Description:
------------
Better support for 64bit integers would be very nice, for example when performing this operation:

$val = 0x1 << 32;
echo $val;

The value "1" is returned, due to the 32bit wrapping applied to the integer. Even when using type casts to float or double it doesn't extend beyond 32bits.

An extra type such as Int64 would be very useful.

Reproduce code:
---------------
$val = 0x1 << 32;
echo $val."\n";

(float)$val = (float)0x1 << 32;
echo $val;

Expected result:
----------------
It should output "4294967296" for both

Actual result:
--------------
It outputs "1" due to 32bit wrapping