ghcide/test/exe/Main.hs

3104 lines
125 KiB
Haskell
Raw Normal View History

2019-08-13 19:23:03 +03:00
-- Copyright (c) 2019 The DAML Authors. All rights reserved.
-- SPDX-License-Identifier: Apache-2.0
{-# LANGUAGE AllowAmbiguousTypes #-}
{-# LANGUAGE DuplicateRecordFields #-}
{-# LANGUAGE ImplicitParams #-}
{-# LANGUAGE PatternSynonyms #-}
{-# LANGUAGE CPP #-}
#include "ghc-api-version.h"
module Main (main) where
import Control.Applicative.Combinators
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
import Control.Exception (bracket, catch)
import qualified Control.Lens as Lens
import Control.Monad
import Control.Monad.IO.Class (liftIO)
import Data.Aeson (FromJSON, Value)
import qualified Data.Binary as Binary
import Data.Foldable
import Data.List.Extra
import Data.Maybe
import Data.Rope.UTF16 (Rope)
import qualified Data.Rope.UTF16 as Rope
import Development.IDE.Core.PositionMapping (fromCurrent, toCurrent)
import Development.IDE.Core.Shake (Q(..))
import Development.IDE.GHC.Util
import qualified Data.Text as T
import Data.Typeable
import Development.IDE.Spans.Common
import Development.IDE.Test
import Development.IDE.Test.Runfiles
import Development.IDE.Types.Location
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
import Development.Shake (getDirectoryFilesIO)
import qualified Experiments as Bench
import Language.Haskell.LSP.Test
import Language.Haskell.LSP.Messages
import Language.Haskell.LSP.Types
import Language.Haskell.LSP.Types.Capabilities
import qualified Language.Haskell.LSP.Types.Lens as Lsp (diagnostics, params, message)
import Language.Haskell.LSP.VFS (applyChange)
import Network.URI
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
import System.Environment.Blank (getEnv, setEnv, unsetEnv)
import System.FilePath
import System.IO.Extra hiding (withTempDir)
import qualified System.IO.Extra
import System.Directory
ShakeSession and shakeEnqueue (#554) * ShakeSession and shakeRunGently Currently we start a new Shake session for every interaction with the Shake database, including type checking, hovers, code actions, completions, etc. Since only one Shake session can ever exist, we abort the active session if any in order to execute the new command in a responsive manner. This is suboptimal in many, many ways: - A hover in module M aborts the typechecking of module M, only to start over! - Read-only commands (hover, code action, completion) need to typecheck all the modules! (or rather, ask Shake to check that the typechecks are current) - There is no way to run non-interfering commands concurrently This is an experiment inspired by the 'ShakeQueue' of @mpickering, and the follow-up discussion in https://github.com/mpickering/ghcide/issues/7 We introduce the concept of the 'ShakeSession' as part of the IDE state. The 'ShakeSession' is initialized by a call to 'shakeRun', and survives until the next call to 'shakeRun'. It is important that the session is restarted as soon as the filesystem changes, to ensure that the database is current. The 'ShakeSession' enables a new command 'shakeRunGently', which appends work to the existing 'ShakeSession'. This command can be called in parallel without any restriction. * Simplify by assuming there is always a ShakeSession * Improved naming and docs * Define runActionSync on top of shakeEnqueue shakeRun is not correct as it never returns anymore * Drive progress reporting from newSession The previous approach reused the shakeProgress thread, which doesn't work anymore as ShakeSession keeps the ShakeDatabase open until the next edit * Deterministic progress messages in tests Dropping the 0.1s sleep to ensure that progress messages during tests are deterministic * Make kick explicit This is required for progress reporting to work, see notes in shakeRun As to whether this is the right thing to do: 1. Less magic, more explicit 2. There's only 2 places where kick is actually used * apply Neil's feedback * avoid a deadlock when the enqueued action throws * Simplify runAction + comments * use a Barrier for clarity A Barrier is a smaller abstraction than an MVar, and the next version of the extra package will come with a suitably small implementation: https://github.com/ndmitchell/extra/commit/98c2a83585d2ca0a9d961dd241c4a967ef87866a * Log timings for code actions, hovers and completions * Rename shakeRun to shakeRestart The action returned by shakeRun now blocks until another call to shakeRun is made, which is a change in behaviour,. but all the current uses of shakeRun ignore this action. Since the new behaviour is not useful, this change simplifies and updates the docs and name accordingly * delete runActionSync as it's just runAction * restart shake session on new component created * requeue pending actions on session restart * hlint * Bumped the delay from 5 to 6 * Add a test for the non-lsp command line * Update exe/Main.hs Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-08 12:36:36 +03:00
import System.Exit (ExitCode(ExitSuccess))
import System.Process.Extra (readCreateProcessWithExitCode, CreateProcess(cwd), proc)
import Test.QuickCheck
import Test.QuickCheck.Instances ()
import Test.Tasty
import Test.Tasty.ExpectedFailure
import Test.Tasty.Ingredients.Rerun
import Test.Tasty.HUnit
import Test.Tasty.QuickCheck
main :: IO ()
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
main = do
-- We mess with env vars so run single-threaded.
setEnv "TASTY_NUM_THREADS" "1" True
defaultMainWithRerun $ testGroup "ghcide"
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
[ testSession "open close" $ do
doc <- createDoc "Testing.hs" "haskell" ""
void (skipManyTill anyMessage message :: Session WorkDoneProgressCreateRequest)
void (skipManyTill anyMessage message :: Session WorkDoneProgressBeginNotification)
closeDoc doc
void (skipManyTill anyMessage message :: Session WorkDoneProgressEndNotification)
, initializeResponseTests
, completionTests
, cppTests
, diagnosticTests
, codeActionTests
, codeLensesTests
, outlineTests
, findDefinitionAndHoverTests
, pluginTests
, preprocessorTests
, thTests
, safeTests
, unitTests
, haddockTests
, positionMappingTests
, watchedFilesTests
, cradleTests
, dependentFileTest
ShakeSession and shakeEnqueue (#554) * ShakeSession and shakeRunGently Currently we start a new Shake session for every interaction with the Shake database, including type checking, hovers, code actions, completions, etc. Since only one Shake session can ever exist, we abort the active session if any in order to execute the new command in a responsive manner. This is suboptimal in many, many ways: - A hover in module M aborts the typechecking of module M, only to start over! - Read-only commands (hover, code action, completion) need to typecheck all the modules! (or rather, ask Shake to check that the typechecks are current) - There is no way to run non-interfering commands concurrently This is an experiment inspired by the 'ShakeQueue' of @mpickering, and the follow-up discussion in https://github.com/mpickering/ghcide/issues/7 We introduce the concept of the 'ShakeSession' as part of the IDE state. The 'ShakeSession' is initialized by a call to 'shakeRun', and survives until the next call to 'shakeRun'. It is important that the session is restarted as soon as the filesystem changes, to ensure that the database is current. The 'ShakeSession' enables a new command 'shakeRunGently', which appends work to the existing 'ShakeSession'. This command can be called in parallel without any restriction. * Simplify by assuming there is always a ShakeSession * Improved naming and docs * Define runActionSync on top of shakeEnqueue shakeRun is not correct as it never returns anymore * Drive progress reporting from newSession The previous approach reused the shakeProgress thread, which doesn't work anymore as ShakeSession keeps the ShakeDatabase open until the next edit * Deterministic progress messages in tests Dropping the 0.1s sleep to ensure that progress messages during tests are deterministic * Make kick explicit This is required for progress reporting to work, see notes in shakeRun As to whether this is the right thing to do: 1. Less magic, more explicit 2. There's only 2 places where kick is actually used * apply Neil's feedback * avoid a deadlock when the enqueued action throws * Simplify runAction + comments * use a Barrier for clarity A Barrier is a smaller abstraction than an MVar, and the next version of the extra package will come with a suitably small implementation: https://github.com/ndmitchell/extra/commit/98c2a83585d2ca0a9d961dd241c4a967ef87866a * Log timings for code actions, hovers and completions * Rename shakeRun to shakeRestart The action returned by shakeRun now blocks until another call to shakeRun is made, which is a change in behaviour,. but all the current uses of shakeRun ignore this action. Since the new behaviour is not useful, this change simplifies and updates the docs and name accordingly * delete runActionSync as it's just runAction * restart shake session on new component created * requeue pending actions on session restart * hlint * Bumped the delay from 5 to 6 * Add a test for the non-lsp command line * Update exe/Main.hs Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-08 12:36:36 +03:00
, nonLspCommandLine
, benchmarkTests
Interface file fixes (#645) * Add test for inconsistent diagnostics * Refactoring ModIfaceFromDisk This started as a pure refactoring to clarify the responsibilities between ModIface and ModIfaceFromDisk, but ended up having some behaviour changes: 1. Regenerate interface when checkOldIface returns something other than UpToDate. This was a bug. 2. Do not generate a diagnostic when regenerating an interface. 2. Previously we conflated stale interface with other errors, and would regenerate in both cases. Now we only regenerate in the first case. Tentative fix for #597 * Split interface tests * Always recompile modules with TH splices Tentative fix for #614 TODO support stability * Fix expectDiagnostics in MacOs * Avoid File does not exist diagnostics for interface files Fixes #642 * Clarify interface tests * hlints * Performance fixes The previous changes were 10X slower, this is 20X faster than those, so 2X faster than upstream, for some benchmarks * formatting * Fix GetModificationTime identity The answer for a GetModification query is independent of the missingFileDiagnostics field (as the diagnostics are not part of the answer) * remove stale comment * Avoid calling ghcSessionDepsDefinition twice * Apply suggestions from code review Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Code review feedback * Address review feedback https://github.com/digital-asset/ghcide/pull/645/files/49b0d9ac65399edf82a7a9cbbb8d8b5420458d8d#r443383239 * Change recomp to direct cradle Co-authored-by: Zubin Duggal <zubin@cmi.ac.in> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-23 12:01:52 +03:00
, ifaceTests
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
]
initializeResponseTests :: TestTree
initializeResponseTests = withResource acquire release tests where
-- these tests document and monitor the evolution of the
-- capabilities announced by the server in the initialize
-- response. Currently the server advertises almost no capabilities
-- at all, in some cases failing to announce capabilities that it
-- actually does provide! Hopefully this will change ...
tests :: IO InitializeResponse -> TestTree
tests getInitializeResponse =
testGroup "initialize response capabilities"
[ chk " text doc sync" _textDocumentSync tds
, chk " hover" _hoverProvider (Just True)
, chk " completion" _completionProvider (Just $ CompletionOptions (Just False) (Just ["."]) Nothing)
, chk "NO signature help" _signatureHelpProvider Nothing
, chk " goto definition" _definitionProvider (Just True)
, chk " goto type definition" _typeDefinitionProvider (Just $ GotoOptionsStatic True)
-- BUG in lsp-test, this test fails, just change the accepted response
-- for now
, chk "NO goto implementation" _implementationProvider (Just $ GotoOptionsStatic True)
, chk "NO find references" _referencesProvider Nothing
, chk "NO doc highlight" _documentHighlightProvider Nothing
, chk " doc symbol" _documentSymbolProvider (Just True)
, chk "NO workspace symbol" _workspaceSymbolProvider Nothing
, chk " code action" _codeActionProvider $ Just $ CodeActionOptionsStatic True
, chk " code lens" _codeLensProvider $ Just $ CodeLensOptions Nothing
, chk "NO doc formatting" _documentFormattingProvider Nothing
, chk "NO doc range formatting"
_documentRangeFormattingProvider Nothing
, chk "NO doc formatting on typing"
_documentOnTypeFormattingProvider Nothing
, chk "NO renaming" _renameProvider (Just $ RenameOptionsStatic False)
, chk "NO doc link" _documentLinkProvider Nothing
, chk "NO color" _colorProvider (Just $ ColorOptionsStatic False)
, chk "NO folding range" _foldingRangeProvider (Just $ FoldingRangeOptionsStatic False)
, che " execute command" _executeCommandProvider (Just $ ExecuteCommandOptions $ List ["typesignature.add"])
, chk " workspace" _workspace (Just $ WorkspaceOptions (Just WorkspaceFolderOptions{_supported = Just True, _changeNotifications = Just ( WorkspaceFolderChangeNotificationsBool True )}))
, chk "NO experimental" _experimental Nothing
] where
tds = Just (TDSOptions (TextDocumentSyncOptions
{ _openClose = Just True
, _change = Just TdSyncIncremental
, _willSave = Nothing
, _willSaveWaitUntil = Nothing
, _save = Just (SaveOptions {_includeText = Nothing})}))
chk :: (Eq a, Show a) => TestName -> (InitializeResponseCapabilitiesInner -> a) -> a -> TestTree
chk title getActual expected =
testCase title $ getInitializeResponse >>= \ir -> expected @=? (getActual . innerCaps) ir
che :: TestName -> (InitializeResponseCapabilitiesInner -> Maybe ExecuteCommandOptions) -> Maybe ExecuteCommandOptions -> TestTree
che title getActual _expected = testCase title doTest
where
doTest = do
ir <- getInitializeResponse
let Just ExecuteCommandOptions {_commands = List [command]} = getActual $ innerCaps ir
True @=? T.isSuffixOf "typesignature.add" command
innerCaps :: InitializeResponse -> InitializeResponseCapabilitiesInner
2020-05-13 15:59:51 +03:00
innerCaps (ResponseMessage _ _ (Right (InitializeResponseCapabilities c))) = c
innerCaps _ = error "this test only expects inner capabilities"
acquire :: IO InitializeResponse
acquire = run initializeResponse
release :: InitializeResponse -> IO ()
release = const $ pure ()
diagnosticTests :: TestTree
diagnosticTests = testGroup "diagnostics"
[ testSessionWait "fix syntax error" $ do
let content = T.unlines [ "module Testing wher" ]
doc <- createDoc "Testing.hs" "haskell" content
expectDiagnostics [("Testing.hs", [(DsError, (0, 15), "parse error")])]
let change = TextDocumentContentChangeEvent
{ _range = Just (Range (Position 0 15) (Position 0 19))
, _rangeLength = Nothing
, _text = "where"
}
changeDoc doc [change]
expectDiagnostics [("Testing.hs", [])]
, testSessionWait "introduce syntax error" $ do
let content = T.unlines [ "module Testing where" ]
doc <- createDoc "Testing.hs" "haskell" content
void $ skipManyTill anyMessage (message :: Session WorkDoneProgressCreateRequest)
void $ skipManyTill anyMessage (message :: Session WorkDoneProgressBeginNotification)
let change = TextDocumentContentChangeEvent
{ _range = Just (Range (Position 0 15) (Position 0 18))
, _rangeLength = Nothing
, _text = "wher"
}
changeDoc doc [change]
expectDiagnostics [("Testing.hs", [(DsError, (0, 15), "parse error")])]
, testSessionWait "variable not in scope" $ do
let content = T.unlines
[ "module Testing where"
, "foo :: Int -> Int -> Int"
, "foo a b = a + ab"
, "bar :: Int -> Int -> Int"
, "bar a b = cd + b"
]
_ <- createDoc "Testing.hs" "haskell" content
expectDiagnostics
[ ( "Testing.hs"
, [ (DsError, (2, 14), "Variable not in scope: ab")
, (DsError, (4, 10), "Variable not in scope: cd")
]
)
]
, testSessionWait "type error" $ do
let content = T.unlines
[ "module Testing where"
, "foo :: Int -> String -> Int"
, "foo a b = a + b"
]
_ <- createDoc "Testing.hs" "haskell" content
expectDiagnostics
[ ( "Testing.hs"
, [(DsError, (2, 14), "Couldn't match type '[Char]' with 'Int'")]
)
]
, testSessionWait "typed hole" $ do
