Fix output for matrix multiple in GLSL code (#228)

When Slang sees a matrix multiplication `M * v` in GLSL code it should (obviously) output GLSL code that also does `M * v`, but there was a bug introduced where the type-checker manages to resolve the operation and recognize it as a matrix-vector multiply, and then the code-generation logic says "oh, I'm generating output for GLSL, and that is reversed from HLSL/Slang, so I'd better reverse these operands!" and outputs `v * M`... which isn't what we want.

I've fixed the problem in an expedient way, by having the front-end resolve the operation to what it believes is an intrinsic multiply operation, rather than a matrix-vector operation. If we ever support cross compilation *from* GLSL (unlikely), we've need to fix this up so that we have both real matrix-vector multiplies and "reversed" multiplies where the operands folow the GLSL convention).

I've added two tests here to confirm the fix. The one under `tests/bugs` catches the actual issue described above, and confirms the fix. The other one under `tests/cross-compile` is just to make sure that we *do* properly reverse the operands to a matrix-vector product when converting from Slang to GLSL.

3 files changed, 67 insertions, 0 deletions

diff --git a/tests/bugs/matrix-mult.glsl b/tests/bugs/matrix-mult.glsl
new file mode 100644
index 000000000..b427dee14
--- /dev/null
+++ b/tests/bugs/matrix-mult.glsl
@@ -0,0 +1,28 @@
+//TEST:COMPARE_GLSL:-profile glsl_fragment_450 -no-checking
+// matrix-mult.glsl
+#version 450
+
+// Confirm that we don't exchange the operands
+// to a matrix-vector multiply when we recognize
+// one in "raw" GLSL code.
+
+#ifdef __SLANG__
+__import empty;
+#endif
+
+layout(binding = 0)
+uniform C
+{
+	mat4x4 m;
+};
+
+layout(location = 0)
+in vec4 v;
+
+layout(location = 0)
+out vec4 r;
+
+void main()
+{
+	r = m * v;
+}
diff --git a/tests/cross-compile/matrix-mult.slang b/tests/cross-compile/matrix-mult.slang
new file mode 100644
index 000000000..180a806e2
--- /dev/null
+++ b/tests/cross-compile/matrix-mult.slang
@@ -0,0 +1,14 @@
+//TEST:CROSS_COMPILE: -profile ps_5_0 -entry main -target spirv-assembly
+
+// Confirm that order of arguments to matrix-vector
+// multiplication gets reversed when generating GLSL.
+
+cbuffer C
+{
+	float4x3 m;
+};
+
+float4 main(float3 v) : SV_Target
+{
+	return mul(m, v);
+}
diff --git a/tests/cross-compile/matrix-mult.slang.glsl b/tests/cross-compile/matrix-mult.slang.glsl
new file mode 100644
index 000000000..0349d854e
--- /dev/null
+++ b/tests/cross-compile/matrix-mult.slang.glsl
@@ -0,0 +1,25 @@
+//TEST_IGNORE_FILE:
+
+layout(binding = 0)
+uniform C
+{
+	mat4x3 m;
+};
+
+vec4 main_(vec3 v)
+{
+	return v * m;
+}
+
+layout(location = 0)
+in vec3 SLANG_in_v;
+
+layout(location = 0)
+out vec4 SLANG_out_main_result;
+
+void main()
+{
+	vec3 v = SLANG_in_v;
+	vec4 main_result = main_(v);
+	SLANG_out_main_result = main_result;
+}