About built-in AAs

[bearophile]

Andrei Alexandrescu:

> Let's please stop this.

OK. This sub-thread about Hash resilience against attacks is a minor detail of my original post.

kennytm has answered my point 4, but I'd like some answers to the first three points; regarding the error messages for not complete hash protocol, the possible removal of opCmp from the hash protocol, and the usage of opCmp instead of opEquals from the current hash protocol.

Bye,
bearophile