------- Additional Comments From tobi at gcc dot gnu dot org 2004-12-08 23:44
-------
The error is precise a->base == 0 is not true, because we don't initialize it to
zero: from the t02.original dump:
unity.data = 0B;
w.data = 0B;
x.data = 0B;
{
real8[] * * D.475;
unity.dtype = 538;
unity.dim[0].lbound = 1;
unity.dim[0].ubound = 2;
unity.dim[0].stride = 1;
unity.dim[1].lbound = 1;
unity.dim[1].ubound = 2;
unity.dim[1].stride = 2;
D.475 = &unity.data;
_gfortran_allocate (D.475, 32, 0);
unity.offset = -3; <----------------------------------------
}
and the same for w and x.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18857
