInv, x;
+   __asm float_rcp bInv.x, b.x;
+   __asm float_rcp bInv.y, b.y;
+   __asm float_rcp bInv.z, b.z;
+   __asm vec4_multiply x, a, bInv;
+   __asm vec4_to_ivec4 __retVal, x;
+}
+
+
+//// Basic ivec4 operators
+
+ivec4 __operator + (const ivec4 a, const ivec4 b)
+{
+   __asm vec4_add __retVal, a, b;
+}
+
+ivec4 __operator - (const ivec4 a, const ivec4 b)
+{
+   __asm vec4_subtract __retVal, a, b;
+}
+
+ivec4 __operator * (const ivec4 a, const ivec4 b)
+{
+   __asm vec4_multiply __retVal, a, b;
+}
+
+ivec4 __operator / (const ivec4 a, const ivec4 b)
+{
+   vec4 bInv, x;
+   __asm float_rcp bInv.x, b.x;
+   __asm float_rcp bInv.y, b.y;
+   __asm float_rcp bInv.z, b.z;
+   __asm float_rcp bInv.w, b.w;
+   __asm vec4_multiply x, a, bInv;
+   __asm vec4_to_ivec4 __retVal, x;
+}
+
+
+//// Basic float operators
+
+float __operator + (const float a, const float b)
+{
+   __asm vec4_add __retVal, a, b;
+}
+
+float __operator - (const float a, const float b)
+{
+   __asm vec4_subtract __retVal, a, b;
+}
+
+float __operator * (const float a, const float b)
+{
+    __asm vec4_multiply __retVal, a, b;
+}
+
+float __operator / (const float a, const float b)
+{
+   float bInv;
+   __asm float_rcp bInv.x, b;
+   __asm vec4_multiply __retVal, a, bInv;
+}
+
+
+//// Basic vec2 operators
+
+vec2 __operator + (const vec2 v, const vec2 u)
+{
+   __asm vec4_add __retVal.xy, v, u;
+}
+
+vec2 __operator - (const vec2 v, const vec2 u)
+{
+    __asm vec4_subtract __retVal.xy, v, u;
+}
+
+vec2 __operator * (const vec2 v, const vec2 u)
+{
+    __asm vec4_multiply __retVal.xy, v, u;
+}
+
+vec2 __operator / (const vec2 v, const vec2 u)
+{
+   vec2 w; // = 1 / u
+   __asm float_rcp w.x, u.x;
+   __asm float_rcp w.y, u.y;
+   __asm vec4_multiply __retVal.xy, v, w;
+}
+
+
+//// Basic vec3 operators
+
+vec3 __operator + (const vec3 v, const vec3 u)
+{
+   __asm vec4_add __retVal.xyz, v, u;
+}
+
+vec3 __operator - (const vec3 v, const vec3 u)
+{
+    __asm vec4_subtract __retVal.xyz, v, u;
+}
+
+vec3 __operator * (const vec3 v, const vec3 u)
+{
+    __asm vec4_multiply __retVal.xyz, v, u;
+}
+
+vec3 __operator / (const vec3 v, const vec3 u)
+{
+   vec3 w; // = 1 / u
+   __asm float_rcp w.x, u.x;
+   __asm float_rcp w.y, u.y;
+   __asm float_rcp w.z, u.z;
+   __asm vec4_multiply __retVal.xyz, v, w;
+}
+
+
+//// Basic vec4 operators
+
+vec4 __operator + (const vec4 v, const vec4 u)
+{
+   __asm vec4_add __retVal, v, u;
+}
+
+vec4 __operator - (const vec4 v, const vec4 u)
+{
+    __asm vec4_subtract __retVal, v, u;
+}
+
+vec4 __operator * (const vec4 v, const vec4 u)
+{
+    __asm vec4_multiply __retVal, v, u;
+}
+
+vec4 __operator / (const vec4 v, const vec4 u)
+{
+   vec4 w; // = 1 / u
+   __asm float_rcp w.x, u.x;
+   __asm float_rcp w.y, u.y;
+   __asm float_rcp w.z, u.z;
+   __asm float_rcp w.w, u.w;
+   __asm vec4_multiply __retVal, v, w;
+}
+
+
+
+
+//// Basic vec2/float operators
+
+vec2 __operator + (const float a, const vec2 u)
+{
+   __asm vec4_add __retVal.xy, a, u.xy;
+}
+
+vec2 __operator + (const vec2 v, const float b)
+{
+   __asm vec4_add __retVal.xy, v.xy, b;
+}
+
+vec2 __operator - (const float a, const vec2 u)
+{
+   __asm vec4_subtract __retVal.xy, a, u.xy;
+}
+
+vec2 __operator - (const vec2 v, const float b)
+{
+   __asm vec4_subtract __retVal.xy, v.xy, b;
+}
+
+vec2 __operator * (const float a, const vec2 u)
+{
+   __asm vec4_multiply __retVal.xy, a, u.xy;
+}
+
+vec2 __operator * (const vec2 v, const float b)
+{
+   __asm vec4_multiply __retVal.xy, v.xy, b;
+}
+
+vec2 __operator / (const float a, const vec2 u)
+{
+   vec2 invU;
+   __asm float_rcp invU.x, u.x;
+   __asm float_rcp invU.y, u.y;
+   __asm vec4_multiply __retVal.xy, a, invU.xy;
+}
+
+vec2 __operator / (const vec2 v, const float b)
+{
+   float invB;
+   __asm float_rcp invB, b;
+   __asm vec4_multiply __retVal.xy, v.xy, invB;
+}
+
+
+//// Basic vec3/float operators
+
+vec3 __operator + (const float a, const vec3 u)
+{
+   __asm vec4_add __retVal.xyz, a, u.xyz;
+}
+
+vec3 __operator + (const vec3 v, const float b)
+{
+   __asm vec4_add __retVal.xyz, v.xyz, b;
+}
+
+vec3 __operator - (const float a, const vec3 u)
+{
+   __asm vec4_subtract __retVal.xyz, a, u.xyz;
+}
+
+vec3 __operator - (const vec3 v, const float b)
+{
+   __asm vec4_subtract __retVal.xyz, v.xyz, b;
+}
+
+vec3 __operator * (const float a, const vec3 u)
+{
+   __asm vec4_multiply __retVal.xyz, a, u.xyz;
+}
+
+vec3 __operator * (const vec3 v, const float b)
+{
+   __asm vec4_multiply __retVal.xyz, v.xyz, b;
+}
+
+vec3 __operator / (const float a, const vec3 u)
+{
+   vec3 invU;
+   __asm float_rcp invU.x, u.x;
+   __asm float_rcp invU.y, u.y;
+   __asm float_rcp invU.z, u.z;
+   __asm vec4_multiply __retVal.xyz, a, invU.xyz;
+}
+
+vec3 __operator / (const vec3 v, const float b)
+{
+   float invB;
+   __asm float_rcp invB, b;
+   __asm vec4_multiply __retVal.xyz, v.xyz, invB;
+}
+
+
+//// Basic vec4/float operators
+
+vec4 __operator + (const float a, const vec4 u)
+{
+   __asm vec4_add __retVal, a, u;
+}
+
+vec4 __operator + (const vec4 v, const float b)
+{
+   __asm vec4_add __retVal, v, b;
+}
+
+vec4 __operator - (const float a, const vec4 u)
+{
+   __asm vec4_subtract __retVal, a, u;
+}
