summaryrefslogtreecommitdiff
path: root/compiler
Commit message (Collapse)AuthorAgeFilesLines
* Account for special GHC.Prim import in warnUnusedPackagesMatthew Pickering2023-04-171-3/+7
| | | | | | | | | | | | The GHC.Prim import is treated quite specially primarily because there isn't an interface file for GHC.Prim. Therefore we record separately in the ModSummary if it's imported or not so we don't go looking for it. This logic hasn't made it's way to `-Wunused-packages` so if you imported GHC.Prim then the warning would complain you didn't use `-package ghc-prim`. Fixes #23212
* Add import lists to few GHC.Driver.Session importsOleg Grenrus2023-04-164-4/+6
| | | | | | Related to https://gitlab.haskell.org/ghc/ghc/-/issues/23261. There are a lot of GHC.Driver.Session which only use DynFlags, but not the parsing code.
* Transfer DFunId_ness onto specialised bindingsSimon Peyton Jones2023-04-161-41/+78
| | | | | | | | | | | | | | | | | | | | Whether a binding is a DFunId or not has consequences for the `-fdicts-strict` flag, essentially if we are doing demand analysis for a DFunId then `-fdicts-strict` does not apply because the constraint solver can create recursive groups of dictionaries. In #22549 this was fixed for the "normal" case, see Note [Do not strictify the argument dictionaries of a dfun]. However the loop still existed if the DFunId was being specialised. The problem was that the specialiser would specialise a DFunId and turn it into a VanillaId and so the demand analyser didn't know to apply special treatment to the binding anymore and the whole recursive group was optimised to bottom. The solution is to transfer over the DFunId-ness of the binding in the specialiser so that the demand analyser knows not to apply the `-fstrict-dicts`. Fixes #22549
* Handle ConcreteTvs in inferResultToTypewip/T23153sheaf2023-04-156-31/+53
| | | | | | | | | | inferResultToType was discarding the ir_frr information, which meant some metavariables ended up being MetaTvs instead of ConcreteTvs. This function now creates new ConcreteTvs as necessary, instead of always creating MetaTvs. Fixes #23154
* Show an error when we cannot default a concrete tyvarKrzysztof Gogolewski2023-04-154-4/+31
| | | | Fixes #23153
* Make approximateWC a bit clevererwip/T22194-flagsSimon Peyton Jones2023-04-141-61/+94
| | | | | | | | | | | This MR fixes #23224: making approximateWC more clever See the long `Note [ApproximateWC]` in GHC.Tc.Solver All this is delicate and ad-hoc -- but it /has/ to be: we are talking about inferring a type for a binding in the presence of GADTs, type families and whatnot: known difficult territory. We just try as hard as we can.
* Improve partial signaturesSimon Peyton Jones2023-04-144-104/+137
| | | | | | | | | | | | | | | | | | | | | | | This MR fixes #23223. The changes are in two places: * GHC.Tc.Bind.checkMonomorphismRestriction See the new `Note [When the MR applies]` We now no longer stupidly attempt to apply the MR when the user specifies a context, e.g. f :: Eq a => _ -> _ * GHC.Tc.Solver.decideQuantification See rewritten `Note [Constraints in partial type signatures]` Fixing this bug apparently breaks three tests: * partial-sigs/should_compile/T11192 * partial-sigs/should_fail/Defaulting1MROff * partial-sigs/should_fail/T11122 However they are all symptoms of #23232, so I'm marking them as expect_broken(23232). I feel happy about this MR. Nice.
