a hash which treats two different strings as identical is a a serious failure.
No, it's not. Collisions in a hash are by design. After all, the possible number of different values in a 8000-character string is much larger than the possible number of 160-but hash values.
That's not what causes the problem here, though. As I'm sure you understand after reading all the other replies. The real problem is that REPLICATE uses the same input as output data type, which in this case has an 8000-byte limit, and that longer values are truncated. Raising a bug is pointless, as this is all both by design and well documented.