... | @@ -64,6 +64,7 @@ data X where |
... | @@ -64,6 +64,7 @@ data X where |
|
- Make better use of the `uo_thing` field, including refactoring `noThing` away and improving term-level error messages.
|
|
- Make better use of the `uo_thing` field, including refactoring `noThing` away and improving term-level error messages.
|
|
|
|
|
|
- Simon also asks that the contents of `uo_thing` should only be `HsSyn`. This would obviate the current zonking/tidying stuff. A quick pass suggests that this will be easy to do.
|
|
- Simon also asks that the contents of `uo_thing` should only be `HsSyn`. This would obviate the current zonking/tidying stuff. A quick pass suggests that this will be easy to do.
|
|
|
|
- Also see [\#13819](https://gitlab.haskell.org//ghc/ghc/issues/13819), where the current treatment of `uo_thing` leads to an outright bug.
|
|
- Take full advantage of `TcTyCon`, getting rid of the dreaded type-checking knot. ([\#13737](https://gitlab.haskell.org//ghc/ghc/issues/13737))
|
|
- Take full advantage of `TcTyCon`, getting rid of the dreaded type-checking knot. ([\#13737](https://gitlab.haskell.org//ghc/ghc/issues/13737))
|
|
- Document why we're not worried about casts in class wanteds. (Short story: any cast should be available for rewriting, and so it will rewrite the kinds.)
|
|
- Document why we're not worried about casts in class wanteds. (Short story: any cast should be available for rewriting, and so it will rewrite the kinds.)
|
|
- Sort out `matchTypeable` (see email) [\#13333](https://gitlab.haskell.org//ghc/ghc/issues/13333)
|
|
- Sort out `matchTypeable` (see email) [\#13333](https://gitlab.haskell.org//ghc/ghc/issues/13333)
|
... | | ... | |