On February 26, 2015 8:13:46 PM CET, Jakub Jelinek wrote:
>Hi!
>
>On the following testcase, there is a redundant & 1 and when
>reassociating,
>we remove the redundant & 1. The outer 2 ops are the same, so changed
>is still false when we reach rewrite_expr_tree with the last 2 ops
>(i.e.
>last st
Hi!
On the following testcase, there is a redundant & 1 and when reassociating,
we remove the redundant & 1. The outer 2 ops are the same, so changed
is still false when we reach rewrite_expr_tree with the last 2 ops (i.e.
last statement). But, as we don't keep that statement as is nor just swap
