March 2026 - ghc-commits

[Git][ghc/ghc][wip/T26930] ghc-internal: Float Generics to near top of module graph
by Teo Camarasu (＠teo) 20 Mar '26

20 Mar '26

Teo Camarasu pushed to branch wip/T26930 at Glasgow Haskell Compiler / GHC Commits: b1e152bd by Teo Camarasu at 2026-03-20T09:14:26+00:00 ghc-internal: Float Generics to near top of module graph We remove GHC.Internal.Generics from the critical path of the `ghc-internal` module graph. GHC.Internal.Generics used to be in the middle of the module graph, but now it is nearer the top (built later). This change thins out the module graph and allows us to get rid of the ByteOrder hs-boot file. We implement this by moving Generics instances from the module where the datatype is defined to the GHC.Internal.Generics module. This trades off increasing the compiled size of GHC.Internal.Generics with reducing the dependency footprint of datatype modules. Not all instances are moved to GHC.Internal.Generics. For instance, `GHC.Internal.Control.Monad.Fix` keeps its instance as it is one of the very last modules compiled in `ghc-internal` and so inverting the relationship here would risk adding GHC.Internal.Generics back onto the critical path. We also don't change modules that are re-exported from the `template-haskell` or `ghc-heap`. This is done to make it easy to eventually move `Generics` to `base` once something like #26657 is implemented. Resolves #26930 Metric Decrease: T21839c - - - - - 25 changed files: - libraries/ghc-internal/src/GHC/Internal/ByteOrder.hs - − libraries/ghc-internal/src/GHC/Internal/ByteOrder.hs-boot - libraries/ghc-internal/src/GHC/Internal/Conc/IO.hs - libraries/ghc-internal/src/GHC/Internal/Data/Foldable.hs - libraries/ghc-internal/src/GHC/Internal/Data/Functor/Const.hs - libraries/ghc-internal/src/GHC/Internal/Data/Functor/Identity.hs - libraries/ghc-internal/src/GHC/Internal/Data/Monoid.hs - libraries/ghc-internal/src/GHC/Internal/Data/Semigroup/Internal.hs - libraries/ghc-internal/src/GHC/Internal/Data/Traversable.hs - libraries/ghc-internal/src/GHC/Internal/Data/Version.hs - libraries/ghc-internal/src/GHC/Internal/Event/Control.hs - libraries/ghc-internal/src/GHC/Internal/Functor/ZipList.hs - libraries/ghc-internal/src/GHC/Internal/Generics.hs - libraries/ghc-internal/src/GHC/Internal/IO/Exception.hs - libraries/ghc-internal/src/GHC/Internal/RTS/Flags.hsc - libraries/ghc-internal/src/GHC/Internal/RTS/Flags/Test.hsc - libraries/ghc-internal/src/GHC/Internal/Read.hs - libraries/ghc-internal/src/GHC/Internal/Unicode/Bits.hs - testsuite/tests/ghci/scripts/ListTuplePunsPpr.stdout - testsuite/tests/ghci/scripts/T10963.stderr - testsuite/tests/ghci/scripts/ghci064.stdout - testsuite/tests/interface-stability/base-exports.stdout - testsuite/tests/interface-stability/base-exports.stdout-javascript-unknown-ghcjs - testsuite/tests/interface-stability/base-exports.stdout-mingw32 - testsuite/tests/interface-stability/base-exports.stdout-ws-32 The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/b1e152bd34fedf647bfe132816bf33a… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/b1e152bd34fedf647bfe132816bf33a… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/marge_bot_batch_merge_job] 17 commits: ghci: Mention active language edition in startup banner
by Marge Bot (＠marge-bot) 20 Mar '26

20 Mar '26

Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC Commits: 07214bc7 by Adam Gundry at 2026-03-20T00:17:35-04:00 ghci: Mention active language edition in startup banner Per GHC proposal 632, this makes the GHCi startup banner include the active language edition, plus an indication of whether this was the default (as opposed to being explicitly selected via an option such as `-XGHC2024`). For example: ``` $ ghci GHCi, version 9.14.1: https://www.haskell.org/ghc/ :? for help Using default language edition: GHC2024 ghci> ``` Fixes #26037. - - - - - 55ca8434 by sheaf at 2026-03-20T00:17:51-04:00 Improve incomplete record selector warnings This commit stops GHC from emitting spurious incomplete record selector warnings for bare selectors/projections such as .fld There are two places we currently emit incomplete record selector warnings: 1. In the desugarer, when we see a record selector or an occurrence of 'getField'. Here, we can use pattern matching information to ensure we don't give false positives. 2. In the typechecker, which might sometimes give false positives but can emit warnings in cases that the pattern match checker would otherwise miss. This is explained in Note [Detecting incomplete record selectors] in GHC.HsToCore.Pmc. Now, we obviously don't want to emit the same error twice, and generally we prefer (1), as those messages contain fewer false positives. So we suppress (2) when we are sure we are going to emit (1); the logic for doing so is in GHC.Tc.Instance.Class.warnIncompleteRecSel, and works by looking at the CtOrigin. Now, the issue was that this logic handled explicit record selectors as well as overloaded record field selectors such as "x.r" (which turns into a simple GetFieldOrigin CtOrigin), but it didn't properly handle record projectors like ".fld" or ".fld1.fld2" (which result in other CtOrigins such as 'RecordFieldProjectionOrigin'). To solve this problem, we re-use the 'isHasFieldOrigin' introduced in fbdc623a (slightly adjusted). On the way, we also had to update the desugarer with special handling for the 'ExpandedThingTc' case in 'ds_app', to make sure that 'ds_app_var' sees all the type arguments to 'getField' in order for it to indeed emit warnings like in (1). Fixes #26686 - - - - - fabf1298 by Cheng Shao at 2026-03-20T00:17:53-04:00 rts: opportunistically grow the MutableByteArray# in-place in resizeMutableByteArray# Following !15234, this patch improves `resizeMutableByteArray#` memory efficiency by growing the `MutableByteArray#` in-place if possible, addressing an old todo comment here. Also adds a new test case `resizeMutableByteArrayInPlace` that stresses this behavior. - - - - - c66a0f2a by Matthew Craven at 2026-03-20T00:17:59-04:00 Change representation of floating point literals This commit changes the representation of floating point literals throughough the compiler, in particular in Core and Cmm. The Rational type is deficient for this purpose, dealing poorly with NaN, +/-Infinity, and negative zero. Instead, the new module GHC.Types.Literal.Floating uses the host Float/Double type to represent NaNs, infinities and negative zero. It also contains a Rational constructor, for the benefit of -fexcess-precision. Other changes: - Remove Note [negative zero] and related code This also removes the restrictions on constant-folding of division by zero, and should make any problems with NaN/Infinity more obvious. - Use -0.0 as the additive identity for Core constant folding rules for floating-point addition, fixing #21227. - Manual worker-wrapper for GHC.Float.rationalToDouble. This