Re: crypt() BC issue

Hi David, On Mon, Jul 21, 2014 at 2:53 PM, David Muir <davidkmuir@gmail.com> wrote: > Prehashing with sha512 means it is no longer blowfish. It is now a > non-vetted DIY algorithm. The whole point of password_hash is to avoid this > type of thing, and should be clearly discouraged in the documentation. > I agree. It's far better if it could handle limitless password length. The problem is "there is no way to achieve this with current implementation". It's a classic example of what not to do. > I agree here, too. We are better to have algorithm that does not enforce user/developer to certain password and recommend "Just use it". However, Using multiple hashes for better security is common technique. An example is SSL. So I would not say one should not. Especially when there is a limitation. In old days, crypt() was unusable securely. There are many users/developers that are used to have static slat. Code like below disables authentication completely. password_hash(hash('sha512', SOME_SECRET_SALT).$password, DEFAULT); This should be prevented. (I would like to prevent it by raising E_NOTICE error) If we would like to recommend "Just use it", we may consider adding SHA512 to password_hash(). Regards, -- Yasuo Ohgaki yohgaki@ohgaki.net