{-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE NoMonomorphismRestriction #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE TemplateHaskell #-} {-# LANGUAGE TupleSections #-} {-# LANGUAGE NamedFieldPuns #-} {-# LANGUAGE ViewPatterns #-} {-# OPTIONS_GHC -Wno-orphans #-} {-# OPTIONS_GHC -Wwarn=deprecations #-} {-# OPTIONS_GHC -fno-cse #-} {-# OPTIONS_GHC -Wno-name-shadowing #-} {-# LANGUAGE FlexibleContexts #-} -- | This module contains all the code needed to output the result which -- is either: `SAFE` or `WARNING` with some reasonable error message when -- something goes wrong. All forms of errors/exceptions should go through -- here. The idea should be to report the error, the source position that -- causes it, generate a suitable .json file and then exit. module Language.Haskell.Liquid.UX.CmdLine ( -- * Get Command Line Configuration getOpts, mkOpts, defConfig, config -- * Update Configuration With Pragma , withPragmas -- * Canonicalize Paths in Config , canonicalizePaths -- * Exit Function , exitWithResult , addErrors -- * Reporting the output of the checking , OutputResult(..) , reportResult -- * Diff check mode , diffcheck -- * Show info about this version of LiquidHaskell , printLiquidHaskellBanner ) where import Prelude hiding (error) import Control.Monad import Control.Monad.IO.Class import Data.Maybe import Data.Functor ((<&>)) import Data.Aeson (encode) import qualified Data.ByteString.Lazy.Char8 as B import Development.GitRev (gitCommitCount) import qualified Paths_liquidhaskell as Meta import System.Directory import System.Exit import System.Environment import System.Console.CmdArgs.Explicit import System.Console.CmdArgs.Implicit hiding (Loud) import qualified System.Console.CmdArgs.Verbosity as CmdArgs import System.Console.CmdArgs.Text import GitHash import Data.List (nub) import System.FilePath (isAbsolute, takeDirectory, ()) import qualified Language.Fixpoint.Types.Config as FC import Language.Fixpoint.Misc import Language.Fixpoint.Types.Names import Language.Fixpoint.Types hiding (panic, Error, Result, saveQuery) import qualified Language.Fixpoint.Types as F import Language.Fixpoint.Solver.Stats as Solver import Language.Haskell.Liquid.UX.Annotate import Language.Haskell.Liquid.UX.Config import Language.Haskell.Liquid.UX.SimpleVersion (simpleVersion) import Liquid.GHC.Misc import Language.Haskell.Liquid.Misc import Language.Haskell.Liquid.Types.PrettyPrint () import Language.Haskell.Liquid.Types hiding (typ) import qualified Language.Haskell.Liquid.UX.ACSS as ACSS import qualified Liquid.GHC.API as GHC import Language.Haskell.TH.Syntax.Compat (fromCode, toCode) import Text.PrettyPrint.HughesPJ hiding (Mode, (<>)) --------------------------------------------------------------------------------- -- Config Magic Numbers---------------------------------------------------------- --------------------------------------------------------------------------------- defaultMaxParams :: Int defaultMaxParams = 2 --------------------------------------------------------------------------------- -- Parsing Command Line---------------------------------------------------------- --------------------------------------------------------------------------------- config :: Mode (CmdArgs Config) config = cmdArgsMode $ Config { loggingVerbosity = enum [ Quiet &= name "quiet" &= help "Minimal logging verbosity" , Normal &= name "normal" &= help "Normal logging verbosity" , CmdArgs.Loud &= name "verbose" &= help "Verbose logging" ] , files = def &= typ "TARGET" &= args &= typFile , idirs = def &= typDir &= help "Paths to Spec Include Directory " , fullcheck = def &= help "Full Checking: check all binders (DEFAULT)" , diffcheck = def &= help "Incremental Checking: only check changed binders" , higherorder = def &= help "Allow higher order binders into the logic" , smtTimeout = def &= help "Timeout of smt queries in msec" , higherorderqs = def &= help "Allow higher order qualifiers to get automatically instantiated" , linear = def &= help "Use uninterpreted integer multiplication and division" , stringTheory = def &= help "Interpretation of Strings