singletons
A framework for generating singleton types
http://www.github.com/goldfirere/singletons
Version on this page:  2.7@rev:1 
LTS Haskell 17.9:  2.7@rev:1 
Stackage Nightly 20210415:  2.7@rev:1 
Latest on Hackage:  3.0 
singletons2.7@sha256:558f88d376f30e8dabde9163a96d2ebcbe2fe533e08d20ab136eea91ad046da3,7605
Module documentation for 2.7
 Data
 Data.Singletons
 Data.Singletons.CustomStar
 Data.Singletons.Decide
 Data.Singletons.Prelude
 Data.Singletons.Prelude.Applicative
 Data.Singletons.Prelude.Base
 Data.Singletons.Prelude.Bool
 Data.Singletons.Prelude.Const
 Data.Singletons.Prelude.Either
 Data.Singletons.Prelude.Enum
 Data.Singletons.Prelude.Eq
 Data.Singletons.Prelude.Foldable
 Data.Singletons.Prelude.Function
 Data.Singletons.Prelude.Functor
 Data.Singletons.Prelude.Identity
 Data.Singletons.Prelude.IsString
 Data.Singletons.Prelude.List
 Data.Singletons.Prelude.Maybe
 Data.Singletons.Prelude.Monad
 Data.Singletons.Prelude.Monoid
 Data.Singletons.Prelude.Num
 Data.Singletons.Prelude.Ord
 Data.Singletons.Prelude.Proxy
 Data.Singletons.Prelude.Semigroup
 Data.Singletons.Prelude.Show
 Data.Singletons.Prelude.Traversable
 Data.Singletons.Prelude.Tuple
 Data.Singletons.Prelude.Void
 Data.Singletons.ShowSing
 Data.Singletons.Sigma
 Data.Singletons.SuppressUnusedWarnings
 Data.Singletons.TH
 Data.Singletons.TypeError
 Data.Singletons.TypeLits
 Data.Singletons.TypeRepTYPE
 Data.Singletons
singletons 2.7
This is the README file for the singletons
library. This file contains all the
documentation for the definitions and functions in the library.
The singletons
library was written by Richard Eisenberg (rae@cs.brynmawr.edu) and
with significant contributions by Jan Stolarek (jan.stolarek@p.lodz.pl) and
Ryan Scott (ryan.gl.scott@gmail.com). There
are two papers that describe the library. Original one, Dependently typed
programming with singletons, is available
here and will
be referenced in this documentation as the “singletons paper”. A followup
paper, Promoting Functions to Type Families in Haskell, is available
here
and will be referenced in this documentation as the
“promotion paper”.
Ryan Scott (ryan.gl.scott@gmail.com) is the active maintainer.
Purpose of the singletons
library
The library contains a definition of singleton types, which allow programmers to use dependently typed techniques to enforce rich constraints among the types in their programs. See the singletons paper for a more thorough introduction.
The package also allows promotion of termlevel functions to typelevel
equivalents and singling functions to dependently typed equivalents.
Accordingly, it exports a Prelude of promoted and singled
functions, mirroring functions and datatypes found in the Prelude
, Data.Bool
,
Data.Maybe
, Data.Either
, Data.Tuple
and Data.List
. See the promotion
paper for a more thorough introduction.
This blog series, authored by Justin Le, offers a tutorial for this library that assumes no knowledge of dependent types.
Compatibility
The singletons
library requires GHC 8.10.1 or greater. Any code that uses the
singleton generation primitives needs to enable a long list of GHC
extensions. This list includes, but is not necessarily limited to, the
following:
DataKinds
DefaultSignatures
EmptyCase
ExistentialQuantification
FlexibleContexts
FlexibleInstances
GADTs
InstanceSigs
KindSignatures
NoCUSKs
NoStarIsType
PolyKinds
RankNTypes
ScopedTypeVariables
StandaloneKindSignatures
TemplateHaskell
TypeApplications
TypeFamilies
TypeOperators
UndecidableInstances
In particular, NoStarIsType
is needed to use the *
type family from the
PNum
class because with StarIsType
enabled, GHC thinks *
is a synonym
for Type
.
You may also want to consider toggling various warning flags:
Wnoredundantconstraints
. The code thatsingletons
generates uses redundant constraints, and there seems to be no way, without a large library redesign, to avoid this.fenablethsplicewarnings
. By default, GHC does not run patternmatch coverage checker warnings on code inside of Template Haskell quotes. This is an extremely common thing to do insingletons
, so you may consider opting in to these warnings.
Modules for singleton types
Data.Singletons
exports all the basic singletons definitions. Import this
module if you are not using Template Haskell and wish only to define your
own singletons.
Data.Singletons.TH
exports all the definitions needed to use the Template
Haskell code to generate new singletons.
Data.Singletons.Prelude
reexports Data.Singletons
along with singleton
definitions for various Prelude
types. This module provides promoted and
singled equivalents of functions from the real Prelude
.
Note that not all functions from original Prelude
could be promoted or
singled.
Data.Singletons.Prelude.*
modules provide promoted and singled equivalents of
definitions found in several commonly used base
library modules, including
(but not limited to) Data.Bool
, Data.Maybe
, Data.Either
, Data.List
,
Data.Tuple
, Data.Void
and GHC.Base
. We also provide promoted and singled
versions of common type classes, including (but not limited to) Eq
, Ord
,
Show
, Enum
, and Bounded
.
Data.Singletons.Decide
exports type classes for propositional equality.
Data.Singletons.TypeLits
exports definitions for working with GHC.TypeLits
.
Functions to generate singletons
The toplevel functions used to generate promoted or singled definitions are
documented in the Data.Singletons.TH
module. The most common case is just
calling singletons
, which I’ll describe here:
singletons :: Q [Dec] > Q [Dec]
This function generates singletons from the definitions given. Because singleton generation requires promotion, this also promotes all of the definitions given to the type level.
Usage example:
$(singletons [d
data Nat = Zero  Succ Nat
pred :: Nat > Nat
pred Zero = Zero
pred (Succ n) = n
])
Definitions used to support singletons
Please refer to the singletons paper for a more indepth explanation of these definitions. Many of the definitions were developed in tandem with Iavor Diatchki.
type Sing :: k > Type
type family Sing
The type family of singleton types. A new instance of this type family is generated for every new singleton type.
class SingI a where
sing :: Sing a
A class used to pass singleton values implicitly. The sing
method produces
an explicit singleton value.
type SomeSing :: Type > Type
data SomeSing k where
SomeSing :: Sing (a :: k) > SomeSing k
The SomeSing
type wraps up an existentiallyquantified singleton. Note that
the type parameter a
does not appear in the SomeSing
type. Thus, this type
can be used when you have a singleton, but you don’t know at compile time what
it will be. SomeSing Thing
is isomorphic to Thing
.
type SingKind :: Type > Constraint
class SingKind k where
type Demote k :: *
fromSing :: Sing (a :: k) > Demote k
toSing :: Demote k > SomeSing k
This class is used to convert a singleton value back to a value in the
original, unrefined ADT. The fromSing
method converts, say, a
singleton Nat
back to an ordinary Nat
. The toSing
method produces
an existentiallyquantified singleton, wrapped up in a SomeSing
.
The Demote
associated
kindindexed type family maps the kind Nat
back to the type Nat
.
type SingInstance :: k > Type
data SingInstance a where
SingInstance :: SingI a => SingInstance a
singInstance :: Sing a > SingInstance a
Sometimes you have an explicit singleton (a Sing
) where you need an implicit
one (a dictionary for SingI
). The SingInstance
type simply wraps a SingI
dictionary, and the singInstance
function produces this dictionary from an
explicit singleton. The singInstance
function runs in constant time, using
a little magic.
Equality classes
There are two different notions of equality applicable to singletons: Boolean equality and propositional equality.

