summaryrefslogtreecommitdiff
path: root/compiler/GHC/Driver
Commit message (Collapse)AuthorAgeFilesLines
* Move 'nextWrapperNum' into 'DsM' and 'TcM'Fendor2021-04-221-5/+0
| | | | | | | | Previously existing in 'DynFlags', 'nextWrapperNum' is a global variable mapping a Module to a number for name generation for FFI calls. This is not the right location for 'nextWrapperNum', as 'DynFlags' should not contain just about any global variable.
* driver: Consider dyn_o files when checking recompilation in -cMatthew Pickering2021-04-222-8/+16
| | | | | | | | | | | | When -dynamic-too is enabled, there are two result files, .o and .dyn_o, therefore we should check both to decide whether to set SourceModified or not. The whole recompilation logic is very messy, a more thorough refactor would be beneficial in this area but this is the minimal patch to fix this more high priority problem. Fixes #17968 and hopefully #17534
* Always generate ModDetails from ModIfaceMatthew Pickering2021-04-143-29/+73
| | | | | | | | | | | | | | | | | | This vastly reduces memory usage when compiling with `--make` mode, from about 900M when compiling Cabal to about 300M. As a matter of uniformity, it also ensures that reading from an interface performs the same as using the in-memory cache. We can also delete all the horrible knot-tying in updateIdInfos. Goes some way to fixing #13586 Accept new output of tests fixing some bugs along the way ------------------------- Metric Decrease: T12545 -------------------------
* Read constants header instead of global platformConstantsSylvain Henry2021-04-103-11/+26
| | | | | | | | | | | | | | | | | | | | | | | | | | With this patch we switch from reading the globally installed platformConstants file to reading the DerivedConstants.h header file that is bundled in the RTS unit. When we build the RTS unit itself, we get it from its includes directories. The new parser is more efficient and strict than the Read instance for PlatformConstants and we get about 2.2MB less allocations in every cases. However it only really shows in tests that don't allocate much, hence the following metric decreases. Metric Decrease: Naperian T10421 T10547 T12150 T12234 T12425 T13035 T18304 T18923 T5837 T6048 T18140
* Remove dynamic-by-default (#16782)Sylvain Henry2021-04-071-12/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | Dynamic-by-default was a mechanism to automatically select the -dynamic way for some targets. It was implemented in a convoluted way: it was defined as a flavour option, hence it couldn't be passed as a global settings (which are produced by `configure` before considering flavours), so a build system rule was used to pass -DDYNAMIC_BY_DEFAULT to the C compiler so that deriveConstants could infer it. * Make build system has it disabled for 8 years (951e28c0625ece7e0db6ac9d4a1e61e2737b10de) * It has never been implemented in Hadrian * Last time someone tried to enable it 1 year ago it didn't work (!2436) * Having this as a global constant impedes making GHC multi-target (see !5427) This commit fully removes support for dynamic-by-default. If someone wants to reimplement something like this, it would probably need to move the logic in the compiler. (Doing this would probably need some refactoring of the way the compiler handles DynFlags: DynFlags are used to store and to pass enabled ways to many parts of the compiler. It can be set by command-line flags, GHC API, global settings. In multi-target GHC, we will use DynFlags to load the target platform and its constants: but at this point with the current DynFlags implementation we can't easily update the existing DynFlags with target-specific options such as dynamic-by-default without overriding ways previously set by the user.)
* Correct warning for deprecated and unrecognised flagsAlfredo Di Napoli2021-04-051-13/+3
| | | | | | | | | | Fixes #19616. This commit changes the `GHC.Driver.Errors.handleFlagWarnings` function to rely on the newly introduced `DiagnosticReason`. This allows us to correctly pretty-print the flags which triggered some warnings and in turn remove the cruft around this function (like the extra filtering and the `shouldPrintWarning` function.
