My memory of the original spec for this was that we should count the number of distinct chunk IDs that are different. Looks like we aren't doing that for some reason.
My memory of the original spec for this was that we should count the number of distinct chunk IDs that are different. Looks like we aren't doing that for some reason.