cp0: repair move of non-tail into 'effect context

Kent noticed the bug, where `make-nontail` used `ignored` where it
should have used `effect`. Also, Kent points out that `make-nontail`
can skip the `$values` wrapped in `effect` contexts in unsafe mode.

original commit: edd9ba3d656f3bd712e5e235b77225f756397077

mats/cp0.ms

@@ -3042,6 +3042,13 @@
          (not (equivalent-expansion?
                (expand/optimize `(lambda (g) ,(mk `(g))))
                '(lambda (g) (g))))
+         ;; When moving into an ignored position, ensure single valued
+         ;; in safe mode:
+         (equivalent-expansion?
+          (expand/optimize `(lambda (g) ,(mk `(g)) 0))
+          (if (= 3 (optimize-level))
+              '(lambda (g) (g) 0)
+              '(lambda (g) (#3%$value (g)) 0)))
          ;; Ditto, but in a nested procedure:
          (not (equivalent-expansion?
                (expand/optimize `(lambda () (lambda (g) ,(mk `(g)))))

s/cp0.ss

@@ -853,12 +853,17 @@
 
     (define make-nontail
       (lambda (ctxt e)
-        (if (context-case ctxt
-              [(tail) (single-valued-without-inspecting-continuation? e)]
-              [(ignored) (single-valued? e)]
-              [else #t])
-            e
-            (build-primcall 3 '$value (list e)))))
+        (context-case ctxt
+          [(tail)
+           (if (single-valued-without-inspecting-continuation? e)
+               e
+               (build-primcall 3 '$value (list e)))]
+          ;; An 'effect, 'ignored, 'value, or 'test position will not
+          ;; have any attachment on the immediate continuation.
+          ;; Also, an 'ignored, 'value, or 'test position will already
+          ;; enforce a single result value
+          [(effect) (safe-single-value e)]
+          [else e])))
 
     (define result-exp
       (lambda (e)