* Introduce SevIgnore Severity to suppress warningsAlfredo Di Napoli2021-04-054-26/+26
| | | | | | | | | | | | | | | | | | | | | | This commit introduces a new `Severity` type constructor called `SevIgnore`, which can be used to classify diagnostic messages which are not meant to be displayed to the user, for example suppressed warnings. This extra constructor allows us to get rid of a bunch of redundant checks when emitting diagnostics, typically in the form of the pattern: ``` when (optM Opt_XXX) $ addDiagnosticTc (WarningWithFlag Opt_XXX) ... ``` Fair warning! Not all checks should be omitted/skipped, as evaluating some data structures used to produce a diagnostic might still be expensive (e.g. zonking, etc). Therefore, a case-by-case analysis must be conducted when deciding if a check can be removed or not. Last but not least, we remove the unnecessary `CmdLine.WarnReason` type, which is now redundant with `DiagnosticReason`.
* Compute Severity of diagnostics at birthAlfredo Di Napoli2021-04-018-100/+74
| | | | | | | | | | | | | | | | | | | | | This commit further expand on the design for #18516 by getting rid of the `defaultReasonSeverity` in favour of a function called `diagReasonSeverity` which correctly takes the `DynFlags` as input. The idea is to compute the `Severity` and the `DiagnosticReason` of each message "at birth", without doing any later re-classifications, which are potentially error prone, as the `DynFlags` might evolve during the course of the program. In preparation for a proper refactoring, now `pprWarning` from the Parser.Ppr module has been renamed to `mkParserWarn`, which now takes a `DynFlags` as input. We also get rid of the reclassification we were performing inside `printOrThrowWarnings`. Last but not least, this commit removes the need for reclassify inside GHC.Tc.Errors, and also simplifies the implementation of `maybeReportError`. Update Haddock submodule
* Move the EPS into UnitEnvSylvain Henry2021-04-015-17/+6
|
* Move HPT in UnitEnvSylvain Henry2021-04-016-43/+28
|
* Move unit DBs in UnitEnvSylvain Henry2021-04-015-19/+7
| | | | | Also make the HomeUnit optional to keep the field strict and prepare for UnitEnvs without a HomeUnit (e.g. in Plugins envs, cf #14335).
* Properly initialise UnitEnvSylvain Henry2021-04-011-1/+3
|
* Encapsulate the EPS IORef in a newtypeSylvain Henry2021-04-014-8/+8
|
* EPA : Rename ApiAnn to EPAnnAlan Zimmerman2021-03-311-3/+3
| | | | | | Follow-up from !2418, see #19579 Updates haddock submodule
* [armv7] PIC by default + [aarch64-linux] T11276 metric increaseMoritz Angermann2021-03-291-0/+1
| | | | | Metric Increase: T11276
* Add `MessageClass`, rework `Severity` and add `DiagnosticReason`.wip/adinapoli-message-class-new-designAlfredo Di Napoli2021-03-2910-137/+118
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Other than that: * Fix T16167,json,json2,T7478,T10637 tests to reflect the introduction of the `MessageClass` type * Remove `makeIntoWarning` * Remove `warningsToMessages` * Refactor GHC.Tc.Errors 1. Refactors GHC.Tc.Errors so that we use `DiagnosticReason` for "choices" (defer types errors, holes, etc); 2. We get rid of `reportWarning` and `reportError` in favour of a general `reportDiagnostic`. * Introduce `DiagnosticReason`, `Severity` is an enum: This big commit makes `Severity` a simple enumeration, and introduces the concept of `DiagnosticReason`, which classifies the /reason/ why we are emitting a particular diagnostic. It also adds a monomorphic `DiagnosticMessage` type which is used for generic messages. * The `Severity` is computed (for now) from the reason, statically. Later improvement will add a `diagReasonSeverity` function to compute the `Severity` taking `DynFlags` into account. * Rename `logWarnings` into `logDiagnostics` * Add note and expand description of the `mkHoleError` function
* Add UnitId to Target recordFendor2021-03-283-4/+10
| | | | | | | | | | In the future, we want `HscEnv` to support multiple home units at the same time. This means, that there will be 'Target's that do not belong to the current 'HomeUnit'. This is an API change without changing behaviour. Update haddock submodule to incorporate API changes.
