summaryrefslogtreecommitdiff
path: root/compiler/GHC
Commit message (Collapse)AuthorAgeFilesLines
...
* Perf: fix strictness in OccurAnalSylvain Henry2021-07-231-168/+169
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch enhances OccurAnal perf by using a dedicated WithUsageDetails datatype instead of a tuple (similarly to what has been done in demand-analysis) with strict fields. OccEnv is also passed strictly with more strict fields as it improves results even more. T9198 flukes isn't reproducible locally (cf https://gitlab.haskell.org/ghc/ghc/-/merge_requests/5667#note_364358) Metric Decrease: ManyConstructors T10421 T12150 T12425 T12707 T13056 T13253 T13253-spj T15164 T16577 T18282 T18698a T18698b T1969 T4801 T5642 T9020 T9233 T9630 T9675 T9961 WWRec T12227 T13035 T18304 T6048 T12234 T783 T20049 Metric Increase: T9198
* Fix #19931John Ericson2021-07-214-9/+32
| | | | | | | | | | | | | | | | | | | | | | | | | The issue was the renderer for x86 addressing modes assumes native size registers, but we were passing in a possibly-smaller index in conjunction with a native-sized base pointer. The easist thing to do is just extend the register first. I also changed the other NGC backends implementing jump tables accordingly. On one hand, I think PowerPC and Sparc don't have the small sub-registers anyways so there is less to worry about. On the other hand, to the extent that's true the zero extension can become a no-op. I should give credit where it's due: @hsyl20 really did all the work for me in https://gitlab.haskell.org/ghc/ghc/-/merge_requests/4717#note_355874, but I was daft and missed the "Oops" and so ended up spending a silly amount of time putting it all back together myself. The unregisterised backend change is a bit different, because here we are translating the actual case not a jump table, and the fix is to handle right-sized literals not addressing modes. But it makes sense to include here too because it's the same change in the subsequent commit that exposes both bugs.
* template-haskell: Add support for default declarationsMario Blažević2021-07-215-9/+17
| | | | Fixes #19373
* Use Ways API instead of Set specific functionsFendor2021-07-2111-26/+33
|
* Refactor SuggestExtension constructor in GhcHintAlfredo Di Napoli2021-07-216-53/+117
| | | | | | | | | | | | | | | | | | This commit refactors the SuggestExtension type constructor of the GhcHint to be more powerful and flexible. In particular, we can now embed extra user information (essentially "sugar") to help clarifying the suggestion. This makes the following possible: Suggested fix: Perhaps you intended to use GADTs or a similar language extension to enable syntax: data T where We can still give to IDEs and tools a `LangExt.Extension` they can use, but in the pretty-printed message we can tell the user a bit more on why such extension is needed. On top of that, we now have the ability to express conjuctions and disjunctons, for those cases where GHC suggests to enable "X or Y" and for the cases where we need "X and Y".
* driver: Fix recompilation for modules importing GHC.PrimMatthew Pickering2021-07-216-15/+33
| | | | | | | | | | | | | The GHC.Prim module is quite special as there is no interface file, therefore it doesn't appear in ms_textual_imports, but the ghc-prim package does appear in the direct package dependencies. This confused the recompilation checking which couldn't find any modules from ghc-prim and concluded that the package was no longer a dependency. The fix is to keep track of whether GHC.Prim is imported separately in the relevant places. Fixes #20084
* Make TmpFs independent of DynFlagsSylvain Henry2021-07-1915-58/+65
| | | | | | | | | This is small step towards #19877. We want to make the Loader/Linker interface more abstract to be easily reused (i.e. don't pass it DynFlags) but the system linker uses TmpFs which required a DynFlags value to get its temp directory. We explicitly pass the temp directory now. Similarly TmpFs was consulting the DynFlags to decide whether to clean or: this is now done by the caller in the driver code.
