On Sun, 1 Jun 2025, Martin Uecker wrote:
> This patch now moves the checking assertion out of the recursion,
> which seems better anyhow. I could not do the check unconditionally
> for all types. because we sometimes call composite_type for with
> mismatching qualifiers.
Calling composite_type with mismatching qualifiers seems like a bug; it's
directly contrary to the comment on composite_type_internal. So either
the code with the call is wrong, or that comment should give a proper
description of the circumstances under which calling with mismatching
qualifiers is OK.
/* Return the composite type of two compatible types.
We assume that comptypes has already been done and returned
nonzero; if that isn't so, this may crash. In particular, we
assume that qualifiers match. */
> @@ -1004,7 +999,15 @@ tree
> composite_type (tree t1, tree t2)
> {
> struct composite_cache cache = { };
> - return composite_type_internal (t1, t2, &cache);
> + tree n = composite_type_internal (t1, t2, &cache);
> + /* For function and arrays there are some exceptions where
> + qualifiers do not match. */
The patch is OK with this comment updated to reference an open bug in
Bugzilla about the cases where composite_type is called with mismatching
qualifiers.
--
Joseph S. Myers
josmy...@redhat.com
