Seeking Control.Lens Combinator
Hello, I'm looking for a combinator along the lines of (&&&) :: Lens' a b -> Lens' a b' -> Lens' a (b,b') I can see how it could lead to lenses that don't follow the laws, but for Lenses which are somehow independent (like _1 and _2), it works perfectly well. Is there a way in lens to specify this "independence"? Thank you, Charles Paul
On Fri, 4 Oct 2013 16:23:23 -0700, Charlie Paul
Hello,
I'm looking for a combinator along the lines of (&&&) :: Lens' a b -> Lens' a b' -> Lens' a (b,b') I can see how it could lead to lenses that don't follow the laws, but for Lenses which are somehow independent (like _1 and _2), it works perfectly well. Is there a way in lens to specify this "independence"?
If you don't mind violating the laws (ie. ensuring you access fields non-independently), then you could do this manually with a splitting/joining “isomorphism” at one end, and joining the other lenses using ‘alongside’. For example: iso (join (,)) (fst *** snd).alongside _1 _2 :: Lens (a,b) (a',b') (a,b) (a',b') Of course, this particular lens is utterly trivial (it's just id). And I'm not sure how much sense it makes to extend this pattern to more complicated cases.
Charles,
I know you specifically asked for a Control.Lens combinator and I don't have one for you, but I take the opportunity to show you how easy this is using fclabels:
tupleUp :: f :-> a
-> f :-> b
-> f :-> (a, b)
tupleUp a b = point $
(,) <$> L.fst >- a -- L.fst, L.snd are lenses from Data.Label.Base
<*> L.snd >- b
In fclabels you can use the Applicative instance (and Alternative for multi-constructor cases) to derive views like this. If you look at the structure it'll become clear how easily this scales up to different types of views.
This particular example uses total monomorphic lenses, but also works for partial lenses and polymorphic updates.
Maybe there is a similar way in Control.Lens to do this.
Sebastiaan
On Oct 5, 2013, at 1:23 AM, Charlie Paul
Hello,
I'm looking for a combinator along the lines of (&&&) :: Lens' a b -> Lens' a b' -> Lens' a (b,b') I can see how it could lead to lenses that don't follow the laws, but for Lenses which are somehow independent (like _1 and _2), it works perfectly well. Is there a way in lens to specify this "independence"?
Thank you,
Charles Paul
participants (3)
-
Charlie Paul -
Niklas Haas -
Sebastiaan Visser