vm-ext: enable JIT in epoch serving mode (Smalltalk 847/847, Datalog 356/356)

register_jit_hook is now installed in the persistent (epoch) serving-mode
branch of sx_server.ml, not just --http/cli/site. Smalltalk-on-SX conformance
under JIT is 847/847 — identical to the no-JIT baseline; Datalog 356/356.
run_tests --jit/no-jit are byte-identical before/after (no regression).

Five distinct root causes fixed (not one "miscompile"):

1. Serving mode never loaded lib/compiler.sx, so JIT used the native
   Sx_compiler.compile stub (arity-0 bytecode, params as GLOBAL_GET →
   "VM undefined: <param>"). Server-mode branch now loads compiler.sx
   before registering the hook, matching http/cli/site.

2. compile-cond / compile-case-clauses / compile-guard-clauses only treated
   keyword :else and true as the catch-all, not the bare symbol `else` that
   the CEK's is-else-clause? accepts → GLOBAL_GET "else". (lib/compiler.sx)

3. OP_DIV produced a float for non-divisible Integer/Integer (1/2 → 0.5)
   instead of the exact Rational the "/" primitive returns. Now delegates to
   the primitive, matching CEK. (sx_vm.ml)

4. OP_EQ / _fast_eq lacked Rational/ListRef cases that the "=" primitive's
   safe_eq has → (= 1/2 1/2) false under JIT. OP_EQ now delegates non-scalars
   to the "=" primitive; _fast_eq gained rational + ListRef. (sx_vm.ml,
   sx_runtime.ml)

5. Continuation-based control flow (Smalltalk ^expr non-local return, block
   escape, exceptions via call/cc) can't run in the stack VM. New data-driven
   exclusion set Sx_types.jit_excluded + `jit-exclude!` primitive, consulted in
   jit_compile_lambda (covers both the CEK hook and vm_call's tiered path).
   lib/smalltalk/eval.sx self-declares its continuation dispatch core
   interpret-only; pure helpers still JIT. The SUnit suite-runner test helper
   pharo-test-class miscompiles mid-loop and is excluded in tests/tokenize.sx.

Also adds SX_JIT_DENY / SX_JIT_ONLY env-var bisection filters to the serving
hook. Known residual documented in plans/jit-bytecode-correctness.md: the hook
re-runs a failed VM execution via CEK (correct result, possible duplicate side
effects); adopting run_tests' propagate-don't-rerun semantics is deferred to
avoid changing shared VM/CEK behavior under this loop.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-06-19 20:36:30 +00:00
parent f026177e63
commit 952ff2289c
9 changed files with 256 additions and 18 deletions

View File

@@ -808,14 +808,31 @@ and run vm =
let b = pop vm and a = pop vm in
push vm (match a, b with
| Integer x, Integer y when y <> 0 && x mod y = 0 -> Integer (x / y)
| Integer x, Integer y -> Number (float_of_int x /. float_of_int y)
(* Non-divisible Integer/Integer must delegate to the "/" primitive:
it returns an exact Rational (e.g. 1/2), matching CEK semantics.
Inlining float division here (0.5) diverges from the interpreter
and breaks numeric equality against rational results. *)
| Number x, Number y -> Number (x /. y)
| Integer x, Number y -> Number (float_of_int x /. y)
| Number x, Integer y -> Number (x /. float_of_int y)
| _ -> (Hashtbl.find Sx_primitives.primitives "/") [a; b])
| 164 (* OP_EQ *) ->
let b = pop vm and a = pop vm in
push vm (Bool (Sx_runtime._fast_eq a b))
(* Trivial scalar cases inline; everything else (Rational, Dict,
Record, Vector, ListRef, nested lists) delegates to the "="
primitive so VM equality matches CEK exactly. _fast_eq is a
stripped-down subset and must not be the source of truth here. *)
push vm (match a, b with
| Integer x, Integer y -> Bool (x = y)
| Number x, Number y -> Bool (x = y)
| Integer x, Number y -> Bool (float_of_int x = y)
| Number x, Integer y -> Bool (x = float_of_int y)
| String x, String y -> Bool (x = y)
| Bool x, Bool y -> Bool (x = y)
| Symbol x, Symbol y -> Bool (x = y)
| Keyword x, Keyword y -> Bool (x = y)
| Nil, Nil -> Bool true
| _ -> (Hashtbl.find Sx_primitives.primitives "=") [a; b])
| 165 (* OP_LT *) ->
let b = pop vm and a = pop vm in
push vm (match a, b with
@@ -1127,6 +1144,11 @@ let jit_compile_lambda (l : lambda) globals =
None
) else if _jit_is_broken_name fn_name then (
None
) else if Hashtbl.mem Sx_types.jit_excluded fn_name then (
(* Guest-declared interpret-only function (continuation-using dispatch
core). Run on the CEK; the stack VM can't escape through a CEK
continuation. See Sx_types.jit_excluded. *)
None
) else
try
_jit_compiling := true;