is intended to prevent the compiler's WW on this function from interfering with constant-folding. This change means that we now avoid allocating a box for the result of a 'realToFrac' call in T10359. - Combine floatDecodeOp and doubleDecodeOp. This change also fixes a bug in doubleDecodeOp wherein it would incorrectly produce an Int# instead of an Int64# literal for the mantissa component with 64-bit targets. - Use Float/Double for assembly immediates, and update the X86 and PowerPC backends to properly handle special values such as NaN and infinity. - Allow 'rational_to' to handle zero denominators, fixing a TODO in GHC.Core.Opt.ConstantFold. Fixes #8364 #9811 #18897 #21227 Progress towards #26919 Metric Decrease: T10359 Co-authored-by: sheaf <sam.derbyshire(a)gmail.com> ------------------------- Metric Decrease: T1969 T5321FD ------------------------- - - - - - 00f06dcd by Zubin Duggal at 2026-03-20T00:18:00-04:00 compiler/ffi: Collapse void pointer chains in capi wrappers New gcc/clang treat -Wincompatible-pointer-types as an error by default. Since C only allows implicit conversion from void*, not void**, capi wrappers for functions taking e.g. abstract** would fail to compile when the Haskell type Ptr (Ptr Abstract) was naively translated to void**. Collapse nested void pointers to a single void* when the pointee type has no known C representation. Fixes #26852 - - - - - b1fd36c8 by Luite Stegeman at 2026-03-20T00:18:13-04:00 Move some functions related to pointer tagging to a separate module - - - - - 2229d09a by Luite Stegeman at 2026-03-20T00:18:13-04:00 Branchless unpacking for enumeration types Change unpacking for enumeration types to go to Word8#/Word16#/Word# directly instead of going through an intermediate unboxed sum. This allows us to do a branchless conversion using DataToTag and TagToEnum. Fixes #26970 - - - - - bc0e4bf9 by Luite Stegeman at 2026-03-20T00:18:18-04:00 bytecode: Carefully SLIDE off the end of a stack chunk The SLIDE bytecode instruction was not checking for stack chunk boundaries and could corrupt the stack underflow frame, leading to crashes. We add a check to use safe writes if we cross the chunk boundary and also handle stack underflow if Sp is advanced past the underflow frame. fix #27001 - - - - - b7f1da12 by Cheng Shao at 2026-03-20T00:18:19-04:00 ghci: serialize BCOByteArray buffer directly when possible This patch changes the `Binary` instances of `BCOByteArray` to directly serialize the underlying buffer when possible, while also taking into account the issue of host-dependent `Word` width. See added comments and amended `Note [BCOByteArray serialization]` for detailed explanation. Closes #27020. - - - - - 268c8294 by Sylvain Henry at 2026-03-20T00:18:40-04:00 JS: replace BigInt with Number arithmetic for 32/64-bit quot/rem (#23597) Replace BigInt-based implementations of quotWord32, remWord32, quotRemWord32, quotRem2Word32, quotWord64, remWord64, quotInt64, and remInt64 with pure Number (double/integer) arithmetic to avoid the overhead of BigInt promotion. - - - - - 29086cac by Sylvain Henry at 2026-03-20T00:18:47-04:00 Core: add constant-folding rules for Addr# eq/ne (#18032) - - - - - 205890dc by Matthew Pickering at 2026-03-20T00:18:48-04:00 Use OsPath rather than FilePath in Downsweep cache This gets us one step closure to uniformly using `OsPath` in the compiler. - - - - - 055fc850 by Cheng Shao at 2026-03-20T00:18:50-04:00 