[Bug middle-end/50640] [4.7 Regression] FAIL: gfortran.dg/select_type_12.f03 -O (internal compiler error)

[hp at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50640

--- Comment #13 from Hans-Peter Nilsson <hp at gcc dot gnu.org> 2011-11-06 17:47:16 UTC ---
(In reply to comment #12)
> second
> I would prefer to see the bug fixed

Certainly.

> rather than hidden by any of the above.

It wouldn't be hidden by xfailing it, as the xfail would refer to this PR.

And this _is_ standard procedure if the fix is more important than errors it
uncovers (which of course isn't known until the offending change is known, but
if the regression is unimportant sometimes it doesn't matter - probably not the
case here as it's an ICE).

But the implied question is unanswered: is there work in progress on this
regression?

I thought the change introducing the regression was known, but apparently not;
would it help to triage it?