th-desugar/CHANGELOG and th-desugar Releases

All Versions

Latest Version

1.12

Avg Release Cycle

Latest Release

Changelog History

Page 1

v1.12 Changes
- 👌 Support GHC 9.0.
- ➕ Add support for explicit specificity. As part of this change, the way th-desugar represents type variable binders has been overhauled:
  - The DTyVarBndr data type is now parameterized by a flag type parameter:
```
data DTyVarBndr flag
  = DPlainTV Name flag
  | DKindedTV Name flag DKind
```
  This can be instantiated to Specificity (for type variable binders that can be specified or inferred) or () (for type variable binders where specificity is irrelevant). DTyVarBndrSpec and DTyVarBndrUnit are also provided as type synonyms for DTyVarBndr Specificity and DTyVarBndr (), respectively.
  - In order to interface with TyVarBndr (the TH counterpart to DTyVarBndr) in a backwards-compatible way, th-desugar now depends on the th-abstraction library.
  - The ForallVisFlag has been removed in favor of the new DForallTelescope data type, which not only distinguishes between invisible and visible foralls but also uses the correct type variable flag for invisible type variables (Specificity) and visible type variables (()).
  - The type of the dsTvb is now different on pre-9.0 versions of GHC:
```
#if __GLASGOW_HASKELL__ >= 900
dsTvb :: DsMonad q => TyVarBndr flag -> q (DTyVarBndr flag)
#else
dsTvb :: DsMonad q => flag -> TyVarBndr -> q (DTyVarBndr flag)
#endif
```
  This is unfortunately required by the fact that prior to GHC 9.0, there is no flag information stored anywhere in a TyVarBndr. If you need to use dsTvb in a backward-compatible way, L.H.TH.Desugar now provides dsTvbSpec and dsTvbUnit functions which specialise dsTvb to particular flag types:
```
dsTvbSpec :: DsMonad q => TyVarBndrSpec -> q DTyVarBndrSpec
dsTvbUnit :: DsMonad q => TyVarBndrUnit -> q DTyVarBndrUnit
```
- The type of the getRecordSelectors function has changed:
```
  -getRecordSelectors :: DsMonad q => DType -> [DCon] -> q [DLetDec]
  +getRecordSelectors :: DsMonad q =>          [DCon] -> q [DLetDec]
```
The old type signature had a DType argument whose sole purpose was to help determine which type variables were existential, as this information was used to filter out "naughty" record selectors, like the example below:
```
  data Some :: (Type -> Type) -> Type where
    MkSome :: { getSome :: f a } -> Some f
```
The old implementation of getRecordSelectors would not include getSome in the returned list, as its type f a mentions an existential type variable, a, that is not mentioned in the return type Some f. The new implementation of getRecordSelectors, on the other hand, makes no attempt to filter out naughty record selectors, so it would include getSome.