Boolean equality is implemented in the type family
(==)
(in thePEq
class) and the(%==
) method (in theSEq
class). See theData.Singletons.Prelude.Eq
module for more information. 
Propositional equality is implemented through the constraint
(~)
, the type(:~:)
, and the classSDecide
. See modulesData.Type.Equality
andData.Singletons.Decide
for more information.
Which one do you need? That depends on your application. Boolean equality has the advantage that your program can take action when two types do not equal, while propositional equality has the advantage that GHC can use the equality of types during type inference.
Instances of SEq
, SDecide
, TestEquality
, and TestCoercion
are generated
when singletons
is called on a datatype that has deriving Eq
. You can also
generate these instances directly through functions exported from
Data.Singletons.TH
.
Show
classes
Promoted and singled versions of the Show
class (PShow
and SShow
,
respectively) are provided in the Data.Singletons.Prelude.Show
module. In
addition, there is a ShowSing
constraint synonym provided in the
Data.Singletons.ShowSing
module:
type ShowSing :: Type > Constraint
type ShowSing k = (forall z. Show (Sing (z :: k))  Approximately
This facilitates the ability to write Show
instances for Sing
instances.
What distinguishes all of these Show
s? Let’s use the False
constructor as
an example. If you used the PShow Bool
instance, then the output of calling
Show_
on False
is "False"
, much like the valuelevel Show Bool
instance
(similarly for the SShow Bool
instance). However, the Show (Sing (z :: Bool))
instance (i.e., ShowSing Bool
) is intended for printing the value of the
singleton constructor SFalse
, so calling show SFalse
yields "SFalse"
.
Instance of PShow
, SShow
, and Show
(for the singleton type) are generated
when singletons
is called on a datatype that has deriving Show
. You can also
generate these instances directly through functions exported from
Data.Singletons.TH
.
A promoted and singled Show
instance is provided for Symbol
, but it is only
a crude approximation of the valuelevel Show
instance for String
. On the
value level, showing String
s escapes special characters (such as double
quotes), but implementing this requires patternmatching on character literals,
something which is currently impossible at the type level. As a consequence, the
typelevel Show
instance for Symbol
s does not do any character escaping.
Errors
The singletons
library provides two different ways to handle errors:

The
Error
type family, fromData.Singletons.TypeLits
:type Error :: a > k type family Error str where {}
This is simply an empty, closed type family, which means that it will fail to reduce regardless of its input. The typical use case is giving it a
Symbol
as an argument, so that something akin toError "This is an error message"
appears in error messages. 
The
TypeError
type family, fromData.Singletons.TypeError
. This is a dropin replacement forTypeError
fromGHC.TypeLits
which can be used at both the type level and the value level (via thetypeError
function).Unlike
Error
,TypeError
will result in an actual compiletime error message, which may be more desirable depending on the use case.
Predefined singletons
The singletons
library defines a number of singleton types and functions
by default. These include (but are not limited to):
Bool
Maybe
Either
Ordering
()
 tuples up to length 7
 lists
These are all available through Data.Singletons.Prelude
. Functions that
operate on these singletons are available from modules such as Data.Singletons.Bool
and Data.Singletons.Maybe
.
Promoting functions
Function promotion allows to generate typelevel equivalents of termlevel definitions. Almost all Haskell source constructs are supported – see the “Supported Haskell constructs” section of this README for a full list.
Promoted definitions are usually generated by calling the promote
function:
$(promote [d
data Nat = Zero  Succ Nat
pred :: Nat > Nat
pred Zero = Zero
pred (Succ n) = n
])
Every promoted function and data constructor definition comes with a set of socalled defunctionalization symbols. These are required to represent partial application at the type level. For more information, refer to the “Promotion and partial application” section below.
Users also have access to Data.Singletons.Prelude
and its submodules (e.g.,
Base
, Bool
, Either
, List
, Maybe
and Tuple
). These provide promoted
versions of function found in GHC’s base
library.
Note that GHC resolves variable names in Template Haskell quotes. You cannot then use an undefined identifier in a quote, making idioms like this not work:
type family Foo a where ...
$(promote [d ... foo x ... ])
In this example, foo
would be out of scope.
Refer to the promotion paper for more details on function promotion.
Promotion and partial application
Promoting higherorder functions proves to be surprisingly tricky. Consider this example:
$(promote [d
map :: (a > b) > [a] > [b]
map _ [] = []
map f (x:xs) = f x : map f xs
])
A naïve attempt to promote map
would be:
type Map :: (a > b) > [a] > [b]
type family Map f xs where
Map _ '[] = '[]
Map f (x:xs) = f x : Map f xs
While this compiles, it is much less useful than we would like. In particular,
common idioms like Map Id xs
will not typecheck, since GHC requires that all
invocations of type families be fully saturated. That is, the use of Id
in
Map Id xs
is rejected since it is not applied to one argument, which the
number of arguments that Id
was defined with. For more information on this
point, refer to the promotion paper.
Not having the ability to partially apply functions at the type level is rather
painful, so we do the next best thing: we defunctionalize all promoted
functions so that we can emulate partial application. For example, if one were
to promote the id
function:
$(promote [d
id :: a > a
id x = x
]
Then in addition to generating the promoted Id
type family, two
defunctionalization symbols will be generated:
type IdSym0 :: a ~> a
type IdSym0 x = x
type IdSym1 (x :: a) = Id a
In general, a function that accepts N arguments generates N+1 defunctionalization symbols when promoted.
IdSym1
is a fully saturated defunctionalization symbol and is usually only
needed when generating code through the Template Haskell machinery. IdSym0
is more interesting: it has the kind a ~> a
, which has a special arrow type
(~>)
. Defunctionalization symbols using the (~>)
kind are typelevel
constants that can be “applied” using a special Apply
type family:
type Apply :: (a ~> b) > a > b
type family Apply f x
Every defunctionalization symbol comes with a corresponding Apply
instance
(except for fully saturated defunctionalization symbols). For instance, here
is the Apply
instance for IdSym0
:
type instance Apply IdSym0 x = IdSym1 x
The (~>)
kind is used when promoting higherorder functions so that partially
applied arguments can be passed to them. For instance, here is our final attempt
at promoting map
:
type Map :: (a ~> b) > [a] > [b]
type family Map f xs where
Map _ '[] = '[]
Map f (x:xs) = Apply f x : Map f xs
Now map id xs
can be promoted to Map IdSym0 xs
, which typechecks without issue.
Defunctionalizing existing type families
The most common way to defunctionalize functions is by promoting them with the
Template Haskell machinery. One can also defunctionalize existing type families,
however, by using genDefunSymbols
. For example:
type MyTypeFamily :: Nat > Bool
type family MyTypeFamily n
$(genDefunSymbols [''MyTypeFamily])
This can be especially useful if MyTypeFamily
needs to be implemented by
hand. Be aware of the following design limitations of genDefunSymbols
:
genDefunSymbols
only works for typelevel declarations. Namely, it only works when given the names of type classes, type families, type synonyms, or data types. Attempting to pass the name of a term level function, class method, data constructor, or record selector will throw an error. Passing the name of a data type to
genDefunSymbols
will cause its data constructors to be defunctionalized but not its record selectors.  Passing the name of a type class to
genDefunSymbols
will cause the class itself to be defunctionalized, but /not/ its associated type families or methods.
Note that the limitations above reflect the current design of
genDefunSymbols
. As a result, they are subject to change in the future.
Defunctionalization and visible dependent quantification
Unlike most other parts of singletons
, which disallow visible dependent
quantification (VDQ), genDefunSymbols
has limited support for VDQ.
Consider this example:
type MyProxy :: forall (k :: Type) > k > Type
type family MyProxy k (a :: k) :: Type where
MyProxy k (a :: k) = Proxy a
$(genDefunSymbols [''MyProxy])
This will generate the following defunctionalization symbols:
type MyProxySym0 :: Type ~> k ~> Type
type MyProxySym1 :: forall (k :: Type) > k ~> Type
type MyProxySym2 k (a :: k) = MyProxy k a
Note that MyProxySym0
is a bit more general than it ought to be, since
there is no dependency between the first kind (Type
) and the second kind
(k
). But this would require the ability to write something like this:
type MyProxySym0 :: forall (k :: Type) ~> k ~> Type
This currently isn’t possible. So for the time being, the kind of
MyProxySym0
will be slightly more general, which means that under rare
circumstances, you may have to provide extra type signatures if you write
code which exploits the dependency in MyProxy
’s kind.
Classes and instances
This is best understood by example. Let’s look at a stripped down Ord
:
class Eq a => Ord a where
compare :: a > a > Ordering
(<) :: a > a > Bool
x < y = case x `compare` y of
LT > True
EQ > False
GT > False
This class gets promoted to a “kind class” thus:
class PEq a => POrd a where
type Compare (x :: a) (y :: a) :: Ordering
type (<) (x :: a) (y :: a) :: Bool
type x < y = ...  promoting `case` is yucky.
Note that default method definitions become default associated type family instances. This works out quite nicely.
We also get this singleton class:
class SEq a => SOrd a where
sCompare :: forall (x :: a) (y :: a). Sing x > Sing y > Sing (Compare x y)
(%<) :: forall (x :: a) (y :: a). Sing x > Sing y > Sing (x < y)
default (%<) :: forall (x :: a) (y :: a).
((x < y) ~ { RHS from (<) above })
=> Sing x > Sing y > Sing (x < y)
x %< y = ...  this is a bit yucky too
Note that a singled class needs to use default
signatures, because
typechecking the default body requires that the default associated type
family instance was used in the promoted class. The extra equality constraint
on the default signature asserts this fact to the type checker.
Instances work roughly similarly.
instance Ord Bool where
compare False False = EQ
compare False True = LT
compare True False = GT
compare True True = EQ
instance POrd Bool where
type Compare 'False 'False = 'EQ
type Compare 'False 'True = 'LT
type Compare 'True 'False = 'GT
type Compare 'True 'True = 'EQ
instance SOrd Bool where
sCompare :: forall (x :: a) (y :: a). Sing x > Sing y > Sing (Compare x y)
sCompare SFalse SFalse = SEQ
sCompare SFalse STrue = SLT
sCompare STrue SFalse = SGT
sCompare STrue STrue = SEQ
The only interesting bit here is the instance signature. It’s not necessary in such a simple scenario, but more complicated functions need to refer to scoped type variables, which the instance signature can bring into scope. The defaults all just work.
On names
The singletons
library has to produce new names for the new constructs it
generates. Here are some examples showing how this is done:

