On Thu, Nov 06, 2014 at 11:10:37AM +0100, Jakub Jelinek wrote:
> Sorry for being slow on this, but in the current algorithm you are using
> the latest vector element from each bb only anyway, so the
> && tree_int_cst_compare (cur_align, align) == 0
> is unnecessary, even if the condition isn't true, and thus in your patch we
> would not pop the old one, we'd still never use the not popped element,
> as the stmt pushed after it would be used instead, or both stmts popped up
> if the domwalk left the bb already.
Ah, yeah, after some drawing on paper, I agree. I should've notice that
earlier :/.
> If you want, commit your patch as is and in test the one liner together with
> some other change.
I'll do that. Thanks,
Marek
