The "punt" logging pattern in the crypto jets was intended to surface unusual error conditions. But urcrypt and libaes-siv don't clearly distinguish between deterministic decryption failure and other errors (invalid calls, allocation failures, &c). to ensure determinism, so the relevant jets "punt" all errors back to the nock interpreter. And +de:sivc is used by ames, which means that receiving any invalid ciphertext or stale keys produces noisy, useless output: two lines of sivc-de-punt per packet.

The PR just removes the logging, but there are two non-trivial next steps which should be done:

• deterministic decryption failure should not punt

Accomplishing this will require review of libaes-siv and the openssl functions it depends on, and may require patches to distinguish error conditions. This is important for many reasons, not the least of which is that punting is expensive (and therefore a DOS vector).

• receipt of bad packets needs better tracking and logging

Other than the (accidentally) frequent "punt" logging, the ames driver keeps a counter of failed packets and prints every 1K (or every time with -v). But if we're receiving lots of invalid packets with a valid ames header, there's a good chance that we or our peer is behind on PKI state. (And if packets from a particular ip:port are consistently invalid, we probably want to stop try to process them for a while.)