wcc/cgen: #117 review — cite #120 at cstage tuple arm (cs-accept/ww-reject divergence)

The #117 emit_slice_data TY_TUPLE arm fires for any foldable tuple row,
but wwstage's checker admits only the (str,*fn) shape and loud-rejects
the rest before cgen. Document the retained acceptance divergence at the
site (rule 7/8) with the #120 pointer. Comment-only, asm-neutral.

cmd/w6c/cgen.c

@@ -15234,7 +15234,18 @@ emit_slice_data(FILE *out, Cg *c, const char *directive, const char *name,
 		/* #117 aggregate-element arm: k tuple rows. Validate every row
 		 * first (two-pass, partial-row safe). The backing is ONE DATAW
 		 * (w6a ignores +off on DATAW), then per-row relocs at the
-		 * row's backing-relative offset. */
+		 * row's backing-relative offset.
+		 *
+		 * #120 (retained divergence): this cgen arm fires for ANY
+		 * foldable TY_TUPLE row, but wwstage's checker today admits
+		 * only the (str,*fn) const-slice shape (fold-6's need) and
+		 * loud-rejects the rest ("let: not assignable") before cgen.
+		 * So non-(str,*fn) foldable tuple const-slices (e.g.
+		 * `[](i64,i64)`, `[](str,i64)`) are cs-accept / ww-checker-
+		 * reject — an acceptance divergence (ww LOUD, never silent),
+		 * pre-train symmetric-loud, now backstopped only by ww's
+		 * checker. Tracked by #120 (#29-kin); align is a checker-layer
+		 * concern, out of this cgen fold's scope (rob NARROW). */
 		int stride = etype ? (int)etype->size : 0;
 		for (Node *e = rhs->list; e; e = e->next) {
 			Node *row = e;