This reason for this change is ultimately because determining which type variables are existentially quantified in the context of Template Haskell is rather challenging in the general case. There are heuristics we could employ to guess which variables are existential, but we have found these heuristics difficult to predict (let alone specify). As a result, we take the slightly less correct (but much easier to explain) approach of returning all record selectors, regardless of whether they are naughty or not.
- 🚚 The conExistentialTvbs function has been removed. It was horribly buggy, especially in the presence of GADT constructors. Moreover, this function was used in the implementation of getRecordSelectors function, so bugs in conExistentialTvbs often affected the results of getRecordSelectors.
- The types of decToTH, letDecToTH, and pragmaToTH have changed:
```
  -decToTH :: DDec -> [Dec]
  +decToTH :: DDec -> Dec

  -letDecToTH :: DLetDec -> Maybe Dec
  +letDecToTH :: DLetDec -> Dec

  -pragmaToTH :: DPragma -> Maybe Pragma
  +pragmaToTH :: DPragma -> Pragma
```
The semantics of pragmaToTH have changed accordingly. Previously, pragmaToTH would return Nothing when the argument is a DPragma that is not supported on an old version of GHC, but now an error will be thrown instead. decToTH and letDecToTH, which transitively invoke pragmaToTH, have had their types updated to accommodate pragmaToTH's type change.
- The type of the substTyVarBndrs function has been simplified to avoid the needless use of continuation-passing style:
```
  -substTyVarBndrs :: Quasi q => DSubst -> [DTyVarBndr flag] -> (DSubst -> [DTyVarBndr flag] -> q a) -> q a
  +substTyVarBndrs :: Quasi q => DSubst -> [DTyVarBndr flag] -> q (DSubst, [DTyVarBndr flag])
```
v1.11 Changes
- 👌 Support GHC 8.10.
- ➕ Add support for visible dependent quantification. As part of this change, the way th-desugar represents forall and constraint types has been overhauled:
  - The existing DForallT constructor has been split into two smaller constructors:
```
 data DType
   = ...
-  | DForallT [DTyVarBndr] DCxt DType
+  | DForallT ForallVisFlag [DTyVarBndr] DType
+  | DConstrainedT DCxt DType
   | ...

+data ForallVisFlag
+  = ForallVis
+  | ForallInvis
```
  The previous design combined foralls and constraints into a single constructor, while the new design puts them in distinct constructors DForallT and DConstrainedT, respectively. The new DForallT constructor also has a ForallVisFlag field to distinguish invisible foralls (e.g., forall a. a) from visible foralls (e.g., forall a -> a).
  - The unravel function has been renamed to unravelDType and now returns (DFunArgs, DType), where DFunArgs is a data type that represents the possible arguments in a function type (see the Haddocks for DFunArgs for more details). There is also an unravelDType counterpart for Types named unravelType, complete with its own FunArgs data type.
  {D}FunArgs also have some supporting operations, including filter{D}VisFunArgs (to obtain only the visible arguments) and ravel{D}Type (to construct a function type using {D}FunArgs and a return {D}Type).
- 👌 Support standalone kind signatures by adding a DKiSigD constructor to DDec.
- ➕ Add dsReifyType, reifyTypeWithLocals_maybe, and reifyTypeWithLocals, which allow looking up the types or kinds of locally declared entities.
- 🛠 Fix a bug in which reifyFixityWithLocals would not look into local fixity declarations inside of type classes.
- 🛠 Fix a bug in which reifyFixityWithLocals would return incorrect results for classes with associated type family defaults.
v1.10 Changes
- 👌 Support GHC 8.8. Drop support for GHC 7.6.
- ➕ Add support for visible kind application, type variable foralls in RULES, and explicit foralls in type family instances. Correspondingly,
  - There is now a DAppKindT constructor in DType.
  - Previously, the DDataInstD constructor had fields of type Name and [DType]. Those have been scrapped in favor of a single field of type DType, representing the application of the data family name (which was previously the Name) to its arguments (which was previously the [DType]).
  DDataInstD also has a new field of type Maybe [DTyVarBndr] to represent its explicitly quantified type variables (if present).
  - Previously, the DTySynEqn constructor had a field of type [DType]. That has been scrapped in favor of a field of type DType, representing the application of the type family name (which DTySynEqn did not used to contain!) to its arguments (which was previously the [DType]).
  DTySynEqn also has a new field of type Maybe [DTyVarBndr] to represent its explicitly quantified type variables (if present).
  - DTySynInstD no longer has a field of type Name, as that is redundant now that each DTySynEqn contains the same Name.
