Version 0.6 of the ghc-typelits-natnormalise plugin emits a b <= a constraints when it finds a solution for a constraint that contain a subtraction a - b. Without this b <= a constraint, the found solution for the equation involving a subtraction might otherwise potentially lead to unsound behaviour.
As a result this change, your package will no longer type-check if you build it with version 0.6 of the plugin. As a work-around you can add the
{-# OPTIONS_GHC -fplugin-opt GHC.TypeLits.Normalise:allow-negated-numbers #-}
pragma to the Mezzo.Model.Prim, Mezzo.Model.Harmony.Chords, Mezzo.Model.Harmony.Functional, and Mezzo.Compose.Combine module. However, I advice you to check: https://hackage.haskell.org/package/ghc-typelits-natnormalise-0.6/docs/GHC-TypeLits-Normalise.html and see if you can fix your code by adding b <= a constraints instead of that above pragma.
Version 0.6 of the
ghc-typelits-natnormaliseplugin emits ab <= aconstraints when it finds a solution for a constraint that contain a subtractiona - b. Without thisb <= aconstraint, the found solution for the equation involving a subtraction might otherwise potentially lead to unsound behaviour.As a result this change, your package will no longer type-check if you build it with version 0.6 of the plugin. As a work-around you can add the
pragma to the
Mezzo.Model.Prim,Mezzo.Model.Harmony.Chords,Mezzo.Model.Harmony.Functional, andMezzo.Compose.Combinemodule. However, I advice you to check: https://hackage.haskell.org/package/ghc-typelits-natnormalise-0.6/docs/GHC-TypeLits-Normalise.html and see if you can fix your code by addingb <= aconstraints instead of that above pragma.