* Don't duplicate constructors in the simplifierSimon Peyton Jones2021-07-191-9/+39
| | | | | | | | | | | | Ticket #20125 showed that the Simplifier could sometimes duplicate a constructor binding. CSE would often eliminate it later, but doing it in the first place was utterly wrong. See Note [Do not duplicate constructor applications] in Simplify.hs I also added a short-cut to Simplify.simplNonRecX for the case when the RHS is trivial. I don't think this will change anything, just make the compiler run a tiny bit faster.
* Rename RecordPuns to NamedFieldPuns in LangExt.ExtensionAlfredo Di Napoli2021-07-196-12/+8
| | | | | | | | | | This commit renames the `RecordPuns` type constructor inside `GHC.LanguageExtensions.Type.hs` to `NamedFieldPuns`. The rationale is that the `RecordPuns` language extension was deprecated a long time ago, but it was still present in the AST, introducing an annoying mismatch between what GHC suggested (i.e. "use NamedFieldPuns") and what that translated into in terms of Haskell types.
* th: Weaken return type of myCoreToStgExprMatthew Pickering2021-07-193-56/+18
| | | | | | | | | | | | | | | | | | | | The previous code assumed properties of the CoreToStg translation, namely that a core let expression which be translated to a single non-recursive top-level STG binding. This assumption was false, as evidenced by #20060. The consequence of this was the need to modify the call sites of `myCoreToStgExpr`, the main one being in hscCompileCoreExpr', which the meant we had to use byteCodeGen instead of stgExprToBCOs to convert the returned value to bytecode. I removed the `stgExprToBCOs` function as it is no longer used in the compiler. There is still some partiallity with this patch (the lookup in hscCompileCoreExpr') but this should be more robust that before. Fixes #20060
* Better sharing of join points (#19996)wip/T19557Simon Peyton Jones2021-07-192-29/+92
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch, provoked by regressions in the text package (#19557), improves sharing of join points. This also fixes the terrible behaviour in #20049. See Note [Duplicating join points] in GHC.Core.Opt.Simplify. * In the StrictArg case of mkDupableContWithDmds, don't use Plan A for data constructors * In postInlineUnconditionally, don't inline JoinIds Avoids inlining join $j x = Just x in case blah of A -> $j x1 B -> $j x2 C -> $j x3 * In mkDupableStrictBind and mkDupableStrictAlt, create join points (much) more often: exprIsTrivial rather than exprIsDupable. This may be much, but we'll see. Metric Decrease: T12545 T13253-spj T13719 T18140 T18282 T18304 T18698a T18698b Metric Increase: T16577 T18923 T9961
* Add Word64#/Int64# primopsSylvain Henry2021-07-1512-33/+623
| | | | | | | | | | | | | | | | | | | | | | | Word64#/Int64# are only used on 32-bit architectures. Before this patch, operations on these types were directly using the FFI. Now we use real primops that are then lowered into ccalls. The advantage of doing this is that we can now perform constant folding on Word64#/Int64# (#19024). Most of this work was done by John Ericson in !3658. However this patch doesn't go as far as e.g. changing Word64 to always be using Word64#. Noticeable performance improvements T9203(normal) run/alloc 89870808.0 66662456.0 -25.8% GOOD haddock.Cabal(normal) run/alloc 14215777340.8 12780374172.0 -10.1% GOOD haddock.base(normal) run/alloc 15420020877.6 13643834480.0 -11.5% GOOD Metric Decrease: T9203 haddock.Cabal haddock.base
* Additional constant-folding rule for binary AND/ORSylvain Henry2021-07-151-10/+51
| | | | | | | | Add a constant folding rule allowing the subsumption of an application if the same argument is applied twice, e.g. (v .&. 0xFF) .&. 0xFF ~~> v .&. 0xFF (v .|. 0xFF) .|. 0xFF ~~> v .|. 0xFF
* Don't panic on 'no skolem info' and add failing testswip/no-skolem-panicZubin Duggal2021-07-132-3/+23
|
* driver: Fix interaction of -Wunused-packages and reexported-modulesMatthew Pickering2021-07-133-18/+36
| | | | | | | | | | | | | | Spurious warnings were previously emitted if an import came from a reexport due to how -Wunused-packages were implemented. Removing the dependency would cause compilation to fail. The fix is to reimplement the warning a bit more directly, by searching for which package each import comes from using the normal module finding functions rather than consulting the EPS. This has the advantage that the check could be performed at any time after downsweep rather than also relying on a populated EPS. Fixes #19518 and #19777
* Add proper GHCHints for most PsMessage constructorswip/adinapoli-issue-20055Alfredo Di Napoli2021-07-126-106/+197
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This commit adds proper hints to most diagnostic types in the `GHC.Parser.Errors.Types` module. By "proper" we mean that previous to this commit the hints were bundled together with the diagnostic message, whereas now we moved most of them as proper `[GhcHint]` in the implementation of `diagnosticHints`. More specifically, this is the list of constructors which now has proper hints: * PsErrIllegalBangPattern * PsWarnOperatorWhitespaceExtConflict * PsErrLambdaCase * PsErrIllegalPatSynExport * PsWarnOperatorWhitespace * PsErrMultiWayIf * PsErrIllegalQualifiedDo * PsErrNumUnderscores * PsErrLinearFunction * PsErrIllegalTraditionalRecordSyntax * PsErrIllegalExplicitNamespace * PsErrOverloadedRecordUpdateNotEnabled * PsErrIllegalDataTypeContext * PsErrSemiColonsInCondExpr * PsErrSemiColonsInCondCmd * PsWarnStarIsType * PsWarnImportPreQualified * PsErrImportPostQualified * PsErrEmptyDoubleQuotes * PsErrIllegalRoleName * PsWarnStarBinder For some reason, this patch increases the peak_megabyte_allocated of the T11545 test to 90 (from a baseline of 80) but that particular test doesn't emit any parsing diagnostic or hint and the metric increase happens only for the `aarch64-linux-deb10`. Metric Increase: T11545
* Don't return unitExpr in dsWhenNoErrssheaf2021-07-104-26/+94
| | | | | | | | | | | | | - fixes #18149 and #14765 dsWhenNoErrs now returns "runtimeError @ty" when disallowed representation polymorphism is detected, where ty is the type of the result CoreExpr. "ty" is passed as an additional argument to dsWhenNoErrs, and is used only in the case of such an error. The calls to dsWhenNoErrs must now compute the type of the CoreExpr they are trying to build, so that an error of the right type can be used in case of a representation polymorphism failure.
* Avoid unsafePerformIO for getProgNameSylvain Henry2021-07-092-15/+2
| | | | | | | | getProgName was used to append the name of the program (e.g. "ghc") to printed error messages in the Show instance of GhcException. It doesn't belong here as GHCi and GHC API users may want to override this behavior by setting a different error handler. So we now call it in the defaultErrorHandler instead.
* Add TcRnIllegalWildcardsInConstructor to TcRnMessageAlfredo Di Napoli2021-07-093-7/+18
|
* Remove redundant patSigErrAlfredo Di Napoli2021-07-091-16/+0
|
* Add TcRnCharLiteralOutOfRange to TcRnMessageAlfredo Di Napoli2021-07-093-6/+18
|
* Add TcRnIllegalViewPattern to TcRnMessageAlfredo Di Napoli2021-07-093-6/+21
|
* Add TcRnDuplicateFieldName to TcRnMessageAlfredo Di Napoli2021-07-093-16/+31
|
* Add TcRnIllegalWildCardsInRecord to TcRnMessageAlfredo Di Napoli2021-07-093-3/+37
|
* Add TcRnIllegalFieldPunning to TcRnMessageAlfredo Di Napoli2021-07-094-7/+27
|
* Add TcRnEmptyRecordUpdate to TcRnMessageAlfredo Di Napoli2021-07-093-4/+18
|
* Add TcRnIllegalPatSynDecl to TcRnMessageAlfredo Di Napoli2021-07-093-4/+23
|
* Add TcRnSimplifierTooManyIterations to TcRnMessageAlfredo Di Napoli2021-07-095-6/+33
|
* Add the TcRnDuplicateWarningDecls to TcRnMessageAlfredo Di Napoli2021-07-093-8/+20
|
* Add the TcRnShadowedName constructor to TcRnMessageAlfredo Di Napoli2021-07-093-16/+46
| | | | | This commit adds the TcRnShadowedName to the TcRnMessage type and it uses it in GHC.Rename.Utils.
* Rename getErrorMessages and getMessages function in parser codewip/adinapoli-issue-19920Alfredo Di Napoli2021-07-085-23/+22
| | | | | | | | | | | This commit renames the `getErrorMessages` and `getMessages` function in the parser code to `getPsErrorMessages` and `getPsMessages`, to avoid import conflicts, as we have already `getErrorMessages` and `getMessages` defined in `GHC.Types.Error`. Fixes #19920. Update haddock submodule
* driver: Convert runPipeline to use a free monadMatthew Pickering2021-07-0710-1929/+1896
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch converts the runPipeline function to be implemented in terms of a free monad rather than the previous CompPipeline. The advantages of this are three-fold: 1. Different parts of the pipeline can return different results, the limits of runPipeline were being pushed already by !5555, this opens up futher fine-grainedism of the pipeline. 2. The same mechanism can be extended to build-plan at the module level so the whole build plan can be expressed in terms of one computation which can then be treated uniformly. 3. The pipeline monad can now be interpreted in different ways, for example, you may want to interpret the `TPhase` action into the monad for your own build system (such as shake). That bit will probably require a bit more work, but this is a step in the right directin. There are a few more modules containing useful functions for interacting with the pipelines. * GHC.Driver.Pipeline: Functions for building pipelines at a high-level * GHC.Driver.Pipeline.Execute: Functions for providing the default interpretation of TPhase, in terms of normal IO. * GHC.Driver.Pipeline.Phases: The home for TPhase, the typed phase data type which dictates what the phases are. * GHC.Driver.Pipeline.Monad: Definitions to do with the TPipelineClass and MonadUse class. Hooks consumers may notice the type of the `phaseHook` has got slightly more restrictive, you can now no longer control the continuation of the pipeline by returning the next phase to execute but only override individual phases. If this is a problem then please open an issue and we will work out a solution. ------------------------- Metric Decrease: T4029 -------------------------
* Fix #19889 - Invalid BMI2 instructions generated.wip/andreask/bim-fixAndreas Klebinger2021-07-062-24/+26
| | | | | When arguments are 8 *or 16* bits wide, then truncate before/after and use the 32bit operation.
* Fix issue 20038 - Change 'variable' -> 'variables'Edward2021-07-061-1/+1
|
* Use target platform in guessOutputFileSylvain Henry2021-07-061-2/+3
|
* Arity: Handle shadowing properlySebastian Graf2021-07-021-9/+51
| | | | | | | | | | | | | | In #20070, we noticed that `findRhsArity` copes badly with shadowing. A simple function like `g_123 x_123 = x_123`, where the labmda binder shadows, already regressed badly. Indeed, the whole `arityType` function wasn't thinking about shadowing *at all*. I rectified that and established the invariant that `ae_join` and `am_sigs` should always be disjoint. That entails deleting bindings from `ae_join` whenever we add something to `am_sigs` and vice versa, which would otherwise be a bug in the making. That *should* fix (but I don't want to close it) #20070.
* Support unlifted datatypes in GHCiLuite Stegeman2021-07-024-42/+60
| | | | fixes #19628
* One-shot changes (#20008)Simon Peyton Jones2021-07-022-0/+21
| | | | | | | | | | | | | | | | | | | I discovered that GHC.Core.Unify.bindTv was getting arity 2, rather than 3, in one of my builds. In HEAD it does get the right arity, but only because CallArity (just) manages to spot it. In my situation it (just) failed to discover this. Best to make it robust, which this patch does. See Note [INLINE pragmas and (>>)] in GHC.Utils.Monad. There a bunch of other modules that probably should have the same treatment: GHC.CmmToAsm.Reg.Linear.State GHC.Tc.Solver.Monad GHC.Tc.Solver.Rewrite GHC.Utils.Monad.State.Lazy GHC.Utils.Monad.State.Strict but doing so is not part of this patch
* float out some docstrings and comment some function parametersMarco Zocca2021-07-011-10/+27
|
* [Parser: Lexer] Fix !6132Moritz Angermann2021-07-011-1/+2
| | | | clang's cpp injects spaces prior to #!/.
* Dynflags: introduce DiagOptsSylvain Henry2021-07-0120-192/+267
| | | | | | | | | | | | | | | | | | | | | | Use DiagOpts for diagnostic options instead of directly querying DynFlags (#17957). Surprising performance improvements on CI: T4801(normal) ghc/alloc 313236344.0 306515216.0 -2.1% GOOD T9961(normal) ghc/alloc 384502736.0 380584384.0 -1.0% GOOD ManyAlternatives(normal) ghc/alloc 797356128.0 786644928.0 -1.3% ManyConstructors(normal) ghc/alloc 4389732432.0 4317740880.0 -1.6% T783(normal) ghc/alloc 408142680.0 402812176.0 -1.3% Metric Decrease: T4801 T9961 T783 ManyAlternatives ManyConstructors Bump haddock submodule
* Remove useless .hs-bootSylvain Henry2021-07-0116-39/+2
|
* Make withException use SDocContext instead of DynFlagsSylvain Henry2021-07-014-9/+14
|
* Use HsExpansion for overloaded list patternssheaf2021-06-2915-229/+413
| | | | Fixes #14380, #19997
* Demand: Better representation (#19050)Sebastian Graf2021-06-292-186/+351
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In #19050, we identified several ways in which we could make more illegal states irrepresentable. This patch introduces a few representation changes around `Demand` and `Card` with a better and earlier-failing API exported through pattern synonyms. Specifically, 1. The old enum definition of `Card` led to severely bloated code of operations on it. I switched to a bit vector representation; much nicer overall IMO. See Note [Bit vector representation for Card]. Most of the gripes with the old representation were related to where which kind of `Card` was allowed and the fact that it doesn't make sense for an absent or bottoming demand to carry a `SubDemand` that describes an evaluation context that is never realised. 2. So I refactored the `Demand` representation so that it has two new data constructors for `AbsDmd` and `BotDmd`. The old `(:*)` data constructor becomes a pattern synonym which expands absent demands as needed, so that it still forms a complete match and a versatile builder. The new `Demand` data constructor now carries a `CardNonAbs` and only occurs in a very limited number of internal call sites. 3. Wherever a full-blown `Card` might end up in a `CardNonAbs` field (like that of `D` or `Call`), I assert the consistency. When the smart builder of `(:*)` is called with an absent `Card`, I assert that the `SubDemand` is the same that we would expand to in the matcher. 4. `Poly` now takes a `CardNonOnce` and encodes the previously noticed invariant that we never produce `Poly C_11` or `Poly C_01`. I made sure that we never construct a `Poly` with `C_11` or `C_01`. Fixes #19050. We lose a tiny bit of anal perf overall, probably because the new `Demand` definition can't be unboxed. The biggest loser is WWRec, where allocations go from 16MB to 26MB in DmdAnal, making up for a total increase of (merely) 1.6%. It's all within acceptance thresholds. There are even two ghc/alloc metric decreases. T11545 decreases by *67%*! Metric Decrease: T11545 T18304
* Fix type and strictness signature of forkOn#Ryan Scott2021-06-281-1/+5
| | | | | | This is a follow-up to #19992, which fixes the type and strictness signature for `fork#`. The `forkOn#` primop also needs analogous changes, which this patch accomplishes.
* Try to simplify zoo of functions in `Tc.Utils.Monad`Alfredo Di Napoli2021-06-2848-788/+1259
| | | | | | | | | | | This commit tries to untangle the zoo of diagnostic-related functions in `Tc.Utils.Monad` so that we can have the interfaces mentions only `TcRnMessage`s while we push the creation of these messages upstream. It also ports TcRnMessage diagnostics to use the new API, in particular this commit switch to use TcRnMessage in the external interfaces of the diagnostic functions, and port the old SDoc to be wrapped into TcRnUnknownMessage.