Defer type errors (#47) * TEST: Degrade type error to warning It will be upgraded again later, but for the time being we want to see whether the proposed mechanism for deferring type errors works at all. As it turns out the first, most obvious approach, does not work: this is documented in the next commit. A second approach was found that does work, and appears in the commit after the next. This test is failing until the second approach is implemented. * Defer type errors (first approach: FAILED) The idea is to set the `-fdefer-type-errors` and `-fwarn-deferred-type-errors` flags, by setting options programatically inside the `Ghc` monad. Deferral of type errors was not observed with this approach. The (less obvious) approach used in the next commit seems to be more successful. * Defer type errors (second approach: SUCCESS) This approach modifies the `ParsedModule` which is passed to `GHC.typecheckedModule` by hie-core's `typecheckModule`. Type warning deferral is now observed at run time, and the tests pass. * TEST: Reinstate severity of type errors So far, type errors have been deferred and reported as warnings. The next step is to ensure that the deferred type errors are reported as errors rather than warnings, once again. This test fails until the implementation arrives in the next commit. * Upgrade severity of deferred Type Errors after typecheck ... and make the test pass again. * Hide helper functions in local scopes * Stop setting Opt_WarnDeferredTypeErrors ... and the tests still pass, thereby confirming @hsenag's hypothesis that this flag is not needed. * TEST: Check that typed holes are reported as errors * TEST: Downgrade severity of typed holes Error -> Warning This test fails, thereby falsifying the hypothesis that `Opt_DeferTypeErrors` implies `Opt_DeferTypedHoles`. * Defer typed holes ... and pass the failing test. * TEST: Reinstate severity of typed holes ... failing the test until the implementation catches up in the next commit. * Upgrade severity of deferred Typed Holes after typecheck ... and pass the test once again. * TEST: Degrade variable out of scope from Error to Warning ... test fails until next commit. * Defer out of scope variables ... passing the test which was changed in the last commit. * TEST: Reinstate severity of out of scope variables ... failing the test, and forcing the implementation to catch up. * Upgrade severity of deferred out of scope vars after typecheck ... passing the test once again. * Add explicit tests for deferrals * Add IdeOption for deferral switching * Improve documentation of optDefer * Add IdeDefer newtype
2019-09-17 15:28:20 +03:00
let content = T.unlines
[ "module Testing where"
, "foo :: Int -> String"
, "foo a = _ a"
]
_ <- createDoc "Testing.hs" "haskell" content
Defer type errors (#47) * TEST: Degrade type error to warning It will be upgraded again later, but for the time being we want to see whether the proposed mechanism for deferring type errors works at all. As it turns out the first, most obvious approach, does not work: this is documented in the next commit. A second approach was found that does work, and appears in the commit after the next. This test is failing until the second approach is implemented. * Defer type errors (first approach: FAILED) The idea is to set the `-fdefer-type-errors` and `-fwarn-deferred-type-errors` flags, by setting options programatically inside the `Ghc` monad. Deferral of type errors was not observed with this approach. The (less obvious) approach used in the next commit seems to be more successful. * Defer type errors (second approach: SUCCESS) This approach modifies the `ParsedModule` which is passed to `GHC.typecheckedModule` by hie-core's `typecheckModule`. Type warning deferral is now observed at run time, and the tests pass. * TEST: Reinstate severity of type errors So far, type errors have been deferred and reported as warnings. The next step is to ensure that the deferred type errors are reported as errors rather than warnings, once again. This test fails until the implementation arrives in the next commit. * Upgrade severity of deferred Type Errors after typecheck ... and make the test pass again. * Hide helper functions in local scopes * Stop setting Opt_WarnDeferredTypeErrors ... and the tests still pass, thereby confirming @hsenag's hypothesis that this flag is not needed. * TEST: Check that typed holes are reported as errors * TEST: Downgrade severity of typed holes Error -> Warning This test fails, thereby falsifying the hypothesis that `Opt_DeferTypeErrors` implies `Opt_DeferTypedHoles`. * Defer typed holes ... and pass the failing test. * TEST: Reinstate severity of typed holes ... failing the test until the implementation catches up in the next commit. * Upgrade severity of deferred Typed Holes after typecheck ... and pass the test once again. * TEST: Degrade variable out of scope from Error to Warning ... test fails until next commit. * Defer out of scope variables ... passing the test which was changed in the last commit. * TEST: Reinstate severity of out of scope variables ... failing the test, and forcing the implementation to catch up. * Upgrade severity of deferred out of scope vars after typecheck ... passing the test once again. * Add explicit tests for deferrals * Add IdeOption for deferral switching * Improve documentation of optDefer * Add IdeDefer newtype
2019-09-17 15:28:20 +03:00
expectDiagnostics
[ ( "Testing.hs"
, [(DsError, (2, 8), "Found hole: _ :: Int -> String")]
)
]
, testGroup "deferral" $
let sourceA a = T.unlines
[ "module A where"
, "a :: Int"
, "a = " <> a]
sourceB = T.unlines
[ "module B where"
, "import A"
, "b :: Float"
, "b = True"]
bMessage = "Couldn't match expected type 'Float' with actual type 'Bool'"
expectedDs aMessage =
[ ("A.hs", [(DsError, (2,4), aMessage)])
, ("B.hs", [(DsError, (3,4), bMessage)])]
deferralTest title binding msg = testSessionWait title $ do
_ <- createDoc "A.hs" "haskell" $ sourceA binding
_ <- createDoc "B.hs" "haskell" sourceB
expectDiagnostics $ expectedDs msg
Defer type errors (#47) * TEST: Degrade type error to warning It will be upgraded again later, but for the time being we want to see whether the proposed mechanism for deferring type errors works at all. As it turns out the first, most obvious approach, does not work: this is documented in the next commit. A second approach was found that does work, and appears in the commit after the next. This test is failing until the second approach is implemented. * Defer type errors (first approach: FAILED) The idea is to set the `-fdefer-type-errors` and `-fwarn-deferred-type-errors` flags, by setting options programatically inside the `Ghc` monad. Deferral of type errors was not observed with this approach. The (less obvious) approach used in the next commit seems to be more successful. * Defer type errors (second approach: SUCCESS) This approach modifies the `ParsedModule` which is passed to `GHC.typecheckedModule` by hie-core's `typecheckModule`. Type warning deferral is now observed at run time, and the tests pass. * TEST: Reinstate severity of type errors So far, type errors have been deferred and reported as warnings. The next step is to ensure that the deferred type errors are reported as errors rather than warnings, once again. This test fails until the implementation arrives in the next commit. * Upgrade severity of deferred Type Errors after typecheck ... and make the test pass again. * Hide helper functions in local scopes * Stop setting Opt_WarnDeferredTypeErrors ... and the tests still pass, thereby confirming @hsenag's hypothesis that this flag is not needed. * TEST: Check that typed holes are reported as errors * TEST: Downgrade severity of typed holes Error -> Warning This test fails, thereby falsifying the hypothesis that `Opt_DeferTypeErrors` implies `Opt_DeferTypedHoles`. * Defer typed holes ... and pass the failing test. * TEST: Reinstate severity of typed holes ... failing the test until the implementation catches up in the next commit. * Upgrade severity of deferred Typed Holes after typecheck ... and pass the test once again. * TEST: Degrade variable out of scope from Error to Warning ... test fails until next commit. * Defer out of scope variables ... passing the test which was changed in the last commit. * TEST: Reinstate severity of out of scope variables ... failing the test, and forcing the implementation to catch up. * Upgrade severity of deferred out of scope vars after typecheck ... passing the test once again. * Add explicit tests for deferrals * Add IdeOption for deferral switching * Improve documentation of optDefer * Add IdeDefer newtype
2019-09-17 15:28:20 +03:00
in
[ deferralTest "type error" "True" "Couldn't match expected type"
, deferralTest "typed hole" "_" "Found hole"
, deferralTest "out of scope var" "unbound" "Variable not in scope"
Defer type errors (#47) * TEST: Degrade type error to warning It will be upgraded again later, but for the time being we want to see whether the proposed mechanism for deferring type errors works at all. As it turns out the first, most obvious approach, does not work: this is documented in the next commit. A second approach was found that does work, and appears in the commit after the next. This test is failing until the second approach is implemented. * Defer type errors (first approach: FAILED) The idea is to set the `-fdefer-type-errors` and `-fwarn-deferred-type-errors` flags, by setting options programatically inside the `Ghc` monad. Deferral of type errors was not observed with this approach. The (less obvious) approach used in the next commit seems to be more successful. * Defer type errors (second approach: SUCCESS) This approach modifies the `ParsedModule` which is passed to `GHC.typecheckedModule` by hie-core's `typecheckModule`. Type warning deferral is now observed at run time, and the tests pass. * TEST: Reinstate severity of type errors So far, type errors have been deferred and reported as warnings. The next step is to ensure that the deferred type errors are reported as errors rather than warnings, once again. This test fails until the implementation arrives in the next commit. * Upgrade severity of deferred Type Errors after typecheck ... and make the test pass again. * Hide helper functions in local scopes * Stop setting Opt_WarnDeferredTypeErrors ... and the tests still pass, thereby confirming @hsenag's hypothesis that this flag is not needed. * TEST: Check that typed holes are reported as errors * TEST: Downgrade severity of typed holes Error -> Warning This test fails, thereby falsifying the hypothesis that `Opt_DeferTypeErrors` implies `Opt_DeferTypedHoles`. * Defer typed holes ... and pass the failing test. * TEST: Reinstate severity of typed holes ... failing the test until the implementation catches up in the next commit. * Upgrade severity of deferred Typed Holes after typecheck ... and pass the test once again. * TEST: Degrade variable out of scope from Error to Warning ... test fails until next commit. * Defer out of scope variables ... passing the test which was changed in the last commit. * TEST: Reinstate severity of out of scope variables ... failing the test, and forcing the implementation to catch up. * Upgrade severity of deferred out of scope vars after typecheck ... passing the test once again. * Add explicit tests for deferrals * Add IdeOption for deferral switching * Improve documentation of optDefer * Add IdeDefer newtype
2019-09-17 15:28:20 +03:00
]
, testSessionWait "remove required module" $ do
let contentA = T.unlines [ "module ModuleA where" ]
docA <- createDoc "ModuleA.hs" "haskell" contentA
let contentB = T.unlines
[ "module ModuleB where"
, "import ModuleA"
]
_ <- createDoc "ModuleB.hs" "haskell" contentB
let change = TextDocumentContentChangeEvent
{ _range = Just (Range (Position 0 0) (Position 0 20))
, _rangeLength = Nothing
, _text = ""
}
changeDoc docA [change]
expectDiagnostics [("ModuleB.hs", [(DsError, (1, 0), "Could not find module")])]
, testSessionWait "add missing module" $ do
let contentB = T.unlines
[ "module ModuleB where"
, "import ModuleA"
]
_ <- createDoc "ModuleB.hs" "haskell" contentB
expectDiagnostics [("ModuleB.hs", [(DsError, (1, 7), "Could not find module")])]
let contentA = T.unlines [ "module ModuleA where" ]
_ <- createDoc "ModuleA.hs" "haskell" contentA
expectDiagnostics [("ModuleB.hs", [])]
, testSessionWait "add missing module (non workspace)" $ do
let contentB = T.unlines
[ "module ModuleB where"
, "import ModuleA"
]
_ <- createDoc "/tmp/ModuleB.hs" "haskell" contentB
expectDiagnostics [("/tmp/ModuleB.hs", [(DsError, (1, 7), "Could not find module")])]
let contentA = T.unlines [ "module ModuleA where" ]
_ <- createDoc "/tmp/ModuleA.hs" "haskell" contentA
expectDiagnostics [("/tmp/ModuleB.hs", [])]
, testSessionWait "cyclic module dependency" $ do
let contentA = T.unlines
[ "module ModuleA where"
, "import ModuleB"
]
let contentB = T.unlines
[ "module ModuleB where"
, "import ModuleA"
]
_ <- createDoc "ModuleA.hs" "haskell" contentA
_ <- createDoc "ModuleB.hs" "haskell" contentB
expectDiagnostics
[ ( "ModuleA.hs"
, [(DsError, (1, 7), "Cyclic module dependency between ModuleA, ModuleB")]
)
, ( "ModuleB.hs"
, [(DsError, (1, 7), "Cyclic module dependency between ModuleA, ModuleB")]
)
]
, testSessionWait "cyclic module dependency with hs-boot" $ do
2019-09-10 12:35:52 +03:00
let contentA = T.unlines
[ "module ModuleA where"
, "import {-# SOURCE #-} ModuleB"
]
let contentB = T.unlines
[ "module ModuleB where"
, "import ModuleA"
]
let contentBboot = T.unlines
[ "module ModuleB where"
]
_ <- createDoc "ModuleA.hs" "haskell" contentA
_ <- createDoc "ModuleB.hs" "haskell" contentB
_ <- createDoc "ModuleB.hs-boot" "haskell" contentBboot
2019-09-10 12:35:52 +03:00
expectDiagnostics []
, testSessionWait "correct reference used with hs-boot" $ do
2019-09-10 12:35:52 +03:00
let contentB = T.unlines
[ "module ModuleB where"
, "import {-# SOURCE #-} ModuleA"
]
let contentA = T.unlines
[ "module ModuleA where"
, "import ModuleB"
, "x = 5"
]
let contentAboot = T.unlines
[ "module ModuleA where"
]
let contentC = T.unlines
[ "module ModuleC where"
, "import ModuleA"
-- this reference will fail if it gets incorrectly
-- resolved to the hs-boot file
, "y = x"
]
_ <- createDoc "ModuleB.hs" "haskell" contentB
_ <- createDoc "ModuleA.hs" "haskell" contentA
_ <- createDoc "ModuleA.hs-boot" "haskell" contentAboot
_ <- createDoc "ModuleC.hs" "haskell" contentC
2019-09-10 12:35:52 +03:00
expectDiagnostics []
, testSessionWait "redundant import" $ do
let contentA = T.unlines ["module ModuleA where"]
let contentB = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import ModuleA"
]
_ <- createDoc "ModuleA.hs" "haskell" contentA
_ <- createDoc "ModuleB.hs" "haskell" contentB
expectDiagnostics
[ ( "ModuleB.hs"
, [(DsWarning, (2, 0), "The import of 'ModuleA' is redundant")]
)
]
, testSessionWait "package imports" $ do
2019-09-12 10:39:13 +03:00
let thisDataListContent = T.unlines
[ "module Data.List where"
, "x :: Integer"
2019-09-12 10:39:13 +03:00
, "x = 123"
]
let mainContent = T.unlines
[ "{-# LANGUAGE PackageImports #-}"
, "module Main where"
, "import qualified \"this\" Data.List as ThisList"
, "import qualified \"base\" Data.List as BaseList"
, "useThis = ThisList.x"
, "useBase = BaseList.map"
, "wrong1 = ThisList.map"
, "wrong2 = BaseList.x"
]
_ <- createDoc "Data/List.hs" "haskell" thisDataListContent
_ <- createDoc "Main.hs" "haskell" mainContent
2019-09-12 10:39:13 +03:00
expectDiagnostics
[ ( "Main.hs"
, [(DsError, (6, 9), "Not in scope: \8216ThisList.map\8217")
,(DsError, (7, 9), "Not in scope: \8216BaseList.x\8217")
]
)
]
, testSessionWait "unqualified warnings" $ do
let fooContent = T.unlines
[ "{-# OPTIONS_GHC -Wredundant-constraints #-}"
, "module Foo where"
, "foo :: Ord a => a -> Int"
, "foo a = 1"
]
_ <- createDoc "Foo.hs" "haskell" fooContent
expectDiagnostics
[ ( "Foo.hs"
-- The test is to make sure that warnings contain unqualified names
-- where appropriate. The warning should use an unqualified name 'Ord', not
-- sometihng like 'GHC.Classes.Ord'. The choice of redundant-constraints to
-- test this is fairly arbitrary.
, [(DsWarning, (2, 0), "Redundant constraint: Ord a")
]
)
]
, testSessionWait "lower-case drive" $ do
let aContent = T.unlines
[ "module A.A where"
, "import A.B ()"
]
bContent = T.unlines
[ "{-# OPTIONS_GHC -Wall #-}"
, "module A.B where"
, "import Data.List"
]
uriB <- getDocUri "A/B.hs"
Just pathB <- pure $ uriToFilePath uriB
uriB <- pure $
let (drive, suffix) = splitDrive pathB
in filePathToUri (joinDrive (lower drive) suffix)
liftIO $ createDirectoryIfMissing True (takeDirectory pathB)
liftIO $ writeFileUTF8 pathB $ T.unpack bContent
uriA <- getDocUri "A/A.hs"
Just pathA <- pure $ uriToFilePath uriA
uriA <- pure $
let (drive, suffix) = splitDrive pathA
in filePathToUri (joinDrive (lower drive) suffix)
let itemA = TextDocumentItem uriA "haskell" 0 aContent
let a = TextDocumentIdentifier uriA
sendNotification TextDocumentDidOpen (DidOpenTextDocumentParams itemA)
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
diagsNot <- skipManyTill anyMessage diagnostic
let PublishDiagnosticsParams fileUri diags = _params (diagsNot :: PublishDiagnosticsNotification)
-- Check that if we put a lower-case drive in for A.A
-- the diagnostics for A.B will also be lower-case.
liftIO $ fileUri @?= uriB
let msg = _message (head (toList diags) :: Diagnostic)
liftIO $ unless ("redundant" `T.isInfixOf` msg) $
assertFailure ("Expected redundant import but got " <> T.unpack msg)
closeDoc a
, testSessionWait "haddock parse error" $ do
let fooContent = T.unlines
[ "module Foo where"
, "foo :: Int"
, "foo = 1 {-|-}"
]
_ <- createDoc "Foo.hs" "haskell" fooContent
expectDiagnostics
[ ( "Foo.hs"
, [(DsWarning, (2, 8), "Haddock parse error on input")
]
)
]
, testSessionWait "strip file path" $ do
let
name = "Testing"
content = T.unlines
[ "module " <> name <> " where"
, "value :: Maybe ()"
, "value = [()]"
]
_ <- createDoc (T.unpack name <> ".hs") "haskell" content
notification <- skipManyTill anyMessage diagnostic
let
offenders =
Lsp.params .
Lsp.diagnostics .
Lens.folded .
Lsp.message .