+
+vec4 __operator - (const vec4 v, const float b)
+{
+   __asm vec4_subtract __retVal, v, b;
+}
+
+vec4 __operator * (const float a, const vec4 u)
+{
+   __asm vec4_multiply __retVal, a, u;
+}
+
+vec4 __operator * (const vec4 v, const float b)
+{
+   __asm vec4_multiply __retVal, v, b;
+}
+
+vec4 __operator / (const float a, const vec4 u)
+{
+   vec4 invU;
+   __asm float_rcp invU.x, u.x;
+   __asm float_rcp invU.y, u.y;
+   __asm float_rcp invU.z, u.z;
+   __asm float_rcp invU.w, u.w;
+   __asm vec4_multiply __retVal, a, invU;
+}
+
+vec4 __operator / (const vec4 v, const float b)
+{
+   float invB;
+   __asm float_rcp invB, b;
+   __asm vec4_multiply __retVal, v, invB;
+}
+
+
+
+//// Basic ivec2/int operators
+
+ivec2 __operator + (const int a, const ivec2 u)
+{
+   __retVal = ivec2(a) + u;
+}
+
+ivec2 __operator + (const ivec2 v, const int b)
+{
+   __retVal = v + ivec2(b);
+}
+
+ivec2 __operator - (const int a, const ivec2 u)
+{
+   __retVal = ivec2(a) - u;
+}
+
+ivec2 __operator - (const ivec2 v, const int b)
+{
+   __retVal = v - ivec2(b);
+}
+
+ivec2 __operator * (const int a, const ivec2 u)
+{
+   __retVal = ivec2(a) * u;
+}
+
+ivec2 __operator * (const ivec2 v, const int b)
+{
+   __retVal = v * ivec2(b);
+}
+
+ivec2 __operator / (const int a, const ivec2 u)
+{
+   __retVal = ivec2(a) / u;
+}
+
+ivec2 __operator / (const ivec2 v, const int b)
+{
+   __retVal = v / ivec2(b);
+}
+
+
+//// Basic ivec3/int operators
+
+ivec3 __operator + (const int a, const ivec3 u)
+{
+   __retVal = ivec3(a) + u;
+}
+
+ivec3 __operator + (const ivec3 v, const int b)
+{
+   __retVal = v + ivec3(b);
+}
+
+ivec3 __operator - (const int a, const ivec3 u)
+{
+   __retVal = ivec3(a) - u;
+}
+
+ivec3 __operator - (const ivec3 v, const int b)
+{
+   __retVal = v - ivec3(b);
+}
+
+ivec3 __operator * (const int a, const ivec3 u)
+{
+   __retVal = ivec3(a) * u;
+}
+
+ivec3 __operator * (const ivec3 v, const int b)
+{
+   __retVal = v * ivec3(b);
+}
+
+ivec3 __operator / (const int a, const ivec3 u)
+{
+   __retVal = ivec3(a) / u;
+}
+
+ivec3 __operator / (const ivec3 v, const int b)
+{
+   __retVal = v / ivec3(b);
+}
+
+
+//// Basic ivec4/int operators
+
+ivec4 __operator + (const int a, const ivec4 u)
+{
+   __retVal = ivec4(a) + u;
+}
+
+ivec4 __operator + (const ivec4 v, const int b)
+{
+   __retVal = v + ivec4(b);
+}
+
+ivec4 __operator - (const int a, const ivec4 u)
+{
+   __retVal = ivec4(a) - u;
+}
+
+ivec4 __operator - (const ivec4 v, const int b)
+{
+   __retVal = v - ivec4(b);
+}
+
+ivec4 __operator * (const int a, const ivec4 u)
+{
+   __retVal = ivec4(a) * u;
+}
+
+ivec4 __operator * (const ivec4 v, const int b)
+{
+   __retVal = v * ivec4(b);
+}
+
+ivec4 __operator / (const int a, const ivec4 u)
+{
+   __retVal = ivec4(a) / u;
+}
+
+ivec4 __operator / (const ivec4 v, const int b)
+{
+   __retVal = v / ivec4(b);
+}
+
+
+
+
+//// Unary negation operator
+
+int __operator - (const int a)
+{
+   __asm vec4_negate __retVal.x, a;
+}
+
+ivec2 __operator - (const ivec2 v)
+{
+   __asm vec4_negate __retVal, v;
+}
+
+ivec3 __operator - (const ivec3 v)
+{
+   __asm vec4_negate __retVal, v;
+}
+
+ivec4 __operator - (const ivec4 v)
+{
+   __asm vec4_negate __retVal, v;
+}
+
+float __operator - (const float a)
+{
+   __asm vec4_negate __retVal.x, a;
+}
+
+vec2 __operator - (const vec2 v)
+{
+   __asm vec4_negate __retVal.xy, v.xy;
+}
+
+vec3 __operator - (const vec3 v)
+{
+   __asm vec4_negate __retVal.xyz, v.xyz;
+}
+
+vec4 __operator - (const vec4 v)
+{
+   __asm vec4_negate __retVal, v;
+}
+
+mat2 __operator - (const mat2 m)
+{
+   __retVal[0] = -m[0];
+   __retVal[1] = -m[1];
+}
+
+mat3 __operator - (const mat3 m)
+{
+   __retVal[0] = -m[0];
+   __retVal[1] = -m[1];
+   __retVal[2] = -m[2];
+}
+
+mat4 __operator - (const mat4 m)
+{
+   __retVal[0] = -m[0];
+   __retVal[1] = -m[1];
+   __retVal[2] = -m[2];
+   __retVal[3] = -m[3];
+}
+
+
+
+//// dot product
+
+float dot(const float a, const float b)
+{
+   __retVal = a * b;
+}
+
+float dot(const vec2 a, const vec2 b)
+{
+   __retVal = a.x * b.x + a.y * b.y;
+}
+
+float dot(const vec3 a, const vec3 b)
+{
+    __asm vec3_dot __retVal, a, b;
+}
+
+float dot(const vec4 a, const vec4 b)
+{
+    __asm vec4_dot __retVal, a, b;
+}
+
+
+
+//// int assignment operators
+
+int __operator += (inout int a, const int b)
+{
+   a = a + b;
+   return a;
+}
+
+int __operator -= (inout int a, const int b)
+{
+   a = a - b;
+   return a;
+}
+
+int __operator *= (inout int a, const int b)
+{
+   a = a * b;
+   return a;
+}
+
+int __operator /= (inout int a, const int b)
+{
+   a = a / b;
+   return a;
+}
+
+
+//// ivec2 assignment operators
+
+ivec2 __operator += (inout ivec2 v, const ivec2 u)
+{
+   v = v + u;
+   return v;
+}
+
+ivec2 __operator -= (inout ivec2 v, const ivec2 u)
+{
+   v = v - u;
+   return v;
+}
+
+ivec2 __operator *= (inout ivec2 v, const ivec2 u)
+{
+   v = v * u;
+   return v;
+}
+
+ivec2 __operator /= (inout ivec2 v, const ivec2 u)
+{
+   v = v / u;
+   return v;
+}
+
+
+//// ivec3 assignment operators
+
+ivec3 __operator += (inout ivec3 v, const ivec3 u)
+{
+   v = v + u;
+   return v;
+}
+
+ivec3 __operator -= (inout ivec3 v, const ivec3 u)
+{
+   v = v - u;
+   return v;