original datatype:
Nat
promoted kind:
Nat
singleton type:
SNat
(which is really a synonym forSing
) 
original datatype:
/\
promoted kind:
/\
singleton type:
%/\

original constructor:
Succ
promoted type:
'Succ
(you can useSucc
when unambiguous)singleton constructor:
SSucc
symbols:
SuccSym0
,SuccSym1

original constructor:
:+:
promoted type:
':+:
singleton constructor:
:%+:
symbols:
:+:@#@$
,:+:@#@$$
,:+:@#@$$$

original value:
pred
promoted type:
Pred
singleton value:
sPred
symbols:
PredSym0
,PredSym1

original value:
+
promoted type:
+
singleton value:
%+
symbols:
+@#@$
,+@#@$$
,+@#@$$$

original class:
Num
promoted class:
PNum
singleton class:
SNum

original class:
~>
promoted class:
#~>
singleton class:
%~>
Special names
There are some special cases, listed below (with asterisks* denoting special treatment):

original datatype:
[]
promoted kind:
[]
singleton type*:
SList

original constructor:
[]
promoted type:
'[]
singleton constructor*:
SNil
symbols*:
NilSym0

original constructor:
:
promoted type:
':
singleton constructor*:
SCons
symbols:
:@#@$
,:@#@$$
,:@#@$$$

original datatype:
(,)
promoted kind:
(,)
singleton type*:
STuple2

original constructor:
(,)
promoted type:
'(,)
singleton constructor*:
STuple2
symbols*:
Tuple2Sym0
,Tuple2Sym1
,Tuple2Sym2
All tuples (including the 0tuple, unit) are treated similarly.

original value:
___foo
promoted type*:
US___foo
(”US
” stands for “underscore”)singleton value*:
___sfoo
symbols*:
US___fooSym0
All functions that begin with leading underscores are treated similarly.
If desired, you can pick your own naming conventions by using the
Data.Singletons.TH.Options
module. Here is an example of how this module can
be used to prefix a singled data constructor with MyS
instead of S
:
import Control.Monad.Trans.Class
import Data.Singletons.TH
import Data.Singletons.TH.Options
import Language.Haskell.TH (Name, mkName, nameBase)
$(let myPrefix :: Name > Name
myPrefix name = mkName ("MyS" ++ nameBase name) in
withOptions defaultOptions{singledDataConName = myPrefix} $
singletons $ lift [d data T = MkT ])
Supported Haskell constructs
Full support
The following constructs are fully supported:
 variables
 tuples
 constructors
 if statements
 infix expressions and types
_
patterns aliased patterns
 lists (including list comprehensions)
do
notation sections
 undefined
 error
 class constraints (though these sometimes fail with
let
,lambda
, andcase
)  literals (for
Nat
andSymbol
), including overloaded number literals  unboxed tuples (which are treated as normal tuples)
 pattern guards
 case
 let
 lambda expressions
!
and~
patterns (silently but successfully ignored during promotion) class and instance declarations
 signatures (e.g.,
(x :: Maybe a)
) in expressions InstanceSigs
Partial support
The following constructs are partially supported:
deriving
 finite arithmetic sequences
 records
 signatures (e.g.,
(x :: Maybe a)
) in patterns  functional dependencies
 type families
See the following sections for more details.
deriving
singletons
is slightly more conservative with respect to deriving
than GHC is.
The only classes that singletons
can derive without an explicit deriving
strategy are the following stock classes:
Eq
Ord
Show
Bounded
Enum
Functor
Foldable
Traversable
To do anything more exotic, one must explicitly indicate one’s intentions by
using the DerivingStrategies
extension. singletons
fully supports the
anyclass
strategy as well as the stock
strategy (at least, for the classes
listed above). singletons
does not support the newtype
or via
strategies,
as there is no equivalent of coerce
at the type level.
Finite arithmetic sequences
singletons
has partial support for arithmetic sequences (which desugar to
methods from the Enum
class under the hood). Finite sequences (e.g.,
[0..42]) are fully supported. However, infinite sequences (e.g., [0..]),
which desugar to calls to enumFromTo
or enumFromThenTo
, are not supported,
as these would require using infinite lists at the type level.
Records
Record selectors are promoted to toplevel functions, as there is no record
syntax at the type level. Record selectors are also singled to toplevel
functions because embedding records directly into singleton data constructors
can result in surprising behavior (see
this bug report for more
details on this point). THgenerated code is not affected by this limitation
since singletons
desugars away most uses of record syntax. On the other hand,
it is not possible to write out code like
SIdentity { sRunIdentity = SIdentity STrue }
by hand.
Signatures in patterns
singletons
can promote basic pattern signatures, such as in the following
examples:
f :: forall a. a > a
f (x :: a) = (x :: a)
g :: forall a. a > a
g (x :: b) = (x :: b)  b is the same as a
What does /not/ work are more advanced uses of pattern signatures that take advantage of the fact that type variables in pattern signatures can alias other types. Here are some examples of functions that one cannot promote:

