safeOp.divPow2

Equivalent to left / pow(2, exp), but faster and works with a wider range of inputs. This is a safer alternative to left >> exp that is still very fast.

Note that (conceptually) rounding occurs after the /, meaning that divPow2(left, -exp) is equivalent to mulPow2(left, exp).

auto divPow2(N left, M exp)
template safeOp(IntFlagPolicy policy)
pure @safe nothrow @nogc
auto
divPow2
(
N
M
)
(
const N left
,
const M exp
)
if (
(
isFloatingPoint!N &&
isScalarType!M
) ||
(
isScalarType!N &&
isFloatingPoint!M
)
)
auto divPow2(N left, M exp)

Examples

import checkedint.sticky : safeOp; // use IntFlagPolicy.sticky

assert(safeOp.divPow2(65536, 8) == 256);
assert(safeOp.divPow2(-100, 100) == 0);
assert(safeOp.divPow2(-23, -5) == -736);

safeOp.divPow2(10_000_000, -10);
assert(IntFlags.local.clear() == IntFlag.posOver);

safeOp.divPow2

Examples

Meta

Source