[Octave-bug-tracker] [bug #50105] __ilu__ and __ichol__ should use liboctave functions instead of feval

[Rik]

Update of bug #50105 (project octave):

                  Status:                    None => Confirmed              
        Operating System:                     BSD => Any                    

    _______________________________________________________

Follow-up Comment #1:

Amen to that.  For reference, here are the function invocations in question.


corefcn/__ichol__.cc:201:      sm = octave::feval ("tril",
arg_list)(0).sparse_matrix_value ();
corefcn/__ichol__.cc:211:      sm = octave::feval ("tril",
arg_list)(0).sparse_complex_matrix_value ();
corefcn/__ichol__.cc:444:        octave::feval ("tril",
arg_list)(0).sparse_matrix_value ();
corefcn/__ichol__.cc:448:      cols_norm = octave::feval ("norm",
arg_list)(0).vector_value ();
corefcn/__ichol__.cc:462:        octave::feval ("tril",
arg_list)(0).sparse_complex_matrix_value ();
corefcn/__ichol__.cc:466:      cols_norm = octave::feval ("norm",
arg_list)(0).complex_vector_value ();
corefcn/__ilu__.cc:163:      retval(1) = octave::feval ("triu",
arg_list)(0).sparse_matrix_value ();
corefcn/__ilu__.cc:165:        octave::feval ("speye", ovl (sm.cols
()))(0).sparse_matrix_value ();
corefcn/__ilu__.cc:168:        octave::feval ("tril",
arg_list)(0).sparse_matrix_value ();
corefcn/__ilu__.cc:176:      retval(1) = octave::feval ("triu",
arg_list)(0).sparse_complex_matrix_value ();
corefcn/__ilu__.cc:178:        octave::feval ("speye", ovl (sm.cols
()))(0).sparse_complex_matrix_value ();
corefcn/__ilu__.cc:181:        octave::feval ("tril",
arg_list)(0).sparse_complex_matrix_value ();
corefcn/__ilu__.cc:497:      SparseMatrix sm_u = octave::feval ("triu",
arg_list)(0).sparse_matrix_value ();
corefcn/__ilu__.cc:499:      SparseMatrix sm_l = octave::feval ("tril",
arg_list)(0).sparse_matrix_value ();
corefcn/__ilu__.cc:501:      rows_norm = octave::feval ("norm",
arg_list)(0).vector_value ();
corefcn/__ilu__.cc:503:      cols_norm = octave::feval ("norm",
arg_list)(0).vector_value ();
corefcn/__ilu__.cc:511:      SparseMatrix eye = octave::feval ("speye",
arg_list)(0).sparse_matrix_value ();
corefcn/__ilu__.cc:519:        octave::feval ("triu",
arg_list)(0).sparse_complex_matrix_value ();
corefcn/__ilu__.cc:522:        octave::feval ("tril",
arg_list)(0).sparse_complex_matrix_value ();
corefcn/__ilu__.cc:524:      rows_norm = octave::feval ("norm",
arg_list)(0).complex_vector_value ();
corefcn/__ilu__.cc:526:      cols_norm = octave::feval ("norm",
arg_list)(0).complex_vector_value ();
corefcn/__ilu__.cc:535:        octave::feval ("speye",
arg_list)(0).sparse_complex_matrix_value ();
corefcn/__ilu__.cc:969:      nnz_u = (octave::feval ("triu",
arg_list)(0).sparse_matrix_value ()).nnz ();
corefcn/__ilu__.cc:971:      nnz_l = (octave::feval ("tril",
arg_list)(0).sparse_matrix_value ()).nnz ();
corefcn/__ilu__.cc:976:      rc_norm = octave::feval ("norm",
arg_list)(0).vector_value ();
corefcn/__ilu__.cc:985:        octave::feval ("speye",
arg_list)(0).sparse_matrix_value ();
corefcn/__ilu__.cc:1018:        octave::feval ("triu",
arg_list)(0).sparse_complex_matrix_value ().nnz ();
corefcn/__ilu__.cc:1021:        octave::feval ("tril",
arg_list)(0).sparse_complex_matrix_value ().nnz ();
corefcn/__ilu__.cc:1026:      rc_norm = octave::feval ("norm",
arg_list)(0).complex_vector_value ();
corefcn/__ilu__.cc:1036:        octave::feval ("speye",
arg_list)(0).sparse_complex_matrix_value ();


Also, see bug #50116 where feval should be replaced with direct calls to C++
functions in libinterp using the Fname_of_function syntax.


    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?50105>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/