First pass at support for cross-compilation

This is a large change that contains many pieces:

- Update the `cross-compile0` test to actually make use of cross compilation.

  Now the `cross-compile0.hlsl` file contains both HLSL and GLSL source code, and then imports code from `cross-compile0.slang`, which provides a "library" (one function) that can be shared between both the HLSL and GLSL version of things.

- Fixed a bug in the support for backslash-escaped newlines.

- Added a new `__import` declaration type (replaces the `using` directive that was still around in a vestigial form)

  An `__import` causes the compiler to look for a Slang source file (currently using the ordinary `#include` lookup logic), and then parse/check the found file as an additional module ("translation unit"), before making its declarations visible in the current scope.

- Refactored the main compilation flow to be simpler. There were the `ShaderCompiler` and `ShaderCompilerImpl` classes that weren't relaly doing anything, but added complexity to the whole workflow.

- The `render-test` application has been heavily modified to better support testing cross-compilation workflows. At the most basic level we are starting to distinguish pass-through vs. rewriter workflows, and are passing various `#define`s down to the compiler(s) to let the source code be customized as needed for each case.

  Several annoying corner cases are caused here by having to support the GLSL compilation model, which really wants each entry point in its own specific translation unit, whereas we really want to keep things nicely contained in single files.

- Added support for `__intrinsic` operations to have target-specific behavior.

  This allows a function to be given a different name for some specific target (so a call gets emitted as a call to that other operation).
More generally, the library writer can put together an arbitrary format string that will be used in place of expressions that call the given function, e.g.:

      __intrinsic(hlsl, "$1 - $0") __intrinsic int foo(int a, int b);

  Given this declaration, a call like `foo(x,y)` will code generate as `x - y` for HLSL, and as `foo(x,y)` for all other targets.

Annoying things still to be dealt with:

- The way that I'm filtering the user-provided options when passing things down to the compilation of dynamically loaded modules is a bit ad hoc. It would be good to have a systematic notion of which options will be inherited and which won't. There is also more code duplication than I'd like, so we risk having the compiler behave differently when compiling a file at the top level, vs. because of `__import`.

- Adding target-specific behavior to intrinsics is all well and good, but the current approach means we can only add this to the original declaration, which limits the ability to easily extend the set of targets.

  A better approach long-term would be to add a more robust notion of target-based overload resolution (which would happen after semantic checking). Then one mechanism would be used to find the right target-specific overload to use for an operation, and then each (target-specific) definition could use a simpler attribute to intercept code-generation behavior.

  Note that we might eventually need a similar notion to deal with stage- or profile-specific functions and the overloading behavior around them, so using this for intrinsics doesn't seem like a bad idea.

1 files changed, 37 insertions, 42 deletions

diff --git a/source/slang/slang-stdlib.cpp b/source/slang/slang-stdlib.cpp
index f74fcd603..494a32e4d 100644
--- a/source/slang/slang-stdlib.cpp
+++ b/source/slang/slang-stdlib.cpp
@@ -8,16 +8,16 @@
 #define LINE_STRING STRINGIZE(__LINE__)
 
 enum { kLibIncludeStringLine = __LINE__+1 };
-const char * LibIncludeStringChunks[] = { R"(
+const char * LibIncludeStringChunks[] = { R"=(
 
 typedef uint UINT;
 
-__generic<T> __intrinsic(Assign) T operator=(out T left, T right);
+__generic<T> __intrinsic_op(Assign) T operator=(out T left, T right);
 
-__generic<T,U> __intrinsic(Sequence) U operator,(T left, U right);
+__generic<T,U> __intrinsic_op(Sequence) U operator,(T left, U right);
 
-__generic<T> __intrinsic(Select) T operator?:(bool condition, T ifTrue, T ifFalse);
-__generic<T, let N : int> __intrinsic(Select) vector<T,N> operator?:(vector<bool,N> condition, vector<T,N> ifTrue, vector<T,N> ifFalse);
+__generic<T> __intrinsic_op(Select) T operator?:(bool condition, T ifTrue, T ifFalse);
+__generic<T, let N : int> __intrinsic_op(Select) vector<T,N> operator?:(vector<bool,N> condition, vector<T,N> ifTrue, vector<T,N> ifFalse);
 
 __generic<T> __magic_type(HLSLAppendStructuredBufferType) struct AppendStructuredBuffer
 {
@@ -234,7 +234,7 @@ __generic<T> __magic_type(HLSLPointStreamType) struct PointStream {};
 __generic<T> __magic_type(HLSLLineStreamType) struct LineStream {};
 __generic<T> __magic_type(HLSLLineStreamType) struct TriangleStream {};
 
-)", R"(
+)=", R"=(
 
 // Note(tfoley): Trying to systematically add all the HLSL builtins
 
@@ -503,7 +503,7 @@ __generic<T : __BuiltinFloatingPointType> __intrinsic T fwidth(T x);
 __generic<T : __BuiltinFloatingPointType, let N : int> __intrinsic vector<T,N> fwidth(vector<T,N> x);
 __generic<T : __BuiltinFloatingPointType, let N : int, let M : int> __intrinsic matrix<T,N,M> fwidth(matrix<T,N,M> x);
 
-)", R"(
+)=", R"=(
 
 // Get number of samples in render target
 __intrinsic uint GetRenderTargetSampleCount();
@@ -615,27 +615,27 @@ __intrinsic uint4 msad4(uint reference, uint2 source, uint4 accum);
 // General inner products
 
 // scalar-scalar
-__generic<T : __BuiltinArithmeticType> __intrinsic(Mul_Scalar_Scalar) T mul(T x, T y);
+__generic<T : __BuiltinArithmeticType> __intrinsic_op(Mul_Scalar_Scalar) T mul(T x, T y);
 
 // scalar-vector and vector-scalar
-__generic<T : __BuiltinArithmeticType, let N : int> __intrinsic(Mul_Vector_Scalar) vector<T,N> mul(vector<T,N> x, T y);
-__generic<T : __BuiltinArithmeticType, let N : int> __intrinsic(Mul_Scalar_Vector) vector<T,N> mul(T x, vector<T,N> y);
+__generic<T : __BuiltinArithmeticType, let N : int> __intrinsic_op(Mul_Vector_Scalar) vector<T,N> mul(vector<T,N> x, T y);
+__generic<T : __BuiltinArithmeticType, let N : int> __intrinsic_op(Mul_Scalar_Vector) vector<T,N> mul(T x, vector<T,N> y);
 
 // scalar-matrix and matrix-scalar
-__generic<T : __BuiltinArithmeticType, let N : int, let M :int> __intrinsic(Mul_Matrix_Scalar) matrix<T,N,M> mul(matrix<T,N,M> x, T y);
-__generic<T : __BuiltinArithmeticType, let N : int, let M :int> __intrinsic(Mul_Scalar_Matrix) matrix<T,N,M> mul(T x, matrix<T,N,M> y);
+__generic<T : __BuiltinArithmeticType, let N : int, let M :int> __intrinsic_op(Mul_Matrix_Scalar) matrix<T,N,M> mul(matrix<T,N,M> x, T y);
+__generic<T : __BuiltinArithmeticType, let N : int, let M :int> __intrinsic_op(Mul_Scalar_Matrix) matrix<T,N,M> mul(T x, matrix<T,N,M> y);
 
 // vector-vector (dot product)
-__generic<T : __BuiltinArithmeticType, let N : int> __intrinsic(InnerProduct_Vector_Vector) T mul(vector<T,N> x, vector<T,N> y);
+__generic<T : __BuiltinArithmeticType, let N : int> __intrinsic_op(InnerProduct_Vector_Vector) T mul(vector<T,N> x, vector<T,N> y);
 
 // vector-matrix
-__generic<T : __BuiltinArithmeticType, let N : int, let M : int> __intrinsic(InnerProduct_Vector_Matrix) vector<T,M> mul(vector<T,N> x, matrix<T,N,M> y);
+__generic<T : __BuiltinArithmeticType, let N : int, let M : int> __intrinsic_op(InnerProduct_Vector_Matrix) vector<T,M> mul(vector<T,N> x, matrix<T,N,M> y);
 