* Add some documentation about redundant constraintsSimon Peyton Jones2023-04-142-5/+28
|
* Major refactor in the handling of equality constraintsSimon Peyton Jones2023-04-1420-1165/+2061
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This MR substantially refactors the way in which the constraint solver deals with equality constraints. The big thing is: * Intead of a pipeline in which we /first/ canonicalise and /then/ interact (the latter including performing unification) the two steps are more closely integreated into one. That avoids the current rather indirect communication between the two steps. The proximate cause for this refactoring is fixing #22194, which involve solving [W] alpha[2] ~ Maybe (F beta[4]) by doing this: alpha[2] := Maybe delta[2] [W] delta[2] ~ F beta[4] That is, we don't promote beta[4]! This is very like introducing a cycle breaker, and was very awkward to do before, but now it is all nice. See GHC.Tc.Utils.Unify Note [Promotion and level-checking] and Note [Family applications in canonical constraints]. The big change is this: * Several canonicalisation checks (occurs-check, cycle-breaking, checking for concreteness) are combined into one new function: GHC.Tc.Utils.Unify.checkTyEqRhs This function is controlled by `TyEqFlags`, which says what to do for foralls, type families etc. * `canEqCanLHSFinish` now sees if unification is possible, and if so, actually does it: see `canEqCanLHSFinish_try_unification`. There are loads of smaller changes: * The on-the-fly unifier `GHC.Tc.Utils.Unify.unifyType` has a cheap-and-cheerful version of `checkTyEqRhs`, called `simpleUnifyCheck`. If `simpleUnifyCheck` succeeds, it can unify, otherwise it defers by emitting a constraint. This is simpler than before. * I simplified the swapping code in `GHC.Tc.Solver.Equality.canEqCanLHS`. Especially the nasty stuff involving `swap_for_occurs` and `canEqTyVarFunEq`. Much nicer now. See Note [Orienting TyVarLHS/TyFamLHS] Note [Orienting TyFamLHS/TyFamLHS] * Added `cteSkolemOccurs`, `cteConcrete`, and `cteCoercionHole` to the problems that can be discovered by `checkTyEqRhs`. * I fixed #23199 `pickQuantifiablePreds`, which actually allows GHC to to accept both cases in #22194 rather than rejecting both. Yet smaller: * Added a `synIsConcrete` flag to `SynonymTyCon` (alongside `synIsFamFree`) to reduce the need for synonym expansion when checking concreteness. Use it in `isConcreteType`. * Renamed `isConcrete` to `isConcreteType` * Defined `GHC.Core.TyCo.FVs.isInjectiveInType` as a more efficient way to find if a particular type variable is used injectively than finding all the injective variables. It is called in `GHC.Tc.Utils.Unify.definitely_poly`, which in turn is used quite a lot. * Moved `rewriterView` to `GHC.Core.Type`, so we can use it from the constraint solver. Fixes #22194, #23199 Compile times decrease by an average of 0.1%; but there is a 7.4% drop in compiler allocation on T15703. Metric Decrease: T15703
* Stop if type constructors have kind errorsSimon Peyton Jones2023-04-141-1/+7
| | | | | | | | | Otherwise we get knock-on errors, such as #23252. This makes GHC fail a bit sooner, and I have not attempted to add recovery code, to add a fake TyCon place of the erroneous one, in an attempt to get more type errors in one pass. We could do that (perhaps) if there was a call for it.
* Make WordQuotRem2Op ok-for-speculation tooSylvain Henry2023-04-132-3/+6
|
* Make Int64/Word64 division ok for speculation too.Sylvain Henry2023-04-131-1/+7
| | | | Only when the divisor is definitely non-zero.
* Add quot folding rule (#22152)Sylvain Henry2023-04-132-16/+89
| | | | | | | (x / l1) / l2 l1 and l2 /= 0 l1*l2 doesn't overflow ==> x / (l1 * l2)
* Add quotRem rules (#22152)Sylvain Henry2023-04-132-0/+65
| | | | | | | | | | | | | | case quotRemInt# x y of (# q, _ #) -> body ====> case quotInt# x y of q -> body case quotRemInt# x y of (# _, r #) -> body ====> case remInt# x y of r -> body
* Allow generation of TTH syntax with THOleg Grenrus2023-04-121-0/+4
| | | | | | | | | | | | | | | | | | | | In other words allow generation of typed splices and brackets with Untyped Template Haskell. That is useful in cases where a library is build with TTH in mind, but we still want to generate some auxiliary declarations, where TTH cannot help us, but untyped TH can. Such example is e.g. `staged-sop` which works with TTH, but we would like to derive `Generic` declarations with TH. An alternative approach is to use `unsafeCodeCoerce`, but then the derived `Generic` instances would be type-checked only at use sites, i.e. much later. Also `-ddump-splices` output is quite ugly: user-written instances would use TTH brackets, not `unsafeCodeCoerce`. This commit doesn't allow generating of untyped template splices and brackets with untyped TH, as I don't know why one would want to do that (instead of merging the splices, e.g.)
* Clarify a couple of Notes about 'nospec'Simon Peyton Jones2023-04-112-5/+12
|
* compiler: apply cmm node-splitting for wasm backendCheng Shao2023-04-112-3/+10
| | | | | This patch applies cmm node-splitting for wasm32 NCG, which is required when handling irreducible CFGs. Fixes #23237.
* compiler: make WasmCodeGenM an instance of MonadUniqueCheng Shao2023-04-112-6/+14
|
* Renamer: don't call addUsedGRE on an exact Namesheaf2023-04-073-16/+16
| | | | | | | | | | When looking up a record field in GHC.Rename.Env.lookupRecFieldOcc, we could end up calling addUsedGRE on an exact Name, which would then lead to a panic in the bestImport function: it would be incapable of processing a GRE which is not local but also not brought into scope by any imports (as it is referred to by its unique instead). Fixes #23240
* Add structured error messages for GHC.Tc.TyCl.UtilsTorsten Schmits2023-04-074-15/+39
| | | | | | | | | Tracking ticket: #20117 MR: !10251 This converts uses of `mkTcRnUnknownMessage` to newly added constructors of `TcRnMessage`.
* Fix unification with oversaturated type familiesKrzysztof Gogolewski2023-04-041-29/+115
| | | | | | unify_ty was incorrectly saying that F x y ~ T x are surely apart, where F x y is an oversaturated type family and T x is a tyconapp. As a result, the simplifier dropped a live case alternative (#23134).
* Make INLINE pragmas for pattern synonyms work with THRyan Scott2023-04-041-1/+4
| | | | | | | | | | Previously, the code for converting `INLINE <name>` pragmas from TH splices used `vNameN`, which assumed that `<name>` must live in the variable namespace. Pattern synonyms, on the other hand, live in the constructor namespace. I've fixed the issue by switching to `vcNameN` instead, which works for both the variable and constructor namespaces. Fixes #23203.
* GHCi.RemoteTypes: fix doc and avoid unsafeCoerce (#23201)Sylvain Henry2023-04-041-4/+2
|
* Relax assertion in varToRecFieldOccsheaf2023-04-042-4/+7
| | | | | | | | | | | | When using Template Haskell, it is possible to re-parent a field OccName belonging to one data constructor to another data constructor. The lsp-types package did this in order to "extend" a data constructor with additional fields. This ran into an assertion in 'varToRecFieldOcc'. This assertion can simply be relaxed, as the resulting splices are perfectly sound. Fixes #23220
* StgToCmm: Upgrade -fcheck-prim-bounds behaviorMatthew Craven2023-04-043-107/+177
| | | | | Fixes #21054. Additionally, we can now check for range overlap when generating Cmm for primops that use memcpy internally.
* Add a few more memcpy-ish primopsMatthew Craven2023-04-043-15/+134
| | | | | | | | | | | | | | | | * copyMutableByteArrayNonOverlapping# * copyAddrToAddr# * copyAddrToAddrNonOverlapping# * setAddrRange# The implementations of copyBytes, moveBytes, and fillBytes in base:Foreign.Marshal.Utils now use these new primops, which can cause us to work a bit harder generating code for them, resulting in the metric increase in T21839c observed by CI on some architectures. But in exchange, we get better code! Metric Increase: T21839c
* Fixes around unsafeCoerce#Krzysztof Gogolewski2023-04-032-54/+16
| | | | | | | | | | | | 1. `unsafeCoerce#` was documented in `GHC.Prim`. But since the overhaul in 74ad75e87317, `unsafeCoerce#` is no longer defined there. I've combined the documentation in `GHC.Prim` with the `Unsafe.Coerce` module. 2. The documentation of `unsafeCoerce#` stated that you should not cast a function to an algebraic type, even if you later cast it back before applying it. But ghci was doing that type of cast, as can be seen with 'ghci -ddump-ds' and typing 'x = not'. I've changed it to use Any following the documentation.