hadrian: fix ghc-in-ghci flavour stage0 shared libraries This patch fixes missing stage0 shared libraries in hadrian ghc-in-ghci flavour, which was accidentally dropped in 669d09f950a6e88b903d9fd8a7571531774d4d5d and resulted in a regression in HLS support on linux/macos. Fixes #27057. - - - - - bd69e307 by Sylvain Henry at 2026-03-20T00:18:56-04:00 JS: install rts/Types.h header file (#27033) It was an omission, making HsFFI.h not usable with GHC using the JS backend. - - - - - ffdaffa6 by Cheng Shao at 2026-03-20T00:18:56-04:00 hadrian: don't compile RTS with -Winline This patch removes `-Winline` from cflags when compiling the RTS, given that: 1. It generates a huge pile of spam and hurts developer experience 2. Whether inlining happens is highly dependent on toolchains, flavours, etc, and it's not really an issue to fix if inlining doesn't happen; it's a hint to the C compiler anyway. Fixes #27060. - - - - - 7ba44a11 by Cheng Shao at 2026-03-20T00:18:56-04:00 hadrian: compile libffi-clib with -Wno-deprecated-declarations This patch adds `-Wno-deprecated-declarations` to cflags of `libffi-clib`, given that it produces noise at compile-time that aren't really our issue to fix anyway, it's from vendored libffi source code. - - - - - 741a58bc by Rodrigo Mesquita at 2026-03-20T00:18:58-04:00 Expose decodeStackWithIpe from ghc-experimental This decoding is useful to the debugger and it wasn't originally exported as an oversight. - - - - - 101 changed files: - compiler/GHC/Builtin/Names.hs - compiler/GHC/ByteCode/Asm.hs - compiler/GHC/Cmm/CommonBlockElim.hs - compiler/GHC/Cmm/Expr.hs - compiler/GHC/Cmm/Opt.hs - compiler/GHC/Cmm/Parser.y - compiler/GHC/Cmm/Type.hs - compiler/GHC/Cmm/Utils.hs - compiler/GHC/CmmToAsm/AArch64/CodeGen.hs - compiler/GHC/CmmToAsm/AArch64/Ppr.hs - compiler/GHC/CmmToAsm/AArch64/Regs.hs - compiler/GHC/CmmToAsm/LA64/CodeGen.hs - compiler/GHC/CmmToAsm/LA64/Ppr.hs - compiler/GHC/CmmToAsm/LA64/Regs.hs - compiler/GHC/CmmToAsm/PPC/CodeGen.hs - compiler/GHC/CmmToAsm/PPC/Ppr.hs - compiler/GHC/CmmToAsm/PPC/Regs.hs - compiler/GHC/CmmToAsm/RV64/CodeGen.hs - compiler/GHC/CmmToAsm/RV64/Ppr.hs - compiler/GHC/CmmToAsm/RV64/Regs.hs - compiler/GHC/CmmToAsm/Wasm/FromCmm.hs - compiler/GHC/CmmToAsm/X86/CodeGen.hs - compiler/GHC/CmmToAsm/X86/Ppr.hs - compiler/GHC/CmmToAsm/X86/Regs.hs - compiler/GHC/CmmToC.hs - compiler/GHC/CmmToLlvm/CodeGen.hs - compiler/GHC/CmmToLlvm/Data.hs - compiler/GHC/Core.hs - compiler/GHC/Core/Make.hs - compiler/GHC/Core/Opt/ConstantFold.hs - compiler/GHC/Driver/Downsweep.hs - compiler/GHC/HsToCore/Expr.hs - compiler/GHC/HsToCore/Foreign/C.hs - compiler/GHC/HsToCore/Match/Literal.hs - compiler/GHC/HsToCore/Pmc.hs - compiler/GHC/HsToCore/Pmc/Solver/Types.hs - + compiler/GHC/Platform/Tag.hs - compiler/GHC/Stg/Unarise.hs - compiler/GHC/StgToByteCode.hs - compiler/GHC/StgToCmm/Closure.hs - compiler/GHC/StgToCmm/Expr.hs - compiler/GHC/StgToCmm/Lit.hs - compiler/GHC/StgToCmm/Prim.hs - compiler/GHC/StgToJS/Literal.hs - compiler/GHC/Tc/Errors.hs - compiler/GHC/Tc/Instance/Class.hs - compiler/GHC/Tc/TyCl/Build.hs - compiler/GHC/Tc/Types/Origin.hs - compiler/GHC/Types/Id/Make.hs - compiler/GHC/Types/Literal.hs - + compiler/GHC/Types/Literal/Floating.hs - compiler/GHC/Types/RepType.hs - compiler/GHC/Utils/Binary.hs - compiler/ghc.cabal.in - docs/users_guide/9.16.1-notes.rst - ghc/GHCi/UI.hs - ghc/Main.hs - hadrian/src/Settings/Flavours/GhcInGhci.hs - hadrian/src/Settings/Packages.hs - libraries/ghc-experimental/ghc-experimental.cabal.in - + libraries/ghc-experimental/src/GHC/Stack/Decode/Experimental.hs - libraries/ghc-internal/src/GHC/Internal/Float.hs - libraries/ghci/GHCi/ResolvedBCO.hs - rts/Interpreter.c - rts/PrimOps.cmm - rts/js/arith.js - rts/rts.cabal - + testsuite/tests/bytecode/T27001.hs - + testsuite/tests/bytecode/T27001.stdout - testsuite/tests/bytecode/all.T - + testsuite/tests/codeGen/should_run/T21227.hs - + testsuite/tests/codeGen/should_run/T21227.stdout - + testsuite/tests/codeGen/should_run/T9811.hs - + testsuite/tests/codeGen/should_run/T9811.stdout - testsuite/tests/codeGen/should_run/all.T - testsuite/tests/count-deps/CountDepsAst.stdout - testsuite/tests/count-deps/CountDepsParser.stdout - + testsuite/tests/ffi/should_compile/T26852.h - + testsuite/tests/ffi/should_compile/T26852.hs - + testsuite/tests/ffi/should_compile/T26852.stderr - testsuite/tests/ffi/should_compile/all.T - testsuite/tests/ghci/should_run/BinaryArray.hs - testsuite/tests/interface-stability/ghc-experimental-exports.stdout - testsuite/tests/interface-stability/ghc-experimental-exports.stdout-mingw32 - + testsuite/tests/javascript/js-c-sources/T27033.hs - + testsuite/tests/javascript/js-c-sources/T27033.stdout - + testsuite/tests/javascript/js-c-sources/T27033_c.c - + testsuite/tests/javascript/js-c-sources/T27033_js.js - testsuite/tests/javascript/js-c-sources/all.T - testsuite/tests/numeric/should_run/T7014.hs - + testsuite/tests/overloadedrecflds/should_compile/T26686.hs - + testsuite/tests/overloadedrecflds/should_compile/T26686.stderr - testsuite/tests/overloadedrecflds/should_compile/all.T - testsuite/tests/rts/all.T - + testsuite/tests/rts/resizeMutableByteArrayInPlace.hs - + testsuite/tests/simplCore/should_compile/T18032.hs - + testsuite/tests/simplCore/should_compile/T18032.stderr - testsuite/tests/simplCore/should_compile/all.T - testsuite/tests/simplStg/should_run/all.T - + testsuite/tests/simplStg/should_run/unpack_enum.hs - + testsuite/tests/simplStg/should_run/unpack_enum.stdout The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/2f21fbdb87e7879a3504849af2b120… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/2f21fbdb87e7879a3504849af2b120… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/spj-reinstallable-base] More.. [skip ci]
by Simon Peyton Jones (＠simonpj) 20 Mar '26

20 Mar '26

Simon Peyton Jones pushed to branch wip/spj-reinstallable-base at Glasgow Haskell Compiler / GHC Commits: 4058ad0b by Simon Peyton Jones at 2026-03-20T00:19:21+00:00 More.. [skip ci] - - - - - 10 changed files: - compiler/GHC/Builtin/Names.hs - compiler/GHC/Builtin/Types.hs - compiler/GHC/Builtin/Utils.hs - compiler/GHC/Iface/Binary.hs - compiler/GHC/Iface/Load.hs - compiler/GHC/Rename/Names.hs - compiler/GHC/Tc/Errors/Hole.hs - compiler/GHC/Types/Name.hs - compiler/GHC/Types/Name/Cache.hs - libraries/base/src/GHC/Weak/Finalize.hs Changes: ===================================== compiler/GHC/Builtin/Names.hs ===================================== @@ -86,7 +86,7 @@ This is accomplished through a combination of mechanisms: Note [Infinite families of known-key names] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Infinite families of known-key things (e.g. tuples and sums) pose a tricky -problem: we can't add them to the knownKeyNames finite map which we use to +problem: we can't add them to the wiredInNames finite map which we use to ensure that, e.g., a