module mod1
type t1
real :: f1
end type t1
type t2
type(t1), pointer :: f2(:)
real, pointer :: f3(:,:)
end type t2
end module mod1
module mod2
use mod1
type(t1), pointer, save :: v(:)
contains
subroutine foo (x)
use mod1
implicit none
type(t2) :: x
integer :: d
d = size (x%f3, 2)
v = x%f2(:)
end subroutine foo
end module mod2
ICEs in fold_convert, because gfortran created two distinct t1 types,
one for the v variable and a different one for x.
I'd say either gfortran should cache somewhere module derived type tree nodes
and reuse them, or it needs a non-default type comparison hook that would
figure which derived types are in fact compatible.
--
Summary: Module derived type issue
Product: gcc
Version: 4.2.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: fortran
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: jakub at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25391