* Remove UniqSupply from NameCacheSylvain Henry2021-03-261-3/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | As suggested by @alexbiehl, this patch replaces the always updated UniqSupply in NameCache with a fixed Char and use it with `uniqFromMask` to generate uniques. This required some refactoring because getting a new unique from the NameCache can't be done in pure code anymore, in particular not in an atomic update function for `atomicModifyIORef`. So we use an MVar instead to store the OrigNameCache field. For some reason, T12545 increases (+1%) on i386 while it decreases on other CI runners. T9630 ghc/peak increases only with the dwarf build on CI (+16%). Metric Decrease: T12425 T12545 T9198 T12234 Metric Increase: T12545 T9630 Update haddock submodule
* Refactor FinderCacheSylvain Henry2021-03-266-28/+48
|
* Refactor NameCacheSylvain Henry2021-03-262-8/+6
| | | | | | | | * Make NameCache the mutable one and replace NameCacheUpdater with it * Remove NameCache related code duplicated into haddock Bump haddock submodule
* Implement -Wmissing-kind-signaturesOleg Grenrus2021-03-252-0/+2
| | | | Fixes #19564
* EPA : Remove ApiAnn from ParsedModuleGHC GitLab CI2021-03-252-11/+5
| | | | | All the comments are now captured in the AST, there is no need for a side-channel structure for them.
* Add compiler linting to CIHécate2021-03-251-4/+3
| | | | | This commit adds the `lint:compiler` Hadrian target to the CI runner. It does also fixes hints in the compiler/ and libraries/base/ codebases.
* Move loader state into InterpSylvain Henry2021-03-233-17/+17
| | | | | | | | | | | | | | | | | | The loader state was stored into HscEnv. As we need to have two interpreters and one loader state per interpreter in #14335, it's natural to make the loader state a field of the Interp type. As a side effect, many functions now only require a Interp parameter instead of HscEnv. Sadly we can't fully free GHC.Linker.Loader of HscEnv yet because the loader is initialised lazily from the HscEnv the first time it is used. This is left as future work. HscEnv may not contain an Interp value (i.e. hsc_interp :: Maybe Interp). So a side effect of the previous side effect is that callers of the modified functions now have to provide an Interp. It is satisfying as it pushes upstream the handling of the case where HscEnv doesn't contain an Interpreter. It is better than raising a panic (less partial functions, "parse, don't validate", etc.).
* Generate GHCi bytecode from STG instead of Core and support unboxedLuite Stegeman2021-03-203-41/+67
| | | | | | tuples and sums. fixes #1257
* GHC Exactprint main commitAlan Zimmerman2021-03-203-16/+18
| | | | | | | | Metric Increase: T10370 parsing001 Updates haddock submodule
* Enhance pass result forcingSylvain Henry2021-03-171-2/+11
| | | | | | | | When we use `withTiming` we need to force the results of each timed pass to better represent the time spent in each phase. This patch forces some results that weren't before. It also retrieve timings for the CoreToStg and WriteIface passes.
* Implement the UnliftedDatatypes extensionSebastian Graf2021-03-141-0/+5
| | | | | | | | | | | | | | | | | | GHC Proposal: 0265-unlifted-datatypes.rst Discussion: https://github.com/ghc-proposals/ghc-proposals/pull/265 Issues: https://gitlab.haskell.org/ghc/ghc/-/issues/19523 Implementation Details: Note [Implementation of UnliftedDatatypes] This patch introduces the `UnliftedDatatypes` extension. When this extension is enabled, GHC relaxes the restrictions around what result kinds are allowed in data declarations. This allows data types for which an unlifted or levity-polymorphic result kind is inferred. The most significant changes are in `GHC.Tc.TyCl`, where `Note [Implementation of UnliftedDatatypes]` describes the details of the implementation. Fixes #19523.
* Use GHC2021 as default languageJoachim Breitner2021-03-101-9/+3
|
* Introduce GHC2021 languageJoachim Breitner2021-03-102-2/+53
| | | | | | This adds support for -XGHC2021, as described in Proposal 0380 [1]. [1] https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-ghc2021.rst
* DynFlags: move temp file management into HscEnv (#17957)Sylvain Henry2021-03-088-178/+173
|
* Implement record dot syntaxwip/joachim/bump-haddockShayne Fletcher2021-03-061-0/+2
|
* Implement riscv64 LLVM backendAndreas Schwab2021-03-051-0/+6
| | | | This enables a registerised build for the riscv64 architecture.
* Add a Template Haskell warning flag -Wimplicit-liftKrzysztof Gogolewski2021-03-042-1/+3
| | | | Part of #17804.