Lens.filtered (T.isInfixOf ("/" <> name <> ".hs:"))
failure msg = liftIO $ assertFailure $ "Expected file path to be stripped but got " <> T.unpack msg
Lens.mapMOf_ offenders failure notification
]
codeActionTests :: TestTree
codeActionTests = testGroup "code actions"
[ renameActionTests
, typeWildCardActionTests
, removeImportTests
, extendImportTests
, suggestImportTests
, addExtensionTests
, fixConstructorImportTests
, importRenameActionTests
, fillTypedHoleTests
, addSigActionTests
, insertNewDefinitionTests
, deleteUnusedDefinitionTests
, addInstanceConstraintTests
, addFunctionConstraintTests
, removeRedundantConstraintsTests
, addTypeAnnotationsToLiteralsTest
]
codeLensesTests :: TestTree
codeLensesTests = testGroup "code lenses"
[ addSigLensesTests
]
watchedFilesTests :: TestTree
watchedFilesTests = testGroup "watched files"
[ testSession' "workspace files" $ \sessionDir -> do
liftIO $ writeFile (sessionDir </> "hie.yaml") "cradle: {direct: {arguments: [\"-isrc\"]}}"
_doc <- createDoc "A.hs" "haskell" "{-#LANGUAGE NoImplicitPrelude #-}\nmodule A where\nimport WatchedFilesMissingModule"
watchedFileRegs <- getWatchedFilesSubscriptionsUntil @PublishDiagnosticsNotification
-- Expect 4 subscriptions (A does not get any because it's VFS):
-- - /path-to-workspace/hie.yaml
-- - /path-to-workspace/WatchedFilesMissingModule.hs
-- - /path-to-workspace/WatchedFilesMissingModule.lhs
-- - /path-to-workspace/src/WatchedFilesMissingModule.hs
-- - /path-to-workspace/src/WatchedFilesMissingModule.lhs
liftIO $ length watchedFileRegs @?= 5
, testSession' "non workspace file" $ \sessionDir -> do
liftIO $ writeFile (sessionDir </> "hie.yaml") "cradle: {direct: {arguments: [\"-i/tmp\"]}}"
_doc <- createDoc "A.hs" "haskell" "{-# LANGUAGE NoImplicitPrelude#-}\nmodule A where\nimport WatchedFilesMissingModule"
watchedFileRegs <- getWatchedFilesSubscriptionsUntil @PublishDiagnosticsNotification
-- Expect 2 subscriptions (/tmp does not get any as it is out of the workspace):
-- - /path-to-workspace/hie.yaml
-- - /path-to-workspace/WatchedFilesMissingModule.hs
-- - /path-to-workspace/WatchedFilesMissingModule.lhs
liftIO $ length watchedFileRegs @?= 3
-- TODO add a test for didChangeWorkspaceFolder
]
renameActionTests :: TestTree
renameActionTests = testGroup "rename actions"
[ testSession "change to local variable name" $ do
let content = T.unlines
[ "module Testing where"
, "foo :: Int -> Int"
, "foo argName = argNme"
]
doc <- createDoc "Testing.hs" "haskell" content
_ <- waitForDiagnostics
action <- findCodeAction doc (Range (Position 2 14) (Position 2 20)) "Replace with argName"
executeCodeAction action
contentAfterAction <- documentContents doc
let expectedContentAfterAction = T.unlines
[ "module Testing where"
, "foo :: Int -> Int"
, "foo argName = argName"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
, testSession "change to name of imported function" $ do
let content = T.unlines
[ "module Testing where"
, "import Data.Maybe (maybeToList)"
, "foo :: Maybe a -> [a]"
, "foo = maybToList"
]
doc <- createDoc "Testing.hs" "haskell" content
_ <- waitForDiagnostics
action <- findCodeAction doc (Range (Position 3 6) (Position 3 16)) "Replace with maybeToList"
executeCodeAction action
contentAfterAction <- documentContents doc
let expectedContentAfterAction = T.unlines
[ "module Testing where"
, "import Data.Maybe (maybeToList)"
, "foo :: Maybe a -> [a]"
, "foo = maybeToList"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
, testSession "suggest multiple local variable names" $ do
let content = T.unlines
[ "module Testing where"
, "foo :: Char -> Char -> Char -> Char"
, "foo argument1 argument2 argument3 = argumentX"
]
doc <- createDoc "Testing.hs" "haskell" content
_ <- waitForDiagnostics
_ <- findCodeActions doc (Range (Position 2 36) (Position 2 45))
["Replace with argument1", "Replace with argument2", "Replace with argument3"]
return()
, testSession "change infix function" $ do
let content = T.unlines
[ "module Testing where"
, "monus :: Int -> Int"
, "monus x y = max 0 (x - y)"
, "foo x y = x `monnus` y"
]
doc <- createDoc "Testing.hs" "haskell" content
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 3 12) (Position 3 20))
[fixTypo] <- pure [action | CACodeAction action@CodeAction{ _title = actionTitle } <- actionsOrCommands, "monus" `T.isInfixOf` actionTitle ]
executeCodeAction fixTypo
contentAfterAction <- documentContents doc
let expectedContentAfterAction = T.unlines
[ "module Testing where"
, "monus :: Int -> Int"
, "monus x y = max 0 (x - y)"
, "foo x y = x `monus` y"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
]
typeWildCardActionTests :: TestTree
typeWildCardActionTests = testGroup "type wildcard actions"
[ testSession "global signature" $ do
let content = T.unlines
[ "module Testing where"
, "func :: _"
, "func x = x"
]
doc <- createDoc "Testing.hs" "haskell" content
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 2 1) (Position 2 10))
let [addSignature] = [action | CACodeAction action@CodeAction { _title = actionTitle } <- actionsOrCommands
, "Use type signature" `T.isInfixOf` actionTitle
]
executeCodeAction addSignature
contentAfterAction <- documentContents doc
let expectedContentAfterAction = T.unlines
[ "module Testing where"
, "func :: (p -> p)"
, "func x = x"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
, testSession "multi-line message" $ do
let content = T.unlines
[ "module Testing where"
, "func :: _"
, "func x y = x + y"
]
doc <- createDoc "Testing.hs" "haskell" content
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 2 1) (Position 2 10))
let [addSignature] = [action | CACodeAction action@CodeAction { _title = actionTitle } <- actionsOrCommands
, "Use type signature" `T.isInfixOf` actionTitle
]
executeCodeAction addSignature
contentAfterAction <- documentContents doc
let expectedContentAfterAction = T.unlines
[ "module Testing where"
, "func :: (Integer -> Integer -> Integer)"
, "func x y = x + y"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
, testSession "local signature" $ do
let content = T.unlines
[ "module Testing where"
, "func :: Int -> Int"
, "func x ="
, " let y :: _"
, " y = x * 2"
, " in y"
]
doc <- createDoc "Testing.hs" "haskell" content
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 4 1) (Position 4 10))
let [addSignature] = [action | CACodeAction action@CodeAction { _title = actionTitle } <- actionsOrCommands
, "Use type signature" `T.isInfixOf` actionTitle
]
executeCodeAction addSignature
contentAfterAction <- documentContents doc
let expectedContentAfterAction = T.unlines
[ "module Testing where"
, "func :: Int -> Int"
, "func x ="
, " let y :: (Int)"
, " y = x * 2"
, " in y"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
]
removeImportTests :: TestTree
removeImportTests = testGroup "remove import actions"
[ testSession "redundant" $ do
let contentA = T.unlines
[ "module ModuleA where"
]
_docA <- createDoc "ModuleA.hs" "haskell" contentA
let contentB = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import ModuleA"
, "stuffB :: Integer"
, "stuffB = 123"
]
docB <- createDoc "ModuleB.hs" "haskell" contentB
_ <- waitForDiagnostics
[CACodeAction action@CodeAction { _title = actionTitle }]
<- getCodeActions docB (Range (Position 2 0) (Position 2 5))
liftIO $ "Remove import" @=? actionTitle
executeCodeAction action
contentAfterAction <- documentContents docB
let expectedContentAfterAction = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "stuffB :: Integer"
, "stuffB = 123"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
, testSession "qualified redundant" $ do
let contentA = T.unlines
[ "module ModuleA where"
]
_docA <- createDoc "ModuleA.hs" "haskell" contentA
let contentB = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import qualified ModuleA"
, "stuffB :: Integer"
, "stuffB = 123"
]
docB <- createDoc "ModuleB.hs" "haskell" contentB
_ <- waitForDiagnostics
[CACodeAction action@CodeAction { _title = actionTitle }]
<- getCodeActions docB (Range (Position 2 0) (Position 2 5))
liftIO $ "Remove import" @=? actionTitle
executeCodeAction action
contentAfterAction <- documentContents docB
let expectedContentAfterAction = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "stuffB :: Integer"
, "stuffB = 123"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
, testSession "redundant binding" $ do
let contentA = T.unlines
[ "module ModuleA where"
, "stuffA = False"
, "stuffB :: Integer"
, "stuffB = 123"
, "stuffC = ()"
]
_docA <- createDoc "ModuleA.hs" "haskell" contentA
let contentB = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import ModuleA (stuffA, stuffB, stuffC, stuffA)"
, "main = print stuffB"
]
docB <- createDoc "ModuleB.hs" "haskell" contentB
_ <- waitForDiagnostics
[CACodeAction action@CodeAction { _title = actionTitle }]
<- getCodeActions docB (Range (Position 2 0) (Position 2 5))
liftIO $ "Remove stuffA, stuffC from import" @=? actionTitle
executeCodeAction action
contentAfterAction <- documentContents docB
let expectedContentAfterAction = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import ModuleA (stuffB)"
, "main = print stuffB"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
, testSession "redundant operator" $ do
let contentA = T.unlines
[ "module ModuleA where"
, "a !! b = a"
, "a <?> b = a"
, "stuffB :: Integer"
, "stuffB = 123"
]
_docA <- createDoc "ModuleA.hs" "haskell" contentA
let contentB = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import qualified ModuleA as A ((<?>), stuffB, (!!))"
, "main = print A.stuffB"
]
docB <- createDoc "ModuleB.hs" "haskell" contentB
_ <- waitForDiagnostics
[CACodeAction action@CodeAction { _title = actionTitle }]
<- getCodeActions docB (Range (Position 2 0) (Position 2 5))
liftIO $ "Remove !!, <?> from import" @=? actionTitle
executeCodeAction action
contentAfterAction <- documentContents docB
let expectedContentAfterAction = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import qualified ModuleA as A (stuffB)"
, "main = print A.stuffB"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
, testSession "redundant all import" $ do
2020-01-02 19:31:51 +03:00
let contentA = T.unlines
[ "module ModuleA where"
, "data A = A"
, "stuffB :: Integer"
, "stuffB = 123"
]
_docA <- createDoc "ModuleA.hs" "haskell" contentA
2020-01-02 19:31:51 +03:00
let contentB = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import ModuleA (A(..), stuffB)"
, "main = print stuffB"
]
docB <- createDoc "ModuleB.hs" "haskell" contentB
2020-01-02 19:31:51 +03:00
_ <- waitForDiagnostics
[CACodeAction action@CodeAction { _title = actionTitle }]
<- getCodeActions docB (Range (Position 2 0) (Position 2 5))
liftIO $ "Remove A from import" @=? actionTitle
executeCodeAction action
contentAfterAction <- documentContents docB
let expectedContentAfterAction = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import ModuleA (stuffB)"
, "main = print stuffB"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
, testSession "redundant constructor import" $ do
let contentA = T.unlines
[ "module ModuleA where"
, "data D = A | B"
, "data E = F"
]
_docA <- createDoc "ModuleA.hs" "haskell" contentA
let contentB = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import ModuleA (D(A,B), E(F))"
, "main = B"
]
docB <- createDoc "ModuleB.hs" "haskell" contentB
_ <- waitForDiagnostics
[CACodeAction action@CodeAction { _title = actionTitle }]
<- getCodeActions docB (Range (Position 2 0) (Position 2 5))
liftIO $ "Remove A, E, F from import" @=? actionTitle
executeCodeAction action
contentAfterAction <- documentContents docB
let expectedContentAfterAction = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import ModuleA (D(B))"
, "main = B"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
, testSession "import containing the identifier Strict" $ do
let contentA = T.unlines
[ "module Strict where"
]
_docA <- createDoc "Strict.hs" "haskell" contentA
let contentB = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
, "import Strict"
]
docB <- createDoc "ModuleB.hs" "haskell" contentB
_ <- waitForDiagnostics
[CACodeAction action@CodeAction { _title = actionTitle }]
<- getCodeActions docB (Range (Position 2 0) (Position 2 5))
liftIO $ "Remove import" @=? actionTitle
executeCodeAction action
contentAfterAction <- documentContents docB
let expectedContentAfterAction = T.unlines
[ "{-# OPTIONS_GHC -Wunused-imports #-}"
, "module ModuleB where"
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
]
extendImportTests :: TestTree
extendImportTests = testGroup "extend import actions"
[ testSession "extend single line import with value" $ template
(T.unlines
[ "module ModuleA where"
, "stuffA :: Double"
, "stuffA = 0.00750"
, "stuffB :: Integer"
, "stuffB = 123"
])
(T.unlines
[ "module ModuleB where"
, "import ModuleA as A (stuffB)"
, "main = print (stuffA, stuffB)"
])
(Range (Position 3 17) (Position 3 18))
"Add stuffA to the import list of ModuleA"
(T.unlines
[ "module ModuleB where"
, "import ModuleA as A (stuffA, stuffB)"
, "main = print (stuffA, stuffB)"
])
, testSession "extend single line import with operator" $ template
(T.unlines
[ "module ModuleA where"
, "(.*) :: Integer -> Integer -> Integer"
, "x .* y = x * y"
, "stuffB :: Integer"
, "stuffB = 123"
])
(T.unlines
[ "module ModuleB where"
, "import ModuleA as A (stuffB)"
, "main = print (stuffB .* stuffB)"
])
(Range (Position 3 17) (Position 3 18))
"Add .* to the import list of ModuleA"
(T.unlines
[ "module ModuleB where"
, "import ModuleA as A ((.*), stuffB)"
, "main = print (stuffB .* stuffB)"
])
, testSession "extend single line import with type" $ template
(T.unlines
[ "module ModuleA where"
, "type A = Double"
])
(T.unlines
[ "module ModuleB where"
, "import ModuleA ()"
, "b :: A"
, "b = 0"
])
(Range (Position 2 5) (Position 2 5))
"Add A to the import list of ModuleA"
(T.unlines
[ "module ModuleB where"
, "import ModuleA (A)"
, "b :: A"
, "b = 0"
])
, (`xfail` "known broken") $ testSession "extend single line import with constructor" $ template
(T.unlines
[ "module ModuleA where"
, "data A = Constructor"
])
(T.unlines
[ "module ModuleB where"
, "import ModuleA (A)"
, "b :: A"
, "b = Constructor"
])
(Range (Position 2 5) (Position 2 5))
"Add Constructor to the import list of ModuleA"
(T.unlines
[ "module ModuleB where"
, "import ModuleA (A(Constructor))"
, "b :: A"
, "b = Constructor"
])
, testSession "extend single line qualified import with value" $ template
(T.unlines
[ "module ModuleA where"
, "stuffA :: Double"
, "stuffA = 0.00750"
, "stuffB :: Integer"
, "stuffB = 123"
])
(T.unlines
[ "module ModuleB where"
, "import qualified ModuleA as A (stuffB)"
, "main = print (A.stuffA, A.stuffB)"
])
(Range (Position 3 17) (Position 3 18))
"Add stuffA to the import list of ModuleA"
(T.unlines
[ "module ModuleB where"
, "import qualified ModuleA as A (stuffA, stuffB)"
, "main = print (A.stuffA, A.stuffB)"
])
, testSession "extend multi line import with value" $ template
(T.unlines
[ "module ModuleA where"
, "stuffA :: Double"
, "stuffA = 0.00750"
, "stuffB :: Integer"
, "stuffB = 123"
])
(T.unlines
[ "module ModuleB where"
, "import ModuleA (stuffB"
, " )"
, "main = print (stuffA, stuffB)"
])
(Range (Position 3 17) (Position 3 18))
"Add stuffA to the import list of ModuleA"
(T.unlines
[ "module ModuleB where"
, "import ModuleA (stuffA, stuffB"
, " )"
, "main = print (stuffA, stuffB)"
])
]
where
template contentA contentB range expectedAction expectedContentB = do
_docA <- createDoc "ModuleA.hs" "haskell" contentA
docB <- createDoc "ModuleB.hs" "haskell" contentB
_ <- waitForDiagnostics
CACodeAction action@CodeAction { _title = actionTitle } : _
<- sortOn (\(CACodeAction CodeAction{_title=x}) -> x) <$>
getCodeActions docB range
liftIO $ expectedAction @=? actionTitle
executeCodeAction action
contentAfterAction <- documentContents docB
liftIO $ expectedContentB @=? contentAfterAction
suggestImportTests :: TestTree
suggestImportTests = testGroup "suggest import actions"
[ testGroup "Dont want suggestion"
[ -- extend import
test False ["Data.List.NonEmpty ()"] "f = nonEmpty" [] "import Data.List.NonEmpty (nonEmpty)"
-- data constructor
, test False [] "f = First" [] "import Data.Monoid (First)"
-- internal module
, test False [] "f :: Typeable a => a" ["f = undefined"] "import Data.Typeable.Internal (Typeable)"
-- package not in scope
, test False [] "f = quickCheck" [] "import Test.QuickCheck (quickCheck)"
]
, testGroup "want suggestion"
[ test True [] "f = nonEmpty" [] "import Data.List.NonEmpty (nonEmpty)"
, test True [] "f = (:|)" [] "import Data.List.NonEmpty (NonEmpty((:|)))"
, test True [] "f :: Natural" ["f = undefined"] "import Numeric.Natural (Natural)"
, test True [] "f :: NonEmpty ()" ["f = () :| []"] "import Data.List.NonEmpty (NonEmpty)"
, test True [] "f = First" [] "import Data.Monoid (First(First))"
, test True [] "f = Endo" [] "import Data.Monoid (Endo(Endo))"
, test True [] "f = Version" [] "import Data.Version (Version(Version))"
, test True [] "f ExitSuccess = ()" [] "import System.Exit (ExitCode(ExitSuccess))"
, test True [] "f = AssertionFailed" [] "import Control.Exception (AssertionFailed(AssertionFailed))"
, test True ["Prelude"] "f = nonEmpty" [] "import Data.List.NonEmpty (nonEmpty)"
, test True [] "f :: Alternative f => f ()" ["f = undefined"] "import Control.Applicative (Alternative)"
