[Bug tree-optimization/109008] [13 Regression] Wrong code in scipy package since r13-3926-gd4c2f1d376da6f

["jakub at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109008

--- Comment #18 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
(In reply to Richard Biener from comment #16)
> (In reply to Jakub Jelinek from comment #14)
> > Created attachment 54599 [details]
> > gcc13-pr109008-wip.patch
> > 
> > I have now this WIP but it doesn't work correctly yet, need to debug it now,
> > just finished writing it and making it compile.
> > Note, after the binary search it is still unfinished, I'd like to then
> > adjust it one bit at a time to narrow it further.  But before working on
> > that it needs to work correctly.
> 
> Looks nicer and more precise (even though the iterating function is ugly
> and you also suffer from the lack of copying of alternate range state,
> but using .union () is a way out here I guess ;)).

I thought the union_ handles it.  I didn't want to simply override the lower or
upper bound on r, because frange::set has various canonicalizations and the
like.

> Can we ignore -frounding-math here?

I believe it isn't ignored, -frounding-math ought to be taken into account in
frange_arithmetic.  The intent was that the testing function/lambda would e.g.
for the
addition when extending lower bound perform rounding etc. towards +inf and when
extending upper bound towards -inf.
But as I said, the patch currently doesn't work correctly, the test on the
nextafter
value at the start of the new function fails even when I'd think it should
pass.