selfhost+cstage+test: graduate *[]T indexing to slice-element type (#20)

Cstage and wwstage share the latent: check.c's N_INDEX bespoke
TY_PTR-over-TY_SLICE clause peeled the slice in `*[]T[i]` and
returned the element of the element, while wwstage's elemsizeof
had no N_TSLICE arm for the post-N_TPTR-peel elem and fell to
the 8B catch-all. Splitting leaves one stage broken on the
exact `*[]T[i]` shape the new 754 sentinel asserts byte-identical
between stages (rule 11). The companion 24B per-element copy
emit is a separate codegen wedge already pinned inline at
cmd/w6c/cgen.c:6518; out-of-scope here and noted in the fixture
header.

cmd/wcc/check.c

@@ -874,8 +874,14 @@ cexpr(Checker *c, Node *n)
 			return n->type = u->sub;
 		if (u && u->kind == TY_STR)
 			return n->type = ty_u8;
+		/* `*[N]T` auto-decays to `[N]T` indexing — drill into the
+		 * inner T so callers see the element type, matching C's
+		 * pointer-to-array semantics. `*[]T` does NOT auto-decay:
+		 * `p[i]` for `p: *[]T` yields `[]T` via the default `*U → U`
+		 * fall-through below (here U is `[]T`). Hare-faithful: a
+		 * pointer-to-slice is a 1D array of slices, not of T. */
 		if (u && u->kind == TY_PTR && u->sub &&
-		    (u->sub->kind == TY_ARRAY || u->sub->kind == TY_SLICE))
+		    u->sub->kind == TY_ARRAY)
 			return n->type = u->sub->sub;
 		/* C-style pointer indexing: p[i] → *(p+i) */
 		if (u && u->kind == TY_PTR && u->sub)