reference to (,) gets assigned the right unique (if this doesn't sound familiar see Note [Known-key names] above). ===================================== compiler/GHC/Builtin/Types.hs ===================================== @@ -284,7 +284,7 @@ names in GHC.Builtin.Names, so they use wTcQual, wDataQual, etc -} --- This list is used only to define GHC.Builtin.Utils.knownKeyNames. That in turn +-- This list is used only to define GHC.Builtin.Utils.wiredInNames. That in turn -- is used to initialise the name environment carried around by the renamer. -- This means that if we look up the name of a TyCon (or its implicit binders) -- that occurs in this list that name will be assigned the wired-in key we ===================================== compiler/GHC/Builtin/Utils.hs ===================================== @@ -23,14 +23,9 @@ module GHC.Builtin.Utils ( oldLookupKnownKeyName, oldLookupKnownNameInfo, - -- ** Internal use - -- | 'knownKeyNames' is exported to seed the original name cache only; - -- if you find yourself wanting to look at it you might consider using - -- 'lookupKnownKeyName' or 'isKnownKeyName'. - knownKeyNames, -- * Miscellaneous - wiredInIds, ghcPrimIds, + wiredInNames, wiredInIds, ghcPrimIds, ghcPrimExports, ghcPrimDeclDocs, @@ -118,9 +113,9 @@ Note [About wired-in things] -- | This list is used to ensure that when you say "Prelude.map" in your source -- code, or in an interface file, you get a Name with the correct known key (See -- Note [Known-key names] in "GHC.Builtin.Names") -knownKeyNames :: [Name] +wiredInNames :: [Name] -- ToDo: rename to wiredInNames -knownKeyNames +wiredInNames | debugIsOn , Just badNamesDoc <- knownKeyNamesOkay all_names = pprPanic "badAllKnownKeyNames" badNamesDoc @@ -214,7 +209,7 @@ oldIsKnownKeyName n = -- in the domain are 'Unique's associated with 'Name's (as opposed -- to some other namespace of 'Unique's). oldKnownKeysMap :: UniqFM Name Name -oldKnownKeysMap = listToIdentityUFM knownKeyNames +oldKnownKeysMap = listToIdentityUFM wiredInNames -- | Given a 'Unique' lookup any associated arbitrary SDoc's to be displayed by -- GHCi's ':info' command. ===================================== compiler/GHC/Iface/Binary.hs ===================================== @@ -714,7 +714,7 @@ serialiseName bh name _ -- all known-key uniques fit in this space. This is asserted by -- GHC.Builtin.Utils.knownKeyNamesOkay. -- --- During serialization we check for known-key things using isKnownKeyName. +-- During serialization we check for known-key things using oldIsKnownKeyName. -- During deserialization we use lookupKnownKeyName to get from the unique back -- to its corresponding Name. ===================================== compiler/GHC/Iface/Load.hs ===================================== @@ -64,7 +64,7 @@ import GHC.Iface.Binary import GHC.Iface.Rename import GHC.Iface.Env import GHC.Iface.Errors as Iface_Errors -import GHC.Iface.Errors.Ppr( defaultIfaceMessageOpts ) +import GHC.Iface.Errors.Ppr( defaultIfaceMessageOpts, missingInterfaceErrorDiagnostic ) import GHC.Tc.Errors.Types import GHC.Tc.Utils.Monad @@ -190,7 +190,7 @@ loadKnownKeyOccMap ; mb_res <- liftIO $ findImportedModule hsc_env kNOWN_KEY_NAMES NoPkgQual ; iface <- case mb_res of Found _ mod -> loadInterfaceWithException doc mod ImportBySystem - fr -> do { hsc_env <- getHscEnv + fr -> do { hsc_env <- getTopEnv ; pprPanic "loadKnownKeyOccMap" $ missingInterfaceErrorDiagnostic defaultIfaceMessageOpts $ cannotFindModule