diff --git a/pkg/blobserver/blobpacked/blobpacked_test.go b/pkg/blobserver/blobpacked/blobpacked_test.go
index 3d682fb4c..8fa9e5489 100644
--- a/pkg/blobserver/blobpacked/blobpacked_test.go
+++ b/pkg/blobserver/blobpacked/blobpacked_test.go
@@ -456,7 +456,8 @@ func TestStreamBlobs(t *testing.T) {
 	s.init()
 
 	all := map[blob.Ref]bool{}
-	for i := 0; i < 10; i++ {
+	const nBlobs = 10
+	for i := 0; i < nBlobs; i++ {
 		b := &test.Blob{strconv.Itoa(i)}
 		b.MustUpload(t, small)
 		all[b.BlobRef()] = true
@@ -482,6 +483,36 @@ func TestStreamBlobs(t *testing.T) {
 	if !reflect.DeepEqual(got, all) {
 		t.Errorf("Got blobs %v; want %v", got, all)
 	}
+	storagetest.TestStreamer(t, s, storagetest.WantN(nBlobs))
+}
+
+func TestStreamBlobs_Loose_Enumerate(t *testing.T) {
+	testStreamBlobsLoose(t, false)
+}
+
+func TestStreamBlobs_Loose_Streamed(t *testing.T) {
+	testStreamBlobsLoose(t, true)
+}
+
+func testStreamBlobsLoose(t *testing.T, streamed bool) {
+	var small blobserver.Storage = new(test.Fetcher)
+	if !streamed {
+		// Hide the BlobStreamer interface impl.
+		small = struct{ blobserver.Storage }{small}
+
+	}
+	s := &storage{
+		small: small,
+		large: new(test.Fetcher), // unused
+		meta:  sorted.NewMemoryKeyValue(),
+		log:   test.NewLogger(t, "blobpacked: "),
+	}
+	s.init()
+	const nBlobs = 10
+	for i := 0; i < nBlobs; i++ {
+		(&test.Blob{strconv.Itoa(i)}).MustUpload(t, small)
+	}
+	storagetest.TestStreamer(t, s, storagetest.WantN(nBlobs))
 }
 
 func TestForeachZipBlob(t *testing.T) {
diff --git a/pkg/blobserver/blobpacked/stream.go b/pkg/blobserver/blobpacked/stream.go
index 8bcc47be8..182ccc368 100644
--- a/pkg/blobserver/blobpacked/stream.go
+++ b/pkg/blobserver/blobpacked/stream.go
@@ -63,54 +63,3 @@ func (st largeBlobStreamer) StreamBlobs(ctx *context.Context, dest chan<- blobse
 	// TODO(bradfitz): implement
 	return nil
 }
-
-// TODO: move some ofthis old pre-NewMultiBlobStreamer code into
-// blobserver. in particular, transparently using enumerate for
-// BlobStreamer when the storage doesn't support it should be provided
-// by the blobserver package. inevitably others will want that.
-/*
-	enumErrc := make(chan error, 1)
-	go func() {
-		defer close(sbc)
-		enumErrc <- blobserver.EnumerateAllFrom(enumCtx, s.small, strings.TrimPrefix(contToken, "after:"), func(sb blob.SizedRef) error {
-			select {
-			case sbc <- sb:
-				return nil
-			case <-enumDone:
-				return context.ErrCanceled
-			}
-		})
-	}()
-	var sent int64
-	var lastRef blob.Ref
-	for {
-		sb, ok := <-sbc
-		if !ok {
-			break
-		}
-		opener := func() types.ReadSeekCloser {
-			return blob.NewLazyReadSeekCloser(s.small, sb.Ref)
-		}
-		select {
-		case dest <- blob.NewBlob(sb.Ref, sb.Size, opener):
-			lastRef = sb.Ref
-			sent += int64(sb.Size)
-		case <-ctx.Done():
-			return "", context.ErrCanceled
-		}
-	}
-
-	enumCtx.Cancel() // redundant if sbc was already closed, but harmless.
-	enumErr := <-enumErrc
-	if enumErr == nil {
-		return "l:", nil
-	}
-
-	// See if we didn't send anything due to enumeration errors.
-	if sent == 0 {
-		enumCtx.Cancel()
-		return "l:", enumErr
-	}
-	return "s:after:" + lastRef.String(), nil
-}
-*/