  - There is now a field of type Maybe [DTyVarBndr] in the DRuleP constructor to represent bound type variables in RULES (if present).
- ➕ Add a field of type Maybe [DTyVarBndr] to DInstanceD and DStandaloneDerivD for optionally quantifying type variables explicitly. If supplied with a Just, this sweetens the instance type to use a ForallT to represent the explicit quantification. This trick is not supported for InstanceD on GHC 8.0 and for StandaloneDerivD on GHC 7.10 or 8.0, so be aware of this limitation if you supply Just for this field.
- ➕ Add support for desugaring implicit params. This does not involve any changes to the th-desugar AST, as:
  - (?x :: a) => ... is desugared to IP "x" a => ....
  - id ?x is desugared to id (ip @"x").
  - let ?x = 42 in ... is desugared to let new_x_val = 42 in bindIP @"x" new_x_val ... (where bindIP is a new utility function exported by Language.Haskell.TH.Desugar on GHC 8.0 or later).
In order to support this desugaring, the type signatures of dsLetDec and dsLetDecs now return ([DLetDec], DExp -> DExp) instead of just [DLetDec], where DExp -> DExp is the expression which binds the values of implicit params (e.g., \z -> bindIP @"x" new_x_val z) if any are bound. (If none are bound, this is simply the id function.)
- 🛠 Fix a bug in which toposortTyVarsOf would error at runtime if given types containing foralls as arguments.
- 🛠 Fix a bug in which fvDType would return incorrect results if given a type containing quantified constraints.
- 🛠 Fix a bug in which expandType would not expand type synonyms in the kinds of type variable binders in foralls.
- 🛠 Fix a bug in which getRecordSelectors would omit record selectors from GADT constructors.
- 🛠 Fix a bug in which toposortTyVarsOf would sometimes not preserve the left-to-right ordering of Names generated with qNewName.
- Locally reified class methods, data constructors, and record selectors now quantify kind variables properly.
- Desugared ADT constructors now quantify kind variables properly.
- ✂ Remove DPred, as it has become too similar to DType. This also means that the DPat constructors, which previously ended with the suffix Pa, can now use the suffix P, mirroring TH.
- The type of applyDType has changed from DType -> [DType] -> DType to DType -> [DTypeArg] -> DType, where DTypeArg is a new data type that encodes whether an argument is a normal type argument (e.g., the Int in Maybe Int) or a visible kind argument (e.g., the @Type in Proxy @Type Char). A TypeArg data type (which is like DTypeArg, but with Types/Kinds instead of DTypes/DKinds) is also provided.
A handful of utility functions for manipulating TypeArgs and DTypeArgs are also exported.
- 🆓 th-desugar functions that compute free variables (e.g., fvDType) now return an OSet, a variant of Set that remembers the order in which elements were inserted. A consequence of this change is that it fixes a bug that causes free variables to be computed in different orders depending on which unique numbers GHC happened to generate internally.
- Substition and type synonym expansion are now more efficient by avoiding the use of syb in inner loops.
v1.9 Changes
- Suppose GHC 8.6.
- ➕ Add support for DerivingVia. Correspondingly, there is now a DDerivStrategy data type.
- ➕ Add support for QuantifiedConstraints. Correspondingly, there is now a DForallPr constructor in DPred to represent quantified constraint types.
- ✂ Remove the DStarT constructor of DType in favor of DConT ''Type. Two utility functions have been added to Language.Haskell.TH.Desugar to ease this transition:
  - isTypeKindName: returns True if the argument Name is that of Type or ★ (or *, to support older GHCs).
  - typeKindName: the name of Type (on GHC 8.0 or later) or * (on older GHCs).
- th-desugar now desugars all data types to GADT syntax. The most significant API-facing changes resulting from this new design are:
  - The DDataD, DDataFamilyD, and DDataFamInstD constructors of DDec now have Maybe DKind fields that either have Just an explicit return kind (e.g., the k -> Type -> Type in data Foo :: k -> Type -> Type) or Nothing (if lacking an explicit return kind).
  - The DCon constructor previously had a field of type Maybe DType, since there was a possibility it could be a GADT (with an explicit return type) or non-GADT (without an explicit return type) constructor. Since all data types are desugared to GADTs now, this field has been changed to be simply a DType.
  - The type signature of dsCon was previously:
```
dsCon :: DsMonad q => Con -> q [DCon]
```
  However, desugaring constructors now needs more information than before, since GADT constructors have richer type signatures. Accordingly, the type of dsCon is now:
```
dsCon :: DsMonad q
      => [DTyVarBndr] -- ^ The universally quantified type variables
