/reports/phase0_compile.md · raw

Phase 0: compile findings (2026-10-09)

Toolchain: opam 2.5.0, switch bb6, OCaml 4.14.2, Coq 8.20.1 with coq-native (COQ_NATIVE_COMPILER_DEFAULT=yes). Activate with eval $(opam env --switch=bb6).

Framework build

Single-proof compiles (from a temp dir, -Q verify BusyCoq -Q verify/BigInt BigInt -native-compiler yes)

File Result Wall time Notes
Example.v OK 2 s nonhalt lemmas live inside modules
Bouncer_v3.v (1LB1RA_0RA0RC_---1RD_1RE1RB_1LF0LE_0RD0LE) OK 244 s needs ulimit -s unlimited (stack overflow at default 8 MB); one Qed takes 158 s
Bouncer_v2.v (1RB0LB_1RC0LC_1LD1RF_---1LE_0RC1LA_0RB1RA) OK 188 s Print Assumptions TM1.nonhalt: Closed under the global context (no axioms)
1RB1LD_1RC1RE_0LA1LB_0LD1LC_1RF0RA_---0RC.v OK 11 s cryptid-style rule reduction: no nonhalt lemma
1RB0LD_1RC1RF_1LA0RA_0LA0LE_1LD1LA_0RB---.v OK 6 s cryptid-style rule reduction: no nonhalt lemma

Gotchas for check_kill.sh

Assumptions baseline

Print Assumptions TM1.nonhalt. for Bouncer_v3.v:

Axioms:
FunctionalExtensionality.functional_extensionality_dep :
  forall (A : Type) (B : A -> Type) (f g : forall x : A, B x),
  (forall x : A, f x = g x) -> f = g

Bouncer_v2.v: Closed under the global context.

Baseline axiom set = { FunctionalExtensionality.functional_extensionality_dep } (the max over accepted proofs seen; a proof with no axioms trivially passes).