h :: a > a > a h (x :: a) (_ :: b) = x
This typechecks by virtue of the fact that
b
aliasesa
. However, the same trick does not work whenh
is promoted to a type family, as a type family would considera
andb
to be distinct type variables. 
i :: Bool > Bool i (x :: a) = x
This typechecks by virtue of the fact that
a
aliasesBool
. Again, this would not work at the type level, as a type family would considera
to be a separate type fromBool
.
Functional dependencies
Inference dependent on functional dependencies is unpredictably bad. The problem is that a use of an associated type family tied to a class with fundeps doesn’t provoke the fundep to kick in. This is GHC’s problem, in the end.
Type families
Promoting functions with types that contain type families is likely to fail due to GHC#12564. Note that promoting type family declarations is fine (and often desired, since that produces defunctionalization symbols for them).
Support for promotion, but not singling
The following constructs are supported for promotion but not singleton generation:
 data constructors with contexts
 overlapping patterns
GADTs
 instances of polykinded type classes
See the following sections for more details.
Data constructors with contexts
For example, the following datatype does not single:
data T a where
MkT :: Show a => a > T a
Constructors like these do not interact well with the current design of the
SingKind
class. But see
this bug report, which
proposes a redesign for SingKind
(in a future version of GHC with certain
bugfixes) which could permit constructors with equality constraints.
Overlapping patterns
Note that overlapping patterns are sometimes not obvious. For example, the
filter
function does not single due to overlapping patterns:
filter :: (a > Bool) > [a] > [a]
filter _pred [] = []
filter pred (x:xs)
 pred x = x : filter pred xs
 otherwise = filter pred xs
Overlap is caused by otherwise
catchall guard, which is always true and thus
overlaps with pred x
guard.
Another nonobvious source of overlapping patterns comes from partial pattern
matches in do
notation. For example:
f :: [()]
f = do
Just () < [Nothing]
return ()
This has overlap because the partial pattern match desugars to the following:
f :: [()]
f = case [Nothing] of
Just () > return ()
_ > fail "Partial pattern match in do notation"
Here, it is more evident that the catchall pattern _
overlaps with the
one above it.
GADTs
Singling GADTs is likely to fail due to the generated SingKind
instances
not typechecking. (See
#150).
However, one can often work around the issue by suppressing the generation
of SingKind
instances by using custom Options
. See the T150
test case
for an example.
Instances of polykinded type classes
Singling instances of polykinded type classes is likely to fail due to
#358.
However, one can often work around the issue by using InstanceSigs
. For
instance, the following code will not single:
class C (f :: k > Type) where
method :: f a
instance C [] where
method = []
Adding a type signature for method
in the C []
is sufficient
to work around the issue, though:
instance C [] where
method :: [a]
method = []
Little to no support
The following constructs are either unsupported or almost never work:
 scoped type variables
 datatypes that store arrows,
Nat
, orSymbol
 rankn types
 promoting
TypeRep
s TypeApplications
See the following sections for more details.
Scoped type variables
Promoting functions that rely on the behavior of ScopedTypeVariables
is very
tricky—see
this GitHub issue for an
extended discussion on the topic. This is not to say that promoting functions
that rely on ScopedTypeVariables
is guaranteed to fail, but it is rather
fragile. To demonstrate how fragile this is, note that the following function
will promote successfully:
f :: forall a. a > a
f x = id x :: a
But this one will not:
g :: forall a. a > a
g x = id (x :: a)
There are usually workarounds one can use instead of ScopedTypeVariables
:

Use pattern signatures:
g :: forall a. a > a g (x :: a) = id (x :: a)

Use local definitions:
g :: forall a. a > a g x = id' a where id' :: a > a id' x = x
Arrows, Nat
, Symbol
, and literals
As described in the promotion paper, promotion of datatypes that store arrows is currently impossible. So if you have a declaration such as
data Foo = Bar (Bool > Maybe Bool)
you will quickly run into errors.
Literals are problematic because we rely on GHC’s builtin support, which
currently is limited. Functions that operate on strings will not work because
type level strings are no longer considered lists of characters. Functions
working over integer literals can be promoted by rewriting them to use
Nat
. Since Nat
does not exist at the term level, it will only be possible to
use the promoted definition, but not the original, termlevel one.
This is the same line of reasoning that forbids the use of Nat
or Symbol
in datatype definitions. But, see this bug
report for a workaround.
Rankn types
singletons
does not support type signatures that have higherrank types.
More precisely, the only types that can be promoted or singled are
vanilla types, where a vanilla function type is a type that:

Only uses a
forall
at the top level, if used at all. That is to say, it does not contain any nested or higherrankforall
s. 
Only uses a context (e.g.,
c => ...
) at the top level, if used at all, and only after the toplevelforall
if one is present. That is to say, it does not contain any nested or higherrank contexts. 
Contains no visible dependent quantification.
Promoting TypeRep
s
The builtin Haskell promotion mechanism does not yet have a full story around
the kind *
(the kind of types that have values). Ideally, promoting some form
of TypeRep
would yield *
, but the implementation of TypeRep
would have to
be updated for this to really work out. In the meantime, users who wish to
experiment with this feature have two options:

The module
Data.Singletons.TypeRepTYPE
has all the definitions possible for making*
the promoted version ofTypeRep
, asTypeRep
is currently implemented. The singleton associated withTypeRep
has one constructor:type instance Sing @(TYPE rep) = TypeRep
(Recall that
type * = TYPE LiftedRep
.) Note that any datatypes that storeTypeRep
s will not generally work as expected; the builtin promotion mechanism will not promoteTypeRep
to*
. 
The module
Data.Singletons.CustomStar
allows the programmer to define a subset of types with which to work. See the Haddock documentation for the functionsingletonStar
for more info.
TypeApplications
singletons
currently cannot handle promoting or singling code that uses
TypeApplications
syntax, so singletons
will simply drop any visible type
applications. For example, id @Bool True
will be promoted to Id True
and
singled to sId STrue
. See
#378 for a discussion
of how singletons
may support TypeApplications
in the future.
On the other hand, singletons
does make an effort to preserve the order of
type variables when promoting and singling certain constructors. These include:
 Kind signatures of promoted toplevel functions
 Type signatures of singled toplevel functions
 Kind signatures of singled data type declarations
 Type signatures of singled data constructors
 Kind signatures of singled class declarations
 Type signatures of singled class methods
For example, consider this type signature:
const2 :: forall b a. a > b > a
The promoted version of const
will have the following kind signature:
type Const2 :: forall b a. a > b > a
The singled version of const2
will have the following type signature:
sConst2 :: forall b a (x :: a) (y :: a). Sing x > Sing y > Sing (Const x y)
Therefore, writing const2 @T1 @T2
works just as well as writing
Const2 @T1 @T2
or sConst2 @T1 @T2
, since the signatures for const2
, Const2
,
and sConst2
all begin with forall b a.
, in that order. Again, it is worth
emphasizing that the TH machinery does not support promoting or singling
const2 @T1 @T2
directly, but you can write the type applications by hand if
you so choose.
singletons
also has limited support for preserving the order of type variables
for the following constructs:

Kind signatures of defunctionalization symbols. The order of type variables is only guaranteed to be preserved if:
 The thing being defunctionalized has a standalone type (or kind) signature.
 The type (or kind) signature of the thing being defunctionalized is a vanilla type. (See the “Rankn types” section above for what “vanilla” means.)
If either of these conditions do not hold,
singletons
will fall back to a slightly different approach to generating defunctionalization symbols that does not guarantee the order of type variables. As an example, consider the following example:data T (x :: a) :: forall b. b > Type $(genDefunSymbols [''T])
The kind of
T
isforall a. a > forall b. b > Type
, which is not vanilla. Currently,singletons
will generate the following defunctionalization symbols forT
:data TSym0 :: a ~> b ~> Type data TSym1 (x :: a) :: b ~> Type
In both symbols, the kind starts with
forall a b.
rather than quantifying theb
after the visible argument of kinda
. These symbols can still be useful even with this flaw, sosingletons
permits generating them regardless. Be aware of this drawback if you try doing something similar yourself! 
Kind signatures of promoted class methods. The order of type variables will often “just work” by happy coincidence, but there are some situations where this does not happen. Consider the following class:
class C (b :: Type) where m :: forall a. a > b > a
The full type of
m
isforall b. C b => forall a. a > b > a
, which bindsb
beforea
. This order is preserved when singlingm
, but not when promotingm
. This is because theC
class is promoted as follows:class PC (b :: Type) where type M (x :: a) (y :: b) :: a
Due to the way GHC kindchecks associated type families, the kind of
M
isforall a b. a > b > a
, which bindsb
aftera
. Moreover, theStandaloneKindSignatures
extension does not provide a way to explicitly declare the full kind of an associated type family, so this limitation is not easy to work around.The defunctionalization symbols for
M
will also follow a similar order of type variables:type MSym0 :: forall a b. a ~> b ~> a type MSym1 :: forall a b. a > b ~> a
Changes
Changelog for singletons project
2.7

Require GHC 8.10.

Record selectors are now singled as toplevel functions. For instance,
$(singletons [d data T = MkT { unT :: Bool } ])
will now generate this:data ST :: T > Type where SMkT :: Sing b > Sing (MkT b) sUnT :: Sing (t :: T) > Sing (UnT t :: Bool) sUnT (SMkT sb) = sb ...
Instead of this:
data ST :: T > Type where SMkT :: { sUnT :: Sing b } > Sing (MkT b)
Note that the new type of
sUnT
is more general than the previous type (Sing (MkT b) > Sing b
).There are two primary reasons for this change:
 Singling record selectors as toplevel functions is consistent with how
promoting records works (note that
MkT
is also a toplevel function). As  Embedding record selectors directly into a singleton data constructor can result in surprising behavior. This can range from simple code using a record selector not typechecking to the inability to define multiple constructors that share the same record name.
See this GitHub issue for an extended discussion on the motivation behind this change.
 Singling record selectors as toplevel functions is consistent with how
promoting records works (note that

The Template Haskell machinery now supports finegrained configuration in the way of an
Options
data type, which lives in the newData.Singletons.TH.Options
module. BesidesOptions
, this module also contains:Options
’ record selectors. Currently, these include options to toggle generating quoted declarations, toggle generatingSingKind
instances, and configure howsingletons
generates the names of promoted or singled types. In the future, there may be additional options. A
defaultOptions
value.  An
mtl
likeOptionsMonad
class for monads that support carryingOption
s. This includesQ
, which usesdefaultOptions
if it is the top of the monad transformer stack.  An
OptionM
monad transformer that turns anyDsMonad
into anOptionsMonad
.  A
withOptions
function which allows passingOptions
to TH functions (e.g.,promote
orsingletons
). See theREADME
for a full example of how to usewithOptions
. Most TH functions are now polymorphic overOptionsMonad
instead ofDsMonad
.

singletons
now does a much better job of preserving the order of type variables in type signatures during promotion and singling. See theSupport for TypeApplications
section of theREADME
for more details.When generating typelevel declarations in particular (e.g., promoted type families or defunctionalization symbols),
singletons
will likely also generate standalone kind signatures to preserve type variable order. As a result, mostsingletons
code that uses Template Haskell will require the use of theStandaloneKindSignatures
extension (and, by extension, theNoCUSKs
extension) to work. 
singletons
now does a more much thorough job of rejecting higherrank types during promotion or singling, assingletons
cannot support them. (Previously,singletons
would sometimes accept them, often changing rank2 types to rank1 types incorrectly in the process.) 
Add the
Data.Singletons.Prelude.Proxy
module. 
Remove the promoted versions of
genericTake
,genericDrop
,genericSplitAt
,genericIndex
, andgenericReplicate
fromData.Singletons.Prelude.List
. These definitions were subtly wrong since (1) they claim to work over anyIntegral
typei
, but in practice would only work onNat
s, and (2) wouldn’t even typecheck if they were singled. 
Export
ApplyTyConAux1
,ApplyTyConAux2
, as well as the record pattern synonyms selectorapplySing2
,applySing3
, etc. fromData.Singletons
. These were unintentionally left out in previous releases. 
Export promoted and singled versions of the
getDown
record selector inData.Singletons.Prelude.Ord
. 
Fix a slew of bugs related to fixity declarations:
 Fixity declarations for data types are no longer singled, as fixity declarations do not serve any purpose for singled data type constructors, which always have exactly one argument.
singletons
now promotes fixity declarations for class names.genPromotions
/genSingletons
now also handle fixity declarations for classes, class methods, data types, and record selectors correctly.singletons
will no longer erroneously try to single fixity declarations for type synonym or type family names. A bug that caused fixity declarations for certain defunctionalization symbols not to be generated has been fixed.
promoteOnly
andsingletonsOnly
will now produce fixity declarations for values with infix names.
2.6

Require GHC 8.8.

Sing
has switched from a data family to a type family. This GitHub issue comment provides a detailed explanation for the motivation behind this change.This has a number of consequences:

Names like
SBool
,SMaybe
, etc. are no longer type synonyms for particular instantiations ofSing
but are instead the names of the singleton data types themselves. In other words, previous versions ofsingletons
would provide this:data instance Sing :: Bool > Type where SFalse :: Sing False STrue :: Sing True type SBool = (Sing :: Bool > Type)
Whereas with
Sing
asatypefamily,singletons
now provides this:data SBool :: Bool > Type where SFalse :: SBool False STrue :: SBool True type instance Sing @Bool = SBool

The
Sing
instance forTYPE rep
inData.Singletons.TypeRepTYPE
is now directly defined astype instance Sing @(TYPE rep) = TypeRep
, without the use of an intermediate newtype as before. 
Due to limitations in the ways that quantified constraints and type families can interact (see this GHC issue), the internals of
ShowSing
has to be tweaked in order to continue to work withSing
asatypefamily. One notable consequence of this is thatShow
instances for singleton types can no longer be derived—they must be written by hand in order to work around this GHC bug. This is unlikely to affect you unless you define ‘Show’ instances for singleton types by hand. For more information, refer to the Haddocks forShowSing'
inData.Singletons.ShowSing
. 
GHC does not permit type class instances to mention type families, which means that it is no longer possible to define instances that mention the
Sing
type constructor. For this reason, aWrappedSing
data type (which is a newtype aroundSing
) was introduced so that one can hang instances off of it.This had one noticeable effect in
singletons
itself: there are no longerTestEquality Sing
orTestCoercion Sing
instances. Instead,singletons
now generates a separateTestEquality
/TestCoercion
instance for every data type that singles a derivedEq
instance. In addition, theData.Singletons.Decide
module now provides topleveldecideEquality
/decideCoercion
functions which provide the behavior oftestEquality
/testCoercion
, but monomorphized toSing
. Finally,TestEquality
/TestCoercion
instances are provided forWrappedSing
.


