docs: cross-link the numeric specializer + op tables (jolt-nzuo)

numeric.clj dbl-spec/lng-spec and backend_scheme.clj dbl-ops/lng-ops must agree —
a spec'd op with no table entry makes emit-numeric splice a nil op string. Document
the contract on both sides. Comment-only; seed re-mints byte-identical, gate green.

The other two ideas in this bead were rejected after inspecting the code:
- collapsing inline.clj local-escapes? onto reduce-ir-children would reintroduce the
  under-reporting hazard its docstring deliberately guards against (default-true is
  load-bearing for scalar-replacement soundness).
- folding numeric recur-kinds/recur-arg-lists into one walk loses the type-env
  threading recur-kinds needs through :let; the parallel split is justified.

jolt-core/jolt/backend_scheme.clj

@@ -80,6 +80,10 @@
 ;; :num-kind :double|:long when every operand is that kind; these are the Chez
 ;; flonum/fixnum ops it lowers to — no generic dispatch, fixnums unboxed. fl?/fx?
 ;; comparisons carry the question mark; fl+/fx+ don't.
+;;
+;; CONTRACT: every op name jolt.passes.numeric/dbl-spec (resp. lng-spec) tags must
+;; have an entry here, or emit-numeric splices a nil op string into the output. Keep
+;; these tables and those specializers in sync.
 (def ^:private dbl-ops
   {"+" "fl+" "-" "fl-" "*" "fl*" "/" "fl/" "min" "flmin" "max" "flmax"
    "<" "fl<?" ">" "fl>?" "<=" "fl<=?" ">=" "fl>=?" "=" "fl=?" "==" "fl=?"})

jolt-core/jolt/passes/numeric.clj

@@ -28,6 +28,8 @@
 
 ;; result kind of a double-specialized op at this name/arity, or nil if N/A.
 ;; arithmetic -> :double; comparison -> :bool (operands specialized, result not numeric).
+;; Every op name dbl-spec / lng-spec returns non-nil for must have a Chez op in
+;; jolt.backend-scheme/dbl-ops resp. lng-ops, or emit-numeric splices a nil op.
 (defn- dbl-spec [nm n]
   (cond
     (and (>= n 1) (contains? #{"+" "-" "*" "/" "min" "max"} nm)) :double