WIP: HLSL: handle clip/cull distance array semantic matching

In HLSL, there are three (TODO: ??) dimensions of clip and cull
distance values:

  * The semantic's value N, ala SV_ClipDistanceN.
  * The array demension, if the value is an array.
  * The vector element, if the value is a vector or array of vectors.

In SPIR-V, clip and cull distance are arrays of scalar floats, always.

This PR currently ignores the semantic N axis, and handles the other
two axes by sequentially copying each vector element of each array member
into sequential floats in the output array.

Fixes: #946

Test/hlsl.clipdistance-5.vert

@@ -0,0 +1,21 @@
+struct VS_INPUT         {
+    float4 Position             : POSITION;
+};
+
+struct VS_OUTPUT        {
+    float4 Position             : SV_Position;
+    float2 ClipRect[2]          : SV_ClipDistance0;  // array of float2 in split struct
+};
+
+VS_OUTPUT main(const VS_INPUT v)
+{
+    VS_OUTPUT           Output;
+    Output.Position     = 0;
+
+    Output.ClipRect[0].x = 1;
+    Output.ClipRect[0].y = 2;
+    Output.ClipRect[1].x = 3;
+    Output.ClipRect[1].y = 4;
+
+    return Output;
+}