* Use a newtype for CHeader and CStub in ForeignStubsMatthew Pickering2021-03-033-11/+12
|
* Add option to give each usage of a data constructor its own info tableMatthew Pickering2021-03-033-75/+3
| | | | | | | | | | | | | The `-fdistinct-constructor-tables` flag will generate a fresh info table for the usage of any data constructor. This is useful for debugging as now by inspecting the info table, you can determine which usage of a constructor caused that allocation rather than the old situation where the info table always mapped to the definition site of the data constructor which is useless. In conjunction with `-hi` and `-finfo-table-map` this gives a more fine grained understanding of where constructor allocations arise from in a program.
* Add -finfo-table-map which maps info tables to source positionsMatthew Pickering2021-03-036-37/+168
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This new flag embeds a lookup table from the address of an info table to information about that info table. The main interface for consulting the map is the `lookupIPE` C function > InfoProvEnt * lookupIPE(StgInfoTable *info) The `InfoProvEnt` has the following structure: > typedef struct InfoProv_{ > char * table_name; > char * closure_desc; > char * ty_desc; > char * label; > char * module; > char * srcloc; > } InfoProv; > > typedef struct InfoProvEnt_ { > StgInfoTable * info; > InfoProv prov; > struct InfoProvEnt_ *link; > } InfoProvEnt; The source positions are approximated in a similar way to the source positions for DWARF debugging information. They are only approximate but in our experience provide a good enough hint about where the problem might be. It is therefore recommended to use this flag in conjunction with `-g<n>` for more accurate locations. The lookup table is also emitted into the eventlog when it is available as it is intended to be used with the `-hi` profiling mode. Using this flag will significantly increase the size of the resulting object file but only by a factor of 2-3x in our experience.
* Add a flag to dump the FastString tableSylvain Henry2021-03-032-0/+4
|
* Reimplement Stream in "yoneda" style for efficiencyMatthew Pickering2021-02-261-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | 'Stream' is implemented in the "yoneda" style for efficiency. By representing a stream in this manner 'fmap' and '>>=' operations are accumulated in the function parameters before being applied once when the stream is destroyed. In the old implementation each usage of 'mapM' and '>>=' would traverse the entire stream in order to apply the substitution at the leaves. It is well-known for free monads that this representation can improve performance, and the test results demonstrate this for GHC as well. The operation mapAccumL is not used in the compiler and can't be implemented efficiently because it requires destroying and rebuilding the stream. I removed one use of mapAccumL_ which has similar problems but the other use was difficult to remove. In the future it may be worth exploring whether the 'Stream' encoding could be modified further to capture the mapAccumL pattern, and likewise defer the passing of accumulation parameter until the stream is finally consumed. The >>= operation for 'Stream' was a hot-spot in the ticky profile for the "ManyConstructors" test which called the 'cg' function many times in "StgToCmm.hs" Metric Decrease: ManyConstructors
* Implement -Wambiguous-fieldsAdam Gundry2021-02-262-0/+3
| | | | | | Fixes #18966. Adds a new warning -Wambiguous-fields for uses of field selectors or record updates that will be rejected in the future, when the DuplicateRecordFields extension is simplified per https://github.com/ghc-proposals/ghc-proposals/pull/366.
* Don't pass homeUnitId at ExternalPackageState creation time (#10827)Sylvain Henry2021-02-221-1/+1
| | | | | It makes the external package state independent of the home unit which is needed to make several home units share the EPS.