+}
+
+ivec3 __operator *= (inout ivec3 v, const ivec3 u)
+{
+   v = v * u;
+   return v;
+}
+
+ivec3 __operator /= (inout ivec3 v, const ivec3 u)
+{
+   v = v / u;
+   return v;
+}
+
+
+//// ivec4 assignment operators
+
+ivec4 __operator += (inout ivec4 v, const ivec4 u)
+{
+   v = v + u;
+   return v;
+}
+
+ivec4 __operator -= (inout ivec4 v, const ivec4 u)
+{
+   v = v - u;
+   return v;
+}
+
+ivec4 __operator *= (inout ivec4 v, const ivec4 u)
+{
+   v = v * u;
+   return v;
+}
+
+ivec4 __operator /= (inout ivec4 v, const ivec4 u)
+{
+   v = v / u;
+   return v;
+}
+
+
+//// float assignment operators
+
+float __operator += (inout float a, const float b)
+{
+   a = a + b;
+   return a;
+}
+
+float __operator -= (inout float a, const float b)
+{
+   a = a - b;
+   return a;
+}
+
+float __operator *= (inout float a, const float b)
+{
+   a = a * b;
+   return a;
+}
+
+float __operator /= (inout float a, const float b)
+{
+   a = a / b;
+   return a;
+}
+
+
+//// vec2 assignment operators
+
+vec2 __operator += (inout vec2 v, const vec2 u)
+{
+   v = v + u;
+   return v;
+}
+
+vec2 __operator -= (inout vec2 v, const vec2 u)
+{
+   v = v - u;
+   return v;
+}
+
+vec2 __operator *= (inout vec2 v, const vec2 u)
+{
+   v = v * u;
+   return v;
+}
+
+vec2 __operator /= (inout vec2 v, const vec2 u)
+{
+   v = v / u;
+   return v;
+}
+
+
+//// vec3 assignment operators
+
+vec3 __operator += (inout vec3 v, const vec3 u)
+{
+   v = v + u;
+   return v;
+}
+
+vec3 __operator -= (inout vec3 v, const vec3 u)
+{
+   v = v - u;
+   return v;
+}
+
+vec3 __operator *= (inout vec3 v, const vec3 u)
+{
+   v = v * u;
+   return v;
+}
+
+vec3 __operator /= (inout vec3 v, const vec3 u)
+{
+   v = v / u;
+   return v;
+}
+
+
+//// vec4 assignment operators
+
+vec4 __operator += (inout vec4 v, const vec4 u)
+{
+   v = v + u;
+   return v;
+}
+
+vec4 __operator -= (inout vec4 v, const vec4 u)
+{
+   v = v - u;
+   return v;
+}
+
+vec4 __operator *= (inout vec4 v, const vec4 u)
+{
+   v = v * u;
+   return v;
+}
+
+vec4 __operator /= (inout vec4 v, const vec4 u)
+{
+   v = v / u;
+   return v;
+}
+
+
+
+//// ivec2/int assignment operators
+
+ivec2 __operator += (inout ivec2 v, const int a)
+{
+   v = v + ivec2(a);
+   return v;
+}
+
+ivec2 __operator -= (inout ivec2 v, const int a)
+{
+   v = v - ivec2(a);
+   return v;
+}
+
+ivec2 __operator *= (inout ivec2 v, const int a)
+{
+   v = v * ivec2(a);
+   return v;
+}
+
+ivec2 __operator /= (inout ivec2 v, const int a)
+{
+   v = v / ivec2(a);
+   return v;
+}
+
+
+//// ivec3/int assignment operators
+
+ivec3 __operator += (inout ivec3 v, const int a)
+{
+   v = v + ivec3(a);
+   return v;
+}
+
+ivec3 __operator -= (inout ivec3 v, const int a)
+{
+   v = v - ivec3(a);
+   return v;
+}
+
+ivec3 __operator *= (inout ivec3 v, const int a)
+{
+   v = v * ivec3(a);
+   return v;
+}
+
+ivec4 __operator /= (inout ivec3 v, const int a)
+{
+   v = v / ivec3(a);
+   return v;
+}
+
+
+//// ivec4/int assignment operators
+
+ivec4 __operator += (inout ivec4 v, const int a)
+{
+   v = v + ivec4(a);
+   return v;
+}
+
+ivec4 __operator -= (inout ivec4 v, const int a)
+{
+   v = v - ivec4(a);
+   return v;
+}
+
+ivec4 __operator *= (inout ivec4 v, const int a)
+{
+   v = v * ivec4(a);
+   return v;
+}
+
+ivec4 __operator /= (inout ivec4 v, const int a)
+{
+   v = v / ivec4(a);
+   return v;
+}
+
+
+
+//// vec2/float assignment operators
+
+vec2 __operator += (inout vec2 v, const float a)
+{
+   v = v + vec2(a);
+   return v;
+}
+
+vec2 __operator -= (inout vec2 v, const float a)
+{
+   v = v - vec2(a);
+   return v;
+}
+
+vec2 __operator *= (inout vec2 v, const float a)
+{
+   v = v * vec2(a);
+   return v;
+}
+
+vec2 __operator /= (inout vec2 v, const float a)
+{
+   v = v / vec2(a);
+   return v;
+}
+
+
+//// vec3/float assignment operators
+
+vec3 __operator += (inout vec3 v, const float a)
+{
+   v = v + vec3(a);
+   return v;
+}
+
+vec3 __operator -= (inout vec3 v, const float a)
+{
+   v = v - vec3(a);
+   return v;
+}
+
+vec3 __operator *= (inout vec3 v, const float a)
+{
+   v = v * vec3(a);
+   return v;
+}
+
+vec3 __operator /= (inout vec3 v, const float a)
+{
+   v = v / vec3(a);
+   return v;
+}
+
+
+//// vec4/float assignment operators
+
+vec4 __operator += (inout vec4 v, const float a)
+{
+   v = v + vec4(a);
+   return v;
+}
+
+vec4 __operator -= (inout vec4 v, const float a)
+{
+   v = v - vec4(a);
+   return v;
+}
+
+vec4 __operator *= (inout vec4 v, const float a)
+{
+   v = v * vec4(a);
+   return v;
+}
+
+vec4 __operator /= (inout vec4 v, const float a)
+{
+   v = v / vec4(a);
+   return v;
+}
+
+
+
+
+
+//// Basic mat2 operations
+
+mat2 __operator + (const mat2 m, const mat2 n)
+{
+   __retVal[0] = m[0] + n[0];
+   __retVal[1] = m[1] + n[1];
+}
+
+mat2 __operator - (const mat2 m, const mat2 n)
+{
+   __retVal[0] = m[0] - n[0];
+   __retVal[1] = m[1] - n[1];
+}
+
+mat2 __operator * (const mat2 m, const mat2 n)
+{
+   __retVal[0] = m[0] * n[0].xx + m[1] * n[0].yy;
+   __retVal[1] = m[0] * n[1].xx + m[1] * n[1].yy;
+}
+
+mat2 __operator / (const mat2 m, const mat2 n)
+{
+   __retVal[0] = m[0] / n[0];
+   __retVal[1] = m[1] / n[1];
+}
+
+
+//// Basic mat3 operations
+
