[PATCH] Don't use slowcompare method unconditionally

[Kelley Cook]

> I don't think so.  I'm using coreutils-5.93, release apparently 3
days 
> ago, and it refuses tail +NNN form by default.
> Of course I can use "_POSIX2_VERSION=199506 tail +16c" instead of
just 
> tail +16c and it will work, but that doesn't sound to me like
"coreutils 
> has already been fixed".

So wouldn't defining a POSIX_TAIL="_POSIX2_VERSION=199506 tail" and
using then using $(POSIX_TAIL) +16c be a more complete idea for bith
broken and non-broken coreutils?