reject P2P message (BIP61) #75

Long overdue, I finally found the time to tidy up my 'gist' into a BIP.

Thanks, I think this can be very useful. Random thoughts:

Should the string somehow be limited? Why is there no "accept" message as well? If there is a reason, I think it should be noted somewhere in this BIP. I think there should be a standard rejection code for "Double spend". That's the most common reason transactions for non-propagating transactions from SPV clients.

The string is limited by the maximum length on P2P messages. If you're worried about use-up-bandwidth denial-of-service attacks, that worry is misplaced (I can send you max-length unknown messages all day long that you ignore as easily as reject messages, but it costs me as much bandwidth as it costs you).

There is no accept message because that would just waste bandwidth in the common case. As for putting that in the spec: not going to do that, specs should be short and to-the-point, not rehash every design decision that went into making them.

rejection code for double-spend: patches welcome, feel free to implement that and submit a new BIP.

Sure, here is my "patch":

| 0x43 || Attempted double spend

Why a new BIP for that simple addition?

re "accept": I thought it might be useful. Many devs will ask this question, as usually you have an accept/reject pair.

@laanwj Did you look at my comments before merging?

@schildbach Yes, but unlike with code, not all comments need to be addressed before putting something in the repository in the first place. The information is not final. You can submit any changes as your own pull request.