 // matrix-vector
-__generic<T : __BuiltinArithmeticType, let N : int, let M : int> __intrinsic(InnerProduct_Matrix_Vector) vector<T,N> mul(matrix<T,N,M> x, vector<T,M> y);
+__generic<T : __BuiltinArithmeticType, let N : int, let M : int> __intrinsic_op(InnerProduct_Matrix_Vector) vector<T,N> mul(matrix<T,N,M> x, vector<T,M> y);
 
 // matrix-matrix
-__generic<T : __BuiltinArithmeticType, let R : int, let N : int, let C : int> __intrinsic(InnerProduct_Matrix_Matrix) matrix<T,R,C> mul(matrix<T,R,N> x, matrix<T,N,C> y);
+__generic<T : __BuiltinArithmeticType, let R : int, let N : int, let C : int> __intrinsic_op(InnerProduct_Matrix_Matrix) matrix<T,R,C> mul(matrix<T,R,N> x, matrix<T,N,C> y);
 
 // noise (deprecated)
 __intrinsic float noise(float x);
@@ -759,9 +759,17 @@ __generic<T : __BuiltinFloatingPointType, let N : int> __intrinsic vector<T,N> r
 __generic<T : __BuiltinFloatingPointType, let N : int, let M : int> __intrinsic matrix<T,N,M> rsqrt(matrix<T,N,M> x);
 
 // Clamp value to [0,1] range
-__generic<T : __BuiltinFloatingPointType> __intrinsic T saturate(T x);
-__generic<T : __BuiltinFloatingPointType, let N : int> __intrinsic vector<T,N> saturate(vector<T,N> x);
-__generic<T : __BuiltinFloatingPointType, let N : int, let M : int> __intrinsic matrix<T,N,M> saturate(matrix<T,N,M> x);
+__generic<T : __BuiltinFloatingPointType>
+__intrinsic(glsl, "clamp($0, 0, 1)") __intrinsic
+T saturate(T x);
+
+__generic<T : __BuiltinFloatingPointType, let N : int>
+__intrinsic(glsl, "clamp($0, 0, 1)") __intrinsic
+vector<T,N> saturate(vector<T,N> x);
+
+__generic<T : __BuiltinFloatingPointType, let N : int, let M : int>
+__intrinsic(glsl, "clamp($0, 0, 1)") __intrinsic
+matrix<T,N,M> saturate(matrix<T,N,M> x);
 
 
 // Extract sign of value
@@ -769,7 +777,7 @@ __generic<T : __BuiltinSignedArithmeticType> __intrinsic int sign(T x);
 __generic<T : __BuiltinSignedArithmeticType, let N : int> __intrinsic vector<int,N> sign(vector<T,N> x);
 __generic<T : __BuiltinSignedArithmeticType, let N : int, let M : int> __intrinsic matrix<int,N,M> sign(matrix<T,N,M> x);
 
-)", R"(
+)=", R"=(
 
 
 // Sine
@@ -853,7 +861,7 @@ __generic<T : __BuiltinFloatingPointType, let N : int> __intrinsic vector<T,N> t
 __generic<T : __BuiltinFloatingPointType, let N : int, let M : int> __intrinsic matrix<T,N,M> trunc(matrix<T,N,M> x);
 
 
-)", R"(
+)=", R"=(
 
 // Shader model 6.0 stuff
 
@@ -930,13 +938,13 @@ __generic<T : __BuiltinType, let N : int> __intrinsic vector<T,N> WaveReadLaneAt
 __generic<T : __BuiltinType, let N : int, let M : int> __intrinsic matrix<T,N,M> WaveReadLaneAt(matrix<T,N,M> expr, int laneIndex);
 
 
-)", R"(
+)=", R"=(
 
 // `typedef`s to help with the fact that HLSL has been sorta-kinda case insensitive at various points
 typedef Texture2D texture2D;
 
