https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92342
--- Comment #8 from Segher Boessenkool <segher at gcc dot gnu.org> ---
(In reply to Richard Earnshaw from comment #6)
> (In reply to Richard Earnshaw from comment #5)
> > So if the AND-based idiom is now preferred, shouldn't the if-then-else
> > variant be transformed into it? Similarly for IOR, when we get
> >
> > (IOR (NEG (<cond-op>)) (reg))
> >
> > from
> >
> > (IF_THEN_ELSE (<cond-op>)
> > (reg)
> > (const_int -1))
>
> except that should be
>
> (IF_THEN_ELSE (<cond-op>')
> (reg)
> (const_int -1))
>
> Where <cond-op>' is the reversed condition.
Or just
(if_then_else (<cond-op>)
(const_int -1)
(reg))
(No order )like constant last) of those two arguments is prescribed, and
this makes sense).