+mat3 __operator + (const mat3 m, const mat3 n)
+{
+   __retVal[0] = m[0] + n[0];
+   __retVal[1] = m[1] + n[1];
+   __retVal[2] = m[2] + n[2];
+}
+
+mat3 __operator - (const mat3 m, const mat3 n)
+{
+   __retVal[0] = m[0] - n[0];
+   __retVal[1] = m[1] - n[1];
+   __retVal[2] = m[2] - n[2];
+}
+
+mat3 __operator * (const mat3 m, const mat3 n)
+{
+   __retVal[0] = m[0] * n[0].xxx + m[1] * n[0].yyy + m[2] * n[0].zzz;
+   __retVal[1] = m[0] * n[1].xxx + m[1] * n[1].yyy + m[2] * n[1].zzz;
+   __retVal[2] = m[0] * n[2].xxx + m[1] * n[2].yyy + m[2] * n[2].zzz;
+}
+
+mat3 __operator / (const mat3 m, const mat3 n)
+{
+    __retVal[0] = m[0] / n[0];
+    __retVal[1] = m[1] / n[1];
+    __retVal[2] = m[2] / n[2];
+}
+
+
+//// Basic mat4 operations
+
+mat4 __operator + (const mat4 m, const mat4 n)
+{
+   __retVal[0] = m[0] + n[0];
+   __retVal[1] = m[1] + n[1];
+   __retVal[2] = m[2] + n[2];
+   __retVal[3] = m[3] + n[3];
+}
+
+mat4 __operator - (const mat4 m, const mat4 n)
+{
+   __retVal[0] = m[0] - n[0];
+   __retVal[1] = m[1] - n[1];
+   __retVal[2] = m[2] - n[2];
+   __retVal[3] = m[3] - n[3];
+}
+
+mat4 __operator * (const mat4 m, const mat4 n)
+{
+   __retVal[0] = m[0] * n[0].xxxx + m[1] * n[0].yyyy + m[2] * n[0].zzzz + m[3] * n[0].wwww;
+   __retVal[1] = m[0] * n[1].xxxx + m[1] * n[1].yyyy + m[2] * n[1].zzzz + m[3] * n[1].wwww;
+   __retVal[2] = m[0] * n[2].xxxx + m[1] * n[2].yyyy + m[2] * n[2].zzzz + m[3] * n[2].wwww;
+   __retVal[3] = m[0] * n[3].xxxx + m[1] * n[3].yyyy + m[2] * n[3].zzzz + m[3] * n[3].wwww;
+}
+
+mat4 __operator / (const mat4 m, const mat4 n)
+{
+    __retVal[0] = m[0] / n[0];
+    __retVal[1] = m[1] / n[1];
+    __retVal[2] = m[2] / n[2];
+    __retVal[3] = m[3] / n[3];
+}
+
+
+//// mat2/float operations
+
+mat2 __operator + (const float a, const mat2 n)
+{
+   __retVal[0] = a + n[0];
+   __retVal[1] = a + n[1];
+}
+
+mat2 __operator + (const mat2 m, const float b)
+{
+   __retVal[0] = m[0] + b;
+   __retVal[1] = m[1] + b;
+}
+
+mat2 __operator - (const float a, const mat2 n)
+{
+   __retVal[0] = a - n[0];
+   __retVal[1] = a - n[1];
+}
+
+mat2 __operator - (const mat2 m, const float b)
+{
+   __retVal[0] = m[0] - b;
+   __retVal[1] = m[1] - b;
+}
+
+mat2 __operator * (const float a, const mat2 n)
+{
+   __retVal[0] = a * n[0];
+   __retVal[1] = a * n[1];
+}
+
+mat2 __operator * (const mat2 m, const float b)
+{
+   __retVal[0] = m[0] * b;
+   __retVal[1] = m[1] * b;
+}
+
+mat2 __operator / (const float a, const mat2 n)
+{
+   __retVal[0] = a / n[0];
+   __retVal[1] = a / n[1];
+}
+
+mat2 __operator / (const mat2 m, const float b)
+{
+   __retVal[0] = m[0] / b;
+   __retVal[1] = m[1] / b;
+}
+
+
+//// mat3/float operations
+
+mat3 __operator + (const float a, const mat3 n)
+{
+   __retVal[0] = a + n[0];
+   __retVal[1] = a + n[1];
+   __retVal[2] = a + n[2];
+}
+
+mat3 __operator + (const mat3 m, const float b)
+{
+   __retVal[0] = m[0] + b;
+   __retVal[1] = m[1] + b;
+   __retVal[2] = m[2] + b;
+}
+
+mat3 __operator - (const float a, const mat3 n)
+{
+   __retVal[0] = a - n[0];
+   __retVal[1] = a - n[1];
+   __retVal[2] = a - n[2];
+}
+
+mat3 __operator - (const mat3 m, const float b)
+{
+   __retVal[0] = m[0] - b;
+   __retVal[1] = m[1] - b;
+   __retVal[2] = m[2] - b;
+}
+
+mat3 __operator * (const float a, const mat3 n)
+{
+   __retVal[0] = a * n[0];
+   __retVal[1] = a * n[1];
+   __retVal[2] = a * n[2];
+}
+
+mat3 __operator * (const mat3 m, const float b)
+{
+   __retVal[0] = m[0] * b;
+   __retVal[1] = m[1] * b;
+   __retVal[2] = m[2] * b;
+}
+
+mat3 __operator / (const float a, const mat3 n)
+{
+   __retVal[0] = a / n[0];
+   __retVal[1] = a / n[1];
+   __retVal[2] = a / n[2];
+}
+
+mat3 __operator / (const mat3 m, const float b)
+{
+   __retVal[0] = m[0] / b;
+   __retVal[1] = m[1] / b;
+   __retVal[2] = m[2] / b;
+}
+
+
+//// mat4/float operations
+
+mat4 __operator + (const float a, const mat4 n)
+{
+   __retVal[0] = a + n[0];
+   __retVal[1] = a + n[1];
+   __retVal[2] = a + n[2];
+   __retVal[3] = a + n[3];
+}
+
+mat4 __operator + (const mat4 m, const float b)
+{
+   __retVal[0] = m[0] + b;
+   __retVal[1] = m[1] + b;
+   __retVal[2] = m[2] + b;
+   __retVal[3] = m[3] + b;
+}
+
+mat4 __operator - (const float a, const mat4 n)
+{
+   __retVal[0] = a - n[0];
+   __retVal[1] = a - n[1];
+   __retVal[2] = a - n[2];
+   __retVal[3] = a - n[3];
+}
+
+mat4 __operator - (const mat4 m, const float b)
+{
+   __retVal[0] = m[0] - b;
+   __retVal[1] = m[1] - b;
+   __retVal[2] = m[2] - b;
+   __retVal[3] = m[3] - b;
+}
+
+mat4 __operator * (const float a, const mat4 n)
+{
+   __retVal[0] = a * n[0];
+   __retVal[1] = a * n[1];
+   __retVal[2] = a * n[2];
+   __retVal[3] = a * n[3];
+}
+
+mat4 __operator * (const mat4 m, const float b)
+{
+   __retVal[0] = m[0] * b;
+   __retVal[1] = m[1] * b;
+   __retVal[2] = m[2] * b;
+   __retVal[3] = m[3] * b;
+}
+
+mat4 __operator / (const float a, const mat4 n)
+{
+   __retVal[0] = a / n[0];
+   __retVal[1] = a / n[1];
+   __retVal[2] = a / n[2];
+   __retVal[3] = a / n[3];
+}
+
+mat4 __operator / (const mat4 m, const float b)
+{
+   __retVal[0] = m[0] / b;
+   __retVal[1] = m[1] / b;
+   __retVal[2] = m[2] / b;
+   __retVal[3] = m[3] / b;
+}
+
+
+