GHC’s behavior surrounding kind inference for local definitions has changed in 8.8, and certain code that
singletons
generates for local definitions may no longer typecheck as a result. While we have taken measures to mitigate the issue onsingletons
’ end, there still exists code that must be patched on the users’ end in order to continue compiling. For instance, here is an example of code that stopped compiling with the switch to GHC 8.8:replicateM_ :: (Applicative m) => Nat > m a > m () replicateM_ cnt0 f = loop cnt0 where loop cnt  cnt <= 0 = pure ()  otherwise = f *> loop (cnt  1)
This produces errors to the effect of:
• Could not deduce (SNum k1) arising from a use of ‘sFromInteger’ from the context: SApplicative m ... • Could not deduce (SOrd k1) arising from a use of ‘%<=’ from the context: SApplicative m ...
The issue is that GHC 8.8 now kindgeneralizes
sLoop
(whereas it did not previously), explaining why the error message mentions a mysterious kind variablek1
that only appeared after kind generalization. The solution is to giveloop
an explicit type signature like so:replicateM_ :: (Applicative m) => Nat > m a > m () +replicateM_ :: forall m a. (Applicative m) => Nat > m a > m () replicateM_ cnt0 f = loop cnt0 where + loop :: Nat > m () loop cnt  cnt <= 0 = pure ()  otherwise = f *> loop (cnt  1)
This general approach should be sufficient to fix any type inference regressions that were introduced between GHC 8.6 and 8.8. If this isn’t the case, please file an issue.

Due to GHC Trac #16133 being fixed,
singletons
generated code now requires explicitly enabling theTypeApplications
extension. (The generated code was always usingTypeApplications
under the hood, but it’s only now that GHC is detecting it.) 
Data.Singletons
now defines a family ofSingI
instances forTyCon1
throughTyCon8
:instance (forall a. SingI a => SingI (f a), ...) => SingI (TyCon1 f) instance (forall a b. (SingI a, SingI b) => SingI (f a b), ...) => SingI (TyCon2 f) ...
As a result,
singletons
no longer generates instances forSingI
instances for applications ofTyCon{N}
to particular type constructors, as they have been superseded by the instances above. 
Changes to
Data.Singletons.Sigma
:
SSigma
, the singleton type forSigma
, is now defined. 
New functions
fstSigma
,sndSigma
,FstSigma
,SndSigma
,currySigma
, anduncurrySigma
have been added. AShow
instance forSigma
has also been added. 
projSigma1
has been redefined to use continuationpassing style to more closely resemble its cousinprojSigma2
. The new type signature ofprojSigma1
is:projSigma1 :: (forall (fst :: s). Sing fst > r) > Sigma s t > r
The old type signature of
projSigma1
can be found in thefstSigma
function. 
Σ
has been redefined such that it is now a partial application ofSigma
, like so:type Σ = Sigma
One benefit of this change is that one no longer needs defunctionalization symbols in order to partially apply
Σ
. As a result,ΣSym0
,ΣSym1
, andΣSym2
have been removed.


In line with corresponding changes in
base4.13
, theFail
/sFail
methods of{P,S}Monad
have been removed in favor of new{P,S}MonadFail
classes introduced in theData.Singletons.Prelude.Monad.Fail
module. These classes are also reexported fromData.Singletons.Prelude
. 
Fix a bug where expressions with explicit signatures involving function types would fail to single.

The infix names
(.)
and(!)
are no longer mapped to(:.)
and(:!)
, as GHC 8.8 learned to parse them at the type level. 
The
Enum
instance forSomeSing
now uses more efficient implementations ofenumFromTo
andenumFromThenTo
that no longer require aSingKind
constraint.
2.5.1
ShowSing
is now a type class (with a single instance) instead of a type synonym. This was changed because definingShowSing
as a type synonym prevents it from working well with recursive types due to an unfortunate GHC bug. For more information, see issue #371. Add an
IsString
instance forSomeSing
.
2.5

The
Data.Promotion.Prelude.*
namespace has been removed. Use the corresponding modules in theData.Singletons.Prelude.*
namespace instead. 
Fix a regression in which certain infix type families, such as
(++)
,($)
,(+)
, and others, did not have the correct fixities. 
The default implementation of the
(==)
type inPEq
was changed from(Data.Type.Equality.==)
to a custom type family,DefaultEq
. The reason for this change is that(Data.Type.Equality.==)
is unable to conclude thata == a
reduces toTrue
for anya
. (As a result, the previous version ofsingletons
regressed in terms of type inference for thePEq
instances forNat
andSymbol
, which used that default.) On the other hand,DefaultEq a a
does reduce toTrue
for alla
. 
Add
Enum Nat
,Show Nat
, andShow Symbol
instances toData.Singletons.TypeLits
. 
Template Haskellgenerated code may require
DataKinds
andPolyKinds
in scenarios which did not previously require it:singletons
now explicitly quantifies all kind variables used in explicitforall
s.singletons
now generatesa ~> b
instead ofTyFun a b > Type
whenever possible.

Since
thdesugar
now desugars all data types to GADT syntax, Template Haskellgenerated code may requireGADTs
in situations that didn’t require it before. 
Overhaul the way derived
Show
instances for singleton types works. Before, there was an awkwardShowSing
class (which was essentially a cargoculted version ofShow
specialized forSing
) that one had to create instances for separately. Now that GHC hasQuantifiedConstraints
, we can scrap this whole class and turnShowSing
into a simple type synonym:type ShowSing k = forall z. Show (Sing (z :: k))
Now, instead of generating a handwritten
ShowSing
andShow
instance for each singleton type, we only generate a single (derived!)Show
instance. As a result of this change, you will likely need to enableQuantifiedConstraints
andStandaloneDeriving
if you single any derivedShow
instances in your code. 
The kind of the type parameter to
SingI
is no longer specified. This only affects you if you were using thesing
method withTypeApplications
. For instance, if you were usingsing @Bool @True
before, then you will now need to now usesing @Bool
instead. 
singletons
now generatesSingI
instances for defunctionalization symbols through Template Haskell. As a result, you may need to enableFlexibleInstances
in more places. 
genDefunSymbols
is now more robust with respect to types that use dependent quantification, such as:type family MyProxy k (a :: k) :: Type where MyProxy k (a :: k) = Proxy a
See the documentation for
genDefunSymbols
for limitations to this. 
Rename
Data.Singletons.TypeRepStar
toData.Singletons.TypeRepTYPE
, and generalize theSing :: Type > Type
instance toSing :: TYPE rep > Type
, allowing it to work over more open kinds. Also renameSomeTypeRepStar
toSomeTypeRepTYPE
, and change its definition accordingly. 
Promoting or singling a type synonym or type family declaration now produces defunctionalization symbols for it. (Previously, promoting or singling a type synonym did nothing whatsoever, and promoting or singling a type family produced an error.)

singletons
now produces fixity declarations for defunctionalization symbols when appropriate. 
Add
(%<=?)
, a singled version of(<=?)
fromGHC.TypeNats
, as well as defunctionalization symbols for(<=?)
, toData.Singletons.TypeLits
. 
Add
Data.Singletons.Prelude.{Semigroup,Monoid}
, which define promoted and singled versions of theSemigroup
andMonoid
type classes, as well as various newtype modifiers.Symbol
is now has promotedSemigroup
andMonoid
instances as well. As a consequence,Data.Singletons.TypeLits
no longer exports(<>)
or(%<>)
, as they are superseded by the corresponding methods fromPSemigroup
andSSemigroup
. 
Add promoted and singled versions of the
Functor
,Foldable
,Traversable
,Applicative
,Alternative
,Monad
,MonadPlus
, andMonadZip
classes. Among other things, this grants the ability to promote or singledo
notation and list comprehensions.Data.Singletons.Prelude.List
now reexports more generalFoldable
/Traversable
functions wherever possible, just asData.List
does.

