That is, some faces have odd sizes, so the completion display may look a
bit odd.

But it does accurately tell you what the result is doing to be, so
perhaps it's OK?  Anybody have any opinions?

diff --git a/lisp/hi-lock.el b/lisp/hi-lock.el
index b70d4a7569..18e5f73369 100644
--- a/lisp/hi-lock.el
+++ b/lisp/hi-lock.el
@@ -727,11 +727,16 @@ hi-lock-read-face-name
 			   (cdr (member last-used-face hi-lock-face-defaults))
 			   hi-lock-face-defaults))
 	 face)
-          (if (and hi-lock-auto-select-face (not current-prefix-arg))
+    (if (and hi-lock-auto-select-face (not current-prefix-arg))
 	(setq face (or (pop hi-lock--unused-faces) (car defaults)))
-      (setq face (completing-read
-		  (format-prompt "Highlight using face" (car defaults))
-		  obarray 'facep t nil 'face-name-history defaults))
+      (let ((completion-extra-properties
+             (list :annotation-function
+                   (lambda (s)
+                     (add-face-text-property
+                      0 (length s) (intern s) t s)))))
+        (setq face (completing-read
+		    (format-prompt "Highlight using face" (car defaults))
+		    obarray 'facep t nil 'face-name-history defaults)))
       ;; Update list of un-used faces.
       (setq hi-lock--unused-faces (remove face hi-lock--unused-faces))
       ;; Grow the list of defaults.


-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no