Tejas Belagod <tejas.bela...@arm.com> writes:
> There is an assumption in many places in c-typeck.cc that GNU vectors sizes
> are
> always known at compile time. SVE vectors now piggy-back on GNU vector code
> so this patch changes one of the places where there is an assumption of
> vectors
> being fixed-length to being variable width.
>
> gcc/ChangeLog:
>
> * c/c-typeck.cc (process_init_element): Fix condition to include
> sizeless types.
> ---
> gcc/c/c-typeck.cc | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/gcc/c/c-typeck.cc b/gcc/c/c-typeck.cc
> index 6673cbf7294..94959f7b8ad 100644
> --- a/gcc/c/c-typeck.cc
> +++ b/gcc/c/c-typeck.cc
> @@ -11969,8 +11969,9 @@ retry:
> tree elttype = TYPE_MAIN_VARIANT (TREE_TYPE (constructor_type));
>
> /* Do a basic check of initializer size. Note that vectors
> - always have a fixed size derived from their type. */
> - if (tree_int_cst_lt (constructor_max_index, constructor_index))
> + may not always have a fixed size derived from their type. */
> + if (known_lt (tree_to_poly_uint64 (constructor_max_index),
> + tree_to_poly_uint64 (constructor_index)))
> {
> pedwarn_init (loc, 0,
> "excess elements in vector initializer");
Others might disagree, but FWIW: I think it would make sense to fold
this into patch 9. The end result there looks good to me, but this
intermediate state looks surprising.
Thanks,
Richard
