[Bug c++/56930] [4.8/4.9 regression] pointless -Wconversion warning with sizeof

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56930

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
So perhaps for the purposes of the -Wconversion warning we also want to
t = fold_nondependent_expr_sfinae (expr, tf_none);
t = maybe_const_value (t);
before warnings_for_convert_and_check call and pass the result of that instead
of expr?