+//// matrix / vector products
+
+vec2 __operator * (const mat2 m, const vec2 v)
+{
+   __retVal = m[0] * v.xx
+            + m[1] * v.yy;
+}
+
+vec2 __operator * (const vec2 v, const mat2 m)
+{
+   __retVal.x = dot(v, m[0]);
+   __retVal.y = dot(v, m[1]);
+}
+
+vec3 __operator * (const mat3 m, const vec3 v)
+{
+   __retVal = m[0] * v.xxx
+            + m[1] * v.yyy
+            + m[2] * v.zzz;
+}
+
+vec3 __operator * (const vec3 v, const mat3 m)
+{
+   __retVal.x = dot(v, m[0]);
+   __retVal.y = dot(v, m[1]);
+   __retVal.z = dot(v, m[2]);
+}
+
+vec4 __operator * (const mat4 m, const vec4 v)
+{
+   __retVal = m[0] * v.xxxx
+            + m[1] * v.yyyy
+            + m[2] * v.zzzz
+            + m[3] * v.wwww;
+}
+
+vec4 __operator * (const vec4 v, const mat4 m)
+{
+   __retVal.x = dot(v, m[0]);
+   __retVal.y = dot(v, m[1]);
+   __retVal.z = dot(v, m[2]);
+   __retVal.w = dot(v, m[3]);
+}
+
+
+
+//// mat2 assignment operators
+
+mat2 __operator += (inout mat2 m, const mat2 n)
+{
+   m[0] = m[0] + n[0];
+   m[1] = m[1] + n[1];
+   return m;
+}
+
+mat2 __operator -= (inout mat2 m, const mat2 n)
+{
+   m[0] = m[0] - n[0];
+   m[1] = m[1] - n[1];
+   return m;
+}
+
+mat2 __operator *= (inout mat2 m, const mat2 n)
+{
+   m = m * n;
+   return m;
+}
+
+mat2 __operator /= (inout mat2 m, const mat2 n)
+{
+   m[0] = m[0] / n[0];
+   m[1] = m[1] / n[1];
+   return m;
+}
+
+
+//// mat3 assignment operators
+
+mat3 __operator += (inout mat3 m, const mat3 n)
+{
+   m[0] = m[0] + n[0];
+   m[1] = m[1] + n[1];
+   m[2] = m[2] + n[2];
+   return m;
+}
+
+mat3 __operator -= (inout mat3 m, const mat3 n)
+{
+   m[0] = m[0] - n[0];
+   m[1] = m[1] - n[1];
+   m[2] = m[2] - n[2];
+   return m;
+}
+
+mat3 __operator *= (inout mat3 m, const mat3 n)
+{
+   m = m * n;
+   return m;
+}
+
+mat3 __operator /= (inout mat3 m, const mat3 n)
+{
+   m[0] = m[0] / n[0];
+   m[1] = m[1] / n[1];
+   m[2] = m[2] / n[2];
+   return m;
+}
+
+
+// mat4 assignment operators
+
+mat4 __operator += (inout mat4 m, const mat4 n)
+{
+   m[0] = m[0] + n[0];
+   m[1] = m[1] + n[1];
+   m[2] = m[2] + n[2];
+   m[3] = m[3] + n[3];
+   return m;
+}
+
+mat4 __operator -= (inout mat4 m, const mat4 n)
+{
+   m[0] = m[0] - n[0];
+   m[1] = m[1] - n[1];
+   m[2] = m[2] - n[2];
+   m[3] = m[3] - n[3];
+   return m;
+}
+
+mat4 __operator *= (inout mat4 m, const mat4 n)
+{
+   m = m * n;
+   return m;
+}
+
+mat4 __operator /= (inout mat4 m, const mat4 n)
+{
+   m[0] = m[0] / n[0];
+   m[1] = m[1] / n[1];
+   m[2] = m[2] / n[2];
+   m[3] = m[3] / n[3];
+   return m;
+}
+
+
+//// mat2/float assignment operators
+
+mat2 __operator += (inout mat2 m, const float a)
+{
+   vec2 v = vec2(a);
+   m[0] = m[0] + v;
+   m[1] = m[1] + v;
+   return m;
+}
+
+mat2 __operator -= (inout mat2 m, const float a)
+{
+   vec2 v = vec2(a);
+   m[0] = m[0] - v;
+   m[1] = m[1] - v;
+   return m;
+}
+
+mat2 __operator *= (inout mat2 m, const float a)
+{
+   vec2 v = vec2(a);
+   m[0] = m[0] * v;
+   m[1] = m[1] * v;
+   return m;
+}
+
+mat2 __operator /= (inout mat2 m, const float a)
+{
+   vec2 v = vec2(1.0 / a);
+   m[0] = m[0] * v;
+   m[1] = m[1] * v;
+   return m;
+}
+
+
+//// mat3/float assignment operators
+
+mat3 __operator += (inout mat3 m, const float a)
+{
+   vec3 v = vec3(a);
+   m[0] = m[0] + v;
+   m[1] = m[1] + v;
+   m[2] = m[2] + v;
+   return m;
+}
+
+mat3 __operator -= (inout mat3 m, const float a)
+{
+   vec3 v = vec3(a);
+   m[0] = m[0] - v;
+   m[1] = m[1] - v;
+   m[2] = m[2] - v;
+   return m;
+}
+
+mat3 __operator *= (inout mat3 m, const float a)
+{
+   vec3 v = vec3(a);
+   m[0] = m[0] * v;
+   m[1] = m[1] * v;
+   m[2] = m[2] * v;
+   return m;
+}
+
+mat3 __operator /= (inout mat3 m, const float a)
+{
+   vec3 v = vec3(1.0 / a);
+   m[0] = m[0] * v;
+   m[1] = m[1] * v;
+   m[2] = m[2] * v;
+   return m;
+}
+
+
+//// mat4/float assignment operators
+
+mat4 __operator += (inout mat4 m, const float a)
+{
+   vec4 v = vec4(a);
+   m[0] = m[0] + v;
+   m[1] = m[1] + v;
+   m[2] = m[2] + v;
+   m[3] = m[3] + v;
+   return m;
+}
+
+mat4 __operator -= (inout mat4 m, const float a)
+{
+   vec4 v = vec4(a);
+   m[0] = m[0] - v;
+   m[1] = m[1] - v;
+   m[2] = m[2] - v;
+   m[3] = m[3] - v;
+   return m;
+}
+
+mat4 __operator *= (inout mat4 m, const float a)
+{
+   vec4 v = vec4(a);
+   m[0] = m[0] * v;
+   m[1] = m[1] * v;
+   m[2] = m[2] * v;
+   m[3] = m[3] * v;
+   return m;
+}
+
+mat4 __operator /= (inout mat4 m, const float a)
+{
+   vec4 v = vec4(1.0 / a);
+   m[0] = m[0] * v;
+   m[1] = m[1] * v;
+   m[2] = m[2] * v;
+   m[3] = m[3] * v;
+   return m;
+}
+
+
+
+//// vec/mat assignment operators
+
+vec2 __operator *= (inout vec2 v, const mat2 m)
+{
+   v = v * m;
+   return v;
+}
+
+vec3 __operator *= (inout vec3 v, const mat3 m)
+{
+   v = v * m;
+   return v;
+}
+
+vec4 __operator *= (inout vec4 v, const mat4 m)
+{
+   v = v * m;
+   return v;
+}
+
+
+
+//// pre-decrement operators
+
+int __operator --(inout int a)
+{
+    a = a - 1;
+   __retVal = a;
+}
+
+ivec2 __operator --(inout ivec2 v)
+{
+   v = v - ivec2(1);
+   __retVal = v;
+}
+
+ivec3 __operator --(inout ivec3 v)
+{
+   v = v - ivec3(1);
+   __retVal = v;
+}
+
+ivec4 __operator --(inout ivec4 v)
+{