hsc_env kNOWN_KEY_NAMES fr } ===================================== compiler/GHC/Rename/Names.hs ===================================== @@ -1951,8 +1951,9 @@ warnUnusedImportDecls gbl_env hsc_src -- We should only warn for unnecessary *user* imports, but deciding -- minimal imports should take generated imports into account - ; let usageUserImports = findImportUsage (excludeGenerated imports) uses - usageAllImports = findImportUsage imports uses + ; rb <- goptM Opt_RebindableKnownKeyNames + ; let usageUserImports = findImportUsage rb (excludeGenerated imports) uses + usageAllImports = findImportUsage rb imports uses ; traceRn "warnUnusedImportDecls" $ (vcat [ text "Uses:" <+> ppr uses @@ -1968,11 +1969,12 @@ warnUnusedImportDecls gbl_env hsc_src excludeGenerated :: [LImportDecl GhcRn] -> [LImportDecl GhcRn] excludeGenerated = filterOut (ideclGenerated . ideclExt . unLoc) -findImportUsage :: [LImportDecl GhcRn] +findImportUsage :: Bool -- True <=> Opt_RebindableKnownKeyNames is on + -> [LImportDecl GhcRn] -> [GlobalRdrElt] -> [ImportDeclUsage] -findImportUsage imports used_gres +findImportUsage rebindable_known_key_names imports used_gres = map unused_decl imports where import_usage :: ImportMap @@ -2021,6 +2023,13 @@ findImportUsage imports used_gres in UnusedNames acc_ns acc_wcs acc_fs' | used = acc + + -- -frebindable-known-key-names is on, and `n` is a known-key name + -- Then don't warn about an unused imports. + | rebindable_known_key_names + , isKnownKeyName n + = acc + | otherwise = UnusedNames (acc_ns `extendNameSet` n) acc_wcs acc_fs where ===================================== compiler/GHC/Tc/Errors/Hole.hs ===================================== @@ -681,7 +681,7 @@ findValidHoleFits tidy_env implics simples h@(Hole { hole_sort = ExprHole _ -- BuiltInSyntax names like (:) and [] builtIns :: EnumSet LangExt.Extension -> [Name] - builtIns exts = filter isBuiltInSyntax (knownKeyNames ++ infFamNames) + builtIns exts = filter isBuiltInSyntax (wiredInNames ++ infFamNames) where -- Tuples and sums of are not included in knownKeyName as there are infinitely many of them. -- See Note [Infinite families of known-key names] in GHC.Builtin.Names. ===================================== compiler/GHC/Types/Name.hs ===================================== @@ -69,7 +69,7 @@ module GHC.Types.Name ( isTyVarName, isTyConName, isDataConName, isValName, isVarName, isDynLinkName, isFieldName, isWiredInName, isWiredIn, isBuiltInSyntax, isTupleTyConName, - isKnownKeyName', isKnownKey, + isKnownKeyName, isKnownKey, isSumTyConName, isUnboxedTupleDataConLikeName, isHoleName, @@ -135,7 +135,7 @@ There are several reasons that GHC might want to refer to a known-key Name: * When desugaring, the desugarer wants to refer to a particular class, type, or function. It does this via (e.g.) - dsLookupTyKnownKeyTyCon :: Unique -> DsM TyCon + dsLookupKnownKeyTyCon :: Unique -> DsM TyCon * In a very similar way, for type-class defauting GHC has built-in defaulting behaviour for Num, IsString, etc. It gets hold of these classes via their known key, via @@ -184,7 +184,9 @@ To implement all this, here are the moving parts: -frebindable-known-key-names -fdefines-known-key-names -* When processing source code, /without/ -frebindable +* Suppose the desugarer calls `dsLookupKnownKeyTyCon` on `rationalTyConKey`. + * In client source code, that is, /without/ -frebindable-known-key-names, + we try to import GHC.KnownKeyNames. Assuming