https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84416
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Last reconfirmed| |2018-02-16
CC| |jsm28 at gcc dot gnu.org,
| |rguenth at gcc dot gnu.org
Ever confirmed|0 |1
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
This is a very old known issue with the representation of the dependence
analysis
lambda vectors. It seems to be facilitated by us using __int128 for literals
automagically for a[b - 18446744073709551607] which in this context really
doesn't make sense.
Joseph, the C FE produces
a[b] = a[(__int128) b + -0xfffffffffffffff7];
but shouldn't array indices be restricted to at most ptrdiff_t / size_t? Is
the
FE the appropriate place to do that?
