hlint-2.0.9: Source code suggestions

Safe HaskellNone




WARNING: This module represents the evolving second version of the HLint API. It will be renamed to drop the "3" in the next major version.

This module provides a way to apply HLint hints. If you want to just run hlint in-process and collect the results see hlint. If you want to approximate the hlint experience with a more structured API try:

(flags, classify, hint) <- autoSettings
Right m <- parseModuleEx flags "MyFile.hs" Nothing
print $ applyHints classify hint [m]



hlint :: [String] -> IO [Idea] Source #

This function takes a list of command line arguments, and returns the given hints. To see a list of arguments type hlint --help at the console. This function writes to the stdout/stderr streams, unless --quiet is specified.

As an example:

do hints <- hlint ["src", "--ignore=Use map","--quiet"]
   when (length hints > 3) $ error "Too many hints!"

Warning: The flags provided by HLint are relatively stable, but do not have the same API stability guarantees as the rest of the strongly-typed API. Do not run this function on a your server with untrusted input.

applyHints :: [Classify] -> Hint -> [(Module_, [Comment])] -> [Idea] Source #

Given a way of classifying results, and a Hint, apply to a set of modules generating a list of Ideas. The Idea values will be ordered within a file.

Given a set of modules, it may be faster pass each to applyHints in a singleton list. When given multiple modules at once this function attempts to find hints between modules, which is slower and often pointless (by default HLint passes modules singularly, using --cross to pass all modules together).

Idea data type

data Idea Source #

An idea suggest by a Hint.





Eq Idea Source # 


(==) :: Idea -> Idea -> Bool #

(/=) :: Idea -> Idea -> Bool #

Ord Idea Source # 


compare :: Idea -> Idea -> Ordering #

(<) :: Idea -> Idea -> Bool #

(<=) :: Idea -> Idea -> Bool #

(>) :: Idea -> Idea -> Bool #

(>=) :: Idea -> Idea -> Bool #

max :: Idea -> Idea -> Idea #

min :: Idea -> Idea -> Idea #

Show Idea Source # 


showsPrec :: Int -> Idea -> ShowS #

show :: Idea -> String #

showList :: [Idea] -> ShowS #

data Severity Source #

How severe an issue is.



The issue has been explicitly ignored and will usually be hidden (pass --show on the command line to see ignored ideas).


Suggestions are things that some people may consider improvements, but some may not.


Warnings are suggestions that are nearly always a good idea to apply.


Available as a setting for the user.

data Note Source #

A note describing the impact of the replacement.



The replacement is increases laziness, for example replacing reverse (reverse x) with x makes the code lazier.


The replacement is decreases laziness, for example replacing (fst x, snd x) with x makes the code stricter.

RemovesError String

The replacement removes errors, for example replacing foldr1 (+) with sum removes an error on [], and might contain the text "on []".

ValidInstance String String

The replacement assumes standard type class lemmas, a hint with the note ValidInstance "Eq" "x" might only be valid if the x variable has a reflexive Eq instance.

Note String

An arbitrary note.


Eq Note Source # 


(==) :: Note -> Note -> Bool #

(/=) :: Note -> Note -> Bool #

Ord Note Source # 


compare :: Note -> Note -> Ordering #

(<) :: Note -> Note -> Bool #

(<=) :: Note -> Note -> Bool #

(>) :: Note -> Note -> Bool #

(>=) :: Note -> Note -> Bool #

max :: Note -> Note -> Note #

min :: Note -> Note -> Note #

Show Note Source # 


showsPrec :: Int -> Note -> ShowS #

show :: Note -> String #

showList :: [Note] -> ShowS #


data Classify Source #

How to classify an Idea. If any matching field is "" then it matches everything.




getHLintDataDir :: IO FilePath Source #

Get the Cabal configured data directory of HLint.

autoSettings :: IO (ParseFlags, [Classify], Hint) Source #

The function produces a tuple containg ParseFlags (for parseModuleEx), and Classify and Hint for applyHints. It approximates the normal HLint configuration steps, roughly:

  1. Use findSettings with readSettingsFile to find and load the HLint settings files.
  2. Use parseFlagsAddFixities and resolveHints to transform the outputs of findSettings.

