------- Comment #5 from janus at gcc dot gnu dot org 2009-11-04 20:39 ------- (In reply to comment #2) > While one might not access (type)%dot_g_g as "dot_g_g" is deferred, using > (class)%dot_g_g is valid. (And using (type)%dot_g_g is not possible as one > cannot use "type(abstract_t)"; thus it might be enough to relax the check.)
It might be okay for type-bound procedures. However, this check should also catch the usage of simple abstract procedures, i.e. when adding the following to the test case in comment #4: print *, dot() This is illegal of course, and is not caught without the check in resolve_function. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41873