+   v = v - ivec4(1);
+   __retVal = v;
+}
+
+
+float __operator --(inout float a)
+{
+   a = a - 1.0;
+   __retVal = a;
+}
+
+vec2 __operator --(inout vec2 v)
+{
+   v = v - vec2(1.0);
+   __retVal = v;
+}
+
+vec3 __operator --(inout vec3 v)
+{
+   v = v - vec3(1.0);
+   __retVal = v;
+}
+
+vec4 __operator --(inout vec4 v)
+{
+   v = v - vec4(1.0);
+   __retVal = v;
+}
+
+
+mat2 __operator --(inout mat2 m)
+{
+   m[0] = m[0] - vec2(1.0);
+   m[1] = m[1] - vec2(1.0);
+   __retVal = m;
+}
+
+mat3 __operator --(inout mat3 m)
+{
+   m[0] = m[0] - vec3(1.0);
+   m[1] = m[1] - vec3(1.0);
+   m[2] = m[2] - vec3(1.0);
+   __retVal = m;
+}
+
+mat4 __operator --(inout mat4 m)
+{
+   m[0] = m[0] - vec4(1.0);
+   m[1] = m[1] - vec4(1.0);
+   m[2] = m[2] - vec4(1.0);
+   m[3] = m[3] - vec4(1.0);
+   __retVal = m;
+}
+
+
+//// pre-increment operators
+
+int __operator ++(inout int a)
+{
+    a = a + 1;
+    __retVal = a;
+}
+
+ivec2 __operator ++(inout ivec2 v)
+{
+   v = v + ivec2(1);
+   __retVal = v;
+}
+
+ivec3 __operator ++(inout ivec3 v)
+{
+   v = v + ivec3(1);
+   __retVal = v;
+}
+
+ivec4 __operator ++(inout ivec4 v)
+{
+   v = v + ivec4(1);
+   __retVal = v;
+}
+
+
+float __operator ++(inout float a)
+{
+    a = a + 1.0;
+    __retVal = a;
+}
+
+vec2 __operator ++(inout vec2 v)
+{
+   v = v + vec2(1.0);
+   __retVal = v;
+}
+
+vec3 __operator ++(inout vec3 v)
+{
+   v = v + vec3(1.0);
+   __retVal = v;
+}
+
+vec4 __operator ++(inout vec4 v)
+{
+   v = v + vec4(1.0);
+   __retVal = v;
+}
+
+
+mat2 __operator ++(inout mat2 m)
+{
+   m[0] = m[0] + vec2(1.0);
+   m[1] = m[1] + vec2(1.0);
+   __retVal = m;
+}
+
+mat3 __operator ++(inout mat3 m)
+{
+   m[0] = m[0] + vec3(1.0);
+   m[1] = m[1] + vec3(1.0);
+   m[2] = m[2] + vec3(1.0);
+   __retVal = m;
+}
+
+mat4 __operator ++(inout mat4 m)
+{
+   m[0] = m[0] + vec4(1.0);
+   m[1] = m[1] + vec4(1.0);
+   m[2] = m[2] + vec4(1.0);
+   m[3] = m[3] + vec4(1.0);
+   __retVal = m;
+}
+
+
+
+//// post-decrement
+
+int __postDecr(inout int a)
+{
+   __retVal = a;
+   a = a - 1;
+}
+
+ivec2 __postDecr(inout ivec2 v)
+{
+   __retVal = v;
+   v = v - ivec2(1);
+}
+
+ivec3 __postDecr(inout ivec3 v)
+{
+   __retVal = v;
+   v = v - ivec3(1);
+}
+
+ivec4 __postDecr(inout ivec4 v)
+{
+   __retVal = v;
+   v = v - ivec4(1);
+}
+
+
+float __postDecr(inout float a)
+{
+   __retVal = a;
+   a = a - 1.0;
+}
+
+vec2 __postDecr(inout vec2 v)
+{
+   __retVal = v;
+   v = v - vec2(1.0);
+}
+
+vec3 __postDecr(inout vec3 v)
+{
+   __retVal = v;
+   v = v - vec3(1.0);
+}
+
+vec4 __postDecr(inout vec4 v)
+{
+   __retVal = v;
+   v = v - vec4(1.0);
+}
+
+
+mat2 __postDecr(inout mat2 m)
+{
+   __retVal = m;
+   m[0] = m[0] - vec2(1.0);
+   m[1] = m[1] - vec2(1.0);
+}
+
+mat3 __postDecr(inout mat3 m)
+{
+   __retVal = m;
+   m[0] = m[0] - vec3(1.0);
+   m[1] = m[1] - vec3(1.0);
+   m[2] = m[2] - vec3(1.0);
+}
+
+mat4 __postDecr(inout mat4 m)
+{
+   __retVal = m;
+   m[0] = m[0] - vec4(1.0);
+   m[1] = m[1] - vec4(1.0);
+   m[2] = m[2] - vec4(1.0);
+   m[3] = m[3] - vec4(1.0);
+}
+
+
+//// post-increment
+
+float __postIncr(inout float a)
+{
+   __retVal = a;
+   a = a + 1;
+}
+
+vec2 __postIncr(inout vec2 v)
+{
+   __retVal = v;
+   v = v + vec2(1.0);
+}
+
+vec3 __postIncr(inout vec3 v)
+{
+   __retVal = v;
+   v = v + vec3(1.0);
+}
+
+vec4 __postIncr(inout vec4 v)
+{
+   __retVal = v;
+   v = v + vec4(1.0);
+}
+
+
+int __postIncr(inout int a)
+{
+   __retVal = a;
+   a = a + 1;
+}
+
+ivec2 __postIncr(inout ivec2 v)
+{
+   __retVal = v;
+   v = v + ivec2(1);
+}
+
+ivec3 __postIncr(inout ivec3 v)
+{
+   __retVal = v;
+   v = v + ivec3(1);
+}
+
+ivec4 __postIncr(inout ivec4 v)
+{
+   __retVal = v;
+   v = v + ivec3(1);
+}
+
+
+mat2 __postIncr(inout mat2 m)
+{
+   mat2 n = m;
+   m[0] = m[0] + vec2(1.0);
+   m[1] = m[1] + vec2(1.0);
+   return n;
+}
+
+mat3 __postIncr(inout mat3 m)
+{
+   mat3 n = m;
+   m[0] = m[0] + vec3(1.0);
+   m[1] = m[1] + vec3(1.0);
+   m[2] = m[2] + vec3(1.0);
+   return n;
+}
+
+mat4 __postIncr(inout mat4 m)
+{
+   mat4 n = m;
+   m[0] = m[0] + vec4(1.0);
+   m[1] = m[1] + vec4(1.0);
+   m[2] = m[2] + vec4(1.0);
+   m[3] = m[3] + vec4(1.0);
+   return n;
+}
+
+
+
+//// inequality operators
+
+
+// XXX are the inequality operators for floats/ints really needed????
+bool __operator < (const float a, const float b)
+{
+   __asm vec4_sgt __retVal.x, b, a;
+}
+
+
+bool __operator < (const int a, const int b) {
+    return float (a) < float (b);
+}
+
+bool __operator > (const float a, const float b) {
+    bool c;
+    __asm float_less c, b, a;
+    return c;
+}
+
+bool __operator > (const int a, const int b) {
+    return float (a) > float (b);
+}
+
+bool __operator >= (const float a, const float b) {
+    bool g, e;
+    __asm float_less  g, b, a;
+    __asm float_equal e, a, b;
+    return g || e;
+}
+
+bool __operator >= (const int a, const int b) {
+    return float (a) >= float (b);
+}
+
+bool __operator <= (const float a, const float b) {
+    bool g, e;
+    __asm float_less  g, a, b;
+    __asm float_equal e, a, b;
+    return g || e;
+}
+
+bool __operator <= (const int a, const int b) {
+    return float (a) <= float (b);
+}
+
+
+
+//
+// MESA-specific extension functions.
