https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69251
Martin Sebor <msebor at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |ASSIGNED
Assignee|unassigned at gcc dot gnu.org |msebor at gcc dot
gnu.org
--- Comment #6 from Martin Sebor <msebor at gcc dot gnu.org> ---
Thanks. I can see the problem(s) now.
Working on a patch (the file compiles with the following otherwise untested
fix).
Index: cp/mangle.c
===================================================================
--- cp/mangle.c (revision 232296)
+++ cp/mangle.c (working copy)
@@ -3296,6 +3296,9 @@ write_array_type (const tree type)
/* The INDEX_TYPE gives the upper and lower bounds of the
array. */
max = TYPE_MAX_VALUE (index_type);
+ if (max == NULL_TREE)
+ max = build_all_ones_cst (size_type_node);
+
if (TREE_CODE (max) == INTEGER_CST)
{
/* The ABI specifies that we should mangle the number of
Index: cp/pt.c
===================================================================
--- cp/pt.c (revision 232296)
+++ cp/pt.c (working copy)
@@ -19108,7 +19108,13 @@ unify_array_domain (tree tparms, tree ta
parm_max = TREE_OPERAND (parm_max, 0);
}
arg_max = TYPE_MAX_VALUE (arg_dom);
- arg_cst = TREE_CODE (arg_max) == INTEGER_CST;
+ if (arg_max)
+ arg_cst = TREE_CODE (arg_max) == INTEGER_CST;
+ else
+ {
+ arg_max = build_all_ones_cst (size_type_node);
+ arg_cst = true;
+ }
if (!arg_cst)
{
/* The ARG_MAX may not be a simple MINUS_EXPR, if we are