* Convert diagnostics in GHC.Rename.HsType to proper TcRnMessageHaskell-mouse2023-04-035-79/+333
| | | | | | | | | | | | | | | | | | I've turned all occurrences of TcRnUnknownMessage in GHC.Rename.HsType module into a proper TcRnMessage. Instead, these TcRnMessage messages were introduced: TcRnDataKindsError TcRnUnusedQuantifiedTypeVar TcRnIllegalKindSignature TcRnUnexpectedPatSigType TcRnSectionPrecedenceError TcRnPrecedenceParsingError TcRnIllegalKind TcRnNegativeNumTypeLiteral TcRnUnexpectedKindVar TcRnBindMultipleVariables TcRnBindVarAlreadyInScope
* cmm: implement parsing of MO_AtomicRMW from hand-written CMM filesBodigrim2023-04-021-0/+6
| | | | Fixes #23206
* TH: revert changes to GadtC & RecGadtCsheaf2023-04-016-29/+51
| | | | | | | Commit 3f374399 included a breaking-change to the template-haskell library when it made the GadtC and RecGadtC constructors take non-empty lists of names. As this has the potential to break many users' packages, we decided to revert these changes for now.
* JS: Linker: use saturated JExprdoyougnu2023-04-012-100/+5
| | | | Follow on to MR!10142 in pursuit of #22736
* Add structured error messages for GHC.Tc.TyClTorsten Schmits2023-04-0111-303/+792
| | | | | | | | | Tracking ticket: #20117 MR: !10183 This converts uses of `mkTcRnUnknownMessage` to newly added constructors of `TcRnMessage`.
* driver: Unit State Data.Map -> GHC.Unique.UniqMapdoyougnu2023-04-0112-159/+199
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | In pursuit of #22426. The driver and unit state are major contributors. This commit also bumps the haddock submodule to reflect the API changes in UniqMap. ------------------------- Metric Decrease: MultiComponentModules MultiComponentModulesRecomp T10421 T10547 T12150 T12234 T12425 T13035 T16875 T18140 T18304 T18698a T18698b T18923 T20049 T5837 T6048 T9198 -------------------------
* Add LANGUAGE GADTs to GHC.Rename.Env sheaf2023-03-301-1/+1
| | | | | We need to enable this extension for the file to compile with ghc 9.2, as we are pattern matching on a GADT and this required the GADT extension to be enabled until 9.4.
* Proxies for head and tail: review suggestionssheaf2023-03-301-6/+5
|
* Fixes to accomodate Data.List.{head,tail} with {-# WARNING #-}Bodigrim2023-03-302-3/+27
|
* js: split JMacro into JS eDSL and JS syntaxdoyougnu2023-03-3031-455/+888
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This commit: Splits JExpr and JStat into two nearly identical DSLs: - GHC.JS.Syntax is the JMacro based DSL without unsaturation, i.e., a value cannot be unsaturated, or, a value of this DSL is a witness that a value of GHC.JS.Unsat has been saturated - GHC.JS.Unsat is the JMacro DSL from GHCJS with Unsaturation. Then all binary and outputable instances are changed to use GHC.JS.Syntax. This moves us closer to closing out #22736 and #22352. See #22736 for roadmap. ------------------------- Metric Increase: CoOpt_Read LargeRecord ManyAlternatives PmSeriesS PmSeriesT PmSeriesV T10421 T10858 T11195 T11374 T11822 T12227 T12707 T13035 T13253 T13253-spj T13379 T14683 T15164 T15703 T16577 T17096 T17516 T17836 T18140 T18282 T18304 T18478 T18698a T18698b T18923 T1969 T19695 T20049 T21839c T3064 T4801 T5321FD T5321Fun T5631 T5642 T783 T9198 T9233 T9630 TcPlugin_RewritePerf WWRec -------------------------
* igre_prompt_env: discard "only-qualified" namessheaf2023-03-294-10/+27
| | | | | | | | | | | | | | | We were unnecessarily carrying around names only available qualified in igre_prompt_env, violating the icReaderEnv invariant. We now get rid of these, as they aren't needed for the shadowing computation that igre_prompt_env exists for. Fixes #23177 ------------------------- Metric Decrease: T14052 T14052Type -------------------------
* Avoid repeatedly shadowing in shadowNamessheaf2023-03-292-40/+154
| | | | | | | This commit refactors GHC.Type.Name.Reader.shadowNames to first accumulate all the shadowing arising from the introduction of a new set of GREs, and then applies all the shadowing to the old GlobalRdrEnv in one go.
