Fix quat packing XYZW usage (#1204)

glm/detail/type_quat.hpp

@@ -89,7 +89,7 @@ namespace glm
 
 		GLM_FUNC_DECL GLM_CONSTEXPR qua(T s, vec<3, T, Q> const& v);
 
-#		ifdef GLM_FORCE_QUAT_DATA_XYZW
+#		ifdef GLM_FORCE_QUAT_CTOR_XYZW
 		GLM_FUNC_DECL GLM_CONSTEXPR qua(T x, T y, T z, T w);
 #		else
 		GLM_FUNC_DECL GLM_CONSTEXPR qua(T w, T x, T y, T z);

glm/detail/type_quat.inl

@@ -141,7 +141,11 @@ namespace detail
 	{}
 
 	template <typename T, qualifier Q>
+#		ifdef GLM_FORCE_QUAT_CTOR_XYZW
+	GLM_FUNC_QUALIFIER GLM_CONSTEXPR qua<T, Q>::qua(T _x, T _y, T _z, T _w)
+#		else
 	GLM_FUNC_QUALIFIER GLM_CONSTEXPR qua<T, Q>::qua(T _w, T _x, T _y, T _z)
+#		endif
 #		ifdef GLM_FORCE_QUAT_DATA_WXYZ
 			: w(_w), x(_x), y(_y), z(_z)
 #		else

glm/detail/type_quat_simd.inl

@@ -161,39 +161,39 @@ namespace detail
 	{
 		static vec<4, float, Q> call(qua<float, Q> const& q, vec<4, float, Q> const& v)
 		{
-#			ifdef GLM_FORCE_QUAT_DATA_XYZW
-				__m128 const q_wwww = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(3, 3, 3, 3));
-				__m128 const q_swp0 = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(3, 0, 2, 1));
-				__m128 const q_swp1 = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(3, 1, 0, 2));
+#			ifdef GLM_FORCE_QUAT_DATA_WXYZ
+				__m128 const q_wwww = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(0, 0, 0, 0));
+				__m128 const q_swp0 = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(0, 1, 3, 2));
+				__m128 const q_swp1 = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(0, 2, 1, 3));
 				__m128 const v_swp0 = _mm_shuffle_ps(v.data, v.data, _MM_SHUFFLE(3, 0, 2, 1));
 				__m128 const v_swp1 = _mm_shuffle_ps(v.data, v.data, _MM_SHUFFLE(3, 1, 0, 2));
 
-				__m128 uv      = _mm_sub_ps(_mm_mul_ps(q_swp0, v_swp1), _mm_mul_ps(q_swp1, v_swp0));
+				__m128 uv = _mm_sub_ps(_mm_mul_ps(q_swp0, v_swp1), _mm_mul_ps(q_swp1, v_swp0));
 				__m128 uv_swp0 = _mm_shuffle_ps(uv, uv, _MM_SHUFFLE(3, 0, 2, 1));
 				__m128 uv_swp1 = _mm_shuffle_ps(uv, uv, _MM_SHUFFLE(3, 1, 0, 2));
-				__m128 uuv     = _mm_sub_ps(_mm_mul_ps(q_swp0, uv_swp1), _mm_mul_ps(q_swp1, uv_swp0));
+				__m128 uuv = _mm_sub_ps(_mm_mul_ps(q_swp0, uv_swp1), _mm_mul_ps(q_swp1, uv_swp0));
 
 				__m128 const two = _mm_set1_ps(2.0f);
-				uv  = _mm_mul_ps(uv, _mm_mul_ps(q_wwww, two));
+				uv = _mm_mul_ps(uv, _mm_mul_ps(q_wwww, two));
 				uuv = _mm_mul_ps(uuv, two);
 
 				vec<4, float, Q> Result;
 				Result.data = _mm_add_ps(v.data, _mm_add_ps(uv, uuv));
 				return Result;
 #			else
-				__m128 const q_wwww = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(0, 0, 0, 0));
-				__m128 const q_swp0 = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(0, 1, 3, 2));
-				__m128 const q_swp1 = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(0, 2, 1, 3));
+				__m128 const q_wwww = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(3, 3, 3, 3));
+				__m128 const q_swp0 = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(3, 0, 2, 1));
+				__m128 const q_swp1 = _mm_shuffle_ps(q.data, q.data, _MM_SHUFFLE(3, 1, 0, 2));
 				__m128 const v_swp0 = _mm_shuffle_ps(v.data, v.data, _MM_SHUFFLE(3, 0, 2, 1));
 				__m128 const v_swp1 = _mm_shuffle_ps(v.data, v.data, _MM_SHUFFLE(3, 1, 0, 2));
 
-				__m128 uv = _mm_sub_ps(_mm_mul_ps(q_swp0, v_swp1), _mm_mul_ps(q_swp1, v_swp0));
+				__m128 uv      = _mm_sub_ps(_mm_mul_ps(q_swp0, v_swp1), _mm_mul_ps(q_swp1, v_swp0));
 				__m128 uv_swp0 = _mm_shuffle_ps(uv, uv, _MM_SHUFFLE(3, 0, 2, 1));
 				__m128 uv_swp1 = _mm_shuffle_ps(uv, uv, _MM_SHUFFLE(3, 1, 0, 2));
-				__m128 uuv = _mm_sub_ps(_mm_mul_ps(q_swp0, uv_swp1), _mm_mul_ps(q_swp1, uv_swp0));
+				__m128 uuv     = _mm_sub_ps(_mm_mul_ps(q_swp0, uv_swp1), _mm_mul_ps(q_swp1, uv_swp0));
 
 				__m128 const two = _mm_set1_ps(2.0f);
-				uv = _mm_mul_ps(uv, _mm_mul_ps(q_wwww, two));
+				uv  = _mm_mul_ps(uv, _mm_mul_ps(q_wwww, two));
 				uuv = _mm_mul_ps(uuv, two);
 
 				vec<4, float, Q> Result;

glm/gtx/matrix_decompose.inl

@@ -173,10 +173,10 @@ namespace detail
 			j = Next[i];
 			k = Next[j];
 
-#           ifdef GLM_FORCE_QUAT_DATA_XYZW
-                int off = 0;
-#           else
+#           ifdef GLM_FORCE_QUAT_DATA_WXYZ
                 int off = 1;
+#           else
+                int off = 0;
 #           endif
 
 			root = sqrt(Row[i][i] - Row[j][j] - Row[k][k] + static_cast<T>(1.0));