by z3" , saveQuery = def &= help "Save fixpoint query to file (slow)" , checks = def &= help "Check a specific (top-level) binder" &= name "check-var" , pruneUnsorted = def &= help "Disable prunning unsorted Predicates" &= name "prune-unsorted" , notermination = def &= help "Disable Termination Check" &= name "no-termination-check" , nopositivity = def &= help "Disable Data Type Positivity Check" &= name "no-positivity-check" , rankNTypes = def &= help "Adds precise reasoning on presence of rankNTypes" &= name "rankNTypes" , noclasscheck = def &= help "Disable Class Instance Check" &= name "no-class-check" , nostructuralterm = def &= name "no-structural-termination" &= help "Disable structural termination check" , gradual = def &= help "Enable gradual refinement type checking" &= name "gradual" , bscope = def &= help "scope of the outer binders on the inner refinements" &= name "bscope" , gdepth = 1 &= help "Size of gradual conretizations, 1 by default" &= name "gradual-depth" , ginteractive = def &= help "Interactive Gradual Solving" &= name "ginteractive" , totalHaskell = def &= help "Check for termination and totality; overrides no-termination flags" &= name "total-Haskell" , nowarnings = def &= help "Don't display warnings, only show errors" &= name "no-warnings" , noannotations = def &= help "Don't create intermediate annotation files" &= name "no-annotations" , checkDerived = def &= help "Check GHC generated binders (e.g. Read, Show instances)" &= name "check-derived" , caseExpandDepth = 2 &= help "Maximum depth at which to expand DEFAULT in case-of (default=2)" &= name "max-case-expand" , notruetypes = def &= help "Disable Trueing Top Level Types" &= name "no-true-types" , nototality = def &= help "Disable totality check" &= name "no-totality" , cores = def &= help "Use m cores to solve logical constraints" , minPartSize = FC.defaultMinPartSize &= help "If solving on multiple cores, ensure that partitions are of at least m size" , maxPartSize = FC.defaultMaxPartSize &= help ("If solving on multiple cores, once there are as many partitions " ++ "as there are cores, don't merge partitions if they will exceed this " ++ "size. Overrides the minpartsize option.") , smtsolver = def &= help "Name of SMT-Solver" , noCheckUnknown = def &= explicit &= name "no-check-unknown" &= help "Don't complain about specifications for unexported and unused values " , maxParams = defaultMaxParams &= help "Restrict qualifier mining to those taking at most `m' parameters (2 by default)" , shortNames = def &= name "short-names" &= help "Print shortened names, i.e. drop all module qualifiers." , shortErrors = def &= name "short-errors" &= help "Don't show long error messages, just line numbers." , cabalDir = def &= name "cabal-dir" &= help "Find and use .cabal to add paths to sources for imported files" , ghcOptions = def &= name "ghc-option" &= typ "OPTION" &= help "Pass this option to GHC" , cFiles = def &= name "c-files" &= typ "OPTION" &= help "Tell GHC to compile and link against these files" , port = defaultPort &= name "port" &= help "Port at which lhi should listen" , exactDC = def &= help "Exact Type for Data Constructors" &= name "exact-data-cons" , noADT = def &= help "Do not generate ADT representations in refinement logic" &= name "no-adt" , expectErrorContaining = def &= help "Expect an error which containing the provided string from verification (can be provided more than once)" &= name "expect-error-containing" , expectAnyError = def &= help "Expect an error, no matter which kind or what it contains" &= name "expect-any-error" , scrapeImports = False &= help "Scrape qualifiers from imported specifications" &= name "scrape-imports" &= explicit , scrapeInternals = False &= help "Scrape qualifiers from auto generated specifications" &= name "scrape-internals" &= explicit , scrapeUsedImports = False &= help "Scrape qualifiers from used, imported specifications" &= name "scrape-used-imports" &= explicit , elimStats = False &= name "elimStats" &= help "Print eliminate stats" , elimBound = Nothing &= name "elimBound" &= help "Maximum chain length for eliminating