Fixed a bug in the eigenvalue_decomposition which could occur when

a symmetric matrix was used along with the LAPACK bindings.

dlib/matrix/matrix_eigenvalue.h

@@ -193,6 +193,7 @@ namespace dlib
 
                 lapack::integer temp;
                 lapack::syevr('V','A','L',tempA,0,0,0,0,-1,temp,d,V,isupz);
+                return;
             }
 #endif
             // Tridiagonalize.

dlib/test/matrix_eig.cpp

@@ -193,6 +193,21 @@ namespace
         test_eigenvalue(10*randm<float,3,3>());
     }
 
+    template <int dims>
+    void test_eigenvalue2()
+    {
+        for (int seed = 0; seed < 10; ++seed)
+        {
+            print_spinner();
+            matrix<double> H = gaussian_randm(dims,dims,seed);
+            H = H*trans(H);
+
+            eigenvalue_decomposition<matrix<double> > eig(H);
+            matrix<double> HH = eig.get_pseudo_v()*diagm(eig.get_real_eigenvalues())*trans(eig.get_pseudo_v());
+            DLIB_TEST_MSG(max(abs(H - HH))<1e-12, "dims: " << dims << "  error: " << max(abs(H - HH))); 
+        }
+    }
+
 // ----------------------------------------------------------------------------------------
 
     class matrix_tester : public tester
@@ -215,6 +230,12 @@ namespace
             matrix_test_double();
             dlog << LINFO << "begin testing with float";
             matrix_test_float();
+
+            test_eigenvalue2<10>();
+            test_eigenvalue2<11>();
+            test_eigenvalue2<3>();
+            test_eigenvalue2<2>();
+            test_eigenvalue2<1>();
         }
     } a;