[Bug fortran/34813] ICE on incorrect nested type constructor (fold-const.c (fold_convert):2629)

["dfranke at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>]

------- Comment #3 from dfranke at gcc dot gnu dot org  2008-03-21 10:20 -------
Last night, I tracked this to resolve.c (resolve_structure_cons), but I was not
100% sure what to do.

With the simplified testcase, Intel complains:
$> ifort pr34813.f90
fortcom: Error: pr34813.f90, line 15: The type of the NULL(MOLD) element in a
structure-constructor differs from the type of the component in the
derived-type-def.
 tree = kd_tree(kd_tree_node(null()))
-----------------------------^
fortcom: Error: pr34813.f90, line 15: This array or function or substring is
invalid in constant expressions.
 tree = kd_tree(kd_tree_node(null()))
----------------^
fortcom: Error: pr34813.f90, line 15: Replacing invalid structure constructor
by integer 1   [<NULL_STRING>]
 tree = kd_tree(kd_tree_node(null()))
^

Interestingly, if it is slightly changed,
-  tree = kd_tree(kd_tree_node(null()))
+  integer, pointer :: i
+  tree = kd_tree(kd_tree_node(null(i)))

Intel gives an ICE while gfortran happily (but wrongly?) accepts it:
$> ifort -warn all pr34813.f90
fortcom: Severe: pr34813.f90, line 12: **Internal compiler error: internal
abort** Please report this error along with the circumstances in which it
occurred in a Software Problem Report.  Note: File and line given may not be
explicit cause of this error.
  tree = kd_tree(kd_tree_node(null(i)))
------------------------------^


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34813