, test True [] "f = empty" [] "import Control.Applicative (Alternative(empty))"
, test True [] "f = (&)" [] "import Data.Function ((&))"
, test True [] "f = NE.nonEmpty" [] "import qualified Data.List.NonEmpty as NE"
, test True [] "f :: Typeable a => a" ["f = undefined"] "import Data.Typeable (Typeable)"
, test True [] "f = pack" [] "import Data.Text (pack)"
, test True [] "f :: Text" ["f = undefined"] "import Data.Text (Text)"
, test True [] "f = [] & id" [] "import Data.Function ((&))"
, test True [] "f = (&) [] id" [] "import Data.Function ((&))"
]
]
where
test wanted imps def other newImp = testSession' (T.unpack def) $ \dir -> do
let before = T.unlines $ "module A where" : ["import " <> x | x <- imps] ++ def : other
after = T.unlines $ "module A where" : ["import " <> x | x <- imps] ++ [newImp] ++ def : other
cradle = "cradle: {direct: {arguments: [-hide-all-packages, -package, base, -package, text, -package-env, -]}}"
liftIO $ writeFileUTF8 (dir </> "hie.yaml") cradle
doc <- createDoc "Test.hs" "haskell" before
_diags <- waitForDiagnostics
let defLine = length imps + 1
range = Range (Position defLine 0) (Position defLine maxBound)
actions <- getCodeActions doc range
if wanted
then do
action <- liftIO $ pickActionWithTitle newImp actions
executeCodeAction action
contentAfterAction <- documentContents doc
liftIO $ after @=? contentAfterAction
else
liftIO $ [_title | CACodeAction CodeAction{_title} <- actions, _title == newImp ] @?= []
addExtensionTests :: TestTree
addExtensionTests = testGroup "add language extension actions"
[ testSession "add NamedFieldPuns language extension" $ template
(T.unlines
[ "module Module where"
, ""
, "data A = A { getA :: Bool }"
, ""
, "f :: A -> Bool"
, "f A { getA } = getA"
])
(Range (Position 0 0) (Position 0 0))
"Add NamedFieldPuns extension"
(T.unlines
[ "{-# LANGUAGE NamedFieldPuns #-}"
, "module Module where"
, ""
, "data A = A { getA :: Bool }"
, ""
, "f :: A -> Bool"
, "f A { getA } = getA"
])
, testSession "add RecordWildCards language extension" $ template
(T.unlines
[ "module Module where"
, ""
, "data A = A { getA :: Bool }"
, ""
, "f :: A -> Bool"
, "f A { .. } = getA"
])
(Range (Position 0 0) (Position 0 0))
"Add RecordWildCards extension"
(T.unlines
[ "{-# LANGUAGE RecordWildCards #-}"
, "module Module where"
, ""
, "data A = A { getA :: Bool }"
, ""
, "f :: A -> Bool"
, "f A { .. } = getA"
])
]
where
template initialContent range expectedAction expectedContents = do
doc <- createDoc "Module.hs" "haskell" initialContent
_ <- waitForDiagnostics
CACodeAction action@CodeAction { _title = actionTitle } : _
<- sortOn (\(CACodeAction CodeAction{_title=x}) -> x) <$>
getCodeActions doc range
liftIO $ expectedAction @=? actionTitle
executeCodeAction action
contentAfterAction <- documentContents doc
liftIO $ expectedContents @=? contentAfterAction
insertNewDefinitionTests :: TestTree
insertNewDefinitionTests = testGroup "insert new definition actions"
[ testSession "insert new function definition" $ do
let txtB =
["foo True = select [True]"
, ""
,"foo False = False"
]
txtB' =
[""
,"someOtherCode = ()"
]
docB <- createDoc "ModuleB.hs" "haskell" (T.unlines $ txtB ++ txtB')
_ <- waitForDiagnostics
CACodeAction action@CodeAction { _title = actionTitle } : _
<- sortOn (\(CACodeAction CodeAction{_title=x}) -> x) <$>
getCodeActions docB (R 1 0 1 50)
liftIO $ actionTitle @?= "Define select :: [Bool] -> Bool"
executeCodeAction action
contentAfterAction <- documentContents docB
liftIO $ contentAfterAction @?= T.unlines (txtB ++
[ ""
, "select :: [Bool] -> Bool"
, "select = error \"not implemented\""
]
++ txtB')
, testSession "define a hole" $ do
let txtB =
["foo True = _select [True]"
, ""
,"foo False = False"
]
txtB' =
[""
,"someOtherCode = ()"
]
docB <- createDoc "ModuleB.hs" "haskell" (T.unlines $ txtB ++ txtB')
_ <- waitForDiagnostics
CACodeAction action@CodeAction { _title = actionTitle } : _
<- sortOn (\(CACodeAction CodeAction{_title=x}) -> x) <$>
getCodeActions docB (R 1 0 1 50)
liftIO $ actionTitle @?= "Define select :: [Bool] -> Bool"
executeCodeAction action
contentAfterAction <- documentContents docB
liftIO $ contentAfterAction @?= T.unlines (
["foo True = select [True]"
, ""
,"foo False = False"
, ""
, "select :: [Bool] -> Bool"
, "select = error \"not implemented\""
]
++ txtB')
]
deleteUnusedDefinitionTests :: TestTree
deleteUnusedDefinitionTests = testGroup "delete unused definition action"
[ testSession "delete unused top level binding" $
testFor
(T.unlines [ "{-# OPTIONS_GHC -Wunused-top-binds #-}"
, "module A (some) where"
, ""
, "f :: Int -> Int"
, "f 1 = let a = 1"
, " in a"
, "f 2 = 2"
, ""
, "some = ()"
])
(4, 0)
"Delete f"
(T.unlines [
"{-# OPTIONS_GHC -Wunused-top-binds #-}"
, "module A (some) where"
, ""
, "some = ()"
])
, testSession "delete unused top level binding defined in infix form" $
testFor
(T.unlines [ "{-# OPTIONS_GHC -Wunused-top-binds #-}"
, "module A (some) where"
, ""
, "myPlus :: Int -> Int -> Int"
, "a `myPlus` b = a + b"
, ""
, "some = ()"
])
(4, 2)
"Delete myPlus"
(T.unlines [
"{-# OPTIONS_GHC -Wunused-top-binds #-}"
, "module A (some) where"
, ""
, "some = ()"
])
]
where
testFor source pos expectedTitle expectedResult = do
docId <- createDoc "A.hs" "haskell" source
expectDiagnostics [ ("A.hs", [(DsWarning, pos, "not used")]) ]
(action, title) <- extractCodeAction docId "Delete"
liftIO $ title @?= expectedTitle
executeCodeAction action
contentAfterAction <- documentContents docId
liftIO $ contentAfterAction @?= expectedResult
extractCodeAction docId actionPrefix = do
[action@CodeAction { _title = actionTitle }] <- findCodeActionsByPrefix docId (R 0 0 0 0) [actionPrefix]
return (action, actionTitle)
addTypeAnnotationsToLiteralsTest :: TestTree
addTypeAnnotationsToLiteralsTest = testGroup "add type annotations to literals to satisfy contraints"
[
testSession "add default type to satisfy one contraint" $
testFor
(T.unlines [ "{-# OPTIONS_GHC -Wtype-defaults #-}"
, "module A () where"
, ""
, "f = 1"
])
[ (DsWarning, (3, 4), "Defaulting the following constraint") ]
"Add type annotation Integer to 1"
(T.unlines [ "{-# OPTIONS_GHC -Wtype-defaults #-}"
, "module A () where"
, ""
, "f = (1 :: Integer)"
])
, testSession "add default type to satisfy one contraint with duplicate literals" $
testFor
(T.unlines [ "{-# OPTIONS_GHC -Wtype-defaults #-}"
, "{-# LANGUAGE OverloadedStrings #-}"
, "module A () where"
, ""
, "import Debug.Trace"
, ""
, "f = seq \"debug\" traceShow \"debug\""
])
[ (DsWarning, (6, 8), "Defaulting the following constraint")
, (DsWarning, (6, 16), "Defaulting the following constraint")
]
"Add type annotation [Char] to \"debug\""
(T.unlines [ "{-# OPTIONS_GHC -Wtype-defaults #-}"
, "{-# LANGUAGE OverloadedStrings #-}"
, "module A () where"
, ""
, "import Debug.Trace"
, ""
, "f = seq (\"debug\" :: [Char]) traceShow \"debug\""
])
, testSession "add default type to satisfy two contraints" $
testFor
(T.unlines [ "{-# OPTIONS_GHC -Wtype-defaults #-}"
, "{-# LANGUAGE OverloadedStrings #-}"
, "module A () where"
, ""
, "import Debug.Trace"
, ""
, "f a = traceShow \"debug\" a"
])
[ (DsWarning, (6, 6), "Defaulting the following constraint") ]
"Add type annotation [Char] to \"debug\""
(T.unlines [ "{-# OPTIONS_GHC -Wtype-defaults #-}"
, "{-# LANGUAGE OverloadedStrings #-}"
, "module A () where"
, ""
, "import Debug.Trace"
, ""
, "f a = traceShow (\"debug\" :: [Char]) a"
])
, testSession "add default type to satisfy two contraints with duplicate literals" $
testFor
(T.unlines [ "{-# OPTIONS_GHC -Wtype-defaults #-}"
, "{-# LANGUAGE OverloadedStrings #-}"
, "module A () where"
, ""
, "import Debug.Trace"
, ""
, "f = seq (\"debug\" :: [Char]) (seq (\"debug\" :: [Char]) (traceShow \"debug\"))"
])
[ (DsWarning, (6, 54), "Defaulting the following constraint") ]
"Add type annotation [Char] to \"debug\""
(T.unlines [ "{-# OPTIONS_GHC -Wtype-defaults #-}"
, "{-# LANGUAGE OverloadedStrings #-}"
, "module A () where"
, ""
, "import Debug.Trace"
, ""
, "f = seq (\"debug\" :: [Char]) (seq (\"debug\" :: [Char]) (traceShow (\"debug\" :: [Char])))"
])
]
where
testFor source diag expectedTitle expectedResult = do
docId <- createDoc "A.hs" "haskell" source
expectDiagnostics [ ("A.hs", diag) ]
(action, title) <- extractCodeAction docId "Add type annotation"
liftIO $ title @?= expectedTitle
executeCodeAction action
contentAfterAction <- documentContents docId
liftIO $ contentAfterAction @?= expectedResult
extractCodeAction docId actionPrefix = do
[action@CodeAction { _title = actionTitle }] <- findCodeActionsByPrefix docId (R 0 0 0 0) [actionPrefix]
return (action, actionTitle)
fixConstructorImportTests :: TestTree
fixConstructorImportTests = testGroup "fix import actions"
[ testSession "fix constructor import" $ template
(T.unlines
[ "module ModuleA where"
, "data A = Constructor"
])
(T.unlines
[ "module ModuleB where"
, "import ModuleA(Constructor)"
])
(Range (Position 1 10) (Position 1 11))
"Fix import of A(Constructor)"
(T.unlines
[ "module ModuleB where"
, "import ModuleA(A(Constructor))"
])
]
where
template contentA contentB range expectedAction expectedContentB = do
_docA <- createDoc "ModuleA.hs" "haskell" contentA
docB <- createDoc "ModuleB.hs" "haskell" contentB
_diags <- waitForDiagnostics
CACodeAction action@CodeAction { _title = actionTitle } : _
<- sortOn (\(CACodeAction CodeAction{_title=x}) -> x) <$>
getCodeActions docB range
liftIO $ expectedAction @=? actionTitle
executeCodeAction action
contentAfterAction <- documentContents docB
liftIO $ expectedContentB @=? contentAfterAction
importRenameActionTests :: TestTree
importRenameActionTests = testGroup "import rename actions"
[ testSession "Data.Mape -> Data.Map" $ check "Map"
, testSession "Data.Mape -> Data.Maybe" $ check "Maybe" ] where
check modname = do
let content = T.unlines
[ "module Testing where"
, "import Data.Mape"
]
doc <- createDoc "Testing.hs" "haskell" content
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 2 8) (Position 2 16))
let [changeToMap] = [action | CACodeAction action@CodeAction{ _title = actionTitle } <- actionsOrCommands, ("Data." <> modname) `T.isInfixOf` actionTitle ]
executeCodeAction changeToMap
contentAfterAction <- documentContents doc
let expectedContentAfterAction = T.unlines
[ "module Testing where"
, "import Data." <> modname
]
liftIO $ expectedContentAfterAction @=? contentAfterAction
fillTypedHoleTests :: TestTree
fillTypedHoleTests = let
sourceCode :: T.Text -> T.Text -> T.Text -> T.Text
sourceCode a b c = T.unlines
[ "module Testing where"
, ""
, "globalConvert :: Int -> String"
, "globalConvert = undefined"
, ""
, "globalInt :: Int"
, "globalInt = 3"
, ""
, "bar :: Int -> Int -> String"
, "bar n parameterInt = " <> a <> " (n + " <> b <> " + " <> c <> ") where"
, " localConvert = (flip replicate) 'x'"
]
check :: T.Text -> T.Text -> T.Text -> T.Text -> T.Text -> T.Text -> T.Text -> TestTree
check actionTitle
oldA oldB oldC
newA newB newC = testSession (T.unpack actionTitle) $ do
let originalCode = sourceCode oldA oldB oldC
let expectedCode = sourceCode newA newB newC
doc <- createDoc "Testing.hs" "haskell" originalCode
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 9 0) (Position 9 maxBound))
chosenAction <- liftIO $ pickActionWithTitle actionTitle actionsOrCommands
executeCodeAction chosenAction
modifiedCode <- documentContents doc
liftIO $ expectedCode @=? modifiedCode
in
testGroup "fill typed holes"
[ check "replace hole `_` with show"
"_" "n" "n"
"show" "n" "n"
, check "replace hole `_` with globalConvert"
"_" "n" "n"
"globalConvert" "n" "n"
#if MIN_GHC_API_VERSION(8,6,0)
, check "replace hole `_convertme` with localConvert"
"_convertme" "n" "n"
"localConvert" "n" "n"
#endif
, check "replace hole `_b` with globalInt"
"_a" "_b" "_c"
"_a" "globalInt" "_c"
, check "replace hole `_c` with globalInt"
"_a" "_b" "_c"
"_a" "_b" "globalInt"
#if MIN_GHC_API_VERSION(8,6,0)
, check "replace hole `_c` with parameterInt"
"_a" "_b" "_c"
"_a" "_b" "parameterInt"
#endif
]
addInstanceConstraintTests :: TestTree
addInstanceConstraintTests = let
missingConstraintSourceCode :: Maybe T.Text -> T.Text
missingConstraintSourceCode mConstraint =
let constraint = maybe "" (<> " => ") mConstraint
in T.unlines
[ "module Testing where"
, ""
, "data Wrap a = Wrap a"
, ""
, "instance " <> constraint <> "Eq (Wrap a) where"
, " (Wrap x) == (Wrap y) = x == y"
]
incompleteConstraintSourceCode :: Maybe T.Text -> T.Text
incompleteConstraintSourceCode mConstraint =
let constraint = maybe "Eq a" (\c -> "(Eq a, " <> c <> ")") mConstraint
in T.unlines
[ "module Testing where"
, ""
, "data Pair a b = Pair a b"
, ""
, "instance " <> constraint <> " => Eq (Pair a b) where"
, " (Pair x y) == (Pair x' y') = x == x' && y == y'"
]
incompleteConstraintSourceCode2 :: Maybe T.Text -> T.Text
incompleteConstraintSourceCode2 mConstraint =
let constraint = maybe "(Eq a, Eq b)" (\c -> "(Eq a, Eq b, " <> c <> ")") mConstraint
in T.unlines
[ "module Testing where"
, ""
, "data Three a b c = Three a b c"
, ""
, "instance " <> constraint <> " => Eq (Three a b c) where"
, " (Three x y z) == (Three x' y' z') = x == x' && y == y' && z == z'"
]
check :: T.Text -> T.Text -> T.Text -> TestTree
check actionTitle originalCode expectedCode = testSession (T.unpack actionTitle) $ do
doc <- createDoc "Testing.hs" "haskell" originalCode
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 6 0) (Position 6 68))
chosenAction <- liftIO $ pickActionWithTitle actionTitle actionsOrCommands
executeCodeAction chosenAction
modifiedCode <- documentContents doc
liftIO $ expectedCode @=? modifiedCode
in testGroup "add instance constraint"
[ check
"Add `Eq a` to the context of the instance declaration"
(missingConstraintSourceCode Nothing)
(missingConstraintSourceCode $ Just "Eq a")
, check
"Add `Eq b` to the context of the instance declaration"
(incompleteConstraintSourceCode Nothing)
(incompleteConstraintSourceCode $ Just "Eq b")
, check
"Add `Eq c` to the context of the instance declaration"
(incompleteConstraintSourceCode2 Nothing)
(incompleteConstraintSourceCode2 $ Just "Eq c")
]
addFunctionConstraintTests :: TestTree
addFunctionConstraintTests = let
missingConstraintSourceCode :: Maybe T.Text -> T.Text
missingConstraintSourceCode mConstraint =
let constraint = maybe "" (<> " => ") mConstraint
in T.unlines
[ "module Testing where"
, ""
, "eq :: " <> constraint <> "a -> a -> Bool"
, "eq x y = x == y"
]
incompleteConstraintSourceCode :: Maybe T.Text -> T.Text
incompleteConstraintSourceCode mConstraint =
let constraint = maybe "Eq a" (\c -> "(Eq a, " <> c <> ")") mConstraint
in T.unlines
[ "module Testing where"
, ""
, "data Pair a b = Pair a b"
, ""
, "eq :: " <> constraint <> " => Pair a b -> Pair a b -> Bool"
, "eq (Pair x y) (Pair x' y') = x == x' && y == y'"
]
incompleteConstraintSourceCode2 :: Maybe T.Text -> T.Text
incompleteConstraintSourceCode2 mConstraint =
let constraint = maybe "(Eq a, Eq b)" (\c -> "(Eq a, Eq b, " <> c <> ")") mConstraint
in T.unlines
[ "module Testing where"
, ""
, "data Three a b c = Three a b c"
, ""
, "eq :: " <> constraint <> " => Three a b c -> Three a b c -> Bool"
, "eq (Three x y z) (Three x' y' z') = x == x' && y == y' && z == z'"
]
check :: T.Text -> T.Text -> T.Text -> TestTree
check actionTitle originalCode expectedCode = testSession (T.unpack actionTitle) $ do
doc <- createDoc "Testing.hs" "haskell" originalCode
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 6 0) (Position 6 maxBound))
chosenAction <- liftIO $ pickActionWithTitle actionTitle actionsOrCommands
executeCodeAction chosenAction
modifiedCode <- documentContents doc
liftIO $ expectedCode @=? modifiedCode
in testGroup "add function constraint"
[ check
"Add `Eq a` to the context of the type signature for `eq`"
(missingConstraintSourceCode Nothing)
(missingConstraintSourceCode $ Just "Eq a")
, check
"Add `Eq b` to the context of the type signature for `eq`"
(incompleteConstraintSourceCode Nothing)
(incompleteConstraintSourceCode $ Just "Eq b")
, check
"Add `Eq c` to the context of the type signature for `eq`"
(incompleteConstraintSourceCode2 Nothing)
(incompleteConstraintSourceCode2 $ Just "Eq c")
]
removeRedundantConstraintsTests :: TestTree
removeRedundantConstraintsTests = let
header =
[ "{-# OPTIONS_GHC -Wredundant-constraints #-}"
, "module Testing where"
, ""
]
redundantConstraintsCode :: Maybe T.Text -> T.Text
redundantConstraintsCode mConstraint =
let constraint = maybe "" (\c -> "" <> c <> " => ") mConstraint
in T.unlines $ header <>
[ "foo :: " <> constraint <> "a -> a"
, "foo = id"
]
redundantMixedConstraintsCode :: Maybe T.Text -> T.Text
redundantMixedConstraintsCode mConstraint =
let constraint = maybe "(Num a, Eq a)" (\c -> "(Num a, Eq a, " <> c <> ")") mConstraint
in T.unlines $ header <>
[ "foo :: " <> constraint <> " => a -> Bool"
, "foo x = x == 1"
]
typeSignatureSpaces :: T.Text
typeSignatureSpaces = T.unlines $ header <>
[ "foo :: (Num a, Eq a, Monoid a) => a -> Bool"
, "foo x = x == 1"
]
typeSignatureMultipleLines :: T.Text
typeSignatureMultipleLines = T.unlines $ header <>
[ "foo :: (Num a, Eq a, Monoid a)"
, "=> a -> Bool"
, "foo x = x == 1"
]
check :: T.Text -> T.Text -> T.Text -> TestTree
check actionTitle originalCode expectedCode = testSession (T.unpack actionTitle) $ do
doc <- createDoc "Testing.hs" "haskell" originalCode
