Use short headers for stateless reset

[martinthomson]

We don't want to prohibit the use of long headers in other versions - in case they want to share the stateless reset design, but there is no sense in sending long headers.

[mikkelfj]

While I agree with the issue, the PR says that other versions might want to reset with a long header and therefore the receiver must inspect long headers anyway. I can see this, but given that, I think it is pointless to require the current version to only send resets for short headers. A general advice ought to suffice since the MUST is not enforceable anyway.

[martinthomson]

The requirement exists to protect those packets from detection.

[mnot]

Please don't close issues until AFTER consensus is declared (not just called).

[ianswett]

I suspect Martin was confused because this issue was tagged has-proposal 15 days ago.

[mnot]

Right. The trigger is has-consensus, not has-proposal.

[ianswett]

But it's understandable Martin would have assumed this was one of the issues that was sent out in the batch of issues a week and a half ago, given this has been tagged with has-proposal for 15 days.