Continued fuzzing and hardening about interledger-rs

[koivunej]

interledger_packet::packet::fuzzed::fuzzed_0_trailing_bytes

Just realized that there are existing test cases in crates/interledger-packet/src/packet.rs which would suggest that supporting additional trailer bytes is a feature. The tests are inside multiple aspect testing test_try_from named #[test] fns. If this is per the RFCs then the fuzzing target should be fixed and the test removed. I can't really see why this should be supported however and there are no comments on the topic either.

[koivunej]

Removed the security label because the above linked PRs haven't contained any security fixes, or even large mallocs, only strictness, or roundtrippability changes.

Strictness: #707, #710, #715, #716
Roundtrippability: #719, #720

The #720 highlights a possible correctness issue with the parsing of leap seconds on every date which should be further investigated.

[koivunej]

The corpus files after minimization in interledger-packet:

$ du -h fuzz/corpus/
492K    fuzz/corpus/packet
436K    fuzz/corpus/address
932K    fuzz/corpus/

btp is around 820K, stream is 724K, ccp is 828K in total. Would probably be great to push them into the repo, have the CI run 100_000 runs of each for good measure or -max_total_time=1s which must be the cputime of the fuzz_target.