* Handle records in the renamersheaf2023-03-29104-3762/+5143
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch moves the field-based logic for disambiguating record updates to the renamer. The type-directed logic, scheduled for removal, remains in the typechecker. To do this properly (and fix the myriad of bugs surrounding the treatment of duplicate record fields), we took the following main steps: 1. Create GREInfo, a renamer-level equivalent to TyThing which stores information pertinent to the renamer. This allows us to uniformly treat imported and local Names in the renamer, as described in Note [GREInfo]. 2. Remove GreName. Instead of a GlobalRdrElt storing GreNames, which distinguished between normal names and field names, we now store simple Names in GlobalRdrElt, along with the new GREInfo information which allows us to recover the FieldLabel for record fields. 3. Add namespacing for record fields, within the OccNames themselves. This allows us to remove the mangling of duplicate field selectors. This change ensures we don't print mangled names to the user in error messages, and allows us to handle duplicate record fields in Template Haskell. 4. Move record disambiguation to the renamer, and operate on the level of data constructors instead, to handle #21443. The error message text for ambiguous record updates has also been changed to reflect that type-directed disambiguation is on the way out. (3) means that OccEnv is now a bit more complex: we first key on the textual name, which gives an inner map keyed on NameSpace: OccEnv a ~ FastStringEnv (UniqFM NameSpace a) Note that this change, along with (2), both increase the memory residency of GlobalRdrEnv = OccEnv [GlobalRdrElt], which causes a few tests to regress somewhat in compile-time allocation. Even though (3) simplified a lot of code (in particular the treatment of field selectors within Template Haskell and in error messages), it came with one important wrinkle: in the situation of -- M.hs-boot module M where { data A; foo :: A -> Int } -- M.hs module M where { data A = MkA { foo :: Int } } we have that M.hs-boot exports a variable foo, which is supposed to match with the record field foo that M exports. To solve this issue, we add a new impedance-matching binding to M foo{var} = foo{fld} This mimics the logic that existed already for impedance-binding DFunIds, but getting it right was a bit tricky. See Note [Record field impedance matching] in GHC.Tc.Module. We also needed to be careful to avoid introducing space leaks in GHCi. So we dehydrate the GlobalRdrEnv before storing it anywhere, e.g. in ModIface. This means stubbing out all the GREInfo fields, with the function forceGlobalRdrEnv. When we read it back in, we rehydrate with rehydrateGlobalRdrEnv. This robustly avoids any space leaks caused by retaining old type environments. Fixes #13352 #14848 #17381 #17551 #19664 #21443 #21444 #21720 #21898 #21946 #21959 #22125 #22160 #23010 #23062 #23063 Updates haddock submodule ------------------------- Metric Increase: MultiComponentModules MultiLayerModules MultiLayerModulesDefsGhci MultiLayerModulesNoCode T13701 T14697 hard_hole_fits -------------------------
* Make exprIsConApp_maybe a bit clevererwip/T23159Simon Peyton Jones2023-03-271-8/+54
| | | | | | | | | | | | | | | | | | | | | | | Addresses #23159. See Note Note [Exploit occ-info in exprIsConApp_maybe] in GHC.Core.SimpleOpt. Compile times go down very slightly, but always go down, never up. Good! Metrics: compile_time/bytes allocated ------------------------------------------------ CoOpt_Singletons(normal) -1.8% T15703(normal) -1.2% GOOD geo. mean -0.1% minimum -1.8% maximum +0.0% Metric Decrease: CoOpt_Singletons T15703
* Improve documentation for resizing of byte arraysBodigrim2023-03-251-11/+35
|
* Improve documentation of atomicModifyMutVar2#Bodigrim2023-03-251-5/+17
|
* codeGen/tsan: Disable instrumentation of unaligned storesBen Gamari2023-03-241-8/+7
| | | | | | | | | | There is some disagreement regarding the prototype of `__tsan_unaligned_write` (specifically whether it takes just the written address, or the address and the value as an argument). Moreover, I have observed crashes which appear to be due to it. Disable instrumentation of unaligned stores as a temporary mitigation. Fixes #23096.
