Adjust hlsl infinity-constant test

Test/hlsl.inf.vert tests parsing and some constant math on
infinities, including (-1.#INF * 0.0).
By IEEE 754 rules, that result is a NaN, but its sign is not significant.
The test output assumes a negative-NaN is in the generated SPIR-V.

However, the math library on some platforms (like macOS 14, a.k.a.
Sonoma) will produce a positive NaN instead.

This PR adjusts the test so it takes the absolute value of the NaN,
to ensure we the emitted SPIR-V has the NaN with a 0 for it sign bit.

Test/hlsl.inf.vert

@@ -5,7 +5,12 @@ float4 main() : SV_Position
     float f3 = +1.#INF;
     float f4 = f2 * 1.#INF + 1.#INF;
     const float f5 = -1.#INF;
-    const float f6 = f5 * 0.0f;
+    // An infinity times zero is a NaN.
+    // In IEEE 754, the sign of a NaN is significant only for
+    // abs, copy, negate, or copySign.  Use abs(.) here to
+    // set the sign bit to zero. Otherwise, some platforms will
+    // have a 1 sign bit and others will have a 0 sign bit.
+    const float f6 = abs(f5 * 0.0f);
 
     return (float4)(f1 + f2 + f3 + f4 + f5 + f6);
-}
+}