Single-pass aser_slot for HTMX path + kernel eval timing + stable hash

Eliminated double-aser for HTMX requests: build OOB wrapper AST
(~shared:layout/oob-sx :content wrapped_ast) and aser_slot in ONE
pass — same pattern as the full-page path. Halves aser_slot calls.

Added kernel-side timing to stderr:
  [aser-slot] eval=3.6s io_flush=0.0s batched=3 result=22235 chars

Results show batch IO works (io_flush=0.0s for 3 highlight calls)
and the bottleneck is pure CEK evaluation time, not IO.

Performance after single-pass fix:
  Home: 0.7s eval (was 2.2s total)
  Reactive: 3.6s eval (was 6.8s total)
  Language: 1.1s eval (was 18.9s total — double-aser eliminated)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

hosts/ocaml/bin/dune

@@ -1,3 +1,3 @@
 (executables
  (names run_tests debug_set sx_server)
- (libraries sx))
+ (libraries sx unix))

hosts/ocaml/bin/sx_server.ml

@@ -736,15 +736,21 @@ let dispatch env cmd =
       let call = List [Symbol "aser";
                         List [Symbol "quote"; expr];
                         Env env] in
+      let t0 = Unix.gettimeofday () in
       let result = Sx_ref.eval_expr call (Env env) in
+      let t1 = Unix.gettimeofday () in
       io_batch_mode := false;
       Hashtbl.remove env.bindings "expand-components?";
       let result_str = match result with
         | String s | SxExpr s -> s
         | _ -> serialize_value result
       in
+      let n_batched = List.length !io_queue in
       (* Flush batched IO: send requests, receive responses, replace placeholders *)
       let final = flush_batched_io result_str in
+      let t2 = Unix.gettimeofday () in
+      Printf.eprintf "[aser-slot] eval=%.1fs io_flush=%.1fs batched=%d result=%d chars\n%!"
+        (t1 -. t0) (t2 -. t1) n_batched (String.length final);
       send (Printf.sprintf "(ok-raw %s)" final)
     with
     | Eval_error msg ->