Add
Data.Singletons.Prelude.{Const,Identity}
, which define promoted and singled version of theConst
andIdentity
data types, respectively. 
Promote and single the
Down
newtype inData.Singletons.Prelude.Ord
. 
To match the
base
library, the promoted/singled versions ofcomparing
andthenCmp
are no longer exported fromData.Singletons.Prelude
. (They continue to live inData.Singletons.Prelude.Ord
.) 
Permit singling of expression and pattern signatures.

Permit promotion and singling of
InstanceSigs
. 
sError
andsUndefined
now haveHasCallStack
constraints, like their counterpartserror
andundefined
. The promoted and singled counterparts toerrorWithoutStackTrace
have also been added in case you do not want this behavior. 
Add
Data.Singletons.TypeError
, which provides a dropin replacement forGHC.TypeLits.TypeError
which can be used at both the value and typelevel.
2.4.1
 Restore the
TyCon1
,TyCon2
, etc. types. It turns out that the newTyCon
doesn’t work with kindpolymorphic tycons.
2.4

Require GHC 8.4.

Demote Nat
is nowNatural
(fromNumeric.Natural
) instead ofInteger
. In accordance with this change,Data.Singletons.TypeLits
now exposesGHC.TypeNats.natVal
(which returns aNatural
) instead ofGHC.TypeLits.natVal
(which returns anInteger
). 
The naming conventions for infix identifiers (e.g.,
(&*)
) have been overhauled.
Infix functions (that are not constructors) are no longer prepended with a colon when promoted to type families. For instance, the promoted version of
(&*)
is now called(&*)
as well, instead of(:&*)
as before.There is one exception to this rule: the
(.)
function, which is promoted as(:.)
. The reason is that one cannot write(.)
at the type level. 
Singletons for infix functions are now always prepended with
%
instead of%:
. 
Singletons for infix classes are now always prepended with
%
instead of:%
. 
Singletons for infix datatypes are now always prepended with a
%
.(Before, there was an unspoken requirement that singling an infix datatype required that name to begin with a colon, and the singleton type would begin with
:%
. But now that infix datatype names can be things like(+)
, this requirement became obsolete.)
The upshot is that most infix names can now be promoted using the same name, and singled by simply prepending the name with
%
. 

The suffix for defunctionalized names of symbolic functions (e.g.,
(+)
) has changed. Before, the promoted type name would be suffixed with some number of dollar signs (e.g.,(+$)
and(+$$)
) to indicate defunctionalization symbols. Now, the promoted type name is first suffixed with@#@
and then followed by dollar signs (e.g.,(+@#@$)
and(+@#@$$)
). Adopting this conventional eliminates naming conflicts that could arise for functions that consisted of solely$
symbols. 
The treatment of
undefined
is less magical. Before, all uses ofundefined
would be promoted toGHC.Exts.Any
and singled toundefined
. Now, there is a properUndefined
type family andsUndefined
singleton function. 
As a consequence of not promoting
undefined
toAny
, there is no need to have a specialany_
function to distinguish the function on lists. The corresponding promoted type, singleton function, and defunctionalization symbols are now namedAny
,sAny
, andAnySym{0,1,2}
. 
Rework the treatment of empty data types:
 Generated
SingKind
instances for empty data types now useEmptyCase
instead of simplyerror
ing.  Derived
PEq
instances for empty data types now returnTrue
instead ofFalse
. DerivedSEq
instances now returnTrue
instead oferror
ing.  Derived
SDecide
instances for empty data types now returnProved bottom
, wherebottom
is a divergent computation, instead oferror
ing.
 Generated

Add
Data.Singletons.Prelude.IsString
andData.Promotion.Prelude.IsString
modules.IsString.fromString
is now used when promoting or singling string literals when theXOverloadedStrings
extension is enabled (similarly to howNum.fromInteger
is currently used when promoting or singling numeric literals). 
Add
Data.Singletons.Prelude.Void
. 
Add promoted and singled versions of
div
,mod
,divMod
,quot
,rem
, andquotRem
toData.Singletons.TypeLits
that utilize the efficientDiv
andMod
type families fromGHC.TypeNats
. Also addsLog2
and defunctionalization symbols forLog2
fromGHC.TypeNats
. 
Add
(<>)
and(%<>)
, the promoted and singled versions ofAppendSymbol
fromGHC.TypeLits
. 
Add
(%^)
, the singleton version ofGHC.TypeLits.^
. 
Add
unlines
andunwords
toData.Singletons.Prelude.List
. 
Add promoted and singled versions of
Show
, includingderiving
support. 
Add a
ShowSing
class, which facilitates the ability to writeShow
instances forSing
instances. 
Permit derived
Ord
instances for empty datatypes. 
Permit standalone
deriving
declarations. 
Permit
DeriveAnyClass
(through theanyclass
keyword ofDerivingStrategies
) 
Add a valuelevel
(@@)
, which is a synonym forapplySing
. 
Add
Eq
,Ord
,Num
,Enum
, andBounded
instances forSomeSing
, which leverage theSEq
,SOrd
,SNum
,SEnum
, andSBounded
instances, respectively, for the underlyingSing
. 
Rework the
Sing (a :: *)
instance inData.Singletons.TypeRepStar
such that it now uses typeindexedTypeable
. The newSing
instance is now:newtype instance Sing :: Type > Type where STypeRep :: TypeRep a > Sing a
Accordingly, the
SingKind
instance has also been changed:instance SingKind Type where type Demote Type = SomeTypeRepStar ... data SomeTypeRepStar where SomeTypeRepStar :: forall (a :: *). !(TypeRep a) > SomeTypeRepStar
Aside from cleaning up some implementation details, this change assures that
toSing
can only be called onTypeRep
s whose kind is of kind*
. The previous implementation did not enforce this, which could lead to segfaults if used carelessly. 
Instead of
error
ing, thetoSing
implementation in theSingKind (k1 ~> k2)
instance now works as one would expect (provided the user adheres to some commonsenseSingKind
laws, which are now documented). 
Add a
demote
function, which is a convenient shorthand forfromSing sing
. 
Add a
Data.Singletons.Sigma
module with aSigma
(dependent pair) data type. 
Export defunctionalization symbols for
Demote
,SameKind,
KindOf,
(~>),
Apply, and
(@@)from
Data.Singletons`. 
Add an explicitly bidirectional pattern synonym
Sing
. Pattern matching onSing
brings aSingI ty
constraint into scope from a singletonSing ty
. 
Add an explicitly bidirectional pattern synonym
FromSing
. Pattern matching on any demoted (base) type gives us the corresponding singleton. 
Add explicitly bidirectional pattern synonyms
SLambda{2..8}
. Pattern matching on any defunctionalized singleton yields a termlevel Haskell function on singletons. 
Remove the family of
TyCon1
,TyCon2
, …, in favor of justTyCon
. GHC 8.4’s type system is powerful enough to allow this nice simplification.
2.3

Documentation clarifiation in
Data.Singletons.TypeLits
, thanks to @ivanm. 
Demote
was no longer a convenient way of callingDemoteRep
and has been removed.DemoteRep
has been renamedDemote
. 
DemoteRep
is now injective. 
Demoting a
Symbol
now givesText
. This is motivated by makingDemoteRep
injective. (IfSymbol
demoted toString
, then there would be a conflict between demoting[Char]
andSymbol
.) 
Generating singletons also now generates fixity declarations for the singletonized definitions, thanks to @intindex.

