> bcrypt(password, salt+pepper) I hope it's obvious that no one should never do this, since the output would contain the "salt+pepper" bits in cleartext alongs

> bcrypt(password, salt+pepper) I hope it's obvious that no one should never do this, since the output would contain the "salt+pepper" bits in cleartext alongside the hash, defeating the entire point of the "pepper": https://www.usenix.org/legacy/event/usenix99/provos/provos_html/node5.html https://www.usenix.org/legacy/event/usenix99/provos/provos_h... In fact, this is a perfect illustration of why it's bad to put secret bits into a crypto function in a place that's not designed to take secret bits. Bcrypt does not treat the salt parameter as a cryptographic secret, and other algorithms might not either. And they might leak it in more subtle ways.