datalog: dl-magic-rewrite worklist now drains across rule chains (239/239)

Real bug: the worklist used (set! queue (rest queue)) to pop the
head, which left queue bound to a fresh empty list as soon as the
last item was popped. Subsequent (append! queue ...) was a no-op
on the empty list — so when the head's rewrite generated new
(rel, adn) pairs to enqueue, they vanished. Multi-relation
programs (e.g. shortest -> path -> edge, or chained derived
relations) only had their head's rules rewritten; downstream
rules silently dropped.

Fix: use an index-based loop (idx 0 → len queue), with append!
adding to the same list. Items added after the current pointer
are picked up in subsequent iterations.

2 new regression tests:
- 4-level chain (a → r1 → r2 → r3 → r4) under magic returns 2
- shortest-path demo via magic equals dl-query (1 result)

lib/datalog/tests/magic.sx

@@ -242,6 +242,32 @@
             (= (len semi) (len magic))))
         true)
 
+      ;; Multi-relation rewrite chain: query r4 → propagate to r3,
+      ;; r2, r1, a. The worklist must process all of them; an
+      ;; earlier bug stopped after only the head pair.
+      (dl-mt-test! "magic chain through 4 rule levels"
+        (let
+          ((db (dl-program
+                 "a(1). a(2). r1(X) :- a(X). r2(X) :- r1(X).
+                  r3(X) :- r2(X). r4(X) :- r3(X).")))
+          (= 2 (len (dl-magic-query db (list (quote r4) (quote X))))))
+        true)
+
+      ;; Shortest-path demo via magic — exercises the rewriter
+      ;; against rules that mix recursive positive lits with an
+      ;; aggregate body literal.
+      (dl-mt-test! "magic on shortest-path demo"
+        (let
+          ((db (dl-program-data
+                 (quote ((edge a b 5) (edge b c 3) (edge a c 10)))
+                 dl-demo-shortest-path-rules)))
+          (let
+            ((semi (dl-query db (quote (shortest a c W))))
+              (magic (dl-magic-query db (quote (shortest a c W)))))
+            (and (= (len semi) (len magic))
+                 (= (len semi) 1))))
+        true)
+
       ;; Magic over a rule whose body contains an aggregate.
       ;; The rewriter passes aggregate body lits through unchanged
       ;; (no propagation generated for them), so semi-naive's count