hlint

# HLint [![Hackage version](https://img.shields.io/hackage/v/hlint.svg?label=Hackage)](https://hackage.haskell.org/package/hlint) [![Stackage version](https://www.stackage.org/package/hlint/badge/nightly?label=Stackage)](https://www.stackage.org/package/hlint) [![Build status](https://img.shields.io/github/actions/workflow/status/ndmitchell/hlint/ci.yml?branch=master)](https://github.com/ndmitchell/hlint/actions)

HLint is a tool for suggesting possible improvements to Haskell code. These suggestions include ideas such as using alternative functions, simplifying code and spotting redundancies. This document is structured as follows:

* [Installing and running HLint](./README.md#installing-and-running-hlint)
* [FAQ](./README.md#faq)
* [Customizing the hints](./README.md#customizing-the-hints)
* [Hacking HLint](./README.md#hacking-hlint)

### Bugs and limitations

Bugs can be reported [on the bug tracker](https://github.com/ndmitchell/hlint/issues). There are some issues that I do not intend to fix:

* HLint operates on each module at a time in isolation, as a result HLint does not know about types or which names are in scope. This decision is deliberate, allowing HLint to parallelise and be used incrementally on code that may not type-check. If fixities are required to parse the code properly, they [can be supplied](./README.md#why-doesnt-hlint-know-the-fixity-for-my-custom--operator).
* The presence of `seq` may cause some hints (i.e. eta-reduction) to change the semantics of a program.
* Some transformed programs may require [additional type signatures](https://stackoverflow.com/questions/16402942/how-can-eta-reduction-of-a-well-typed-function-result-in-a-type-error/), particularly if the transformations trigger the monomorphism restriction or involve rank-2 types. In rare cases, there might be [nowhere to write](https://stackoverflow.com/questions/19758828/eta-reduce-is-not-always-held-in-haskell) the required type signature.
* Sometimes HLint will change the code in a way that causes values to default to different types, which may change the behaviour.
* HLint assumes duplicate identical expressions within in a single expression are used at the same type.
* The `RebindableSyntax` extension can cause HLint to suggest incorrect changes.
* HLint can be configured with knowledge of C Pre Processor flags, but it can only see one conditional set of code at a time.
* HLint turns on many language extensions so it can parse more documents, occasionally some break otherwise legal syntax - e.g. `{-#INLINE foo#-}` doesn't work with `MagicHash`, `foo $bar` means something different with `TemplateHaskell`. These extensions can be disabled with `-XNoMagicHash` or `-XNoTemplateHaskell` etc.
* HLint doesn't run any custom preprocessors, e.g. [markdown-unlit](https://hackage.haskell.org/package/markdown-unlit) or [record-dot-preprocessor](https://hackage.haskell.org/package/record-dot-preprocessor), so code making use of them will usually fail to parse.
* Some