This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

When to ingore signed vs. unsigned comparison?


In tackling the missed folding of &a[2]-1 to &a[1], I'm
getting unsigned indices instead of signed ones.  For
further allowing &a[1] == &a[1] to be folded, I need a way
to say if I can ignore that 1 (lhs) is unsinged and 1 (rhs)
is signed.

I.e. we don't fold &a[1] == &a[1U] at the moment - how to fix
this?  I tried using may_negate_without_overflow_p, but that
does another thing.  Is there some utility function that would
help me doing this?  Any hints?

Thanks,
Richard.

--
Richard Guenther <richard dot guenther at uni-tuebingen dot de>
WWW: http://www.tat.physik.uni-tuebingen.de/~rguenth/


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]