On Wed, 10 Jan 2024 14:58:37 GMT, Raffaello Giulietti <rgiulie...@openjdk.org> wrote:
>> src/java.base/share/classes/java/math/BigInteger.java line 3998: >> >>> 3996: int i = ArraysSupport.mismatch(m1, m2, len1); >>> 3997: if (i != -1) >>> 3998: return Integer.compareUnsigned(m1[i], m2[i]) < 0 ? -1 : 1; >> >> Just an observation. The (Java and intrinsic) implementation of >> Integer.compareUnsigned already returns -1, 0, 1. >> Returning `Integer.compareUnsigned(m1[i], m2[i])` would yield the same >> result without the tertiary expression. > > Yes, that's what was proposed > [here](https://github.com/openjdk/jdk/pull/14630#discussion_r1242245875) some > time ago. > > But the spec of `compareUnsigned()` does _not_ guarantee a -1, 0, 1 result, > so there's a (minimal) risk when returning its value directly. (For some > reason, `BigInteger` specifies a -1, 0, 1 outcome.) In expressing intent, perhaps `Integer.signum(Integer.compareUnsigned(m1[i], [m2[i]))`. Though it may all be for nothing if C2 can optimize it away. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/14630#discussion_r1447521230
