fix: add missing transposition to adjugate for mat2 and mat4

test/gtx/gtx_matrix_operation.cpp

@@ -1,9 +1,86 @@
 #define GLM_ENABLE_EXPERIMENTAL
+#include <glm/glm.hpp>
+#include <glm/gtc/epsilon.hpp>
 #include <glm/gtx/matrix_operation.hpp>
+#include <limits>
 
-int main()
+int test_adjugate()
 {
-	int Error(0);
+	int Error = 0;
+
+	const float epsilon = std::numeric_limits<float>::epsilon();
+
+	// mat2
+	const glm::mat2 m2 {
+		2, 3,
+		1, 5
+	};
+
+	const glm::mat2 eam2 {
+		5, -3,
+		-1, 2
+	};
+
+	const glm::mat2 am2 = glm::adjugate(m2);
+
+	Error += glm::all(glm::bvec2 {
+		glm::all(glm::epsilonEqual(am2[0], eam2[0], epsilon)),
+		glm::all(glm::epsilonEqual(am2[1], eam2[1], epsilon))
+	}) ? 0 : 1;
+
+	// mat3
+	const glm::mat3 m3 {
+		2, 3, 3,
+		1, 5, 4,
+		4, 6, 8
+	};
+
+	const glm::mat3 eam3 {
+		16, -6, -3,
+		8, 4, -5,
+		-14, 0, 7
+	};
+
+	const glm::mat3 am3 = glm::adjugate(m3);
+
+	Error += glm::all(glm::bvec3 {
+		glm::all(glm::epsilonEqual(am3[0], eam3[0], epsilon)),
+		glm::all(glm::epsilonEqual(am3[1], eam3[1], epsilon)),
+		glm::all(glm::epsilonEqual(am3[2], eam3[2], epsilon))
+	}) ? 0 : 1;
+
+	// mat4
+	const glm::mat4 m4 {
+		2, 3, 3, 1,
+		1, 5, 4, 3,
+		4, 6, 8, 5,
+		-2, -3, -3, 4
+	};
+
+	const glm::mat4 eam4 {
+		97, -30, -15, 17,
+		45, 20, -25, 5,
+		-91, 0, 35, -21,
+		14, 0, 0, 14
+	};
+
+	const glm::mat4 am4 = glm::adjugate(m4);
+
+	Error += glm::all(glm::bvec4 {
+		glm::all(glm::epsilonEqual(am4[0], eam4[0], epsilon)),
+		glm::all(glm::epsilonEqual(am4[1], eam4[1], epsilon)),
+		glm::all(glm::epsilonEqual(am4[2], eam4[2], epsilon)),
+		glm::all(glm::epsilonEqual(am4[3], eam4[3], epsilon)),
+	}) ? 0 : 1;
+
+	return Error;
+}
+
+int main()
+{
+	int Error = 0;
+
+	Error += test_adjugate();
 
 	return Error;
 }