This repository was archived by the owner on Aug 3, 2024. It is now read-only.
Miscellaneous improvements to Convert
#1020
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This makes 'TypeFamilies3' pass.
This mirrors similar logic in 'toIfaceTypeX' which converts 'Type' into
'IfaceType'. The idea is to check the first argument of 'FunTy{} :: Type'
to see if it classified as a given/wanted in the typechecker (see 'isPredTy').
Then, we use the same logic as 'ForAllTy{} :: Type' to try to extract an
'HsQualTy'.
This makes the 'ImplicitParams' test pass again.
This makes sure that, for example, a context containing only an implicit parameter still ends up being parenthesized.
This means that the kinds do not get completely swallowed up in
signatures where they matter, like
foo :: forall proxy (a :: ()). proxy a
It also means that _all_ higher-kinded type variables now have
explicit kinds.
The idea is to hide kind signatures for type variables where the kind can be inferred from at least one use of the type variable in the given type. This seems to strike a nice balance between maintaining existing behaviour and showing something that isn't incorrect.
* filter out more 'forall's in class decls/instance decls * extract default class methods too * filter out more 'forall's in pattern synonym decls
When deciding if we want to have an explicit 'forall', one of the things to check is if the deduced type variable order would match the explicit one. Turns out that we need to roll our own function for that. One example of where this got fixed in the Prelude: fst :: forall a b. (a, b) -> a turned back into fst :: (a, b) -> a
This is especially nice for bundled pattern synonyms, where the
synonyms go back to looking (almost always) like plain constructors.
In 'BundledPatterns', this renders
pattern (:>) :: a -> Vec n a -> Vec (n + 1) a
instead of
pattern (:>) :: forall a (n :: Nat). a -> Vec n a -> Vec (n + 1) a
since the kind of `n` can be inferred from its use in `Vec n a`.
Class decls produced by 'tyThingToLHsDecl' now contain associated type defaults when appropriate.
Fixes the 'Semigroup' method: stimes :: forall b. Integral b => b -> a -> a Into stimes :: Integral b => b -> a -> a
This should make it clearer why we need to define our copies of some of these functions.
This fixes the empty forall on the 'pattern Blub'. Before: pattern Blub :: forall. () => forall x. Show x => x -> BlubType After: pattern Blub :: () => Show x => x -> BlubType
'synifyDataTyConReturnKind' was behaving badly with kind signatures involving foralls and kind tyvars. Thankfully, 'TyCon's already have 'tyConResKind' for figuring out the return kind of type constructors. Consider `data (a :: *) >< b = Empty` with `-XPolyKinds`. Here is the header we generate: Before: data a >< (b :: k) :: forall k. Type -> k -> Type After: data a >< (b :: k)
...when the other method is let through.
Misc issues: * when synifying, avoid producing empty contexts * when synifying, guess at the fixity from the name * don't skip printing type variables with kinds in hoogle
simonpj
reviewed
Feb 4, 2019
| synifyType s forallty@(ForAllTy _tv _ty) = | ||
| let (tvs, ctx, tau) = tcSplitSigmaTyPreserveSynonyms forallty | ||
| synifyType s vs funty@(FunTy t1 t2) | ||
| | isPredTy t1 = synifyForAllType s vs funty |
There was a problem hiding this comment.
NB: I'm working on a patch (Trac #16185) that will add a flag to FunTy to make this test much easier.
I don't wan to to block this change, just call to attention that wen #16185 lands we can simplify code here slightly.
There was a problem hiding this comment.
Glad to hear, that seems like a useful change!
I've subscribed to the ticket in the meantime. Thanks for the head's up.
alanz
pushed a commit
that referenced
this pull request
Mar 25, 2020
Now that Haddock is moving towards working entirely over `.hi` and `.hie` files,
all declarations and types are going to be synthesized via the `Convert` module.
In preparation for this change, here are a bunch of fixes to this module:
* Add kind annotations to type variables in `forall`'s whose kind is not `Type`,
unless the kind can be inferred from some later use of the variable. See
`implicitForAll` and `noKindTyVars` in particular if you wish to dive into this.
* Properly detect `HsQualTy` in `synifyType`. This is done by following suit with
what GHC's `toIfaceTypeX` does and checking the first argument of
`FunTy{} :: Type` to see if it classified as a given/wanted in the typechecker
(see `isPredTy`).
* Beef up the logic around figuring out when an explicit `forall` is needed. This
includes: observing if any of the type variables will need kind signatures, if the
inferred type variable order _without_ a forall will still match the one GHC
claims, and some other small things.
* Add some (not yet used) functionality for default levity polymorphic type
signatures. This functionality similar to `fprint-explicit-runtime-reps`.
Couple other smaller fixes only worth mentioning:
* Show the family result signature only when it isn't `Type`
* Fix rendering of implicit parameters in the LaTeX and Hoogle backends
* Better handling of the return kind of polykinded H98 data declarations
* Class decls produced by `tyThingToLHsDecl` now contain associated type
defaults and default method signatures when appropriate
* Filter out more `forall`'s in pattern synonyms
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Now that Haddock is moving towards working entirely over
.hiand.hiefiles,all declarations and types are going to be synthesized via the
Convertmodule.In preparation for this change, here are a bunch of fixes to this module:
Add kind annotations to type variables in
forall's whose kind is notType,unless the kind can be inferred from some later use of the variable. See
implicitForAllandnoKindTyVarsin particular if you wish to dive into this.Properly detect
HsQualTyinsynifyType. This is done by following suit withwhat GHC's
toIfaceTypeXdoes and checking the first argument ofFunTy{} :: Typeto see if it classified as a given/wanted in the typechecker(see
isPredTy).Beef up the logic around figuring out when an explicit
forallis needed. Thisincludes: observing if any of the type variables will need kind signatures, if the
inferred type variable order without a forall will still match the one GHC
claims, and some other small things.
Add some (not yet used) functionality for default levity polymorphic type
signatures. This functionality similar to
fprint-explicit-runtime-reps.Couple other smaller fixes only worth mentioning:
TypetyThingToLHsDeclnow contain associated typedefaults and default method signatures when appropriate
forall's in pattern synonyms