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 4 0) (Position 4 maxBound))
chosenAction <- liftIO $ pickActionWithTitle actionTitle actionsOrCommands
executeCodeAction chosenAction
modifiedCode <- documentContents doc
liftIO $ expectedCode @=? modifiedCode
checkPeculiarFormatting :: String -> T.Text -> TestTree
checkPeculiarFormatting title code = testSession title $ do
doc <- createDoc "Testing.hs" "haskell" code
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 4 0) (Position 4 maxBound))
liftIO $ assertBool "Found some actions" (null actionsOrCommands)
in testGroup "remove redundant function constraints"
[ check
"Remove redundant constraint `Eq a` from the context of the type signature for `foo`"
(redundantConstraintsCode $ Just "Eq a")
(redundantConstraintsCode Nothing)
, check
"Remove redundant constraints `(Eq a, Monoid a)` from the context of the type signature for `foo`"
(redundantConstraintsCode $ Just "(Eq a, Monoid a)")
(redundantConstraintsCode Nothing)
, check
"Remove redundant constraints `(Monoid a, Show a)` from the context of the type signature for `foo`"
(redundantMixedConstraintsCode $ Just "Monoid a, Show a")
(redundantMixedConstraintsCode Nothing)
, checkPeculiarFormatting
"should do nothing when constraints contain an arbitrary number of spaces"
typeSignatureSpaces
, checkPeculiarFormatting
"should do nothing when constraints contain line feeds"
typeSignatureMultipleLines
]
addSigActionTests :: TestTree
addSigActionTests = let
header = "{-# OPTIONS_GHC -Wmissing-signatures -Wmissing-pattern-synonym-signatures #-}"
moduleH = "{-# LANGUAGE PatternSynonyms #-}\nmodule Sigs where"
before def = T.unlines [header, moduleH, def]
after' def sig = T.unlines [header, moduleH, sig, def]
def >:: sig = testSession (T.unpack def) $ do
let originalCode = before def
let expectedCode = after' def sig
doc <- createDoc "Sigs.hs" "haskell" originalCode
_ <- waitForDiagnostics
actionsOrCommands <- getCodeActions doc (Range (Position 3 1) (Position 3 maxBound))
chosenAction <- liftIO $ pickActionWithTitle ("add signature: " <> sig) actionsOrCommands
executeCodeAction chosenAction
modifiedCode <- documentContents doc
liftIO $ expectedCode @=? modifiedCode
in
testGroup "add signature"
[ "abc = True" >:: "abc :: Bool"
, "foo a b = a + b" >:: "foo :: Num a => a -> a -> a"
, "bar a b = show $ a + b" >:: "bar :: (Show a, Num a) => a -> a -> String"
, "(!!!) a b = a > b" >:: "(!!!) :: Ord a => a -> a -> Bool"
, "a >>>> b = a + b" >:: "(>>>>) :: Num a => a -> a -> a"
, "a `haha` b = a b" >:: "haha :: (t1 -> t2) -> t1 -> t2"
, "pattern Some a = Just a" >:: "pattern Some :: a -> Maybe a"
]
addSigLensesTests :: TestTree
addSigLensesTests = let
missing = "{-# OPTIONS_GHC -Wmissing-signatures -Wmissing-pattern-synonym-signatures -Wunused-matches #-}"
notMissing = "{-# OPTIONS_GHC -Wunused-matches #-}"
moduleH = "{-# LANGUAGE PatternSynonyms #-}\nmodule Sigs where"
other = T.unlines ["f :: Integer -> Integer", "f x = 3"]
before withMissing def
= T.unlines $ (if withMissing then (missing :) else (notMissing :)) [moduleH, def, other]
after' withMissing def sig
= T.unlines $ (if withMissing then (missing :) else (notMissing :)) [moduleH, sig, def, other]
sigSession withMissing def sig = testSession (T.unpack def) $ do
let originalCode = before withMissing def
let expectedCode = after' withMissing def sig
doc <- createDoc "Sigs.hs" "haskell" originalCode
[CodeLens {_command = Just c}] <- getCodeLenses doc
executeCommand c
modifiedCode <- getDocumentEdit doc
liftIO $ expectedCode @=? modifiedCode
in
testGroup "add signature"
[ testGroup title
[ sigSession enableWarnings "abc = True" "abc :: Bool"
, sigSession enableWarnings "foo a b = a + b" "foo :: Num a => a -> a -> a"
, sigSession enableWarnings "bar a b = show $ a + b" "bar :: (Show a, Num a) => a -> a -> String"
, sigSession enableWarnings "(!!!) a b = a > b" "(!!!) :: Ord a => a -> a -> Bool"
, sigSession enableWarnings "a >>>> b = a + b" "(>>>>) :: Num a => a -> a -> a"
, sigSession enableWarnings "a `haha` b = a b" "haha :: (t1 -> t2) -> t1 -> t2"
, sigSession enableWarnings "pattern Some a = Just a" "pattern Some :: a -> Maybe a"
]
| (title, enableWarnings) <-
[("with warnings enabled", True)
,("with warnings disabled", False)
]
]
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
checkDefs :: [Location] -> Session [Expect] -> Session ()
checkDefs defs mkExpectations = traverse_ check =<< mkExpectations where
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
check (ExpectRange expectedRange) = do
assertNDefinitionsFound 1 defs
assertRangeCorrect (head defs) expectedRange
check (ExpectLocation expectedLocation) = do
assertNDefinitionsFound 1 defs
liftIO $ do
canonActualLoc <- canonicalizeLocation (head defs)
canonExpectedLoc <- canonicalizeLocation expectedLocation
canonActualLoc @?= canonExpectedLoc
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
check ExpectNoDefinitions = do
assertNDefinitionsFound 0 defs
check ExpectExternFail = liftIO $ assertFailure "Expecting to fail to find in external file"
check _ = pure () -- all other expectations not relevant to getDefinition
assertNDefinitionsFound :: Int -> [a] -> Session ()
assertNDefinitionsFound n defs = liftIO $ assertEqual "number of definitions" n (length defs)
assertRangeCorrect Location{_range = foundRange} expectedRange =
liftIO $ expectedRange @=? foundRange
canonicalizeLocation :: Location -> IO Location
Interface file fixes (#645) * Add test for inconsistent diagnostics * Refactoring ModIfaceFromDisk This started as a pure refactoring to clarify the responsibilities between ModIface and ModIfaceFromDisk, but ended up having some behaviour changes: 1. Regenerate interface when checkOldIface returns something other than UpToDate. This was a bug. 2. Do not generate a diagnostic when regenerating an interface. 2. Previously we conflated stale interface with other errors, and would regenerate in both cases. Now we only regenerate in the first case. Tentative fix for #597 * Split interface tests * Always recompile modules with TH splices Tentative fix for #614 TODO support stability * Fix expectDiagnostics in MacOs * Avoid File does not exist diagnostics for interface files Fixes #642 * Clarify interface tests * hlints * Performance fixes The previous changes were 10X slower, this is 20X faster than those, so 2X faster than upstream, for some benchmarks * formatting * Fix GetModificationTime identity The answer for a GetModification query is independent of the missingFileDiagnostics field (as the diagnostics are not part of the answer) * remove stale comment * Avoid calling ghcSessionDepsDefinition twice * Apply suggestions from code review Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Code review feedback * Address review feedback https://github.com/digital-asset/ghcide/pull/645/files/49b0d9ac65399edf82a7a9cbbb8d8b5420458d8d#r443383239 * Change recomp to direct cradle Co-authored-by: Zubin Duggal <zubin@cmi.ac.in> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-23 12:01:52 +03:00
canonicalizeLocation (Location uri range) = Location <$> canonicalizeUri uri <*> pure range
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
findDefinitionAndHoverTests :: TestTree
findDefinitionAndHoverTests = let
tst (get, check) pos targetRange title = testSessionWithExtraFiles "hover" title $ \dir -> do
-- Dirty the cache to check that definitions work even in the presence of iface files
liftIO $ runInDir dir $ do
let fooPath = dir </> "Foo.hs"
fooSource <- liftIO $ readFileUtf8 fooPath
fooDoc <- createDoc fooPath "haskell" fooSource
_ <- getHover fooDoc $ Position 4 3
closeDoc fooDoc
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
doc <- openTestDataDoc (dir </> sourceFilePath)
found <- get doc pos
check found targetRange
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
checkHover :: Maybe Hover -> Session [Expect] -> Session ()
checkHover hover expectations = traverse_ check =<< expectations where
check expected =
case hover of
Nothing -> unless (expected == ExpectNoHover) $ liftIO $ assertFailure "no hover found"
Just Hover{_contents = (HoverContents MarkupContent{_value = msg})
,_range = rangeInHover } ->
case expected of
ExpectRange expectedRange -> checkHoverRange expectedRange rangeInHover msg
ExpectHoverRange expectedRange -> checkHoverRange expectedRange rangeInHover msg
ExpectHoverText snippets -> liftIO $ traverse_ (`assertFoundIn` msg) snippets
ExpectNoHover -> liftIO $ assertFailure $ "Expected no hover but got " <> show hover
_ -> pure () -- all other expectations not relevant to hover
_ -> liftIO $ assertFailure $ "test not expecting this kind of hover info" <> show hover
extractLineColFromHoverMsg :: T.Text -> [T.Text]
extractLineColFromHoverMsg = T.splitOn ":" . head . T.splitOn "*" . last . T.splitOn (sourceFileName <> ":")
checkHoverRange :: Range -> Maybe Range -> T.Text -> Session ()
checkHoverRange expectedRange rangeInHover msg =
let
lineCol = extractLineColFromHoverMsg msg
-- looks like hovers use 1-based numbering while definitions use 0-based
-- turns out that they are stored 1-based in RealSrcLoc by GHC itself.
adjust Position{_line = l, _character = c} =
Position{_line = l + 1, _character = c + 1}
in
case map (read . T.unpack) lineCol of
[l,c] -> liftIO $ (adjust $ _start expectedRange) @=? Position l c
_ -> liftIO $ assertFailure $
"expected: " <> show ("[...]" <> sourceFileName <> ":<LINE>:<COL>**[...]", Just expectedRange) <>
"\n but got: " <> show (msg, rangeInHover)
assertFoundIn :: T.Text -> T.Text -> Assertion
assertFoundIn part whole = assertBool
(T.unpack $ "failed to find: `" <> part <> "` in hover message:\n" <> whole)
(part `T.isInfixOf` whole)
sourceFilePath = T.unpack sourceFileName
sourceFileName = "GotoHover.hs"
mkFindTests tests = testGroup "get"
[ testGroup "definition" $ mapMaybe fst tests
, testGroup "hover" $ mapMaybe snd tests
, checkFileCompiles sourceFilePath
, testGroup "type-definition" typeDefinitionTests ]
typeDefinitionTests = [ tst (getTypeDefinitions, checkDefs) aaaL14 (pure tcData) "Saturated data con"
, tst (getTypeDefinitions, checkDefs) opL16 (pure [ExpectNoDefinitions]) "Polymorphic variable"]
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
test runDef runHover look expect = testM runDef runHover look (return expect)
testM runDef runHover look expect title =
( runDef $ tst def look expect title
, runHover $ tst hover look expect title ) where
def = (getDefinitions, checkDefs)
hover = (getHover , checkHover)
-- search locations expectations on results
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
fffL4 = _start fffR ; fffR = mkRange 8 4 8 7 ; fff = [ExpectRange fffR]
fffL8 = Position 12 4 ;
fffL14 = Position 18 7 ;
aaaL14 = Position 18 20 ; aaa = [mkR 11 0 11 3]
dcL7 = Position 11 11 ; tcDC = [mkR 7 23 9 16]
dcL12 = Position 16 11 ;
xtcL5 = Position 9 11 ; xtc = [ExpectExternFail, ExpectHoverText ["Int", "Defined in GHC.Types"]]
tcL6 = Position 10 11 ; tcData = [mkR 7 0 9 16, ExpectHoverText ["TypeConstructor", "GotoHover.hs:8:1"]]
vvL16 = Position 20 12 ; vv = [mkR 20 4 20 6]
opL16 = Position 20 15 ; op = [mkR 21 2 21 4]
opL18 = Position 22 22 ; opp = [mkR 22 13 22 17]
aL18 = Position 22 20 ; apmp = [mkR 22 10 22 11]
b'L19 = Position 23 13 ; bp = [mkR 23 6 23 7]
xvL20 = Position 24 8 ; xvMsg = [ExpectExternFail, ExpectHoverText ["Data.Text.pack", ":: String -> Text"]]
clL23 = Position 27 11 ; cls = [mkR 25 0 26 20, ExpectHoverText ["MyClass", "GotoHover.hs:26:1"]]
clL25 = Position 29 9
eclL15 = Position 19 8 ; ecls = [ExpectExternFail, ExpectHoverText ["Num", "Defined in GHC.Num"]]
dnbL29 = Position 33 18 ; dnb = [ExpectHoverText [":: ()"], mkR 33 12 33 21]
dnbL30 = Position 34 23
lcbL33 = Position 37 26 ; lcb = [ExpectHoverText [":: Char"], mkR 37 26 37 27]
lclL33 = Position 37 22
mclL36 = Position 40 1 ; mcl = [mkR 40 0 40 14]
mclL37 = Position 41 1
spaceL37 = Position 41 24 ; space = [ExpectNoDefinitions, ExpectHoverText [":: Char"]]
docL41 = Position 45 1 ; doc = [ExpectHoverText ["Recognizable docs: kpqz"]]
2020-01-22 12:40:26 +03:00
; constr = [ExpectHoverText ["Monad m"]]
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
eitL40 = Position 44 28 ; kindE = [ExpectHoverText [":: * -> * -> *\n"]]
intL40 = Position 44 34 ; kindI = [ExpectHoverText [":: *\n"]]
tvrL40 = Position 44 37 ; kindV = [ExpectHoverText [":: * -> *\n"]]
intL41 = Position 45 20 ; litI = [ExpectHoverText ["7518"]]
chrL36 = Position 41 24 ; litC = [ExpectHoverText ["'f'"]]
txtL8 = Position 12 14 ; litT = [ExpectHoverText ["\"dfgy\""]]
lstL43 = Position 47 12 ; litL = [ExpectHoverText ["[8391 :: Int, 6268]"]]
outL45 = Position 49 3 ; outSig = [ExpectHoverText ["outer", "Bool"], mkR 46 0 46 5]
innL48 = Position 52 5 ; innSig = [ExpectHoverText ["inner", "Char"], mkR 49 2 49 7]
cccL17 = Position 17 11 ; docLink = [ExpectHoverText ["[Documentation](file://"]]
#if MIN_GHC_API_VERSION(8,6,0)
imported = Position 56 13 ; importedSig = getDocUri "Foo.hs" >>= \foo -> return [ExpectHoverText ["foo", "Foo", "Haddock"], mkL foo 5 0 5 3]
reexported = Position 55 14 ; reexportedSig = getDocUri "Bar.hs" >>= \bar -> return [ExpectHoverText ["Bar", "Bar", "Haddock"], mkL bar 3 0 3 14]
#else
imported = Position 56 13 ; importedSig = getDocUri "Foo.hs" >>= \foo -> return [ExpectHoverText ["foo", "Foo"], mkL foo 5 0 5 3]
reexported = Position 55 14 ; reexportedSig = getDocUri "Bar.hs" >>= \bar -> return [ExpectHoverText ["Bar", "Bar"], mkL bar 3 0 3 14]
#endif
in
mkFindTests
-- def hover look expect
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
[ test yes yes fffL4 fff "field in record definition"
, test broken broken fffL8 fff "field in record construction #71"
, test yes yes fffL14 fff "field name used as accessor" -- 120 in Calculate.hs
, test yes yes aaaL14 aaa "top-level name" -- 120
, test yes yes dcL7 tcDC "data constructor record #247"
, test yes yes dcL12 tcDC "data constructor plain" -- 121
, test yes yes tcL6 tcData "type constructor #248" -- 147
, test broken yes xtcL5 xtc "type constructor external #248,249"
, test broken yes xvL20 xvMsg "value external package #249" -- 120
, test yes yes vvL16 vv "plain parameter" -- 120
, test yes yes aL18 apmp "pattern match name" -- 120
, test yes yes opL16 op "top-level operator" -- 120, 123
, test yes yes opL18 opp "parameter operator" -- 120
, test yes yes b'L19 bp "name in backticks" -- 120
, test yes yes clL23 cls "class in instance declaration #250"
, test yes yes clL25 cls "class in signature #250" -- 147
, test broken yes eclL15 ecls "external class in signature #249,250"
, test yes yes dnbL29 dnb "do-notation bind" -- 137
, test yes yes dnbL30 dnb "do-notation lookup"
, test yes yes lcbL33 lcb "listcomp bind" -- 137
, test yes yes lclL33 lcb "listcomp lookup"
, test yes yes mclL36 mcl "top-level fn 1st clause"
, test yes yes mclL37 mcl "top-level fn 2nd clause #246"
, test yes yes spaceL37 space "top-level fn on space #315"
, test no broken docL41 doc "documentation #7"
, test no yes eitL40 kindE "kind of Either #273"
, test no yes intL40 kindI "kind of Int #273"
, test no broken tvrL40 kindV "kind of (* -> *) type variable #273"
, test no yes intL41 litI "literal Int in hover info #274"
, test no yes chrL36 litC "literal Char in hover info #274"
, test no yes txtL8 litT "literal Text in hover info #274"
, test no yes lstL43 litL "literal List in hover info #274"
, test no yes docL41 constr "type constraint in hover info #283"
, test broken broken outL45 outSig "top-level signature #310"
, test broken broken innL48 innSig "inner signature #310"
, test no yes cccL17 docLink "Haddock html links"
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
, testM yes yes imported importedSig "Imported symbol"
, testM yes yes reexported reexportedSig "Imported symbol (reexported)"
]
where yes, broken :: (TestTree -> Maybe TestTree)
yes = Just -- test should run and pass
broken = Just . (`xfail` "known broken")
2019-12-17 18:27:55 +03:00
no = const Nothing -- don't run this test at all
checkFileCompiles :: FilePath -> TestTree
checkFileCompiles fp =
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
testSessionWithExtraFiles "hover" ("Does " ++ fp ++ " compile") $ \dir -> do
void (openTestDataDoc (dir </> fp))
expectNoMoreDiagnostics 0.5
pluginTests :: TestTree
pluginTests = (`xfail8101` "known broken (#556)")
$ testSessionWait "plugins" $ do
let content =
T.unlines
[ "{-# OPTIONS_GHC -fplugin GHC.TypeLits.KnownNat.Solver #-}"
, "{-# LANGUAGE DataKinds, ScopedTypeVariables, TypeOperators #-}"
, "module Testing where"
, "import Data.Proxy"
, "import GHC.TypeLits"
-- This function fails without plugins being initialized.
, "f :: forall n. KnownNat n => Proxy n -> Integer"
, "f _ = natVal (Proxy :: Proxy n) + natVal (Proxy :: Proxy (n+2))"
, "foo :: Int -> Int -> Int"
, "foo a b = a + c"
]
_ <- createDoc "Testing.hs" "haskell" content
expectDiagnostics
[ ( "Testing.hs",
[(DsError, (8, 14), "Variable not in scope: c")]
)
]
cppTests :: TestTree
cppTests =
testGroup "cpp"
[ testCase "cpp-error" $ do
let content =
T.unlines
[ "{-# LANGUAGE CPP #-}",
"module Testing where",
"#ifdef FOO",
"foo = 42"
]
-- The error locations differ depending on which C-preprocessor is used.
-- Some give the column number and others don't (hence -1). Assert either
-- of them.
