17704, "lydia-duncan", "Fix bug with 'bigint.pow' in BigInteger module when exponent is negative", "2021-05-12T21:38:55Z"

When the exponent is a positive integer, we use the uint version, but when the exponent is negative we send the base as the mod value to work around the fact that GMP doesn't provide a pow version for integers and this wouldn't yield the same result if `y == -1`

or if the base is negative. It looks like this function was added along with a bunch of other int versions - I don't see a review and the PR did not add tests of this function specifically.