nx_msg_sync_gate.nx
buildroot/runtime/nx_msg_sync_gate.nx
about
nx_msg_sync_gate.nx -- REFEREE for C2 (nx_msg_sync, contract msy_devsync). END-TO-END: builds a
server room and device replicas in /tmp/msgate (nothing production is touched), forks BOTH promoted
elfs -- nx_chat_store to author fixtures and READ BACK, nx_msg_sync as the subject -- and proves the
pre-declared done-rule: "the same history converges on a second device; a partition-printed
reconcile measures zero divergence."
THE ANTI-VACUITY DESIGN: convergence is NEVER taken from the syncer's own receipt. T3/T5 compare
nx_chat_store fetch output on the server root vs the device root BYTE FOR BYTE -- the consumer is
the referee, so a syncer that lied about pulling would fail the tooth it cannot see.
NEGATIVE CONTROLS: split-brain divergence refuses (seq named, nothing written) · device-ahead
refuses · absent source never conjures · verify itself goes RED on the diverged replica.
Fixture rooms are epoch-suffixed; every verb is its own fork, so persistence across process death
is structural. license_tier: ORIGINAL expect_exit: 0
dependencies 3 imports · 0 importers
imports: nx_syscalls.nxnx_tool_run.nxnx_gate_verdict.nx
imported by: nobody (leaf or entry point)
call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown
structs
| none |
consts
| 16 | const MSG_CAP: i64 = 65536 |
functions
| 17 | func msg_len(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } return n } |
| 18 | func msg_write(path: *u8, s: *u8) -> i64 |
| 27 | func msg_count(buf: *u8, n: i64, needle: *u8) -> i64 |
| 40 | func main(argc: i64, argv: *i64) -> i64 |