This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug fortran/47455] [4.6 Regression][OOP] internal compiler error: in fold_convert_loc, at fold-const.c:2028
- From: "burnus at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Mon, 31 Jan 2011 22:18:51 +0000
- Subject: [Bug fortran/47455] [4.6 Regression][OOP] internal compiler error: in fold_convert_loc, at fold-const.c:2028
- Auto-submitted: auto-generated
- References: <bug-47455-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47455
--- Comment #10 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-01-31 22:18:41 UTC ---
(In reply to comment #7)
> The test in comment #6 segfault also.
That's a test-case problem.
this%x = find_x(this)
is invalid if the LHS is not associated - and a RHS null() is also problematic.
Use:
this%x => find_x(this)
instead (for all assignments in the file).
> The ICEs disappear with the patch in comment #6, but the test in comment #4
> fails at
> if(any (this%x%i /= [5, 7])) call abort() ! FAILS
I think you mean comment 5. In comment 4, the line:
this%y = this%find_y()
fails; the LHS is a nonpointer, nonallocatable DT. The RHS returns an
(allocated) allocatable; the type is type(tx) w/ allocatable but unallocated
components.
I think the test case is valid
Regarding comment 5: As written there, a deep copy is missing