[Bug fortran/57126] Matching host-associated generic vs. same-named use-associated generic

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57126

Tobias Burnus <burnus at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |burnus at gcc dot gnu.org
            Summary|Wrongly rejects a generic   |Matching host-associated
                   |call because it doesn't     |generic vs. same-named
                   |match the intrinsic while   |use-associated generic
                   |another specific matches    |

--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> 2013-04-30 16:44:36 UTC ---
For clarification:

- ABS and SIZE are both generic names, either having an intrinsic and a user
specific procedure.
- In "sub", the specific functions of "SIZE" do not match the rejected calls.
- The specific functions of "ABS" do match - and both have the same generic
name

See:
    USE MOO, ONLY: RENAME => ABS
    CONTAINS
        SUBROUTINE SUB
            USE MOO, ONLY: RENAME => SIZE, T


Thus, the question is whether the generic "rename" are merged or whether the
inner "rename" overrides the host-associated "rename".

If they are merged, the question is how ambiguity is handled.