* Revert "Make reallyUnsafePtrEquality# levity-polymorphic"Matthew Pickering2021-06-271-1/+1
| | | | | | | | | | | | | | | | | | | | | | | This reverts commit d1f59540e8b7be96b55ab4b286539a70bc75416c. This commit breaks the build of unordered-containers ``` [3 of 9] Compiling Data.HashMap.Internal.Array ( Data/HashMap/Internal/Array.hs, dist/build/Data/HashMap/Internal/Array.o, dist/build/Data/HashMap/Internal/Array.dyn_o ) *** Parser [Data.HashMap.Internal.Array]: Parser [Data.HashMap.Internal.Array]: alloc=21043544 time=13.621 *** Renamer/typechecker [Data.HashMap.Internal.Array]: Renamer/typechecker [Data.HashMap.Internal.Array]: alloc=151218672 time=187.083 *** Desugar [Data.HashMap.Internal.Array]: ghc: panic! (the 'impossible' happened) GHC version 9.3.20210625: expectJust splitFunTy CallStack (from HasCallStack): error, called at compiler/GHC/Data/Maybe.hs:68:27 in ghc:GHC.Data.Maybe expectJust, called at compiler/GHC/Core/Type.hs:1247:14 in ghc:GHC.Core.Type ``` Revert containers submodule update
* Inliner: Regard LitRubbish as TrivArg and not ConLikeSebastian Graf2021-06-272-2/+7
| | | | | | | | | | | | | | | | | | | | | | Part of fixing #19766 required the emission of `LitRubbish` as absent filler in places where we used `absentError` before. In WWRec we have the situation that such bindings occur in the argument to functions. With `LitRubbish` we inlined those functions, because 1. The absent binding was regarded as ConLike. So I fixed `exprIsHNFLike` to respond `False` to `LitRubbish`. 2. The other source of inlining was that after inlining such an absent binding, `LitRubbish` itself was regarded `ValueArg` by `interestingArg`, leading to more inlining. It now responds `TrivArg` to `LitRubbish`. Fixes #20035. There's one slight 1.6% ghc/alloc regression left in T15164 that is due to an additional specialisation `$s$cget`. I've no idea why that happens; the Core output before is identical and has the call site that we specialise for. Metric Decrease: WWRec
* WorkWrap: Make mkWWstr and mkWWcpr generate fewer let bindingsSebastian Graf2021-06-272-75/+69
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In https://gitlab.haskell.org/ghc/ghc/-/merge_requests/5814#note_355144, Simon noted that `mkWWstr` and `mkWWcpr` could generate fewer let bindings and be implemented less indirectly by returning the rebuilt expressions directly, e.g. instead of ``` f :: (Int, Int) -> Int f (x, y) = x+y ==> f :: (Int, Int) -> Int f p = case p of (x, y) -> case x of I# x' -> case y of I# y' -> case $wf x' y' of r' -> let r = I# r' -- immediately returned in r f :: Int# -> Int# -> Int# $wf x' y' = let x = I# x' in -- only used in p let y = I# y' in -- only used in p let p = (x, y) in -- only used in the App below case (\(x,y) -> x+y) p of I# r' -> r' ``` we know generate ``` f :: (Int, Int) -> Int f p = case p of (x, y) -> case x of I# x' -> case y of I# y' -> case $wf x' y' of r' -> I# r' -- 1 fewer let f :: Int# -> Int# -> Int# $wf x' y' = case (\(x,y) -> x+y) (I# x, I# y) of I# r' -> -- 3 fewer lets r' ``` Which is much nicer and makes it easier to comprehend the output of worker-wrapper pre-Simplification as well as puts less strain on the Simplifier. I had to drop support for #18983, but we found that it's broken anyway. Simon is working on a patch that provides a bit more justification.
View Lorry Controller Status