KVars" , noslice = False &= name "noSlice" &= help "Disable non-concrete KVar slicing" , noLiftedImport = False &= name "no-lifted-imports" &= help "Disable loading lifted specifications (for legacy libs)" , json = False &= name "json" &= help "Print results in JSON (for editor integration)" , counterExamples = False &= name "counter-examples" &= help "Attempt to generate counter-examples to type errors (experimental!)" , timeBinds = False &= name "time-binds" &= help "Solve each (top-level) asserted type signature separately & time solving." , untidyCore = False &= name "untidy-core" &= help "Print fully qualified identifier names in verbose mode" , eliminate = FC.Some &= name "eliminate" &= help "Use elimination for 'all' (use TRUE for cut-kvars), 'some' (use quals for cut-kvars) or 'none' (use quals for all kvars)." , noPatternInline = False &= name "no-pattern-inline" &= help "Don't inline special patterns (e.g. `>>=` and `return`) during constraint generation." , noSimplifyCore = False &= name "no-simplify-core" &= help "Don't simplify GHC core before constraint generation" -- PLE-OPT , autoInstantiate -- PLE-OPT = def -- PLE-OPT &= help "How to instantiate axiomatized functions `smtinstances` for SMT instantiation, `liquidinstances` for terminating instantiation" -- PLE-OPT &= name "automatic-instances" , proofLogicEval = def &= help "Enable Proof-by-Logical-Evaluation" &= name "ple" , pleWithUndecidedGuards = def &= help "Unfold invocations with undecided guards in PLE" &= name "ple-with-undecided-guards" &= explicit , oldPLE = def &= help "Enable Proof-by-Logical-Evaluation" &= name "oldple" , interpreter = def &= help "Use an interpreter to assist PLE in solving constraints" &= name "interpreter" , proofLogicEvalLocal = def &= help "Enable Proof-by-Logical-Evaluation locally, per function" &= name "ple-local" , extensionality = def &= help "Enable extensional interpretation of function equality" &= name "extensionality" , nopolyinfer = def &= help "No inference of polymorphic type application. Gives imprecision, but speedup." &= name "fast" , reflection = def &= help "Enable reflection of Haskell functions and theorem proving" &= name "reflection" , compileSpec = def &= name "compile-spec" &= help "Only compile specifications (into .bspec file); skip verification" , noCheckImports = def &= name "no-check-imports" &= help "Do not check the transitive imports; only check the target files." , typedHoles = def &= name "typed-holes" &= help "Use (refinement) typed-holes [currently warns on '_x' variables]" , typeclass = def &= help "Enable Typeclass" &= name "typeclass" , auxInline = def &= help "Enable inlining of class methods" &= name "aux-inline" , maxMatchDepth = def &= name "max-match-depth" &= help "Define the number of expressions to pattern match on (typed-holes must be on to use this flag)." , maxAppDepth = def &= name "max-app-depth" , maxArgsDepth = def &= name "max-args-depth" , rwTerminationCheck = def &= name "rw-termination-check" &= help ( "Enable the rewrite divergence checker. " ++ "Can speed up verification if rewriting terminates, but can also cause divergence." ) , skipModule = def &= name "skip-module" &= help "Completely skip this module, don't even compile any specifications in it." , noLazyPLE = def &= name "no-lazy-ple" &= help "Don't use Lazy PLE" , fuel = Nothing &= help "Maximum fuel (per-function unfoldings) for PLE" , environmentReduction = def &= explicit &= name "environment-reduction" &= help "perform environment reduction (disabled by default)" , noEnvironmentReduction = def &= explicit &= name "no-environment-reduction" &= help "Don't perform environment reduction" , inlineANFBindings = False &= explicit &= name "inline-anf-bindings" &= help (unwords [ "Inline ANF bindings." , "Sometimes improves performance and sometimes worsens it." , "Disabled by --no-environment-reduction" ]) , pandocHtml = False &= name "pandoc-html" &= help "Use pandoc to generate html." } &= program "liquid" &= help "Refinement Types for Haskell" &= summary copyright &= details [ "LiquidHaskell is a Refinement Type based verifier