Text Hasher: know exactly which bytes you fingerprint

The digest represents bytes, not an abstract piece of text. Define Unicode normalization, character encoding, newline policy and whether surrounding whitespace is significant before hashing. Store the canonicalization and algorithm version with the digest. Hashing low-entropy text does not hide it from guessing, and an unkeyed hash does not prove who supplied it. Use a MAC for authenticated integrity and a salted password-hashing function for passwords. Preserve the original length and reject inputs above a clear processing limit.