Be consistent about returning an error

It's mooted in https://github.com/fluxcd/toolkit/discussions/164 that
a distinct metric is used for not completing reconciliation, as
opposed to an unexpected error. Until that discussion has run its
course, we should just do what the other controllers do, and that's
returning an error when the controller is unable to reconcile to
completion.

This also adds a comment noting the purpose of the redundant `Requeue:
true` fields, for the avoidance of confusion later.

controllers/imagerepository_controller.go

@@ -64,8 +64,14 @@ type ImageRepositoryReconciler struct {
 func (r *ImageRepositoryReconciler) Reconcile(req ctrl.Request) (ctrl.Result, error) {
 	ctx := context.Background()
 
+	// NB: In general, if an error is returned then controller-runtime
+	// will requeue the request with back-off. In the following this
+	// is usually made explicit by _also_ returning
+	// `ctrl.Result{Requeue: true}`.
+
 	var imageRepo imagev1alpha1.ImageRepository
 	if err := r.Get(ctx, req.NamespacedName, &imageRepo); err != nil {
+		// _Might_ get requeued
 		return ctrl.Result{}, client.IgnoreNotFound(err)
 	}
 
@@ -83,7 +89,7 @@ func (r *ImageRepositoryReconciler) Reconcile(req ctrl.Request) (ctrl.Result, er
 			return ctrl.Result{Requeue: true}, err
 		}
 		log.Error(err, "Unable to parse image name", "imageName", imageRepo.Spec.Image)
-		return ctrl.Result{}, nil
+		return ctrl.Result{Requeue: true}, err
 	}
 
 	imageRepo.Status.CanonicalImageName = ref.Context().String()