On Sat, Nov 7, 2020 at 3:42 AM Luc Van Oostenryck
<luc.vanoostenryck@xxxxxxxxx> wrote:
>
> These patch add a few simplifications when the condition of a
> select is another select and have constants as operands.
Looks good to me.
I think there's a couple of others, like
SEL(x, x, 0) -> x
which looks insane, but I think these are the kinds of things that end
up showing up when you inline things that return errors, and then you
test them for zero etc.
I haven't thought them all through, and didn't check your cases deeply
and just took your word for them, but it looked all sane to me.
Linus
