code wiki / (root) / nx_ship_fleet_gate.nx

nx_ship_fleet_gate.nx

buildroot/runtime/nx_ship_fleet_gate.nx

4116 B59 linesdepth 3pulls 4 transitivereach 0 importersview sourcekind gate/prooftopic ship
docsdependenciesstructsconstsfunctions

about

nx_ship_fleet_gate.nx -- bite-proves the PURE adaptive-concurrency core in nx_shipfleet_lib.nx: the AIMD cap state machine (additive-increase +1 on progress, multiplicative-decrease halve on a storm, ceilinged by the polite budget, floored at 1) and the effective-width gate (pause on serve-first or admission-deny, else min(budget,cap)). The DRIVER's fork/wait4 pool is I/O and cannot be unit-tested here; the load-bearing DECISIONS are pure and are all tested. Every tooth is an equality on a pure function, so a mutant that (a) grows on a storm, (b) ignores the ceiling, (c) drops below 1, or (d) launches while serving/refused is KILLED. license_tier: ORIGINAL evidence -> stdout + verdict line (exit code carries the verdict via gv_verdict).

dependencies 3 imports · 0 importers

nx_syscalls.nx nx_gate_verdict.nx nx_shipfleet_lib.nx nx_ship_fleet_gate.nx

imports: nx_syscalls.nxnx_gate_verdict.nxnx_shipfleet_lib.nx

imported by: nobody (leaf or entry point)

call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown

main gv_ctr sys_mmap nxa_die sys_write sys_exit nxa_lock_take nxa_lock_addr sys_write ↻ nxa_lock_give nxa_lock_addr ↻ nxa_report_overrun sys_write ↻ nxa_dump_printable sys_write ↻ nxa_dump_sizes sys_write ↻ gv_check gv_puts sys_write ↻ sf_aimd_next sf_effective_width sf_launch_slots gv_verdict gv_note_bare_rate gv_bare_rate gv_at gv_obj_has_n gv_at ↻ gv_puts ↻ gv_num sys_mmap ↻ sys_write ↻ sys_munmap gv_puts ↻ gv_num ↻ gv_journal sys_openat_append sys_mmap ↻ gv_catn

structs

none

consts

none

functions

13func main() -> i64