for Haskell" , "" , "To check a Haskell file foo.hs, type:" , " liquid foo.hs " ] defaultPort :: Int defaultPort = 3000 getOpts :: [String] -> IO Config getOpts as = do cfg0 <- envCfg cfg1 <- mkOpts =<< cmdArgsRun' config { modeValue = (modeValue config) { cmdArgsValue = cfg0 } } as cfg <- fixConfig cfg1 setVerbosity (loggingVerbosity cfg) when (json cfg) $ setVerbosity Quiet withSmtSolver cfg -- | Shows the LiquidHaskell banner, that includes things like the copyright, the -- git commit and the version. printLiquidHaskellBanner :: IO () printLiquidHaskellBanner = whenNormal $ putStrLn copyright cmdArgsRun' :: Mode (CmdArgs a) -> [String] -> IO a cmdArgsRun' md as = case parseResult of Left e -> putStrLn (helpMsg e) >> exitFailure Right a -> cmdArgsApply a where helpMsg e = showText defaultWrap $ helpText [e] HelpFormatDefault md parseResult = process md (wideHelp as) wideHelp = map (\a -> if a == "--help" || a == "-help" then "--help=120" else a) -------------------------------------------------------------------------------- withSmtSolver :: Config -> IO Config -------------------------------------------------------------------------------- withSmtSolver cfg = case smtsolver cfg of Just smt -> do found <- findSmtSolver smt case found of Just _ -> return cfg Nothing -> panic Nothing (missingSmtError smt) Nothing -> do smts <- mapM findSmtSolver [FC.Z3, FC.Cvc4, FC.Mathsat] case catMaybes smts of (s:_) -> return (cfg {smtsolver = Just s}) _ -> panic Nothing noSmtError where noSmtError = "LiquidHaskell requires an SMT Solver, i.e. z3, cvc4, or mathsat to be installed." missingSmtError smt = "Could not find SMT solver '" ++ show smt ++ "'. Is it on your PATH?" findSmtSolver :: FC.SMTSolver -> IO (Maybe FC.SMTSolver) findSmtSolver smt = maybe Nothing (const $ Just smt) <$> findExecutable (show smt) fixConfig :: Config -> IO Config fixConfig cfg = do pwd <- getCurrentDirectory cfg <- canonicalizePaths pwd cfg return $ canonConfig cfg -- | Attempt to canonicalize all `FilePath's in the `Config' so we don't have -- to worry about relative paths. canonicalizePaths :: FilePath -> Config -> IO Config canonicalizePaths pwd cfg = do tgt <- canonicalizePath pwd isdir <- doesDirectoryExist tgt is <- mapM (canonicalize tgt isdir) $ idirs cfg cs <- mapM (canonicalize tgt isdir) $ cFiles cfg return $ cfg { idirs = is, cFiles = cs } canonicalize :: FilePath -> Bool -> FilePath -> IO FilePath canonicalize tgt isdir f | isAbsolute f = return f | isdir = canonicalizePath (tgt f) | otherwise = canonicalizePath (takeDirectory tgt f) envCfg :: IO Config envCfg = do so <- lookupEnv "LIQUIDHASKELL_OPTS" case so of Nothing -> return defConfig Just s -> parsePragma $ envLoc s where envLoc = Loc l l l = safeSourcePos "ENVIRONMENT" 1 1 copyright :: String copyright = concat $ concat [ ["LiquidHaskell "] , [$(simpleVersion Meta.version)] , [gitInfo] -- , [" (" ++ _commitCount ++ " commits)" | _commitCount /= ("1"::String) && -- _commitCount /= ("UNKNOWN" :: String)] , ["\nCopyright 2013-19 Regents of the University of California. All Rights Reserved.\n"] ] where _commitCount = $gitCommitCount gitInfo :: String gitInfo = msg where giTry :: Either String GitInfo giTry = $$(fromCode (toCode tGitInfoCwdTry)) msg = case giTry of Left _ -> " no git information" Right gi -> gitMsg gi gitMsg :: GitInfo -> String gitMsg gi = concat [ " [", giBranch gi, "@", giHash gi , " (", giCommitDate gi, ")" -- , " (", show (giCommitCount gi), " commits in HEAD)" , "] " ] -- [NOTE:searchpath] -- 1. we used to add the directory containing the file to the searchpath, -- but this is bad because GHC does NOT do this, and it causes spurious -- "duplicate module" errors in the following scenario -- > tree -- . -- ├── Bar.hs -- └── Foo -- ├── Bar.hs -- └── Goo.hs -- If we run `liquid Foo/Goo.hs` and that imports Bar, GHC will not know -- whether we mean to import Bar.hs or Foo/Bar.hs -- 2. tests fail if you flip order of idirs' mkOpts :: Config -> IO Config mkOpts cfg = do let files' = sortNub $ files cfg id0 <- getIncludeDir return $ cfg { files = files' -- See NOTE [searchpath] , idirs = [id0 gHC_VERSION, id0] ++ idirs cfg } -------------------------------------------------------------------------------- -- | Updating options -------------------------------------------------------------------------------- canonConfig :: Config -> Config canonConfig cfg = cfg { diffcheck = diffcheck cfg && not (fullcheck cfg) -- , eliminate = if higherOrderFlag cfg then FC.All else eliminate cfg } -------------------------------------------------------------------------------- withPragmas :: MonadIO m => Config -> FilePath -> [Located String] -> (Config -> m a) -> m a -------------------------------------------------------------------------------- withPragmas cfg fp ps action = do cfg' <- liftIO $ (processPragmas cfg ps >>= canonicalizePaths fp) <&> canonConfig -- As the verbosity is set /globally/ via the cmdargs lib, re-set it. liftIO $ setVerbosity (loggingVerbosity cfg') res <- action cfg' liftIO $ setVerbosity (loggingVerbosity cfg) -- restore the original verbosity. pure res where processPragmas :: Config -> [Located String] -> IO Config processPragmas c pragmas = withArgs (val <$> pragmas) $ cmdArgsRun config { modeValue = (modeValue config) { cmdArgsValue = c } } -- | Note that this function doesn't process list arguments properly, like -- 'cFiles' or 'expectErrorContaining' -- TODO: This is only used to parse the contents of the env var LIQUIDHASKELL_OPTS -- so it should be able to parse multiple arguments instead. See issue #1990. withPragma :: Config -> Located String -> IO Config withPragma c s = withArgs [val s] $ cmdArgsRun config { modeValue = (modeValue config) { cmdArgsValue = c } } parsePragma :: Located String -> IO Config parsePragma = withPragma defConfig defConfig :: Config defConfig = Config { loggingVerbosity = Quiet , files = def , idirs = def , fullcheck = def , linear = def , stringTheory = def , higherorder = def , smtTimeout = def , higherorderqs = def , diffcheck = def , saveQuery = def , checks = def , nostructuralterm = def , noCheckUnknown = def , notermination = False , nopositivity = False , rankNTypes = False , noclasscheck = False , gradual = False , bscope = False , gdepth = 1 , ginteractive = False , totalHaskell = def -- True , nowarnings = def , noannotations = def , checkDerived = False , caseExpandDepth = 2 , notruetypes = def , nototality = False , pruneUnsorted = def , exactDC = def , noADT = def , expectErrorContaining = def , expectAnyError = False , cores = def , minPartSize = FC.defaultMinPartSize , maxPartSize = FC.defaultMaxPartSize , maxParams = defaultMaxParams , smtsolver = def , shortNames = def , shortErrors = def , cabalDir = def , ghcOptions = def , cFiles = def , port = defaultPort , scrapeInternals = False , scrapeImports = False , scrapeUsedImports = False , elimStats = False , elimBound = Nothing , json = False , counterExamples = False , timeBinds = False , untidyCore = False , eliminate = FC.Some , noPatternInline = False , noSimplifyCore = False -- PLE-OPT , autoInstantiate = def , noslice = False , noLiftedImport = False , proofLogicEval = False , pleWithUndecidedGuards = False , oldPLE = False , interpreter = False , proofLogicEvalLocal = False , reflection = False , extensionality = False , nopolyinfer = False , compileSpec = False , noCheckImports = False , typedHoles = False , typeclass = False , auxInline = False , maxMatchDepth = 4 , maxAppDepth = 2 , maxArgsDepth = 1 , rwTerminationCheck = False , skipModule = False , noLazyPLE = False , fuel = Nothing , environmentReduction = False , noEnvironmentReduction = False , inlineANFBindings = False , pandocHtml = False } -- | Write the annotations (i.e. the files in the \".liquid\" hidden folder) and -- report the result of the checking using a supplied function, or using an -- implicit JSON function, if @json@ flag is set. reportResult :: MonadIO m => (OutputResult -> m ()) -> Config -> [FilePath] -> Output Doc -> m () reportResult logResultFull cfg targets out = do annm <- {-# SCC "annotate" #-} liftIO $ annotate cfg targets out liftIO $ whenNormal $ donePhase Loud "annotate" if json cfg then liftIO $ reportResultJson annm else do let r = o_result out