(run $ expectError content (2, -1))
`catch` ( \e -> do
let _ = e :: HUnitFailure
run $ expectError content (2, 1)
)
, testSessionWait "cpp-ghcide" $ do
_ <- createDoc "A.hs" "haskell" $ T.unlines
["{-# LANGUAGE CPP #-}"
,"main ="
,"#ifdef __GHCIDE__"
," worked"
,"#else"
," failed"
,"#endif"
]
expectDiagnostics [("A.hs", [(DsError, (3, 2), "Variable not in scope: worked")])]
]
where
expectError :: T.Text -> Cursor -> Session ()
expectError content cursor = do
_ <- createDoc "Testing.hs" "haskell" content
expectDiagnostics
[ ( "Testing.hs",
[(DsError, cursor, "error: unterminated")]
)
]
expectNoMoreDiagnostics 0.5
preprocessorTests :: TestTree
preprocessorTests = testSessionWait "preprocessor" $ do
let content =
T.unlines
[ "{-# OPTIONS_GHC -F -pgmF=ghcide-test-preprocessor #-}"
, "module Testing where"
, "y = x + z" -- plugin replaces x with y, making this have only one diagnostic
]
_ <- createDoc "Testing.hs" "haskell" content
expectDiagnostics
[ ( "Testing.hs",
[(DsError, (2, 8), "Variable not in scope: z")]
)
]
safeTests :: TestTree
safeTests =
testGroup
"SafeHaskell"
[ -- Test for https://github.com/digital-asset/ghcide/issues/424
testSessionWait "load" $ do
let sourceA =
T.unlines
["{-# LANGUAGE Trustworthy #-}"
,"module A where"
,"import System.IO.Unsafe"
,"import System.IO"
,"trustWorthyId :: a -> a"
,"trustWorthyId i = unsafePerformIO $ do"
," putStrLn \"I'm safe\""
," return i"]
sourceB =
T.unlines
["{-# LANGUAGE Safe #-}"
,"module B where"
,"import A"
,"safeId :: a -> a"
,"safeId = trustWorthyId"
]
_ <- createDoc "A.hs" "haskell" sourceA
_ <- createDoc "B.hs" "haskell" sourceB
expectNoMoreDiagnostics 1 ]
thTests :: TestTree
thTests =
testGroup
"TemplateHaskell"
[ -- Test for https://github.com/digital-asset/ghcide/pull/212
testSessionWait "load" $ do
let sourceA =
T.unlines
[ "{-# LANGUAGE PackageImports #-}",
"{-# LANGUAGE TemplateHaskell #-}",
"module A where",
"import \"template-haskell\" Language.Haskell.TH",
"a :: Integer",
"a = $(litE $ IntegerL 3)"
]
sourceB =
T.unlines
[ "{-# LANGUAGE PackageImports #-}",
"{-# LANGUAGE TemplateHaskell #-}",
"module B where",
"import A",
"import \"template-haskell\" Language.Haskell.TH",
"b :: Integer",
"b = $(litE $ IntegerL $ a) + n"
]
_ <- createDoc "A.hs" "haskell" sourceA
_ <- createDoc "B.hs" "haskell" sourceB
expectDiagnostics [ ( "B.hs", [(DsError, (6, 29), "Variable not in scope: n")] ) ]
, testSessionWait "newtype-closure" $ do
let sourceA =
T.unlines
[ "{-# LANGUAGE DeriveDataTypeable #-}"
,"{-# LANGUAGE TemplateHaskell #-}"
,"module A (a) where"
,"import Data.Data"
,"import Language.Haskell.TH"
,"newtype A = A () deriving (Data)"
,"a :: ExpQ"
,"a = [| 0 |]"]
let sourceB =
T.unlines
[ "{-# LANGUAGE TemplateHaskell #-}"
,"module B where"
,"import A"
,"b :: Int"
,"b = $( a )" ]
_ <- createDoc "A.hs" "haskell" sourceA
_ <- createDoc "B.hs" "haskell" sourceB
return ()
, thReloadingTest `xfail` "expect broken (#672)"
]
-- | test that TH is reevaluated on typecheck
thReloadingTest :: TestTree
thReloadingTest = testCase "reloading-th-test" $ withoutStackEnv $ runWithExtraFiles "TH" $ \dir -> do
let aPath = dir </> "THA.hs"
bPath = dir </> "THB.hs"
cPath = dir </> "THC.hs"
aSource <- liftIO $ readFileUtf8 aPath -- th = [d|a :: ()|]
bSource <- liftIO $ readFileUtf8 bPath -- $th
cSource <- liftIO $ readFileUtf8 cPath -- c = a :: ()
adoc <- createDoc aPath "haskell" aSource
bdoc <- createDoc bPath "haskell" bSource
cdoc <- createDoc cPath "haskell" cSource
expectDiagnostics [("THB.hs", [(DsWarning, (4,0), "Top-level binding")])]
-- Change th from () to Bool
let aSource' = T.unlines $ init (T.lines aSource) ++ ["th_a = [d| a = False|]"]
changeDoc adoc [TextDocumentContentChangeEvent Nothing Nothing aSource']
-- generate an artificial warning to avoid timing out if the TH change does not propagate
changeDoc cdoc [TextDocumentContentChangeEvent Nothing Nothing $ cSource <> "\nfoo=()"]
-- Check that the change propagates to C
expectDiagnostics
[("THC.hs", [(DsError, (4, 4), "Couldn't match expected type '()' with actual type 'Bool'")])
,("THC.hs", [(DsWarning, (6,0), "Top-level binding")])
]
closeDoc adoc
closeDoc bdoc
closeDoc cdoc
completionTests :: TestTree
completionTests
= testGroup "completion"
[ testGroup "non local" nonLocalCompletionTests
, testGroup "local" localCompletionTests
, testGroup "other" otherCompletionTests
]
completionTest :: String -> [T.Text] -> Position -> [(T.Text, CompletionItemKind, Bool, Bool)] -> TestTree
completionTest name src pos expected = testSessionWait name $ do
docId <- createDoc "A.hs" "haskell" (T.unlines src)
compls <- getCompletions docId pos
let compls' = [ (_label, _kind) | CompletionItem{..} <- compls]
liftIO $ do
compls' @?= [ (l, Just k) | (l,k,_,_) <- expected]
forM_ (zip compls expected) $ \(CompletionItem{..}, (_,_,expectedSig, expectedDocs)) -> do
when expectedSig $
assertBool ("Missing type signature: " <> T.unpack _label) (isJust _detail)
when expectedDocs $
assertBool ("Missing docs: " <> T.unpack _label) (isJust _documentation)
localCompletionTests :: [TestTree]
localCompletionTests = [
completionTest
"variable"
["bar = xx", "-- | haddock", "xxx :: ()", "xxx = ()", "-- | haddock", "data Xxx = XxxCon"]
(Position 0 8)
[("xxx", CiFunction, True, True),
("XxxCon", CiConstructor, False, True)
],
completionTest
"constructor"
["bar = xx", "-- | haddock", "xxx :: ()", "xxx = ()", "-- | haddock", "data Xxx = XxxCon"]
(Position 0 8)
[("xxx", CiFunction, True, True),
("XxxCon", CiConstructor, False, True)
],
completionTest
"class method"
["bar = xx", "class Xxx a where", "-- | haddock", "xxx :: ()", "xxx = ()"]
(Position 0 8)
[("xxx", CiFunction, True, True)],
completionTest
"type"
["bar :: Xx", "xxx = ()", "-- | haddock", "data Xxx = XxxCon"]
(Position 0 9)
[("Xxx", CiStruct, False, True)],
completionTest
"class"
["bar :: Xx", "xxx = ()", "-- | haddock", "class Xxx a"]
(Position 0 9)
[("Xxx", CiClass, False, True)]
]
nonLocalCompletionTests :: [TestTree]
nonLocalCompletionTests =
[ completionTest
"variable"
["module A where", "f = hea"]
(Position 1 7)
[("head", CiFunction, True, True)],
completionTest
"constructor"
["module A where", "f = Tru"]
(Position 1 7)
[ ("True", CiConstructor, True, True),
("truncate", CiFunction, True, True)
],
completionTest
"type"
["{-# OPTIONS_GHC -Wall #-}", "module A () where", "f :: Bo", "f = True"]
(Position 2 7)
[ ("Bounded", CiClass, True, True),
("Bool", CiStruct, True, True)
],
completionTest
"qualified"
["{-# OPTIONS_GHC -Wunused-binds #-}", "module A () where", "f = Prelude.hea"]
(Position 2 15)
[ ("head", CiFunction, True, True)
]
]
otherCompletionTests :: [TestTree]
otherCompletionTests = [
completionTest
"keyword"
["module A where", "f = newty"]
(Position 1 9)
[("newtype", CiKeyword, False, False)],
completionTest
"type context"
[ "{-# OPTIONS_GHC -Wunused-binds #-}",
"module A () where",
"f = f",
"g :: Intege"
]
-- At this point the module parses but does not typecheck.
-- This should be sufficient to detect that we are in a
-- type context and only show the completion to the type.
(Position 3 11)
[("Integer", CiStruct, True, True)]
]
outlineTests :: TestTree
outlineTests = testGroup
"outline"
[ testSessionWait "type class" $ do
let source = T.unlines ["module A where", "class A a where a :: a -> Bool"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[ moduleSymbol
"A"
(R 0 7 0 8)
[ classSymbol "A a"
(R 1 0 1 30)
[docSymbol' "a" SkMethod (R 1 16 1 30) (R 1 16 1 17)]
]
]
, testSessionWait "type class instance " $ do
let source = T.unlines ["class A a where", "instance A () where"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[ classSymbol "A a" (R 0 0 0 15) []
, docSymbol "A ()" SkInterface (R 1 0 1 19)
]
, testSessionWait "type family" $ do
let source = T.unlines ["{-# language TypeFamilies #-}", "type family A"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left [docSymbolD "A" "type family" SkClass (R 1 0 1 13)]
, testSessionWait "type family instance " $ do
let source = T.unlines
[ "{-# language TypeFamilies #-}"
, "type family A a"
, "type instance A () = ()"
]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[ docSymbolD "A a" "type family" SkClass (R 1 0 1 15)
, docSymbol "A ()" SkInterface (R 2 0 2 23)
]
, testSessionWait "data family" $ do
let source = T.unlines ["{-# language TypeFamilies #-}", "data family A"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left [docSymbolD "A" "data family" SkClass (R 1 0 1 11)]
, testSessionWait "data family instance " $ do
let source = T.unlines
[ "{-# language TypeFamilies #-}"
, "data family A a"
, "data instance A () = A ()"
]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[ docSymbolD "A a" "data family" SkClass (R 1 0 1 11)
, docSymbol "A ()" SkInterface (R 2 0 2 25)
]
, testSessionWait "constant" $ do
let source = T.unlines ["a = ()"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[docSymbol "a" SkFunction (R 0 0 0 6)]
, testSessionWait "pattern" $ do
let source = T.unlines ["Just foo = Just 21"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[docSymbol "Just foo" SkFunction (R 0 0 0 18)]
, testSessionWait "pattern with type signature" $ do
let source = T.unlines ["{-# language ScopedTypeVariables #-}", "a :: () = ()"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[docSymbol "a :: ()" SkFunction (R 1 0 1 12)]
, testSessionWait "function" $ do
let source = T.unlines ["a x = ()"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left [docSymbol "a" SkFunction (R 0 0 0 8)]
, testSessionWait "type synonym" $ do
let source = T.unlines ["type A = Bool"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[docSymbol' "A" SkTypeParameter (R 0 0 0 13) (R 0 5 0 6)]
, testSessionWait "datatype" $ do
let source = T.unlines ["data A = C"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[ docSymbolWithChildren "A"
SkStruct
(R 0 0 0 10)
[docSymbol "C" SkConstructor (R 0 9 0 10)]
]
, testSessionWait "record fields" $ do
let source = T.unlines ["data A = B {", " x :: Int", " , y :: Int}"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @=? Left
[ docSymbolWithChildren "A" SkStruct (R 0 0 2 13)
[ docSymbolWithChildren' "B" SkConstructor (R 0 9 2 13) (R 0 9 0 10)
[ docSymbol "x" SkField (R 1 2 1 3)
, docSymbol "y" SkField (R 2 4 2 5)
]
]
]
, testSessionWait "import" $ do
let source = T.unlines ["import Data.Maybe"]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[docSymbolWithChildren "imports"
SkModule
(R 0 0 0 17)
[ docSymbol "import Data.Maybe" SkModule (R 0 0 0 17)
]
]
, testSessionWait "multiple import" $ do
let source = T.unlines ["", "import Data.Maybe", "", "import Control.Exception", ""]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left
[docSymbolWithChildren "imports"
SkModule
(R 1 0 3 24)
[ docSymbol "import Data.Maybe" SkModule (R 1 0 1 17)
, docSymbol "import Control.Exception" SkModule (R 3 0 3 24)
]
]
, testSessionWait "foreign import" $ do
let source = T.unlines
[ "{-# language ForeignFunctionInterface #-}"
, "foreign import ccall \"a\" a :: Int"
]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left [docSymbolD "a" "import" SkObject (R 1 0 1 33)]
, testSessionWait "foreign export" $ do
let source = T.unlines
[ "{-# language ForeignFunctionInterface #-}"
, "foreign export ccall odd :: Int -> Bool"
]
docId <- createDoc "A.hs" "haskell" source
symbols <- getDocumentSymbols docId
liftIO $ symbols @?= Left [docSymbolD "odd" "export" SkObject (R 1 0 1 39)]
]
where
docSymbol name kind loc =
DocumentSymbol name Nothing kind Nothing loc loc Nothing
docSymbol' name kind loc selectionLoc =
DocumentSymbol name Nothing kind Nothing loc selectionLoc Nothing
docSymbolD name detail kind loc =
DocumentSymbol name (Just detail) kind Nothing loc loc Nothing
docSymbolWithChildren name kind loc cc =
DocumentSymbol name Nothing kind Nothing loc loc (Just $ List cc)
docSymbolWithChildren' name kind loc selectionLoc cc =
DocumentSymbol name Nothing kind Nothing loc selectionLoc (Just $ List cc)
moduleSymbol name loc cc = DocumentSymbol name
Nothing
SkFile
Nothing
(R 0 0 maxBound 0)
loc
(Just $ List cc)
classSymbol name loc cc = DocumentSymbol name
(Just "class")
SkClass
Nothing
loc
loc
(Just $ List cc)
pattern R :: Int -> Int -> Int -> Int -> Range
pattern R x y x' y' = Range (Position x y) (Position x' y')
xfail :: TestTree -> String -> TestTree
xfail = flip expectFailBecause
xfail8101 :: TestTree -> String -> TestTree
#if MIN_GHC_API_VERSION(8,10,0)
xfail8101 = flip expectFailBecause
#else
xfail8101 t _ = t
#endif
expectFailCabal :: String -> TestTree -> TestTree
#ifdef STACK
expectFailCabal _ = id
#else
expectFailCabal = expectFailBecause
#endif
data Expect
= ExpectRange Range -- Both gotoDef and hover should report this range
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
| ExpectLocation Location
-- | ExpectDefRange Range -- Only gotoDef should report this range
| ExpectHoverRange Range -- Only hover should report this range
| ExpectHoverText [T.Text] -- the hover message must contain these snippets
| ExpectExternFail -- definition lookup in other file expected to fail
| ExpectNoDefinitions
| ExpectNoHover
-- | ExpectExtern -- TODO: as above, but expected to succeed: need some more info in here, once we have some working examples
deriving Eq
mkR :: Int -> Int -> Int -> Int -> Expect
mkR startLine startColumn endLine endColumn = ExpectRange $ mkRange startLine startColumn endLine endColumn
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
mkL :: Uri -> Int -> Int -> Int -> Int -> Expect
mkL uri startLine startColumn endLine endColumn = ExpectLocation $ Location uri $ mkRange startLine startColumn endLine endColumn
haddockTests :: TestTree
haddockTests
= testGroup "haddock"
[ testCase "Num" $ checkHaddock
(unlines
[ "However, '(+)' and '(*)' are"
, "customarily expected to define a ring and have the following properties:"
, ""
, "[__Associativity of (+)__]: @(x + y) + z@ = @x + (y + z)@"
, "[__Commutativity of (+)__]: @x + y@ = @y + x@"
, "[__@fromInteger 0@ is the additive identity__]: @x + fromInteger 0@ = @x@"
]
)
(unlines
[ ""
, ""
, "However, `(+)` and `(*)` are"
, "customarily expected to define a ring and have the following properties: "
, "+ ****Associativity of (+)****: `(x + y) + z` = `x + (y + z)`"
, "+ ****Commutativity of (+)****: `x + y` = `y + x`"
, "+ ****`fromInteger 0` is the additive identity****: `x + fromInteger 0` = `x`"
]
)
, testCase "unsafePerformIO" $ checkHaddock
(unlines
[ "may require"
, "different precautions:"
, ""
, " * Use @{\\-\\# NOINLINE foo \\#-\\}@ as a pragma on any function @foo@"
, " that calls 'unsafePerformIO'. If the call is inlined,"
, " the I\\/O may be performed more than once."
, ""
, " * Use the compiler flag @-fno-cse@ to prevent common sub-expression"
, " elimination being performed on the module."
, ""
]
)
(unlines
[ ""
, ""
, "may require"
, "different precautions: "
, "+ Use `{-# NOINLINE foo #-}` as a pragma on any function `foo` "
, " that calls `unsafePerformIO` . If the call is inlined,"
, " the I/O may be performed more than once."
, ""
, "+ Use the compiler flag `-fno-cse` to prevent common sub-expression"
, " elimination being performed on the module."
, ""
]
)
]
where
checkHaddock s txt = spanDocToMarkdownForTest s @?= txt
cradleTests :: TestTree
cradleTests = testGroup "cradle"
[testGroup "dependencies" [sessionDepsArePickedUp]
,testGroup "loading" [loadCradleOnlyonce]
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
,testGroup "multi" [simpleMultiTest, simpleMultiTest2]
]
loadCradleOnlyonce :: TestTree
loadCradleOnlyonce = testGroup "load cradle only once"
[ testSession' "implicit" implicit
, testSession' "direct" direct
]
where
direct dir = do
liftIO $ writeFileUTF8 (dir </> "hie.yaml")
"cradle: {direct: {arguments: []}}"
test dir
implicit dir = test dir
test _dir = do
doc <- createDoc "B.hs" "haskell" "module B where\nimport Data.Foo"
msgs <- someTill (skipManyTill anyMessage cradleLoadedMessage) (skipManyTill anyMessage (message @PublishDiagnosticsNotification))
liftIO $ length msgs @?= 1
changeDoc doc [TextDocumentContentChangeEvent Nothing Nothing "module B where\nimport Data.Maybe"]
msgs <- manyTill (skipManyTill anyMessage cradleLoadedMessage) (skipManyTill anyMessage (message @PublishDiagnosticsNotification))
liftIO $ length msgs @?= 0
_ <- createDoc "A.hs" "haskell" "module A where\nimport LoadCradleBar"
msgs <- manyTill (skipManyTill anyMessage cradleLoadedMessage) (skipManyTill anyMessage (message @PublishDiagnosticsNotification))
liftIO $ length msgs @?= 0
dependentFileTest :: TestTree
dependentFileTest = testGroup "addDependentFile"
[testGroup "file-changed" [testSession' "test" test]
]
where
test dir = do
-- If the file contains B then no type error
-- otherwise type error
liftIO $ writeFile (dir </> "dep-file.txt") "A"
let fooContent = T.unlines
[ "{-# LANGUAGE TemplateHaskell #-}"
, "module Foo where"
, "import Language.Haskell.TH.Syntax"
, "foo :: Int"
, "foo = 1 + $(do"
, " qAddDependentFile \"dep-file.txt\""
, " f <- qRunIO (readFile \"dep-file.txt\")"
, " if f == \"B\" then [| 1 |] else lift f)"
]
let bazContent = T.unlines ["module Baz where", "import Foo"]
_ <-createDoc "Foo.hs" "haskell" fooContent
doc <- createDoc "Baz.hs" "haskell" bazContent
expectDiagnostics
[("Foo.hs", [(DsError, (4, 6), "Couldn't match expected type")])]
-- Now modify the dependent file
liftIO $ writeFile (dir </> "dep-file.txt") "B"
let change = TextDocumentContentChangeEvent
{ _range = Just (Range (Position 2 0) (Position 2 6))
, _rangeLength = Nothing
, _text = "f = ()"
}
-- Modifying Baz will now trigger Foo to be rebuilt as well
changeDoc doc [change]
expectDiagnostics [("Foo.hs", [])]
cradleLoadedMessage :: Session FromServerMessage
cradleLoadedMessage = satisfy $ \case
NotCustomServer (NotificationMessage _ (CustomServerMethod m) _) -> m == cradleLoadedMethod
_ -> False
cradleLoadedMethod :: T.Text
cradleLoadedMethod = "ghcide/cradle/loaded"
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
-- Stack sets this which trips up cabal in the multi-component tests.