 #line default
-)" };
+)=" };
 
 
 using namespace CoreLib::Basic;
@@ -1522,19 +1530,6 @@ namespace Slang
             sb << "__intrinsic mat4 operator * (mat4, mat4);\n";
 #endif
 
-#if 0
-            sb << "__intrinsic(And) bool operator && (bool, bool);\n";
-            sb << "__intrinsic(Or) bool operator || (bool, bool);\n";
-
-            for (auto type : intTypes)
-            {
-                sb << "__intrinsic(And) bool operator && (bool, " << type << ");\n";
-                sb << "__intrinsic(Or) bool operator || (bool, " << type << ");\n";
-                sb << "__intrinsic(And) bool operator && (" << type << ", bool);\n";
-                sb << "__intrinsic(Or) bool operator || (" << type << ", bool);\n";
-            }
-#endif
-
             for (auto op : unaryOps)
             {
                 for (auto type : kBaseTypes)
@@ -1547,17 +1542,17 @@ namespace Slang
 
                     // scalar version
                     sb << fixity;
-                    sb << "__intrinsic(" << int(op.opCode) << ") " << type.name << " operator" << op.opName << "(" << qual << type.name << " value);\n";
+                    sb << "__intrinsic_op(" << int(op.opCode) << ") " << type.name << " operator" << op.opName << "(" << qual << type.name << " value);\n";
 
                     // vector version
                     sb << "__generic<let N : int> ";
                     sb << fixity;
-                    sb << "__intrinsic(" << int(op.opCode) << ") vector<" << type.name << ",N> operator" << op.opName << "(" << qual << "vector<" << type.name << ",N> value);\n";
+                    sb << "__intrinsic_op(" << int(op.opCode) << ") vector<" << type.name << ",N> operator" << op.opName << "(" << qual << "vector<" << type.name << ",N> value);\n";
 
                     // matrix version
                     sb << "__generic<let N : int, let M : int> ";
                     sb << fixity;
-                    sb << "__intrinsic(" << int(op.opCode) << ") matrix<" << type.name << ",N,M> operator" << op.opName << "(" << qual << "matrix<" << type.name << ",N,M> value);\n";
+                    sb << "__intrinsic_op(" << int(op.opCode) << ") matrix<" << type.name << ",N,M> operator" << op.opName << "(" << qual << "matrix<" << type.name << ",N,M> value);\n";
                 }
             }
 
@@ -1580,15 +1575,15 @@ namespace Slang
                     // TODO: handle `SHIFT`
 
                     // scalar version
-                    sb << "__intrinsic(" << int(op.opCode) << ") " << resultType << " operator" << op.opName << "(" << leftQual << leftType << " left, " << rightType << " right);\n";
+                    sb << "__intrinsic_op(" << int(op.opCode) << ") " << resultType << " operator" << op.opName << "(" << leftQual << leftType << " left, " << rightType << " right);\n";
 
                     // vector version
                     sb << "__generic<let N : int> ";
-                    sb << "__intrinsic(" << int(op.opCode) << ") vector<" << resultType << ",N> operator" << op.opName << "(" << leftQual << "vector<" << leftType << ",N> left, vector<" << rightType << ",N> right);\n";
+                    sb << "__intrinsic_op(" << int(op.opCode) << ") vector<" << resultType << ",N> operator" << op.opName << "(" << leftQual << "vector<" << leftType << ",N> left, vector<" << rightType << ",N> right);\n";
 
                     // matrix version
                     sb << "__generic<let N : int, let M : int> ";
-                    sb << "__intrinsic(" << int(op.opCode) << ") matrix<" << resultType << ",N,M> operator" << op.opName << "(" << leftQual << "matrix<" << leftType << ",N,M> left, matrix<" << rightType << ",N,M> right);\n";
+                    sb << "__intrinsic_op(" << int(op.opCode) << ") matrix<" << resultType << ",N,M> operator" << op.opName << "(" << leftQual << "matrix<" << leftType << ",N,M> left, matrix<" << rightType << ",N,M> right);\n";
                 }
             }