------- Comment #5 from taschna at uni-muenster dot de 2006-07-31 07:49 ------- (In reply to comment #3) Steve,
> [...]It's more a question of "why is it a NULL pointer?" not > whether we can work around the NULL pointer. i finally found Paul's mail corresponding to patch revision 86128: http://gcc.gnu.org/ml/fortran/2004-08/msg00102.html There he explains the presence of the NULL: > For each entry point we generate a thunk function which tailcalls the master > functions, passing NULL for any arguments which don't exist for that > function. That seems to be the reason for his change in * trans-array.c (gfc_trans_dummy_array_bias): Treat all args as optional when multiple entry points are present. which I copied for my patch. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25818