* nativeGen/AArch64: Fix bitmask immediate predicateBen Gamari2023-03-241-15/+35
| | | | | | | | | Previously the predicate for determining whether a logical instruction operand could be encoded as a bitmask immediate was far too conservative. This meant that, e.g., pointer untagged required five instructions whereas it should only require one. Fixes #23030.
* Allow WARNING pragmas to be controlled with custom categoriesAdam Gundry2023-03-2416-60/+387
| | | | | | | | | | | | | | | | Closes #17209. This implements GHC Proposal 541, allowing a WARNING pragma to be annotated with a category like so: {-# WARNING in "x-partial" head "This function is undefined on empty lists." #-} The user can then enable, disable and set the severity of such warnings using command-line flags `-Wx-partial`, `-Werror=x-partial` and so on. There is a new warning group `-Wextended-warnings` containing all these warnings. Warnings without a category are treated as if the category was `deprecations`, and are (still) controlled by the flags `-Wdeprecations` and `-Wwarnings-deprecations`. Updates Haddock submodule.
* Add structured error messages for GHC.Tc.TyCl.PatSynTorsten Schmits2023-03-234-57/+139
| | | | | | | | | Tracking ticket: #20117 MR: !10158 This converts uses of `mkTcRnUnknownMessage` to newly added constructors of `TcRnMessage`.
* Add structured error messages for GHC.Tc.Utils.TcMTypeTorsten Schmits2023-03-227-60/+129
| | | | | | | | | Tracking ticket: #20119 MR: !10138 This converts uses of `mkTcRnUnknownMessage` to newly added constructors of `TcRnMessage`.
* Refactor the constraint solver pipelinewip/T23070Simon Peyton Jones2023-03-2213-4114/+4256
| | | | | | | | | | | | | | | | The big change is to put the entire type-equality solver into GHC.Tc.Solver.Equality, rather than scattering it over Canonical and Interact. Other changes * EqCt becomes its own data type, a bit like QCInst. This is great because EqualCtList is then just [EqCt] * New module GHC.Tc.Solver.Dict has come of the class-contraint solver. In due course it will be all. One step at a time. This MR is intended to have zero change in behaviour: it is a pure refactor. It opens the way to subsequent tidying up, we believe.
* Be more careful about quantificationSimon Peyton Jones2023-03-227-160/+215
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This MR is driven by #23051. It does several things: * It is guided by the generalisation plan described in #20686. But it is still far from a complete implementation of that plan. * Add Note [Inferred type with escaping kind] to GHC.Tc.Gen.Bind. This explains that we don't (yet, pending #20686) directly prevent generalising over escaping kinds. * In `GHC.Tc.Utils.TcMType.defaultTyVar` we default RuntimeRep and Multiplicity variables, beause we don't want to quantify over them. We want to do the same for a Concrete tyvar, but there is nothing sensible to default it to (unless it has kind RuntimeRep, in which case it'll be caught by an earlier case). So we promote instead. * Pure refactoring in GHC.Tc.Solver: * Rename decideMonoTyVars to decidePromotedTyVars, since that's what it does. * Move the actual promotion of the tyvars-to-promote from `defaultTyVarsAndSimplify` to `decidePromotedTyVars`. This is a no-op; just tidies up the code. E.g then we don't need to return the promoted tyvars from `decidePromotedTyVars`. * A little refactoring in `defaultTyVarsAndSimplify`, but no change in behaviour. * When making a TauTv unification variable into a ConcreteTv (in GHC.Tc.Utils.Concrete.makeTypeConcrete), preserve the occ-name of the type variable. This just improves error messages. * Kill off dead code: GHC.Tc.Utils.TcMType.newConcreteHole
View Lorry Controller Status