* Move Hooks into HscEnvSylvain Henry2021-02-227-64/+95
|
* directory: ensure xdg compliance (Fix #6077)Gauvain 'GovanifY' Roussel-Tarbouriech2021-02-181-1/+1
|
* Implement NoFieldSelectors extension (ghc-proposals 160)Adam Gundry2021-02-161-0/+16
| | | | | | | | | | | | | | | | | | | | | Fixes #5972. This adds an extension NoFieldSelectors to disable the generation of selector functions corresponding to record fields. When this extension is enabled, record field selectors are not accessible as functions, but users are still able to use them for record construction, pattern matching and updates. See Note [NoFieldSelectors] in GHC.Rename.Env for details. Defining the same field multiple times requires the DuplicateRecordFields extension to be enabled, even when NoFieldSelectors is in use. Along the way, this fixes the use of non-imported DuplicateRecordFields in GHCi with -fimplicit-import-qualified (fixes #18729). Moreover, it extends DisambiguateRecordFields to ignore non-fields when looking up fields in record updates (fixes #18999), as described by Note [DisambiguateRecordFields for updates]. Co-authored-by: Simon Hafner <hafnersimon@gmail.com> Co-authored-by: Fumiaki Kinoshita <fumiexcel@gmail.com>
* Fix over-eager inlining in SimpleOptSimon Peyton Jones2021-02-141-0/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In GHC.Core.SimpleOpt, I found that its inlining could duplicate an arbitary redex inside a lambda! Consider (\xyz. x+y). The occurrence-analysis treats the lamdda as a group, and says that both x and y occur once, even though the occur under the lambda-z. See Note [Occurrence analysis for lambda binders] in OccurAnal. When the lambda is under-applied in a call, the Simplifier is careful to zap the occ-info on x,y, because they appear under the \z. (See the call to zapLamBndrs in simplExprF1.) But SimpleOpt missed this test, resulting in #19347. So this patch * commons up the binder-zapping in GHC.Core.Utils.zapLamBndrs. * Calls this new function from GHC.Core.Opt.Simplify * Adds a call to zapLamBndrs to GHC.Core.SimpleOpt.simple_app This change makes test T12990 regress somewhat, but it was always very delicate, so I'm going to put up with that. In this voyage I also discovered a small, rather unrelated infelicity in the Simplifier: * In GHC.Core.Opt.Simplify.simplNonRecX we should apply isStrictId to the OutId not the InId. See Note [Dark corner with levity polymorphism] It may never "bite", because SimpleOpt should have inlined all the levity-polymorphic compulsory inlnings already, but somehow it bit me at one point and it's generally a more solid thing to do. Fixing the main bug increases runtime allocation in test perf/should_run/T12990, for (acceptable) reasons explained in a comement on Metric Increase: T12990
* Remove deprecated -XGenerics and -XMonoPatBindsKrzysztof Gogolewski2021-02-131-15/+0
| | | | | They have no effect since 2011 (GHC 7.2/7.4), commits cb698570b2b and 49dbe60558.
* Always set `safeInferred`, not only when it turns `False`Joachim Breitner2021-02-131-4/+3
| | | | | | | | | | | | | | | previously, `safeFlagCheck` would be happy to switch the `safeFlag` to `False`, but not the other way around. This meant that after :set -XGeneralizedNewtypeDeriving :set -XNoGeneralizedNewtypeDeriving in GHCi all loaded files would be still be infered as unsafe. This fixes #19243. This is a corner case, but somewhat relevant once ghci by default starts with `GeneralizedNewtypeDeriving` on (due to GHC2021).
* Refactor LoggerSylvain Henry2021-02-1312-453/+438
| | | | | | | | | | | | | | | | | | | | | Before this patch, the only way to override GHC's default logging behavior was to set `log_action`, `dump_action` and `trace_action` fields in DynFlags. This patch introduces a new Logger abstraction and stores it in HscEnv instead. This is part of #17957 (avoid storing state in DynFlags). DynFlags are duplicated and updated per-module (because of OPTIONS_GHC pragma), so we shouldn't store global state in them. This patch also fixes a race in parallel "--make" mode which updated the `generatedDumps` IORef concurrently. Bump haddock submodule The increase in MultilayerModules is tracked in #19293. Metric Increase: MultiLayerModules
* Reduce inlining in deeply-nested casesSimon Peyton Jones2021-02-091-0/+5
| | | | | | | | | | | | This adds a new heuristic, controllable via two new flags to better tune inlining behaviour. The new flags are -funfolding-case-threshold and -funfolding-case-scaling which are document both in the user guide and in Note [Avoid inlining into deeply nested cases]. Co-authored-by: Andreas Klebinger <klebinger.andreas@gmx.at>
* Introduce the DecoratedSDoc typeAlfredo Di Napoli2021-02-013-6/+6
| | | | | This commit introduces a DecoratedSDoc type which replaces the old ErrDoc, and hopefully better reflects the intent.
View Lorry Controller Status