## Changelog for ghc-typelits-knownnat-0.3

## 0.3 *May 15th 2017*

- GHC 8.2.1 support: Underlying representation for
`KnownNat`

in GHC 8.2 is `Natural`

, meaning users of this plugin will need to update their code to use `Natural`

for GHC 8.2 as well.

## 0.2.4 *April 10th 2017*

- New features:
- Derive constraints for unary functions via a
`KnownNat1`

instance; thanks to @nshepperd #11
- Use type-substituted [G]iven KnownNats (partial solve for #13)

## 0.2.3 *January 15th 2017*

- Solve normalised literal constraints, i.e.:
`KnownNat (((addrSize + 1) - (addrSize - 1))) ~ KnownNat 2`

## 0.2.2 *September 29th 2016*

- New features:
- Derive smaller constraints from larger constraints when they differ by a single variable, i.e.
`KnownNat (a + b), KnownNat b`

implies `KnownNat a`

.

## 0.2.1 *August 19th 2016*

- Fixes bugs:
- Source location of derived wanted constraints is, erroneously, always set to line 1, column 1

## 0.2 *August 17th 2016*

- New features:
- Handle
`GHC.TypeLits.-`

- Handle custom, user-defined, type-level operations
- Thanks to Gabor Greif (@ggreif): derive smaller from larger constraints, i.e.
`KnownNat (n+1)`

implies `KnownNat n`

## 0.1.2

- New features: Solve "complex" KnownNat constraints involving arbitrary type-functions, as long as there is a given KnownNat constraint for this type functions.

## 0.1.1 *August 11th 2016*

- Fixes bug: panic on a non-given KnownNat constraint variable

## 0.1 *August 10th 2016*