--- Comment #1 from jv244 at cam dot ac dot uk 2010-04-23 09:21 ---
The error also appears with:
gfortran -fbounds-check -O1 -funswitch-loops test.f90
--
jv244 at cam dot ac dot uk changed:
What|Removed |Added
--- Comment #2 from jv244 at cam dot ac dot uk 2010-04-23 09:25 ---
reduced testcase:
IMPLICIT NONE
INTEGER, PARAMETER :: sp=4, dp=8
TYPE cp_fm_type
REAL(KIND=sp), DIMENSION(:,:), POINTER :: local_data_sp
REAL(KIND=dp), DIMENSION(:,:), POINTER :: local_data
INTEGER
--- Comment #3 from jv244 at cam dot ac dot uk 2010-04-23 10:30 ---
It looks like in the .optimized dump these:
a_sp$offset = amat-local_data_sp.offset;
a_sp$dim$1$stride = amat-local_data_sp.dim[1].stride;
a_sp$dim$1$ubound = amat-local_data_sp.dim[1].ubound;
a_sp$dim$1$lbound
--- Comment #4 from rguenth at gcc dot gnu dot org 2010-04-23 12:29 ---
Hm, I can't reproduce this.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43866
--- Comment #5 from jv244 at cam dot ac dot uk 2010-04-23 13:32 ---
(In reply to comment #4)
Hm, I can't reproduce this.
I see, the reduced testcase (comment #2) indeed doesn't fail with trunk
anymore, but the original does (but only at -O3 -fbounds-check).
--
