1.4.1.914
Oh, I think I caught something there, in that some operations caused the duplicated files to "vanish" - when the dupe: criteria no longer applied.
I believe that's what I saw.
Anyhow, thinking I actually liked the old behavior better, where the user had to actually affect an update, like minimally changing the search term, then immediately reverting that.
As it was, you may have been left with something "wrong", wrong in that it no longer belonged on the (dupe:) list, but by remaining it also left context as to what had happened. Now, that context just "vanishes", if you will. (And with brain cells fading
, that context, even though "wrong", can be helpful 
.)