hotfix for segfault on mkl (#12637)

* probably fixing the segfault

* Trigger CircleCI [doc build]

* move isfinite out one level[doc build]

* slightly faster finiteness check

sklearn/covariance/graph_lasso_.py

@@ -243,6 +243,9 @@ def graphical_lasso(emp_cov, alpha, cov_init=None, mode='cd', tol=1e-4,
                 coefs = np.dot(sub_covariance, coefs)
                 covariance_[idx, indices != idx] = coefs
                 covariance_[indices != idx, idx] = coefs
+            if not np.isfinite(precision_.sum()):
+                raise FloatingPointError('The system is too ill-conditioned '
+                                         'for this solver')
             d_gap = _dual_gap(emp_cov, precision_, alpha)
             cost = _objective(emp_cov, precision_, alpha)
             if verbose: