type Traversal' = Simple Traversal

type Lens' = Simple Lens

A Traversal can be used directly as a Setter or a Fold (but not as a Lens) and provides the ability to both read and update multiple fields, subject to some relatively weak Traversal laws.

These have also been known as multilenses, but they have the signature and spirit of

traverse :: Traversable f => Traversal (f a) (f b) a b

and the more evocative name suggests their application.

Most of the time the Traversal you will want to use is just traverse, but you can also pass any Lens or Iso as a Traversal, and composition of a Traversal (or Lens or Iso) with a Traversal (or Lens or Iso) using (.) forms a valid Traversal.

The laws for a Traversal t follow from the laws for Traversable as stated in "The Essence of the Iterator Pattern".

t pure ≡ pure
fmap (t f) . t g ≡ getCompose . t (Compose . fmap f . g)

One consequence of this requirement is that a Traversal needs to leave the same number of elements as a candidate for subsequent Traversal that it started with. Another testament to the strength of these laws is that the caveat expressed in section 5.5 of the "Essence of the Iterator Pattern" about exotic Traversable instances that traverse the same entry multiple times was actually already ruled out by the second law in that same paper!

A Lens is actually a lens family as described in http://comonad.com/reader/2012/mirrored-lenses/.

With great power comes great responsibility and a Lens is subject to the three common sense Lens laws:

1) You get back what you put in:

view l (set l v s)  ≡ v

2) Putting back what you got doesn't change anything:

set l (view l s) s  ≡ s

3) Setting twice is the same as setting once:

set l v' (set l v s) ≡ set l v' s

These laws are strong enough that the 4 type parameters of a Lens cannot vary fully independently. For more on how they interact, read the "Why is it a Lens Family?" section of http://comonad.com/reader/2012/mirrored-lenses/.

There are some emergent properties of these laws:

1) set l s must be injective for every s This is a consequence of law #1

2) set l must be surjective, because of law #2, which indicates that it is possible to obtain any v from some s such that set s v = s

3) Given just the first two laws you can prove a weaker form of law #3 where the values v that you are setting match:

set l v (set l v s) ≡ set l v s

Every Lens can be used directly as a Setter or Traversal.

You can also use a Lens for Getting as if it were a Fold or Getter.

Since every Lens is a valid Traversal, the Traversal laws are required of any Lens you create:

l pure ≡ pure
fmap (l f) . l g ≡ getCompose . l (Compose . fmap f . g)

type Lens s t a b = forall f. Functor f => LensLike f s t a b

View the value pointed to by a Getter, Iso or Lens or the result of folding over all the results of a Fold or Traversal that points at a monoidal value.

view . to ≡ id

>>> view (to f) a
f a

>>> view _2 (1,"hello")
"hello"

>>> view (to succ) 5
6

>>> view (_2._1) ("hello",("world","!!!"))
"world"

As view is commonly used to access the target of a Getter or obtain a monoidal summary of the targets of a Fold, It may be useful to think of it as having one of these more restricted signatures:

view ::             Getter s a     -> s -> a
view :: Monoid m => Fold s m       -> s -> m
view ::             Iso' s a       -> s -> a
view ::             Lens' s a      -> s -> a
view :: Monoid m => Traversal' s m -> s -> m

In a more general setting, such as when working with a Monad transformer stack you can use:

view :: MonadReader s m             => Getter s a     -> m a
view :: (MonadReader s m, Monoid a) => Fold s a       -> m a
view :: MonadReader s m             => Iso' s a       -> m a
view :: MonadReader s m             => Lens' s a      -> m a
view :: (MonadReader s m, Monoid a) => Traversal' s a -> m a

Use the target of a Lens, Iso, or Getter in the current state, or use a summary of a Fold or Traversal that points to a monoidal value.

>>> evalState (use _1) (a,b)
a

>>> evalState (use _1) ("hello","world")
"hello"

use :: MonadState s m             => Getter s a     -> m a
use :: (MonadState s m, Monoid r) => Fold s r       -> m r
use :: MonadState s m             => Iso' s a       -> m a
use :: MonadState s m             => Lens' s a      -> m a
use :: (MonadState s m, Monoid r) => Traversal' s r -> m r

Replace the target of a Lens or all of the targets of a Setter or Traversal with a constant value.

(<$) ≡ set mapped

>>> set _2 "hello" (1,())
(1,"hello")

>>> set mapped () [1,2,3,4]
[(),(),(),()]

Note: Attempting to set a Fold or Getter will fail at compile time with an relatively nice error message.

set :: Setter s t a b    -> b -> s -> t
set :: Iso s t a b       -> b -> s -> t
set :: Lens s t a b      -> b -> s -> t
set :: Traversal s t a b -> b -> s -> t

Modify the target of a Lens or all the targets of a Setter or Traversal with a function.

fmap ≡ over mapped
fmapDefault ≡ over traverse
sets . over ≡ id
over . sets ≡ id

Given any valid Setter l, you can also rely on the law:

over l f . over l g = over l (f . g)

e.g.

>>> over mapped f (over mapped g [a,b,c]) == over mapped (f . g) [a,b,c]
True

Another way to view over is to say that it transforms a Setter into a "semantic editor combinator".

>>> over mapped f (Just a)
Just (f a)

>>> over mapped (*10) [1,2,3]
[10,20,30]

>>> over _1 f (a,b)
(f a,b)

>>> over _1 show (10,20)
("10",20)

over :: Setter s t a b -> (a -> b) -> s -> t
over :: ASetter s t a b -> (a -> b) -> s -> t

Access the 1st field of a tuple (and possibly change its type).

>>> (1,2)^._1
1

>>> _1 .~ "hello" $ (1,2)
("hello",2)

>>> (1,2) & _1 .~ "hello"
("hello",2)

>>> _1 putStrLn ("hello","world")
hello
((),"world")

This can also be used on larger tuples as well:

>>> (1,2,3,4,5) & _1 +~ 41
(42,2,3,4,5)

_1 :: Lens (a,b) (a',b) a a'
_1 :: Lens (a,b,c) (a',b,c) a a'
_1 :: Lens (a,b,c,d) (a',b,c,d) a a'
...
_1 :: Lens (a,b,c,d,e,f,g,h,i) (a',b,c,d,e,f,g,h,i) a a'

Access the 2nd field of a tuple.

>>> _2 .~ "hello" $ (1,(),3,4)
(1,"hello",3,4)

>>> (1,2,3,4) & _2 *~ 3
(1,6,3,4)

>>> _2 print (1,2)
2
(1,())

anyOf _2 :: (s -> Bool) -> (a, s) -> Bool
traverse . _2 :: (Applicative f, Traversable t) => (a -> f b) -> t (s, a) -> f (t (s, b))
foldMapOf (traverse . _2) :: (Traversable t, Monoid m) => (s -> m) -> t (b, s) -> m

& is a reverse application operator. This provides notational convenience. Its precedence is one higher than that of the forward application operator $, which allows & to be nested in $.

>>> 5 & (+1) & show
"6"

Since: base-4.8.0.0

View the value pointed to by a Getter or Lens or the result of folding over all the results of a Fold or Traversal that points at a monoidal values.

This is the same operation as view with the arguments flipped.

The fixity and semantics are such that subsequent field accesses can be performed with (.).

>>> (a,b)^._2
b

>>> ("hello","world")^._2
"world"

>>> import Data.Complex
>>> ((0, 1 :+ 2), 3)^._1._2.to magnitude
2.23606797749979

(^.) ::             s -> Getter s a     -> a
(^.) :: Monoid m => s -> Fold s m       -> m
(^.) ::             s -> Iso' s a       -> a
(^.) ::             s -> Lens' s a      -> a
(^.) :: Monoid m => s -> Traversal' s m -> m

Replace the target of a Lens or all of the targets of a Setter or Traversal with a constant value.

This is an infix version of set, provided for consistency with (.=).

f <$ a ≡ mapped .~ f $ a

>>> (a,b,c,d) & _4 .~ e
(a,b,c,e)

>>> (42,"world") & _1 .~ "hello"
("hello","world")

>>> (a,b) & both .~ c
(c,c)

(.~) :: Setter s t a b    -> b -> s -> t
(.~) :: Iso s t a b       -> b -> s -> t
(.~) :: Lens s t a b      -> b -> s -> t
(.~) :: Traversal s t a b -> b -> s -> t

Modifies the target of a Lens or all of the targets of a Setter or Traversal with a user supplied function.

This is an infix version of over.

fmap f ≡ mapped %~ f
fmapDefault f ≡ traverse %~ f

>>> (a,b,c) & _3 %~ f
(a,b,f c)

>>> (a,b) & both %~ f
(f a,f b)

>>> _2 %~ length $ (1,"hello")
(1,5)

>>> traverse %~ f $ [a,b,c]
[f a,f b,f c]

>>> traverse %~ even $ [1,2,3]
[False,True,False]

>>> traverse.traverse %~ length $ [["hello","world"],["!!!"]]
[[5,5],[3]]

(%~) :: Setter s t a b    -> (a -> b) -> s -> t
(%~) :: Iso s t a b       -> (a -> b) -> s -> t
(%~) :: Lens s t a b      -> (a -> b) -> s -> t
(%~) :: Traversal s t a b -> (a -> b) -> s -> t

Flipped version of <$>.

(<&>) = flip fmap

Examples

>>> Just 2 <&> (+1)
Just 3

>>> [1,2,3] <&> (+1)
[2,3,4]

>>> Right 3 <&> (+1)
Right 4

Since: base-4.11.0.0

Retrieve the first value targeted by a Fold or Traversal (or Just the result from a Getter or Lens). See also firstOf and ^?, which are similar with some subtle differences (explained below).

listToMaybe . toList ≡ preview folded

preview = view . pre

Unlike ^?, this function uses a MonadReader to read the value to be focused in on. This allows one to pass the value as the last argument by using the MonadReader instance for (->) s However, it may also be used as part of some deeply nested transformer stack.

preview uses a monoidal value to obtain the result. This means that it generally has good performance, but can occasionally cause space leaks or even stack overflows on some data types. There is another function, firstOf, which avoids these issues at the cost of a slight constant performance cost and a little less flexibility.

It may be helpful to think of preview as having one of the following more specialized types:

preview :: Getter s a     -> s -> Maybe a
preview :: Fold s a       -> s -> Maybe a
preview :: Lens' s a      -> s -> Maybe a
preview :: Iso' s a       -> s -> Maybe a
preview :: Traversal' s a -> s -> Maybe a

preview :: MonadReader s m => Getter s a     -> m (Maybe a)
preview :: MonadReader s m => Fold s a       -> m (Maybe a)
preview :: MonadReader s m => Lens' s a      -> m (Maybe a)
preview :: MonadReader s m => Iso' s a       -> m (Maybe a)
preview :: MonadReader s m => Traversal' s a -> m (Maybe a)

Perform a safe head of a Fold or Traversal or retrieve Just the result from a Getter or Lens.

When using a Traversal as a partial Lens, or a Fold as a partial Getter this can be a convenient way to extract the optional value.

Note: if you get stack overflows due to this, you may want to use firstOf instead, which can deal more gracefully with heavily left-biased trees. This is because ^? works by using the First monoid, which can occasionally cause space leaks.

>>> Left 4 ^?_Left
Just 4

>>> Right 4 ^?_Left
Nothing

>>> "world" ^? ix 3
Just 'l'

>>> "world" ^? ix 20
Nothing

This operator works as an infix version of preview.

(^?) ≡ flip preview

It may be helpful to think of ^? as having one of the following more specialized types:

(^?) :: s -> Getter s a     -> Maybe a
(^?) :: s -> Fold s a       -> Maybe a
(^?) :: s -> Lens' s a      -> Maybe a
(^?) :: s -> Iso' s a       -> Maybe a
(^?) :: s -> Traversal' s a -> Maybe a

Access the 5th field of a tuple.

Access the 4th field of a tuple.

Access the 3rd field of a tuple.

A convenient infix (flipped) version of toListOf.

>>> [[1,2],[3]]^..id
[[[1,2],[3]]]
>>> [[1,2],[3]]^..traverse
[[1,2],[3]]
>>> [[1,2],[3]]^..traverse.traverse
[1,2,3]

>>> (1,2)^..both
[1,2]

toList xs ≡ xs ^.. folded
(^..) ≡ flip toListOf

(^..) :: s -> Getter s a     -> a :: s -> Fold s a       -> a :: s -> Lens' s a      -> a :: s -> Iso' s a       -> a :: s -> Traversal' s a -> a :: s -> Prism' s a     -> [a]

Retrieve the first value targeted by a Fold or Traversal (or Just the result from a Getter or Lens) into the current state.

preuse = use . pre

preuse :: MonadState s m => Getter s a     -> m (Maybe a)
preuse :: MonadState s m => Fold s a       -> m (Maybe a)
preuse :: MonadState s m => Lens' s a      -> m (Maybe a)
preuse :: MonadState s m => Iso' s a       -> m (Maybe a)
preuse :: MonadState s m => Traversal' s a -> m (Maybe a)

The kind of types with lifted values. For example Int :: Type.

CallStacks are a lightweight method of obtaining a partial call-stack at any point in the program.

A function can request its call-site with the HasCallStack constraint. For example, we can define

putStrLnWithCallStack :: HasCallStack => String -> IO ()

as a variant of putStrLn that will get its call-site and print it, along with the string given as argument. We can access the call-stack inside putStrLnWithCallStack with callStack.

>>> :{
putStrLnWithCallStack :: HasCallStack => String -> IO ()
putStrLnWithCallStack msg = do
  putStrLn msg
  putStrLn (prettyCallStack callStack)
:}

Thus, if we call putStrLnWithCallStack we will get a formatted call-stack alongside our string.

>>> putStrLnWithCallStack "hello"
hello
CallStack (from HasCallStack):
  putStrLnWithCallStack, called at <interactive>:... in interactive:Ghci...

GHC solves HasCallStack constraints in three steps:

1. If there is a CallStack in scope -- i.e. the enclosing function has a HasCallStack constraint -- GHC will append the new call-site to the existing CallStack.
2. If there is no CallStack in scope -- e.g. in the GHCi session above -- and the enclosing definition does not have an explicit type signature, GHC will infer a HasCallStack constraint for the enclosing definition (subject to the monomorphism restriction).
3. If there is no CallStack in scope and the enclosing definition has an explicit type signature, GHC will solve the HasCallStack constraint for the singleton CallStack containing just the current call-site.

CallStacks do not interact with the RTS and do not require compilation with -prof. On the other hand, as they are built up explicitly via the HasCallStack constraints, they will generally not contain as much information as the simulated call-stacks maintained by the RTS.

A CallStack is a [(String, SrcLoc)]. The String is the name of function that was called, the SrcLoc is the call-site. The list is ordered with the most recently called function at the head.

NOTE: The intrepid user may notice that HasCallStack is just an alias for an implicit parameter ?callStack :: CallStack. This is an implementation detail and should not be considered part of the CallStack API, we may decide to change the implementation in the future.

Since: base-4.8.1.0

An MVar (pronounced "em-var") is a synchronising variable, used for communication between concurrent threads. It can be thought of as a box, which may be empty or full.

The class Typeable allows a concrete representation of a type to be calculated.

Representable types of kind *. This class is derivable in GHC with the DeriveGeneric flag on.

A Generic instance must satisfy the following laws:

from . to ≡ id
to . from ≡ id