liftIO $ writeCheckVars $ o_vars out cr <- liftIO $ resultWithContext r let outputResult = resDocs tidy cr -- For now, always print the \"header\" with colours, irrespective to the logger -- passed as input. liftIO $ printHeader (colorResult r) (orHeader outputResult) logResultFull outputResult where tidy :: F.Tidy tidy = if shortErrors cfg then F.Lossy else F.Full printHeader :: Moods -> Doc -> IO () printHeader mood d = colorPhaseLn mood "" (render d) ------------------------------------------------------------------------ exitWithResult :: Config -> [FilePath] -> Output Doc -> IO () ------------------------------------------------------------------------ exitWithResult cfg targets out = void $ reportResult writeResultStdout cfg targets out reportResultJson :: ACSS.AnnMap -> IO () reportResultJson annm = do putStrLn "LIQUID" B.putStrLn . encode . annErrors $ annm resultWithContext :: F.FixResult UserError -> IO (FixResult CError) resultWithContext (F.Unsafe s es) = F.Unsafe s <$> errorsWithContext es resultWithContext (F.Safe stats) = return (F.Safe stats) resultWithContext (F.Crash es s) = do let (errs, msgs) = unzip es errs' <- errorsWithContext errs return (F.Crash (zip errs' msgs) s) instance Show (CtxError Doc) where show = showpp writeCheckVars :: Symbolic a => Maybe [a] -> IO () writeCheckVars Nothing = return () writeCheckVars (Just []) = colorPhaseLn Loud "Checked Binders: None" "" writeCheckVars (Just ns) = colorPhaseLn Loud "Checked Binders:" "" >> forM_ ns (putStrLn . symbolString . dropModuleNames . symbol) type CError = CtxError Doc data OutputResult = OutputResult { orHeader :: Doc -- ^ The \"header\" like \"LIQUID: SAFE\", or \"LIQUID: UNSAFE\". , orMessages :: [(GHC.SrcSpan, Doc)] -- ^ The list of pretty-printable messages (typically errors) together with their -- source locations. } -- | Writes the result of this LiquidHaskell run to /stdout/. writeResultStdout :: OutputResult -> IO () writeResultStdout (orMessages -> messages) = do forM_ messages $ \(sSpan, doc) -> putStrLn (render $ mkErrorDoc sSpan doc {- pprint sSpan <> (text ": error: " <+> doc)-}) mkErrorDoc :: PPrint a => a -> Doc -> Doc mkErrorDoc sSpan doc = -- Gross on screen, nice for Ghcid -- pprint sSpan <> (text ": error: " <+> doc) -- Nice on screen, invisible in Ghcid ... (pprint sSpan <> text ": error: ") $+$ nest 4 doc -- | Given a 'FixResult' parameterised over a 'CError', this function returns the \"header\" to show to -- the user (i.e. \"SAFE\" or \"UNSAFE\") plus a list of 'Doc's together with the 'SrcSpan' they refer to. resDocs :: F.Tidy -> F.FixResult CError -> OutputResult resDocs _ (F.Safe stats) = OutputResult { orHeader = text $ "LIQUID: SAFE (" <> show (Solver.numChck stats) <> " constraints checked)" , orMessages = mempty } resDocs _k (F.Crash [] s) = OutputResult { orHeader = text "LIQUID: ERROR" , orMessages = [(GHC.noSrcSpan, text s)] } resDocs k (F.Crash xs s) = OutputResult { orHeader = text "LIQUID: ERROR:" <+> text s , orMessages = map (cErrToSpanned k . errToFCrash) xs } resDocs k (F.Unsafe _ xs) = OutputResult { orHeader = text "LIQUID: UNSAFE" , orMessages = map (cErrToSpanned k) (nub xs) } -- | Renders a 'CError' into a 'Doc' and its associated 'SrcSpan'. cErrToSpanned :: F.Tidy -> CError -> (GHC.SrcSpan, Doc) cErrToSpanned k CtxError{ctErr} = (pos ctErr, pprintTidy k ctErr) errToFCrash :: (CError, Maybe String) -> CError errToFCrash (ce, Just msg) = ce { ctErr = ErrOther (pos (ctErr ce)) (fixMessageDoc msg) } errToFCrash (ce, Nothing) = ce { ctErr = tx $ ctErr ce} where tx (ErrSubType l m _ g t t') = ErrFCrash l m g t t' tx e = F.notracepp "errToFCrash?" e fixMessageDoc :: String -> Doc fixMessageDoc msg = vcat (text <$> lines msg) {- TODO: Never used, do I need to exist? reportUrl = text "Please submit a bug report at: https://github.com/ucsd-progsys/liquidhaskell" -} addErrors :: FixResult a -> [a] -> FixResult a addErrors r [] = r addErrors (Safe s) errs = Unsafe s errs addErrors (Unsafe s xs) errs = Unsafe s (xs ++ errs) addErrors r _ = r instance Fixpoint (F.FixResult CError) where toFix = vcat . map snd . orMessages . resDocs F.Full