Hi there,
Karsten and I have been working on metrics-related tasks and one thing that's got a potential to cause us pain going forward is that of "p"-lines in consensuses. Tor has added them with consensus method 5 but only declared them to be optional. In practice, every consensus ever produced that used method 5 or later has added these lines, and I'd like to require them to be included going forward. The reason is that without looking at the consensus method, there's an ambiguity wrt exits: A missing line either means "reject 1-65535" (method 5 or later) or "no information" (anything before that).
I believe this should cause no compatibility issues and wouldn't require a new consensus method. Is there some reason against doing this now?
Cheers Sebastian