On Mon, 9 Sep 2019, Martin Liška wrote:
> On 9/9/19 3:42 PM, Richard Biener wrote:
> > There is no newly created GIMPLE?
>
> Hm, I thought from the beginning that maybe_fold_comparisons_from_match_pd
> can come up with new temporary expressions that need to be inserted into
> GIMPLE stream? But that's probably handled in ifcombine with:
>
> t = force_gimple_operand_gsi_1 (&gsi, t, is_gimple_condexpr, NULL,
> true,
> GSI_SAME_STMT);
> ?
No, that case is done when forcing short-circuiting when there was no
simplification. When there was a simplification we do
if (result_inv)
t = fold_build1 (TRUTH_NOT_EXPR, TREE_TYPE (t), t);
t = canonicalize_cond_expr_cond (t);
if (!t)
return false;
so when it is not a condition suitable for direct replacement into
gimple_cond_set_condition_from_tree (inner_cond, t);
we fail.
Richard.
