LaundroMat <[EMAIL PROTECTED]> writes: > Would it be wise to calculate the hash value based on say for instance > the first Mb? Is there a much larger chance of collusion this way (I > suppose not). If it's helpful, the files would primarily be media > (video) files.
The usual purpose of using this type of hash is to detect corruption and/or tampering. So you want to hash the whole file, not just part of it. If you're not worried about intentional tampering, md5 should be somewhat faster than sha, but there are some attacks against it and you shouldn't use it for high security applications where you want security against forgery. It should still have almost no chance of accidental collisions. -- http://mail.python.org/mailman/listinfo/python-list