Though more an implementation detail: singletons no longer uses kindlevel proxies anywhere, thanks again to @intindex.

Support for promoting higherkinded type variables, thanks for @intindex.

Data.Singletons.TypeLits
now exports defunctionalization symbols forKnownNat
andKnownSymbol
. 
Better type inference support around constraints, as tracked in Issue #176.

Type synonym definitions are now ignored, as they should be.

Show
instances forSNat
andSSymbol
, thanks to @cumber. 
The
singFun
andunSingFun
functions no longer use proxies, preferringTypeApplications
.
2.2

With
TypeInType
, we no longer kindKProxy
. @intindex has very helpfully removed the use ofKProxy
fromsingletons
. 
Drop support for GHC 7.x.

Remove
bugInGHC
. That function was intended to work around GHC’s difficulty in detecting exhaustiveness of GADT pattern matches. GHC 8 comes with a much better exhaustiveness checker, and so this function is no longer necessary.
2.1

Require
thdesugar
>= 1.6 
Work with GHC 8. GHC 8 gives the opportunity to simplify some pieces of singletons, but these opportunities are not yet fully realized. For example, injective type families means that we no longer need
Sing
to be a data family; it could be a type family. This might drastically simplify the way functions are singletonized. But not yet! 
singletons
now outputs a few more type/kind annotations to help GHC do type inference. There may be a few more programs accepted than before. (This is the fix for #136.)
2.0.1
 Lots more functions in
Data.Singletons.Prelude.List
:filter
,find
,elemIndex
,elemIndices
,findIndex
,findIndices
,intersect
,intersectBy
,takeWhile
,dropWhile
,dropWhileEnd
,span
,break
,take
,drop
,splitAt
,group
,maximum
,minimum
,insert
,sort
,groupBy
,lookup
,partition
,sum
,product
,length
,replicate
,transpose
,(!!)
,nub
,nubBy
,unionBy
,union
,genericLength
2.0.0.2
 Fix fixity of
*
.
2.0.0.1
 Make haddock work.
2.0

Instance promotion now works properly – it was quite buggy in 1.0.

Classes and instances can now be singletonized.

Limited support for functional dependencies.

We now have promoted and singletonized versions of
Enum
, as well asBounded
. 
Deriving
Enum
is also now supported. 
Ditto for
Num
, which includes an instance forNat
, naturally. 
Promoting a literal number now uses overloaded literals at the type level, using a typelevel
FromInteger
in the typelevelNum
class. 
Better support for dealing with constraints. Some previouslyunsingletonizable functions that have constrained parameters now work.

No more orphan
Quasi
instances! 
Support for functions of arity 8 (instead of the old limit, 7).

Full support for fixity declarations.

A raft of bugfixes.

Drop support for GHC 7.8. You must have GHC 7.10.2.
1.1.2.1
Fix bug #116, thus allowing locallydeclared symbols to be used in GHC 7.10.
1.1.2
 No more GHC 7.8.2 support – you must have GHC 7.8.3.
1.1.1
Update testsuite to work with thdesugar1.5.2. No functional changes.
1.1
This is a maintenance release to support building (but not testing, due to
GHC bug #10058) with 7.10. This release also targets thdesugar1.5. Some
types changed (using thdesugar’s new DsMonad
instead of Quasi
), but
clients generally won’t need to make any changes, unless they, too, generalize
over Quasi
.
1.0
This is a complete rewrite of the package.

A much wider array of surface syntax is now accepted for promotion and singletonization, including
let
,case
, partiallyapplied functions, and anonymous functions,where
, sections, among others. 
Classes and instances can be promoted (but not singletonized).

Derivation of promoted instances for
Ord
andBounded
.
This release can be seen as a “technology preview”. More features are coming soon.
This version drops GHC 7.6 support.
0.10.0
Template Haskell names are now more hygienic. In other words, singletons
won’t try to gobble up something happened to be named Sing
in your project.
(Note that the Template Haskell names are not completely hygienic; names
generated during singleton generation can still cause conflicts.)
If a function to be promoted or singletonized is missing a type signature, that is now an error, not a warning.
Added a new external module Data.Singletons.TypeLits, which contain the singletons for GHC.TypeLits. Some convenience functions are also provided.
The extension EmptyCase
is no longer needed. This caused pain when trying
to support both GHC 7.6.3 and 7.8.
0.9.3
Fix export list of Data.Singletons.TH, again again.
Add SEq
instances for Nat
and Symbol
.
0.9.2
Fix export list of Data.Singletons.TH, again.
0.9.1
Fix export list of Data.Singletons.TH.
0.9.0
Make compatible with GHC HEAD, but HEAD reports core lint errors sometimes.
Change module structure significantly. If you want to derive your own
singletons, you should import Data.Singletons.TH
. The module
Data.Singletons
now exports functions only for the use of singletons.
New modules Data.Singletons.Bool
, ...Maybe
, ...Either
, and ...List
are just like their equivalents from Data.
, except for List
, which is
quite lacking in features.
For singleton equality, use Data.Singletons.Eq
.
For propositional singleton equality, use Data.Singletons.Decide
.
New module Data.Singletons.Prelude
is meant to mirror the Haskell Prelude,
but with singleton definitions.
Streamline representation of singletons, resulting in exponential speedup at execution. (This has not been rigorously measured, but the data structures are now exponentially smaller.)
Add internal support for TypeLits, because the TypeLits module no longer exports singleton definitions.
Add support for existential singletons, through the toSing
method of
SingKind
.
Remove the SingE
class, bundling its functionality into SingKind
.
Thus, the SingRep
synonym has also been removed.
Name change: KindIs
becomes KProxy
.
Add support for singletonizing calls to error
.
Add support for singletonizing empty data definitions.
0.8.6
Make compatible with GHC HEAD, but HEAD reports core lint errors sometimes.
0.8.5
Bug fix to make singletons compatible with GHC 7.6.1.
Added git info to cabal file.
0.8.4
Update to work with latest version of GHC (7.7.20130114).
Now use branched type family instances to allow for promotion of functions with overlapping patterns.
Permit promotion of functions with constraints by omitting constraints.
0.8.3
Update to work with latest version of GHC (7.7.20121031).
Removed use of Any to simulate kind classes; now using KindOf and OfKind from GHC.TypeLits.
Made compatible with GHC.TypeLits.
0.8.2
Added this changelog
Update to work with latest version of GHC (7.6.1). (There was a change to Template Haskell).
Moved library into Data.Singletons.
0.8.1
Update to work with latest version of GHC. (There was a change to Template Haskell).
Updated dependencies in cabal to include the newer version of TH.
0.8
Initial public release