kram - simd - add equals calls.

All comparison ops == and != ops return short4/int4/long4 just like <=, <, >, >=

libkram/vectormath/double234.h

@@ -439,6 +439,18 @@ SIMD_CALL double3 cross(double3 x, double3 y) {
     return x.yzx * y.zxy - x.zxy * y.yzx;
 }
 
+// equal
+// == and != return a int234 vector, so need these to match other vecs
+SIMD_CALL bool equal(double2 x, double2 y) {
+    return all(x == y);
+}
+SIMD_CALL bool equal(double3 x, double3 y) {
+    return all(x == y);
+}
+SIMD_CALL bool equal(double4 x, double4 y) {
+    return all(x == y);
+}
+
 // equal_abs
 SIMD_CALL bool equal_abs(double2 x, double2 y, double tol) {
     return all((abs(x - y) <= tol));
@@ -518,6 +530,20 @@ SIMD_CALL double4 fix_nan(double4 x, double4 replace) {
 }
 
 
+// fast conversions where possible
+// need non-const too
+SIMD_CALL const double3& as_double3(const double4& m) {
+    return reinterpret_cast<const double3&>(m);
+}
+SIMD_CALL const double3* as_double3(const double4* m) {
+    return reinterpret_cast<const double3*>(m);
+}
+
+// this one is dangerous, since w is undefined
+//SIMD_CALL const double4& as_double4(const double3& m) {
+//    return reinterpret_cast<const double4&>(m);
+//}
+
 
 //-------------------
 // Functions
@@ -757,6 +783,9 @@ macroMatrixOps(double4x4);
 SIMD_CALL const double3x3& as_double3x3(const double4x4& m) {
     return reinterpret_cast<const double3x3&>(m);
 }
+SIMD_CALL const double3x3* as_double3x3(const double4x4* m) {
+    return reinterpret_cast<const double3x3*>(m);
+}
 
 } // SIMD_NAMESPACE
 

libkram/vectormath/float234.h

@@ -325,6 +325,18 @@ SIMD_CALL float3 cross(float3 x, float3 y) {
     return x.yzx * y.zxy - x.zxy * y.yzx;
 }
 
+// equal
+// == and != return a int234 vector, so need these to match other vecs
+SIMD_CALL bool equal(float2 x, float2 y) {
+    return all(x == y);
+}
+SIMD_CALL bool equal(float3 x, float3 y) {
+    return all(x == y);
+}
+SIMD_CALL bool equal(float4 x, float4 y) {
+    return all(x == y);
+}
+
 // equal_abs
 SIMD_CALL bool equal_abs(float2 x, float2 y, float tol) {
     return all((abs(x - y) <= tol));
@@ -505,9 +517,18 @@ SIMD_CALL float4 float4m(float3 v, float w = 1.0f) {
 }
 
 // fast conversions where possible
+// need non-const too
 SIMD_CALL const float3& as_float3(const float4& m) {
     return reinterpret_cast<const float3&>(m);
 }
+SIMD_CALL const float3* as_float3(const float4* m) {
+    return reinterpret_cast<const float3*>(m);
+}
+
+// this one is dangerous, since w is undefined
+//SIMD_CALL const float4& as_float4(const float3& m) {
+//    return reinterpret_cast<const float4&>(m);
+//}
 
 // power series
 macroVectorRepeatFnDecl(float, log)
@@ -841,6 +862,20 @@ SIMD_CALL float4x4 float4x4m(float3 axis, float radians) {
     return float4x4m(quatf(axis, radians));
 }
 
+// These sizes are positive and do not include inversion
+SIMD_CALL float decompose_size(const float4x4& m) {
+    return length(m[0]);
+}
+SIMD_CALL float3 decompose_scale(const float4x4& m) {
+    return float3m(length(m[0]),
+                   length(m[1]),
+                   length(m[2]));
+}
+SIMD_CALL float decompose_scale_max(const float4x4& m) {
+    return reduce_max(decomposeScale(m));
+}
+
+
 float3x3 float3x3m(quatf qq);
 
 float4x4 float4x4_tr(float3 t, quatf r);