                      --   (used if desugaring a non-GADT constructor)
      -> DType        -- ^ The original data declaration's type
                      --   (used if desugaring a non-GADT constructor).
      -> Con -> q [DCon]
```
  The instance Desugar [Con] [DCon] has also been removed, as the previous implementation of desugar (concatMapM dsCon) no longer has enough information to work.
Some other utility functions have also been added as part of this change:
- A conExistentialTvbs function has been introduced to determine the existentially quantified type variables of a DCon. Note that this function is not 100% accurate—refer to the documentation for conExistentialTvbs for more information.
- A mkExtraDKindBinders function has been introduced to turn a data type's return kind into explicit, fresh type variable binders.
- A toposortTyVarsOf function, which finds the free variables of a list of DTypes and returns them in a well scoped list that has been sorted in reverse topological order.
  - th-desugar now desugars partial pattern matches in do-notation and list/monad comprehensions to the appropriate invocation of fail. (Previously, these were incorrectly desugared into case expressions with incomplete patterns.)
  - ➕ Add a mkDLamEFromDPats function for constructing a DLamE expression using a list of DPat arguments and a DExp body.
  - ➕ Add an unravel function for decomposing a function type into its forall'd type variables, its context, its argument types, and its result type.
  - Export a substTyVarBndrs function from Language.Haskell.TH.Desugar.Subst, which substitutes over type variable binders in a capture-avoiding fashion.
  - getDataD, dataConNameToDataName, and dataConNameToCon from Language.Haskell.TH.Desugar.Reify now look up local declarations. As a result, the contexts in their type signatures have been strengthened from Quasi to DsMonad.
  - Export a dTyVarBndrToDType function which converts a DTyVarBndr to a DType, which preserves its kind.
  - Previously, th-desugar would silently accept illegal uses of record construction with fields that did not belong to the constructor, such as Identity { notAField = "wat" }. This is now an error.
v1.8 Changes
- 👌 Support GHC 8.4.
- substTy now properly substitutes into kind signatures.
- 🔦 Expose fvDType, which computes the free variables of a DType.
- Incorporate a DDeclaredInfix field into DNormalC to indicate if it is a constructor that was declared infix.
- Implement lookupValueNameWithLocals, lookupTypeNameWithLocals, mkDataNameWithLocals, and mkTypeNameWithLocals, counterparts to lookupValueName, lookupTypeName, mkDataName, and mkTypeName which have access to local Template Haskell declarations.
- Implement reifyNameSpace to determine a Name's NameSpace.
- Export reifyFixityWithLocals from Language.Haskell.TH.Desugar.
- Export matchTy (among other goodies) from new module Language.Haskell.TH.Subst. This function matches a type template against a target.
v1.7 Changes
- 👌 Support for TH's support for TypeApplications, thanks to @RyanGlScott.
- 👌 Support for unboxed sums, thanks to @RyanGlScott.
- 👌 Support for COMPLETE pragmas.
- getRecordSelectors now requires a list of DCons as an argument. This makes it easier to return correct record selector bindings in the event that a record selector appears in multiple constructors. (See goldfirere/singletons#180 for an example of where the old behavior of getRecordSelectors went wrong.)
- 👍 Better type family expansion (expanding an open type family with variables works now).
v1.6 Changes
- Work with GHC 8, with thanks to @christiaanb for getting this change going. This means that several core datatypes have changed: partcularly, we now have DTypeFamilyHead and fixities are now reified separately from other things.
- 🔀 DKind is merged with DType.
- Generic instances for everything.
v1.5.5 Changes
- 🛠 Fix issue #34. This means that desugaring (twice) is idempotent over expressions, after the second time. That is, if you desugar an expression, sweeten it, desugar again, sweeten again, and then desugar a third time, you get the same result as when you desugared the second time. (The extra round-trip is necessary there to make the output smaller in certain common cases.)
v1.5.4 Changes
- ➕ Added expandUnsoundly
v1.5.4.1 Changes
- 🛠 Fix issue #32, concerning reification of classes with default methods.

th-desugar changelog

Desugars Template Haskell abstract syntax to a simpler format without changing semantics

Changelog History

Page 1

v1.12 Changes

v1.11 Changes

v1.10 Changes

v1.9 Changes

v1.8 Changes

v1.7 Changes

v1.6 Changes

v1.5.5 Changes

v1.5.4 Changes

v1.5.4.1 Changes

th-desugar changelog

Desugars Template Haskell abstract syntax to a simpler format without changing semantics

Changelog History Page 1

Changelog History

Page 1