-- However, our plugin tests rely on those env vars so we unset it locally.
withoutStackEnv :: IO a -> IO a
withoutStackEnv s =
bracket
(mapM getEnv vars >>= \prevState -> mapM_ unsetEnv vars >> pure prevState)
(\prevState -> mapM_ (\(var, value) -> restore var value) (zip vars prevState))
(const s)
where vars =
[ "GHC_PACKAGE_PATH"
, "GHC_ENVIRONMENT"
, "HASKELL_DIST_DIR"
, "HASKELL_PACKAGE_SANDBOX"
, "HASKELL_PACKAGE_SANDBOXES"
]
restore var Nothing = unsetEnv var
restore var (Just val) = setEnv var val True
simpleMultiTest :: TestTree
simpleMultiTest = testCase "simple-multi-test" $ withoutStackEnv $ runWithExtraFiles "multi" $ \dir -> do
let aPath = dir </> "a/A.hs"
bPath = dir </> "b/B.hs"
aSource <- liftIO $ readFileUtf8 aPath
(TextDocumentIdentifier adoc) <- createDoc aPath "haskell" aSource
expectNoMoreDiagnostics 0.5
bSource <- liftIO $ readFileUtf8 bPath
bdoc <- createDoc bPath "haskell" bSource
expectNoMoreDiagnostics 0.5
locs <- getDefinitions bdoc (Position 2 7)
let fooL = mkL adoc 2 0 2 3
checkDefs locs (pure [fooL])
expectNoMoreDiagnostics 0.5
-- Like simpleMultiTest but open the files in the other order
simpleMultiTest2 :: TestTree
simpleMultiTest2 = testCase "simple-multi-test2" $ withoutStackEnv $ runWithExtraFiles "multi" $ \dir -> do
let aPath = dir </> "a/A.hs"
bPath = dir </> "b/B.hs"
bSource <- liftIO $ readFileUtf8 bPath
bdoc <- createDoc bPath "haskell" bSource
expectNoMoreDiagnostics 5
aSource <- liftIO $ readFileUtf8 aPath
(TextDocumentIdentifier adoc) <- createDoc aPath "haskell" aSource
-- Need to have some delay here or the test fails
ShakeSession and shakeEnqueue (#554) * ShakeSession and shakeRunGently Currently we start a new Shake session for every interaction with the Shake database, including type checking, hovers, code actions, completions, etc. Since only one Shake session can ever exist, we abort the active session if any in order to execute the new command in a responsive manner. This is suboptimal in many, many ways: - A hover in module M aborts the typechecking of module M, only to start over! - Read-only commands (hover, code action, completion) need to typecheck all the modules! (or rather, ask Shake to check that the typechecks are current) - There is no way to run non-interfering commands concurrently This is an experiment inspired by the 'ShakeQueue' of @mpickering, and the follow-up discussion in https://github.com/mpickering/ghcide/issues/7 We introduce the concept of the 'ShakeSession' as part of the IDE state. The 'ShakeSession' is initialized by a call to 'shakeRun', and survives until the next call to 'shakeRun'. It is important that the session is restarted as soon as the filesystem changes, to ensure that the database is current. The 'ShakeSession' enables a new command 'shakeRunGently', which appends work to the existing 'ShakeSession'. This command can be called in parallel without any restriction. * Simplify by assuming there is always a ShakeSession * Improved naming and docs * Define runActionSync on top of shakeEnqueue shakeRun is not correct as it never returns anymore * Drive progress reporting from newSession The previous approach reused the shakeProgress thread, which doesn't work anymore as ShakeSession keeps the ShakeDatabase open until the next edit * Deterministic progress messages in tests Dropping the 0.1s sleep to ensure that progress messages during tests are deterministic * Make kick explicit This is required for progress reporting to work, see notes in shakeRun As to whether this is the right thing to do: 1. Less magic, more explicit 2. There's only 2 places where kick is actually used * apply Neil's feedback * avoid a deadlock when the enqueued action throws * Simplify runAction + comments * use a Barrier for clarity A Barrier is a smaller abstraction than an MVar, and the next version of the extra package will come with a suitably small implementation: https://github.com/ndmitchell/extra/commit/98c2a83585d2ca0a9d961dd241c4a967ef87866a * Log timings for code actions, hovers and completions * Rename shakeRun to shakeRestart The action returned by shakeRun now blocks until another call to shakeRun is made, which is a change in behaviour,. but all the current uses of shakeRun ignore this action. Since the new behaviour is not useful, this change simplifies and updates the docs and name accordingly * delete runActionSync as it's just runAction * restart shake session on new component created * requeue pending actions on session restart * hlint * Bumped the delay from 5 to 6 * Add a test for the non-lsp command line * Update exe/Main.hs Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-08 12:36:36 +03:00
expectNoMoreDiagnostics 6
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
locs <- getDefinitions bdoc (Position 2 7)
let fooL = mkL adoc 2 0 2 3
checkDefs locs (pure [fooL])
expectNoMoreDiagnostics 0.5
Interface file fixes (#645) * Add test for inconsistent diagnostics * Refactoring ModIfaceFromDisk This started as a pure refactoring to clarify the responsibilities between ModIface and ModIfaceFromDisk, but ended up having some behaviour changes: 1. Regenerate interface when checkOldIface returns something other than UpToDate. This was a bug. 2. Do not generate a diagnostic when regenerating an interface. 2. Previously we conflated stale interface with other errors, and would regenerate in both cases. Now we only regenerate in the first case. Tentative fix for #597 * Split interface tests * Always recompile modules with TH splices Tentative fix for #614 TODO support stability * Fix expectDiagnostics in MacOs * Avoid File does not exist diagnostics for interface files Fixes #642 * Clarify interface tests * hlints * Performance fixes The previous changes were 10X slower, this is 20X faster than those, so 2X faster than upstream, for some benchmarks * formatting * Fix GetModificationTime identity The answer for a GetModification query is independent of the missingFileDiagnostics field (as the diagnostics are not part of the answer) * remove stale comment * Avoid calling ghcSessionDepsDefinition twice * Apply suggestions from code review Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Code review feedback * Address review feedback https://github.com/digital-asset/ghcide/pull/645/files/49b0d9ac65399edf82a7a9cbbb8d8b5420458d8d#r443383239 * Change recomp to direct cradle Co-authored-by: Zubin Duggal <zubin@cmi.ac.in> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-23 12:01:52 +03:00
ifaceTests :: TestTree
ifaceTests = testGroup "Interface loading tests"
[ -- https://github.com/digital-asset/ghcide/pull/645/
ifaceErrorTest
, ifaceErrorTest2
, ifaceErrorTest3
, ifaceTHTest
Interface file fixes (#645) * Add test for inconsistent diagnostics * Refactoring ModIfaceFromDisk This started as a pure refactoring to clarify the responsibilities between ModIface and ModIfaceFromDisk, but ended up having some behaviour changes: 1. Regenerate interface when checkOldIface returns something other than UpToDate. This was a bug. 2. Do not generate a diagnostic when regenerating an interface. 2. Previously we conflated stale interface with other errors, and would regenerate in both cases. Now we only regenerate in the first case. Tentative fix for #597 * Split interface tests * Always recompile modules with TH splices Tentative fix for #614 TODO support stability * Fix expectDiagnostics in MacOs * Avoid File does not exist diagnostics for interface files Fixes #642 * Clarify interface tests * hlints * Performance fixes The previous changes were 10X slower, this is 20X faster than those, so 2X faster than upstream, for some benchmarks * formatting * Fix GetModificationTime identity The answer for a GetModification query is independent of the missingFileDiagnostics field (as the diagnostics are not part of the answer) * remove stale comment * Avoid calling ghcSessionDepsDefinition twice * Apply suggestions from code review Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Code review feedback * Address review feedback https://github.com/digital-asset/ghcide/pull/645/files/49b0d9ac65399edf82a7a9cbbb8d8b5420458d8d#r443383239 * Change recomp to direct cradle Co-authored-by: Zubin Duggal <zubin@cmi.ac.in> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-23 12:01:52 +03:00
]
-- | test that TH reevaluates across interfaces
ifaceTHTest :: TestTree
ifaceTHTest = testCase "iface-th-test" $ withoutStackEnv $ runWithExtraFiles "TH" $ \dir -> do
let aPath = dir </> "THA.hs"
bPath = dir </> "THB.hs"
cPath = dir </> "THC.hs"
aSource <- liftIO $ readFileUtf8 aPath -- [TH] a :: ()
_bSource <- liftIO $ readFileUtf8 bPath -- a :: ()
cSource <- liftIO $ readFileUtf8 cPath -- c = a :: ()
cdoc <- createDoc cPath "haskell" cSource
-- Change [TH]a from () to Bool
liftIO $ writeFileUTF8 aPath (unlines $ init (lines $ T.unpack aSource) ++ ["th_a = [d| a = False|]"])
-- Check that the change propogates to C
changeDoc cdoc [TextDocumentContentChangeEvent Nothing Nothing cSource]
expectDiagnostics
[("THC.hs", [(DsError, (4, 4), "Couldn't match expected type '()' with actual type 'Bool'")])
,("THB.hs", [(DsWarning, (4,0), "Top-level binding")])]
closeDoc cdoc
Interface file fixes (#645) * Add test for inconsistent diagnostics * Refactoring ModIfaceFromDisk This started as a pure refactoring to clarify the responsibilities between ModIface and ModIfaceFromDisk, but ended up having some behaviour changes: 1. Regenerate interface when checkOldIface returns something other than UpToDate. This was a bug. 2. Do not generate a diagnostic when regenerating an interface. 2. Previously we conflated stale interface with other errors, and would regenerate in both cases. Now we only regenerate in the first case. Tentative fix for #597 * Split interface tests * Always recompile modules with TH splices Tentative fix for #614 TODO support stability * Fix expectDiagnostics in MacOs * Avoid File does not exist diagnostics for interface files Fixes #642 * Clarify interface tests * hlints * Performance fixes The previous changes were 10X slower, this is 20X faster than those, so 2X faster than upstream, for some benchmarks * formatting * Fix GetModificationTime identity The answer for a GetModification query is independent of the missingFileDiagnostics field (as the diagnostics are not part of the answer) * remove stale comment * Avoid calling ghcSessionDepsDefinition twice * Apply suggestions from code review Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Code review feedback * Address review feedback https://github.com/digital-asset/ghcide/pull/645/files/49b0d9ac65399edf82a7a9cbbb8d8b5420458d8d#r443383239 * Change recomp to direct cradle Co-authored-by: Zubin Duggal <zubin@cmi.ac.in> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-23 12:01:52 +03:00
ifaceErrorTest :: TestTree
ifaceErrorTest = testCase "iface-error-test-1" $ withoutStackEnv $ runWithExtraFiles "recomp" $ \dir -> do
let aPath = dir </> "A.hs"
bPath = dir </> "B.hs"
pPath = dir </> "P.hs"
aSource <- liftIO $ readFileUtf8 aPath -- x = y :: Int
bSource <- liftIO $ readFileUtf8 bPath -- y :: Int
pSource <- liftIO $ readFileUtf8 pPath -- bar = x :: Int
bdoc <- createDoc bPath "haskell" bSource
pdoc <- createDoc pPath "haskell" pSource
expectDiagnostics [("P.hs", [(DsWarning,(4,0), "Top-level binding")]) -- So what we know P has been loaded
]
-- Change y from Int to B
changeDoc bdoc [TextDocumentContentChangeEvent Nothing Nothing $ T.unlines ["module B where", "y :: Bool", "y = undefined"]]
-- Check that the error propogates to A
adoc <- createDoc aPath "haskell" aSource
expectDiagnostics
[("A.hs", [(DsError, (5, 4), "Couldn't match expected type 'Int' with actual type 'Bool'")])]
closeDoc adoc -- Close A
changeDoc pdoc [TextDocumentContentChangeEvent Nothing Nothing $ pSource <> "\nfoo = y :: Bool" ]
-- Now in P we have
-- bar = x :: Int
-- foo = y :: Bool
-- HOWEVER, in A...
-- x = y :: Int
-- This is clearly inconsistent, and the expected outcome a bit surprising:
-- - The diagnostic for A has already been received. Ghcide does not repeat diagnostics
-- - P is being typechecked with the last successful artifacts for A.
expectDiagnostics [("P.hs", [(DsWarning,(4,0), "Top-level binding")])
,("P.hs", [(DsWarning,(6,0), "Top-level binding")])
]
expectNoMoreDiagnostics 2
ifaceErrorTest2 :: TestTree
ifaceErrorTest2 = testCase "iface-error-test-2" $ withoutStackEnv $ runWithExtraFiles "recomp" $ \dir -> do
let bPath = dir </> "B.hs"
pPath = dir </> "P.hs"
bSource <- liftIO $ readFileUtf8 bPath -- y :: Int
pSource <- liftIO $ readFileUtf8 pPath -- bar = x :: Int
bdoc <- createDoc bPath "haskell" bSource
pdoc <- createDoc pPath "haskell" pSource
expectDiagnostics [("P.hs", [(DsWarning,(4,0), "Top-level binding")]) -- So that we know P has been loaded
]
-- Change y from Int to B
changeDoc bdoc [TextDocumentContentChangeEvent Nothing Nothing $ T.unlines ["module B where", "y :: Bool", "y = undefined"]]
-- Add a new definition to P
changeDoc pdoc [TextDocumentContentChangeEvent Nothing Nothing $ pSource <> "\nfoo = y :: Bool" ]
-- Now in P we have
-- bar = x :: Int
-- foo = y :: Bool
-- HOWEVER, in A...
-- x = y :: Int
expectDiagnostics
-- As in the other test, P is being typechecked with the last successful artifacts for A
-- (ot thanks to -fdeferred-type-errors)
[("A.hs", [(DsError, (5, 4), "Couldn't match expected type 'Int' with actual type 'Bool'")])
,("P.hs", [(DsWarning,(4,0), "Top-level binding")])
,("P.hs", [(DsWarning,(6,0), "Top-level binding")])
]
expectNoMoreDiagnostics 2
ifaceErrorTest3 :: TestTree
ifaceErrorTest3 = testCase "iface-error-test-3" $ withoutStackEnv $ runWithExtraFiles "recomp" $ \dir -> do
let bPath = dir </> "B.hs"
pPath = dir </> "P.hs"
bSource <- liftIO $ readFileUtf8 bPath -- y :: Int
pSource <- liftIO $ readFileUtf8 pPath -- bar = x :: Int
bdoc <- createDoc bPath "haskell" bSource
-- Change y from Int to B
changeDoc bdoc [TextDocumentContentChangeEvent Nothing Nothing $ T.unlines ["module B where", "y :: Bool", "y = undefined"]]
-- P should not typecheck, as there are no last valid artifacts for A
_pdoc <- createDoc pPath "haskell" pSource
-- In this example the interface file for A should not exist (modulo the cache folder)
-- Despite that P still type checks, as we can generate an interface file for A thanks to -fdeferred-type-errors
expectDiagnostics
[("A.hs", [(DsError, (5, 4), "Couldn't match expected type 'Int' with actual type 'Bool'")])
,("P.hs", [(DsWarning,(4,0), "Top-level binding")])
]
expectNoMoreDiagnostics 2
sessionDepsArePickedUp :: TestTree
sessionDepsArePickedUp = testSession'
"session-deps-are-picked-up"
$ \dir -> do
liftIO $
writeFileUTF8
(dir </> "hie.yaml")
"cradle: {direct: {arguments: []}}"
-- Open without OverloadedStrings and expect an error.
doc <- createDoc "Foo.hs" "haskell" fooContent
expectDiagnostics
[("Foo.hs", [(DsError, (3, 6), "Couldn't match expected type")])]
-- Update hie.yaml to enable OverloadedStrings.
liftIO $
writeFileUTF8
(dir </> "hie.yaml")
"cradle: {direct: {arguments: [-XOverloadedStrings]}}"
-- Send change event.
let change =
TextDocumentContentChangeEvent
{ _range = Just (Range (Position 4 0) (Position 4 0)),
_rangeLength = Nothing,
_text = "\n"
}
changeDoc doc [change]
-- Now no errors.
expectDiagnostics [("Foo.hs", [])]
where
fooContent =
T.unlines
[ "module Foo where",
"import Data.Text",
"foo :: Text",
"foo = \"hello\""
]
ShakeSession and shakeEnqueue (#554) * ShakeSession and shakeRunGently Currently we start a new Shake session for every interaction with the Shake database, including type checking, hovers, code actions, completions, etc. Since only one Shake session can ever exist, we abort the active session if any in order to execute the new command in a responsive manner. This is suboptimal in many, many ways: - A hover in module M aborts the typechecking of module M, only to start over! - Read-only commands (hover, code action, completion) need to typecheck all the modules! (or rather, ask Shake to check that the typechecks are current) - There is no way to run non-interfering commands concurrently This is an experiment inspired by the 'ShakeQueue' of @mpickering, and the follow-up discussion in https://github.com/mpickering/ghcide/issues/7 We introduce the concept of the 'ShakeSession' as part of the IDE state. The 'ShakeSession' is initialized by a call to 'shakeRun', and survives until the next call to 'shakeRun'. It is important that the session is restarted as soon as the filesystem changes, to ensure that the database is current. The 'ShakeSession' enables a new command 'shakeRunGently', which appends work to the existing 'ShakeSession'. This command can be called in parallel without any restriction. * Simplify by assuming there is always a ShakeSession * Improved naming and docs * Define runActionSync on top of shakeEnqueue shakeRun is not correct as it never returns anymore * Drive progress reporting from newSession The previous approach reused the shakeProgress thread, which doesn't work anymore as ShakeSession keeps the ShakeDatabase open until the next edit * Deterministic progress messages in tests Dropping the 0.1s sleep to ensure that progress messages during tests are deterministic * Make kick explicit This is required for progress reporting to work, see notes in shakeRun As to whether this is the right thing to do: 1. Less magic, more explicit 2. There's only 2 places where kick is actually used * apply Neil's feedback * avoid a deadlock when the enqueued action throws * Simplify runAction + comments * use a Barrier for clarity A Barrier is a smaller abstraction than an MVar, and the next version of the extra package will come with a suitably small implementation: https://github.com/ndmitchell/extra/commit/98c2a83585d2ca0a9d961dd241c4a967ef87866a * Log timings for code actions, hovers and completions * Rename shakeRun to shakeRestart The action returned by shakeRun now blocks until another call to shakeRun is made, which is a change in behaviour,. but all the current uses of shakeRun ignore this action. Since the new behaviour is not useful, this change simplifies and updates the docs and name accordingly * delete runActionSync as it's just runAction * restart shake session on new component created * requeue pending actions on session restart * hlint * Bumped the delay from 5 to 6 * Add a test for the non-lsp command line * Update exe/Main.hs Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-08 12:36:36 +03:00
-- A test to ensure that the command line ghcide workflow stays working
nonLspCommandLine :: TestTree
nonLspCommandLine = testGroup "ghcide command line"
[ testCase "works" $ withTempDir $ \dir -> do
ghcide <- locateGhcideExecutable
copyTestDataFiles dir "multi"
let cmd = (proc ghcide ["a/A.hs"]){cwd = Just dir}
setEnv "HOME" "/homeless-shelter" False
(ec, _, _) <- withoutStackEnv $ readCreateProcessWithExitCode cmd ""
ec @=? ExitSuccess
]
benchmarkTests :: TestTree
-- These tests require stack and will fail with cabal test
benchmarkTests =
let ?config = Bench.defConfig
{ Bench.verbosity = Bench.Quiet
, Bench.repetitions = Just 3
, Bench.buildTool = Bench.Stack
} in
withResource Bench.setup id $ \_ -> testGroup "benchmark experiments"
[ expectFailCabal "Requires stack" $ testCase (Bench.name e) $ do
res <- Bench.runBench runInDir e
assertBool "did not successfully complete 5 repetitions" $ Bench.success res
| e <- Bench.experiments
, Bench.name e /= "edit" -- the edit experiment does not ever fail
]
----------------------------------------------------------------------
-- Utils
testSession :: String -> Session () -> TestTree
testSession name = testCase name . run
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
testSessionWithExtraFiles :: FilePath -> String -> (FilePath -> Session ()) -> TestTree
testSessionWithExtraFiles prefix name = testCase name . runWithExtraFiles prefix
testSession' :: String -> (FilePath -> Session ()) -> TestTree
testSession' name = testCase name . run'
testSessionWait :: String -> Session () -> TestTree
testSessionWait name = testSession name .