+//
+
+void printMESA (const float f) {
+    __asm float_print f;
+}
+
+void printMESA (const int i) {
+    __asm int_print i;
+}
+
+void printMESA (const bool b) {
+    __asm bool_print b;
+}
+
+void printMESA (const vec2 v) {
+    printMESA (v.x);
+    printMESA (v.y);
+}
+
+void printMESA (const vec3 v) {
+    printMESA (v.x);
+    printMESA (v.y);
+    printMESA (v.z);
+}
+
+void printMESA (const vec4 v) {
+    printMESA (v.x);
+    printMESA (v.y);
+    printMESA (v.z);
+    printMESA (v.w);
+}
+
+void printMESA (const ivec2 v) {
+    printMESA (v.x);
+    printMESA (v.y);
+}
+
+void printMESA (const ivec3 v) {
+    printMESA (v.x);
+    printMESA (v.y);
+    printMESA (v.z);
+}
+
+void printMESA (const ivec4 v) {
+    printMESA (v.x);
+    printMESA (v.y);
+    printMESA (v.z);
+    printMESA (v.w);
+}
+
+void printMESA (const bvec2 v) {
+    printMESA (v.x);
+    printMESA (v.y);
+}
+
+void printMESA (const bvec3 v) {
+    printMESA (v.x);
+    printMESA (v.y);
+    printMESA (v.z);
+}
+
+void printMESA (const bvec4 v) {
+    printMESA (v.x);
+    printMESA (v.y);
+    printMESA (v.z);
+    printMESA (v.w);
+}
+
+void printMESA (const mat2 m) {
+    printMESA (m[0]);
+    printMESA (m[1]);
+}
+
+void printMESA (const mat3 m) {
+    printMESA (m[0]);
+    printMESA (m[1]);
+    printMESA (m[2]);
+}
+
+void printMESA (const mat4 m) {
+    printMESA (m[0]);
+    printMESA (m[1]);
+    printMESA (m[2]);
+    printMESA (m[3]);
+}
+
+void printMESA (const sampler1D e) {
+    __asm int_print e;
+}
+
+void printMESA (const sampler2D e) {
+    __asm int_print e;
+}
+
+void printMESA (const sampler3D e) {
+    __asm int_print e;
+}
+
+void printMESA (const samplerCube e) {
+    __asm int_print e;
+}
+
+void printMESA (const sampler1DShadow e) {
+    __asm int_print e;
+}
+
+void printMESA (const sampler2DShadow e) {
+    __asm int_print e;
+}
+
diff --git a/src/mesa/slang/library/slang_fragment_builtin.gc b/src/mesa/slang/library/slang_fragment_builtin.gc
new file mode 100644
index 0000000000..54a80ea0e0
--- /dev/null
+++ b/src/mesa/slang/library/slang_fragment_builtin.gc
@@ -0,0 +1,299 @@
+/*
+ * Mesa 3-D graphics library
+ * Version:  6.5
+ *
+ * Copyright (C) 2006  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+//
+// From Shader Spec, ver. 1.10, rev. 59
+//
+
+__fixed_input vec4 gl_FragCoord;
+__fixed_input bool gl_FrontFacing;
+__fixed_output vec4 gl_FragColor;
+__fixed_output vec4 gl_FragData[gl_MaxDrawBuffers];
+__fixed_output float gl_FragDepth;
+
+varying vec4 gl_Color;
+varying vec4 gl_SecondaryColor;
+varying vec4 gl_TexCoord[gl_MaxTextureCoords];
+varying float gl_FogFragCoord;
+
+
+
+//// 8.7 Texture Lookup Functions (with bias)
+
+vec4 texture1D(const sampler1D sampler, const float coord, const float bias)
+{
+   vec4 coord4;
+   coord4.x = coord;
+   coord4.w = bias;
+   __asm vec4_tex_1d_bias __retVal, sampler, coord4;
+}
+
+vec4 texture1DProj(const sampler1D sampler, const vec2 coord, const float bias)
+{
+   // do projection here (there's no vec4_texbp1d instruction)
+   vec4 pcoord;
+   pcoord.x = coord.x / coord.y;
+   pcoord.w = bias;
+   __asm vec4_tex_1d_bias __retVal, sampler, pcoord;
+}
+
+vec4 texture1DProj(const sampler1D sampler, const vec4 coord, const float bias)
+{
+   // do projection here (there's no vec4_texbp1d instruction)
+   vec4 pcoord;
+   pcoord.x = coord.x / coord.z;
+   pcoord.w = bias;
+   __asm vec4_tex_1d_bias __retVal, sampler, pcoord;
+}
+
+
+
+
+vec4 texture2D(const sampler2D sampler, const vec2 coord, const float bias)
+{
+   vec4 coord4;
+   coord4.xy = coord.xy;
+   coord4.w = bias;
+   __asm vec4_tex_2d_bias __retVal, sampler, coord4;
+}
+
+vec4 texture2DProj(const sampler2D sampler, const vec3 coord, const float bias)
+{
+   // do projection here (there's no vec4_texbp2d instruction)
+   vec4 pcoord;
+   pcoord.xy = coord.xy / coord.z;
+   pcoord.w = bias;
+   __asm vec4_tex_2d_bias __retVal, sampler, pcoord;
+}
+
+vec4 texture2DProj(const sampler2D sampler, const vec4 coord, const float bias)
+{
+   // do projection here (there's no vec4_texbp2d instruction)
+   vec4 pcoord;
+   pcoord.xy = coord.xy / coord.w;
+   pcoord.w = bias;
+   __asm vec4_tex_2d_bias __retVal, sampler, pcoord;
+}
+
+
+
+
+vec4 texture3D(const sampler3D sampler, const vec3 coord, const float bias)
+{
+   vec4 coord4;
+   coord4.xyz = coord.xyz;
+   coord4.w = bias;
+   __asm vec4_tex_3d_bias __retVal, sampler, coord4;
+}
+
+vec4 texture3DProj(const sampler3D sampler, const vec4 coord, const float bias)
+{
+   // do projection here (there's no vec4_texbp3d instruction)
+   vec4 pcoord;
+   pcoord.xyz = coord.xyz / coord.w;
+   pcoord.w = bias;
+   __asm vec4_tex_3d_bias __retVal, sampler, pcoord;
+}
+
+
+
+
+vec4 textureCube(const samplerCube sampler, const vec3 coord, const float bias)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   coord4.w = bias;
+   __asm vec4_tex_cube __retVal, sampler, coord4;
+}
+
+
+
+vec4 shadow1D(const sampler1DShadow sampler, const vec3 coord, const float bias)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   coord4.w = bias;
+   __asm vec4_tex_1d_bias_shadow __retVal, sampler, coord4;
+}
+
+vec4 shadow1DProj(const sampler1DShadow sampler, const vec4 coord, const float bias)
+{
+   vec4 pcoord;
+   pcoord.x = coord.x / coord.w;
+   pcoord.z = coord.z;
+   pcoord.w = bias;
+   __asm vec4_tex_1d_bias_shadow __retVal, sampler, pcoord;
+}
+
+vec4 shadow2D(const sampler2DShadow sampler, const vec3 coord, const float bias)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   coord4.w = bias;
+   __asm vec4_tex_2d_bias_shadow __retVal, sampler, coord4;
+}
+
+vec4 shadow2DProj(const sampler2DShadow sampler, const vec4 coord, const float bias)
+{
+   vec4 pcoord;
+   pcoord.xy = coord.xy / coord.w;
+   pcoord.z = coord.z;
+   pcoord.w = bias;
+   __asm vec4_tex_2d_bias_shadow __retVal, sampler, pcoord;
+}
+
+
+
+//// GL_EXT_texture_array
+
+vec4 texture1DArray(const sampler1DArray sampler, const vec2 coord)
+{
+   vec4 coord4;
+   coord4.xy = coord;
+   __asm vec4_tex_1d_array __retVal, sampler, coord4;
+}
+
+vec4 texture1DArray(const sampler1DArray sampler, const vec2 coord, const float bias)
+{
+   vec4 coord4;
+   coord4.xy = coord;
+   coord4.w = bias;
+   __asm vec4_tex_1d_array_bias __retVal, sampler, coord4;
+}
+
+vec4 texure2DArray(const sampler2DArray sampler, const vec3 coord)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   __asm vec4_tex_2d_array __retVal, sampler, coord4;
+}
+
+vec4 texture2DArray(const sampler2DArray sampler, const vec3 coord, const float bias)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   coord4.w = bias;
+   __asm vec4_tex_2d_array_bias __retVal, sampler, coord4;
+}
+
+vec4 shadow1DArray(const sampler1DArrayShadow sampler, const vec2 coord)
+{
+   vec4 coord4;
+   coord4.xy = coord;
+   __asm vec4_tex_1d_array_shadow __retVal, sampler, coord4;
+}
+