If you want to do anything custom (e.g. using a different data directory, storing intermediate outputs, loading hints from a database) you are expected to copy and paste this function, then change it to your needs.

argsSettings :: [String] -> IO (ParseFlags, [Classify], Hint) Source #

A version of autoSettings which respects some of the arguments supported by HLint. If arguments unrecognised by HLint are used it will result in an error. Arguments which have no representation in the return type are silently ignored.

findSettings :: (String -> IO (FilePath, Maybe String)) -> Maybe String -> IO ([Fixity], [Classify], [Either HintBuiltin HintRule]) Source #

Given a function to load a module (typically readSettingsFile), and a module to start from (defaults to hlint.yaml) find the information from all settings files.

readSettingsFile :: Maybe FilePath -> String -> IO (FilePath, Maybe String) Source #

Given a directory (or Nothing to imply getHLintDataDir), and a module name (e.g. HLint.Default), find the settings file associated with it, returning the name of the file, and (optionally) the contents.

This function looks for all settings files starting with HLint. in the directory argument, and all other files relative to the current directory.


data HintRule Source #

A LHS ==> RHS style hint rule.




data Hint Source #

Functions to generate hints, combined using the Monoid instance.




  • hintModules :: [Setting] -> [(Scope, Module_)] -> [Idea]

    Given a list of modules (and their scope information) generate some Ideas.

  • hintModule :: [Setting] -> Scope -> Module_ -> [Idea]

    Given a single module and its scope information generate some Ideas.

  • hintDecl :: [Setting] -> Scope -> Module_ -> Decl_ -> [Idea]

    Given a declaration (with a module and scope) generate some Ideas. This function will be partially applied with one module/scope, then used on multiple Decl values.

  • hintComment :: [Setting] -> Comment -> [Idea]

    Given a comment generate some Ideas.



data Scope Source #

Data type representing the modules in scope within a module. Created with scopeCreate and queried with scopeMatch and scopeMove. Note that the mempty Scope is not equivalent to scopeCreate on an empty module, due to the implicit import of Prelude.


scopeCreate :: Module_ -> Scope Source #

Create a Scope value from a module, based on the modules imports.

scopeMatch :: (Scope, QName S) -> (Scope, QName S) -> Bool Source #

Given a two names in scopes, could they possibly refer to the same thing. This property is reflexive.

scopeMove :: (Scope, QName S) -> Scope -> QName S Source #

Given a name in a scope, and a new scope, create a name for the new scope that will refer to the same thing. If the resulting name is ambiguous, it picks a plausible candidate.


parseModuleEx :: ParseFlags -> FilePath -> Maybe String -> IO (Either ParseError (Module_, [Comment])) Source #

Parse a Haskell module. Applies the C pre processor, and uses best-guess fixity resolution if there are ambiguities. The filename - is treated as stdin. Requires some flags (often defaultParseFlags), the filename, and optionally the contents of that file.

parseFlagsAddFixities :: [Fixity] -> ParseFlags -> ParseFlags Source #

Given some fixities, add them to the existing fixities in ParseFlags.

data ParseError Source #

A parse error from parseModuleEx.




data ParseFlags Source #

Created with defaultParseFlags, used by parseModuleEx.




  • cppFlags :: CppFlags

    How the file is preprocessed (defaults to NoCpp).

  • hseFlags :: ParseMode

    How the file is parsed (defaults to all fixities in the base package and most non-conflicting extensions).

data CppFlags Source #

What C pre processor should be used.



No pre processing is done.


Lines prefixed with # are stripped.

Cpphs CpphsOptions

The cpphs library is used.