-- Check that any diagnostics produced were already consumed by the test case.
--
-- If in future we add test cases where we don't care about checking the diagnostics,
-- this could move elsewhere.
--
-- Experimentally, 0.5s seems to be long enough to wait for any final diagnostics to appear.
( >> expectNoMoreDiagnostics 0.5)
pickActionWithTitle :: T.Text -> [CAResult] -> IO CodeAction
pickActionWithTitle title actions = do
assertBool ("Found no matching actions: " <> show titles) (not $ null matches)
return $ head matches
where
titles =
[ actionTitle
| CACodeAction CodeAction { _title = actionTitle } <- actions
]
matches =
[ action
| CACodeAction action@CodeAction { _title = actionTitle } <- actions
, title == actionTitle
]
mkRange :: Int -> Int -> Int -> Int -> Range
mkRange a b c d = Range (Position a b) (Position c d)
run :: Session a -> IO a
run s = withTempDir $ \dir -> runInDir dir s
Multi Component (#522) * Multi component support In this commit we add support for loading multiple components into one ghcide session. The current behaviour is that each component is loaded lazily into the session. When a file from an unrecognised component is loaded, the cradle is consulted again to get a new set of options for the new component. This will cause all the currently loaded files to be reloaded into a new HscEnv which is shared by all the currently known components. The result of this is that functions such as go-to definition work between components if they have been loaded into the same session but you have to open at least one file from each component before it will work. Only minimal changes are needed to the internals to ghcide to make the file searching logic look in include directories for all currently loaded components. The main changes are in exe/Main.hs which has been heavily rewritten to avoid shake indirections. A global map is created which maps a filepath to the HscEnv which should be used to compile it. When a new component is created this map is completely refreshed so each path maps to a new Which paths belong to a componenent is determined by the targets listed by the cradle. Therefore it is important that each cradle also lists all the targets for the cradle. There are some other choices here as well which are less accurate such as mapping via include directories which is the aproach that I implemented in haskell-ide-engine. The commit has been tested so far with cabal and hadrian. Also deleted the .ghci file which was causing errors during testing and seemed broken anyway. Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> * Final tweaks? * Fix 8.4 build * Add multi-component test * Fix hlint * Add cabal to CI images * Modify path * Set PATH in the right place (hopefully) * Always generate interface files and hie files * Use correct DynFlags in mkImportDirs You have to use the DynFlags for the file we are currently compiling to get the right packages in the package db so that lookupPackage doesn't always fail. * Revert "Always generate interface files and hie files" This reverts commit 820aa241890c4498c566e29b0823a803fb2fd297. * remove traces * Another test * lint * Unset env vars set my stack * Fix extra-source-files As usual, stack doesn’t understand Cabal properly and doesn’t seem to like ** wildcards so I’ve enumerated it manually. * Unset env locally Co-authored-by: Alan Zimmerman <alan.zimm@gmail.com> Co-authored-by: fendor <power.walross@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-06-02 15:44:16 +03:00
runWithExtraFiles :: FilePath -> (FilePath -> Session a) -> IO a
runWithExtraFiles prefix s = withTempDir $ \dir -> do
copyTestDataFiles dir prefix
runInDir dir (s dir)
copyTestDataFiles :: FilePath -> FilePath -> IO ()
copyTestDataFiles dir prefix = do
-- Copy all the test data files to the temporary workspace
testDataFiles <- getDirectoryFilesIO ("test/data" </> prefix) ["//*"]
for_ testDataFiles $ \f -> do
createDirectoryIfMissing True $ dir </> takeDirectory f
copyFile ("test/data" </> prefix </> f) (dir </> f)
run' :: (FilePath -> Session a) -> IO a
run' s = withTempDir $ \dir -> runInDir dir (s dir)
runInDir :: FilePath -> Session a -> IO a
runInDir dir s = do
ghcideExe <- locateGhcideExecutable
-- Temporarily hack around https://github.com/mpickering/hie-bios/pull/56
-- since the package import test creates "Data/List.hs", which otherwise has no physical home
createDirectoryIfMissing True $ dir ++ "/Data"
Support for interface files (#457) * Rules for loading interface files * Typechecking with interface files * Add a note in the README about the optimal project setup * Improve support for hs-boot files The branch was failing to load GHC because the module graph was missing edges between a .hs file and its .hs-boot file. This means the .hs-boot file was getting added into the HPT after the .hs file which led to confusing errors about variables being out of scope. The fix is to maintain a map from hs-boot to hs files and then add an edge for this case when calling `transitiveDependencies`. Also tidy up some code in setupEnv which I assume was attempting to fix this but in an incorrect manner. Add the -boot suffix when looking for hi-boot files. For some reason, the `hi` path is not set to the right thing for `hs-boot` files. I don't know why not perhaps it is ok to use an existing `.hi` file in place of an `hs-boot` file. More investigation needed. My experience is that GHC is quite bad a recompilation avoidance for hs-boot files anyway. For example: https://gitlab.haskell.org/ghc/ghc/issues/17434 Add the -boot suffix when writing interface files * Generate .hi and .hie files during type checking * Refactor GetModIface to not retain TypeChecked module in memory This improves memory performance on a cold cache. * Trailing whitespace * Turn debug log messages into diagnostics * Implement "hie" files for ghc-8.6.5 This means that the .hi files patch can also be used with 8.6.5 * Add tests for hover/definition on imported symbols * hlints * Generate .hie files when missing * Fix subtle bug in setDefaultHieDir * Simplify optimal project setup in README * Move interface loading diagnostics behind --test flag Reusing the --test flag for this seems harmless, I cannot justify introducing a new flag * Avoid expensive interface file generation for files of interest * avoid redundant arguments (thanks Moritz K) * qualify a DAML only comment * Skip module source when generating hie file thanks Moritz Kiefer for noting that we don't care for the generated .hie files to embed module sources * runGhcEnv <-> evalGhcEnv * Apply suggestions from code review Thanks Moritz Kiefer Co-Authored-By: Moritz Kiefer <moritz.kiefer@purelyfunctional.org> * Add suggested Show instance Co-Authored-By: Matthew Pickering <matthewtpickering@gmail.com> * Use Control.Exception.Safe This is to avoid accidentally capturing asynchronous exceptions * Rename atomicFileUpdate * Fix a flaky test We have to be careful with module naming in tests to avoid interference of .hi files across tests * Undo formatting of D.IDE.GHC.Util * follow changes in master Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com> Co-authored-by: Moritz Kiefer <moritz.kiefer@purelyfunctional.org>
2020-03-24 14:39:53 +03:00
let cmd = unwords [ghcideExe, "--lsp", "--test", "--cwd", dir]
-- HIE calls getXgdDirectory which assumes that HOME is set.
-- Only sets HOME if it wasn't already set.
setEnv "HOME" "/homeless-shelter" False
let lspTestCaps = fullCaps { _window = Just $ WindowClientCapabilities $ Just True }
runSessionWithConfig conf cmd lspTestCaps dir s
where
conf = defaultConfig
-- If you uncomment this you can see all logging
-- which can be quite useful for debugging.
-- { logStdErr = True, logColor = False }
-- If you really want to, you can also see all messages
-- { logMessages = True, logColor = False }
openTestDataDoc :: FilePath -> Session TextDocumentIdentifier
openTestDataDoc path = do
source <- liftIO $ readFileUtf8 $ "test/data" </> path
createDoc path "haskell" source
findCodeActions :: TextDocumentIdentifier -> Range -> [T.Text] -> Session [CodeAction]
findCodeActions = findCodeActions' (==) "is not a superset of"
findCodeActionsByPrefix :: TextDocumentIdentifier -> Range -> [T.Text] -> Session [CodeAction]
findCodeActionsByPrefix = findCodeActions' T.isPrefixOf "is not prefix of"
findCodeActions' :: (T.Text -> T.Text -> Bool) -> String -> TextDocumentIdentifier -> Range -> [T.Text] -> Session [CodeAction]
findCodeActions' op errMsg doc range expectedTitles = do
actions <- getCodeActions doc range
let matches = sequence
[ listToMaybe
[ action
| CACodeAction action@CodeAction { _title = actionTitle } <- actions
, expectedTitle `op` actionTitle]
| expectedTitle <- expectedTitles]
let msg = show
[ actionTitle
| CACodeAction CodeAction { _title = actionTitle } <- actions
]
++ " " <> errMsg <> " "
++ show expectedTitles
liftIO $ case matches of
Nothing -> assertFailure msg
Just _ -> pure ()
return (fromJust matches)
findCodeAction :: TextDocumentIdentifier -> Range -> T.Text -> Session CodeAction
findCodeAction doc range t = head <$> findCodeActions doc range [t]
unitTests :: TestTree
unitTests = do
testGroup "Unit"
[ testCase "empty file path does NOT work with the empty String literal" $
uriToFilePath' (fromNormalizedUri $ filePathToUri' "") @?= Just "."
, testCase "empty file path works using toNormalizedFilePath'" $
uriToFilePath' (fromNormalizedUri $ filePathToUri' (toNormalizedFilePath' "")) @?= Just ""
, testCase "empty path URI" $ do
Just URI{..} <- pure $ parseURI (T.unpack $ getUri $ fromNormalizedUri emptyPathUri)
uriScheme @?= "file:"
uriPath @?= ""
, testCase "from empty path URI" $ do
let uri = Uri "file://"
uriToFilePath' uri @?= Just ""
, testCase "Key with empty file path roundtrips via Binary" $
Binary.decode (Binary.encode (Q ((), emptyFilePath))) @?= Q ((), emptyFilePath)
]
positionMappingTests :: TestTree
positionMappingTests =
testGroup "position mapping"
[ testGroup "toCurrent"
[ testCase "before" $
toCurrent
(Range (Position 0 1) (Position 0 3))
"ab"
(Position 0 0) @?= Just (Position 0 0)
, testCase "after, same line, same length" $
toCurrent
(Range (Position 0 1) (Position 0 3))
"ab"
(Position 0 3) @?= Just (Position 0 3)
, testCase "after, same line, increased length" $
toCurrent
(Range (Position 0 1) (Position 0 3))
"abc"
(Position 0 3) @?= Just (Position 0 4)
, testCase "after, same line, decreased length" $
toCurrent
(Range (Position 0 1) (Position 0 3))
"a"
(Position 0 3) @?= Just (Position 0 2)
, testCase "after, next line, no newline" $
toCurrent
(Range (Position 0 1) (Position 0 3))
"abc"
(Position 1 3) @?= Just (Position 1 3)
, testCase "after, next line, newline" $
toCurrent
(Range (Position 0 1) (Position 0 3))
"abc\ndef"
(Position 1 0) @?= Just (Position 2 0)
, testCase "after, same line, newline" $
toCurrent
(Range (Position 0 1) (Position 0 3))
"abc\nd"
(Position 0 4) @?= Just (Position 1 2)
, testCase "after, same line, newline + newline at end" $
toCurrent
(Range (Position 0 1) (Position 0 3))
"abc\nd\n"
(Position 0 4) @?= Just (Position 2 1)
, testCase "after, same line, newline + newline at end" $
toCurrent
(Range (Position 0 1) (Position 0 1))
"abc"
(Position 0 1) @?= Just (Position 0 4)
]
, testGroup "fromCurrent"
[ testCase "before" $
fromCurrent
(Range (Position 0 1) (Position 0 3))
"ab"
(Position 0 0) @?= Just (Position 0 0)
, testCase "after, same line, same length" $
fromCurrent
(Range (Position 0 1) (Position 0 3))
"ab"
(Position 0 3) @?= Just (Position 0 3)
, testCase "after, same line, increased length" $
fromCurrent
(Range (Position 0 1) (Position 0 3))
"abc"
(Position 0 4) @?= Just (Position 0 3)
, testCase "after, same line, decreased length" $
fromCurrent
(Range (Position 0 1) (Position 0 3))
"a"
(Position 0 2) @?= Just (Position 0 3)
, testCase "after, next line, no newline" $
fromCurrent
(Range (Position 0 1) (Position 0 3))
"abc"
(Position 1 3) @?= Just (Position 1 3)
, testCase "after, next line, newline" $
fromCurrent
(Range (Position 0 1) (Position 0 3))
"abc\ndef"
(Position 2 0) @?= Just (Position 1 0)
, testCase "after, same line, newline" $
fromCurrent
(Range (Position 0 1) (Position 0 3))
"abc\nd"
(Position 1 2) @?= Just (Position 0 4)
, testCase "after, same line, newline + newline at end" $
fromCurrent
(Range (Position 0 1) (Position 0 3))
"abc\nd\n"
(Position 2 1) @?= Just (Position 0 4)
, testCase "after, same line, newline + newline at end" $
fromCurrent
(Range (Position 0 1) (Position 0 1))
"abc"
(Position 0 4) @?= Just (Position 0 1)
]
, adjustOption (\(QuickCheckTests i) -> QuickCheckTests (max 1000 i)) $ testGroup "properties"
[ testProperty "fromCurrent r t <=< toCurrent r t" $ do
-- Note that it is important to use suchThatMap on all values at once
-- instead of only using it on the position. Otherwise you can get
-- into situations where there is no position that can be mapped back
-- for the edit which will result in QuickCheck looping forever.
let gen = do
rope <- genRope
range <- genRange rope
PrintableText replacement <- arbitrary
oldPos <- genPosition rope
pure (range, replacement, oldPos)
forAll
(suchThatMap gen
(\(range, replacement, oldPos) -> (range, replacement, oldPos,) <$> toCurrent range replacement oldPos)) $
\(range, replacement, oldPos, newPos) ->
fromCurrent range replacement newPos === Just oldPos
, testProperty "toCurrent r t <=< fromCurrent r t" $ do
let gen = do
rope <- genRope
range <- genRange rope
PrintableText replacement <- arbitrary
let newRope = applyChange rope (TextDocumentContentChangeEvent (Just range) Nothing replacement)
newPos <- genPosition newRope
pure (range, replacement, newPos)
forAll
(suchThatMap gen
(\(range, replacement, newPos) -> (range, replacement, newPos,) <$> fromCurrent range replacement newPos)) $
\(range, replacement, newPos, oldPos) ->
toCurrent range replacement oldPos === Just newPos
]
]
newtype PrintableText = PrintableText { getPrintableText :: T.Text }
deriving Show
instance Arbitrary PrintableText where
arbitrary = PrintableText . T.pack . getPrintableString <$> arbitrary
genRope :: Gen Rope
genRope = Rope.fromText . getPrintableText <$> arbitrary
genPosition :: Rope -> Gen Position
genPosition r = do
row <- choose (0, max 0 $ rows - 1)
let columns = Rope.columns (nthLine row r)
column <- choose (0, max 0 $ columns - 1)
pure $ Position row column
where rows = Rope.rows r
genRange :: Rope -> Gen Range
genRange r = do
startPos@(Position startLine startColumn) <- genPosition r
let maxLineDiff = max 0 $ rows - 1 - startLine
endLine <- choose (startLine, startLine + maxLineDiff)
let columns = Rope.columns (nthLine endLine r)
endColumn <-
if startLine == endLine
then choose (startColumn, columns)
else choose (0, max 0 $ columns - 1)
pure $ Range startPos (Position endLine endColumn)
where rows = Rope.rows r
-- | Get the ith line of a rope, starting from 0. Trailing newline not included.
nthLine :: Int -> Rope -> Rope
nthLine i r
| i < 0 = error $ "Negative line number: " <> show i
| i == 0 && Rope.rows r == 0 = r
| i >= Rope.rows r = error $ "Row number out of bounds: " <> show i <> "/" <> show (Rope.rows r)
| otherwise = Rope.takeWhile (/= '\n') $ fst $ Rope.splitAtLine 1 $ snd $ Rope.splitAtLine (i - 1) r
getWatchedFilesSubscriptionsUntil :: forall end . (FromJSON end, Typeable end) => Session [Maybe Value]
getWatchedFilesSubscriptionsUntil = do
msgs <- manyTill (Just <$> message @RegisterCapabilityRequest <|> Nothing <$ anyMessage) (message @end)
return
[ args
| Just RequestMessage{_params = RegistrationParams (List regs)} <- msgs
, Registration _id WorkspaceDidChangeWatchedFiles args <- regs
]
-- | Version of 'System.IO.Extra.withTempDir' that canonicalizes the path
-- Which we need to do on macOS since the $TMPDIR can be in @/private/var@ or
-- @/var@
withTempDir :: (FilePath -> IO a) -> IO a
withTempDir f = System.IO.Extra.withTempDir $ \dir -> do
dir' <- canonicalizePath dir
f dir'