http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45743
Richard Guenther <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |ice-on-invalid-code
Priority|P1 |P4
Status|UNCONFIRMED |NEW
Last reconfirmed| |2010.11.04 11:46:57
Component|tree-optimization |fortran
Ever Confirmed|0 |1
--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-11-04
11:46:57 UTC ---
I wonder why there is a dg-error directive but only a warning is emitted. This
sounds like both a fortran frontend error and a testsuite bug (it should
have a FAIL because of the missing error?).
Another bug is that the frontend seems to emit multiple decls for
the EXTERNAL R symbol dependent on the call site. And it makes that
decls use a type dependent on the actual call despite knowing better
(and issuing a warning).