Hi!
In r11-6895 handling of empty bases has been fixed such that non-lval
stores of empty classes are not added when the type of *valp doesn't
match the type of the initializer, but as this testcase shows it is
done only when *valp is non-NULL. If it is NULL, we still shouldn't
add empty class constructors if the type of the constructor elt *valp
points to doesn't match.
Bootstrapped/regtested on x86_64-linux and i686-linux, tested also by
Liu Hao on 10 branch, preapproved by Jason in the PR, committed to
trunk.
2021-04-16 Jakub Jelinek <ja...@redhat.com>
PR c++/100111
* constexpr.c (cxx_eval_store_expression): Don't add CONSTRUCTORs
for empty classes into *valp when types don't match even when *valp
is NULL.
* g++.dg/cpp0x/constexpr-100111.C: New test.
--- gcc/cp/constexpr.c.jj 2021-04-08 17:19:12.235414181 +0200
+++ gcc/cp/constexpr.c 2021-04-16 11:06:27.138019389 +0200
@@ -5518,6 +5518,14 @@ cxx_eval_store_expression (const constex
CONSTRUCTOR_NO_CLEARING (*valp)
= CONSTRUCTOR_NO_CLEARING (init);
}
+ else if (TREE_CODE (init) == CONSTRUCTOR
+ && !same_type_ignoring_top_level_qualifiers_p (TREE_TYPE (init),
+ type))
+ {
+ /* See above on initialization of empty bases. */
+ gcc_assert (is_empty_class (TREE_TYPE (init)) && !lval);
+ return init;
+ }
else
*valp = init;
--- gcc/testsuite/g++.dg/cpp0x/constexpr-100111.C.jj 2021-04-16
11:33:20.193882609 +0200
+++ gcc/testsuite/g++.dg/cpp0x/constexpr-100111.C 2021-04-16
11:32:46.039266786 +0200
@@ -0,0 +1,7 @@
+// PR c++/100111
+// { dg-do compile { target c++11 } }
+// { dg-options "-fno-elide-constructors" }
+
+struct A {};
+struct B : A { int b; constexpr B (A x) : A(x), b() {} };
+struct C { B c; constexpr C () : c({}) {} } d;
Jakub
