On Tue, Jul 21, 2020 at 05:43:00PM +0200, Tobias Burnus wrote:
> --- a/gcc/c-family/c-omp.c
> +++ b/gcc/c-family/c-omp.c
> @@ -106,6 +106,18 @@ c_finish_omp_taskgroup (location_t loc, tree body, tree
> clauses)
> tree
> c_finish_omp_critical (location_t loc, tree body, tree name, tree clauses)
> {
> + gcc_assert (!clauses || OMP_CLAUSE_CODE (clauses) == OMP_CLAUSE_HINT);
> + if (name == NULL_TREE
> + && clauses != NULL_TREE
> + && TREE_CODE (OMP_CLAUSE_HINT_EXPR (clauses)) == INTEGER_CST
> + && integer_nonzerop (OMP_CLAUSE_HINT_EXPR (clauses)))
I still don't see the point of the INTEGER_CST, integer_nonzerop does that
(well, it allows also COMPLEX_CST, but the clauses checking should complain
if the expression is complex).
So, ok for trunk with that line removed.
Thanks and sorry for not getting everything at once.
Jakub
