zeekayandHanzo Dev 578de7ef2b pop: freeze the over-long key, so the width is refused from both sides ...

The corpus froze a 47-byte key and both a 95- and a 97-byte proof, but no
49-byte key — so the one width a port could get wrong was the one the standard
did not carry. The clause is a width equality, not a lower bound: a decoder that
guarded only the short side reads the first 48 bytes of a longer buffer and
checks a proof against a PREFIX of what the registrant published. That is a
second spelling of an already-seated key, and two spellings of one key is two
seats and two votes on one signature.

The case is generated by the oracle and self-checked against pop.Verify, like
every other row, and lands as an ENCODING refusal before any pairing.

Go already held it as a unit assertion and Rust held its own; Rust's is dropped
here, because a case the frozen corpus carries for all three does not also want
a private restatement that has to be kept in step.

Co-authored-by: Hanzo Dev <[email protected]>

2026-09-02 02:22:56 -07:00

..

pop.json

pop: freeze the over-long key, so the width is refused from both sides

2026-09-02 02:22:56 -07:00