+vec4 shadow1DArray(const sampler1DArrayShadow sampler, const vec2 coord, const float bias)
+{
+   vec4 coord4;
+   coord4.xy = coord;
+   coord4.w = bias;
+   __asm vec4_tex_1d_array_bias_shadow __retVal, sampler, coord4;
+}
+
+vec4 shadow2DArray(const sampler2DArrayShadow sampler, const vec3 coord)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   __asm vec4_tex_2d_array_shadow __retVal, sampler, coord4;
+}
+
+vec4 shadow2DArray(const sampler2DArrayShadow sampler, const vec3 coord, const float bias)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   coord4.w = bias;
+   __asm vec4_tex_2d_array_bias_shadow __retVal, sampler, coord4;
+}
+
+
+
+//
+// 8.8 Fragment Processing Functions
+//
+
+float dFdx(const float p)
+{
+   __asm vec4_ddx __retVal.x, p.xxxx;
+}
+
+vec2 dFdx(const vec2 p)
+{
+   __asm vec4_ddx __retVal.xy, p.xyyy;
+}
+
+vec3 dFdx(const vec3 p)
+{
+   __asm vec4_ddx __retVal.xyz, p.xyzz;
+}
+
+vec4 dFdx(const vec4 p)
+{
+   __asm vec4_ddx __retVal, p;
+}
+
+float dFdy(const float p)
+{
+   __asm vec4_ddy __retVal.x, p.xxxx;
+}
+
+vec2 dFdy(const vec2 p)
+{
+   __asm vec4_ddy __retVal.xy, p.xyyy;
+}
+
+vec3 dFdy(const vec3 p)
+{
+   __asm vec4_ddy __retVal.xyz, p.xyzz;
+}
+
+vec4 dFdy(const vec4 p)
+{
+   __asm vec4_ddy __retVal, p;
+}
+
+float fwidth (const float p)
+{
+   // XXX hand-write with __asm
+   return abs(dFdx(p)) + abs(dFdy(p));
+}
+
+vec2 fwidth(const vec2 p)
+{
+   // XXX hand-write with __asm
+   return abs(dFdx(p)) + abs(dFdy(p));
+}
+
+vec3 fwidth(const vec3 p)
+{
+   // XXX hand-write with __asm
+   return abs(dFdx(p)) + abs(dFdy(p));
+}
+
+vec4 fwidth(const vec4 p)
+{
+   // XXX hand-write with __asm
+   return abs(dFdx(p)) + abs(dFdy(p));
+}
+
diff --git a/src/mesa/slang/library/slang_vertex_builtin.gc b/src/mesa/slang/library/slang_vertex_builtin.gc
new file mode 100644
index 0000000000..0c67c2ef20
--- /dev/null
+++ b/src/mesa/slang/library/slang_vertex_builtin.gc
@@ -0,0 +1,210 @@
+/*
+ * Mesa 3-D graphics library
+ * Version:  6.5
+ *
+ * Copyright (C) 2006  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+//
+// From Shader Spec, ver. 1.10, rev. 59
+//
+
+__fixed_output vec4 gl_Position;
+__fixed_output float gl_PointSize;
+__fixed_output vec4 gl_ClipVertex;
+
+attribute vec4 gl_Color;
+attribute vec4 gl_SecondaryColor;
+attribute vec3 gl_Normal;
+attribute vec4 gl_Vertex;
+attribute vec4 gl_MultiTexCoord0;
+attribute vec4 gl_MultiTexCoord1;
+attribute vec4 gl_MultiTexCoord2;
+attribute vec4 gl_MultiTexCoord3;
+attribute vec4 gl_MultiTexCoord4;
+attribute vec4 gl_MultiTexCoord5;
+attribute vec4 gl_MultiTexCoord6;
+attribute vec4 gl_MultiTexCoord7;
+attribute float gl_FogCoord;
+
+varying vec4 gl_FrontColor;
+varying vec4 gl_BackColor;
+varying vec4 gl_FrontSecondaryColor;
+varying vec4 gl_BackSecondaryColor;
+varying vec4 gl_TexCoord[gl_MaxTextureCoords];
+varying float gl_FogFragCoord;
+
+//
+// Geometric Functions
+//
+
+vec4 ftransform()
+{
+   __retVal = gl_ModelViewProjectionMatrix[0] * gl_Vertex.xxxx
+            + gl_ModelViewProjectionMatrix[1] * gl_Vertex.yyyy
+            + gl_ModelViewProjectionMatrix[2] * gl_Vertex.zzzz
+            + gl_ModelViewProjectionMatrix[3] * gl_Vertex.wwww;
+}
+
+
+
+//
+// 8.7 Texture Lookup Functions
+// These are pretty much identical to the ones in slang_fragment_builtin.gc
+// When used in a vertex program, the texture sample instructions should not
+// be using a LOD term so it's effectively zero.  Adding 'lod' to that does
+// what we want.
+//
+
+vec4 texture1DLod(const sampler1D sampler, const float coord, const float lod)
+{
+   vec4 coord4;
+   coord4.x = coord;
+   coord4.w = lod;
+   __asm vec4_tex_1d_bias __retVal, sampler, coord4;
+}
+
+vec4 texture1DProjLod(const sampler1D sampler, const vec2 coord, const float lod)
+{
+   vec4 pcoord;
+   pcoord.x = coord.x / coord.y;
+   pcoord.w = lod;
+   __asm vec4_tex_1d_bias __retVal, sampler, pcoord;
+}
+
+vec4 texture1DProjLod(const sampler1D sampler, const vec4 coord, const float lod)
+{
+   vec4 pcoord;
+   pcoord.x = coord.x / coord.z;
+   pcoord.w = lod;
+   __asm vec4_tex_1d_bias __retVal, sampler, pcoord;
+}
+
+
+
+vec4 texture2DLod(const sampler2D sampler, const vec2 coord, const float lod)
+{
+   vec4 coord4;
+   coord4.xy = coord.xy;
+   coord4.w = lod;
+   __asm vec4_tex_2d_bias __retVal, sampler, coord4;
+}
+
+vec4 texture2DProjLod(const sampler2D sampler, const vec3 coord, const float lod)
+{
+   vec4 pcoord;
+   pcoord.xy = coord.xy / coord.z;
+   pcoord.w = lod;
+   __asm vec4_tex_2d_bias __retVal, sampler, pcoord;
+}
+
+vec4 texture2DProjLod(const sampler2D sampler, const vec4 coord, const float lod)
+{
+   vec4 pcoord;
+   pcoord.xy = coord.xy / coord.z;
+   pcoord.w = lod;
+   __asm vec4_tex_2d_bias __retVal, sampler, pcoord;
+}
+
+
+vec4 texture3DLod(const sampler3D sampler, const vec3 coord, const float lod)
+{
+   vec4 coord4;
+   coord4.xyz = coord.xyz;
+   coord4.w = lod;
+   __asm vec4_tex_3d_bias __retVal, sampler, coord4;
+}
+
+vec4 texture3DProjLod(const sampler3D sampler, const vec4 coord, const float lod)
+{
+   // do projection here (there's no vec4_tex_3d_bias_proj instruction)
+   vec4 pcoord;
+   pcoord.xyz = coord.xyz / coord.w;
+   pcoord.w = lod;
+   __asm vec4_tex_3d_bias __retVal, sampler, pcoord;
+}
+
+
+vec4 textureCubeLod(const samplerCube sampler, const vec3 coord, const float lod)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   coord4.w = lod;
+   __asm vec4_tex_cube __retVal, sampler, coord4;
+}
+
+
+vec4 shadow1DLod(const sampler1DShadow sampler, const vec3 coord, const float lod)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   coord4.w = lod;
+   __asm vec4_tex_1d_bias_shadow __retVal, sampler, coord4;
+}
+
+vec4 shadow1DProjLod(const sampler1DShadow sampler, const vec4 coord,
+                     const float lod)
+{
+   vec4 pcoord;
+   pcoord.x = coord.x / coord.w;
+   pcoord.z = coord.z;
+   pcoord.w = lod;
+   __asm vec4_tex_1d_bias_shadow __retVal, sampler, pcoord;
+}
+
+
+vec4 shadow2DLod(const sampler2DShadow sampler, const vec3 coord, const float lod)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   coord4.w = lod;
+   __asm vec4_tex_2d_bias_shadow __retVal, sampler, coord4;
+}
+
+vec4 shadow2DProjLod(const sampler2DShadow sampler, const vec4 coord,
+                     const float lod)
+{
+   vec4 pcoord;
+   pcoord.xy = coord.xy / coord.w;
+   pcoord.z = coord.z;
+   pcoord.w = lod;
+   __asm vec4_tex_2d_bias_shadow __retVal, sampler, pcoord;
+}
+
+
+//// GL_EXT_texture_array
+
+vec4 texture1DArrayLod(const sampler1DArray sampler, const vec2 coord, const float lod)
+{
+   vec4 coord4;
+   coord4.xy = coord;
+   coord4.w = lod;
+   __asm vec4_tex_1d_array_bias __retVal, sampler, coord4;
+}
+
+
+vec4 texture2DArrayLod(const sampler2DArray sampler, const vec3 coord, const float lod)
+{
+   vec4 coord4;
+   coord4.xyz = coord;
+   coord4.w = lod;
+   __asm vec4_tex_2d_array_bias __retVal, sampler, coord4;
+}
+