this is succesful, Note [About the NameSorts] ~~~~~~~~~~~~~~~~~~~~~~~~~~ @@ -394,20 +396,19 @@ isInternalName :: Name -> Bool isExternalName :: Name -> Bool isSystemName :: Name -> Bool isWiredInName :: Name -> Bool -isKnownKeyName' :: Name -> Bool +isKnownKeyName :: Name -> Bool isWiredInName (Name {n_sort = WiredIn _ _ _}) = True isWiredInName _ = False --- TODO: disambiguate with builtin utils -isKnownKeyName' (Name {n_sort = KnownKey _}) = True -isKnownKeyName' _ = False +isKnownKeyName (Name {n_sort = KnownKey _}) = True +isKnownKeyName _ = False isWiredIn :: NamedThing thing => thing -> Bool isWiredIn = isWiredInName . getName isKnownKey :: NamedThing thing => thing -> Bool -isKnownKey = isKnownKeyName' . getName +isKnownKey = isKnownKeyName . getName wiredInNameTyThing_maybe :: Name -> Maybe TyThing wiredInNameTyThing_maybe (Name {n_sort = WiredIn _ thing _}) = Just thing ===================================== compiler/GHC/Types/Name/Cache.hs ===================================== @@ -91,7 +91,7 @@ So we deal with them in lookupOrigNameCache by means of isInfiniteFamilyOrigName At the same time, simple finite built-in names (`[]`, `:`, `->`) can be put in the OrigNameCache without any issues (they end up there because they're -knownKeyNames). It doesn't matter that they're built-in syntax. +wiredInNames). It doesn't matter that they're built-in syntax. One might wonder: what's the point of having any built-in syntax in the OrigNameCache at all? Good question; after all, @@ -200,7 +200,7 @@ updateNameCache name_cache !_mod !_occ upd_fn {-# NOINLINE knownKeysOrigNameCache #-} knownKeysOrigNameCache :: OrigNameCache -knownKeysOrigNameCache = initOrigNames knownKeyNames +knownKeysOrigNameCache = initOrigNames wiredInNames isKnownOrigName_maybe :: Module -> OccName -> Maybe Name isKnownOrigName_maybe = lookupOrigNameCache knownKeysOrigNameCache ===================================== libraries/base/src/GHC/Weak/Finalize.hs ===================================== @@ -21,6 +21,7 @@ import GHC.Internal.IO.Handle.Types (Handle) import GHC.Internal.IO.Handle.Text (hPutStrLn) import GHC.Internal.Prim (Array#, State#, RealWorld) import GHC.Internal.Types (Int, IO) +import GHC.Internal.Num( Num ) -- A necessary known-key name {-# DEPRECATED runFinalizerBatch View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/4058ad0b6765f350dfe57bacef369d4… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/4058ad0b6765f350dfe57bacef369d4… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T26930] ghc-internal: Float Generics to near top of module graph
by Teo Camarasu (＠teo) 20 Mar '26

20 Mar '26

Teo Camarasu pushed to branch wip/T26930 at Glasgow Haskell Compiler / GHC Commits: 9cfe99bc by Teo Camarasu at 2026-03-20T00:08:28+00:00 ghc-internal: Float Generics to near top of module graph We remove GHC.Internal.Generics from the critical path of the `ghc-internal` module graph. GHC.Internal.Generics used to be in the middle of the module graph, but now it is nearer the top (built later). This change thins out the module graph and allows us to get rid of the ByteOrder hs-boot file. We implement this by moving Generics instances from the module where the datatype is defined to the GHC.Internal.Generics module. This trades off increasing the compiled size of GHC.Internal.Generics with reducing the dependency footprint of datatype modules. Not all instances are moved to GHC.Internal.Generics. For instance, `GHC.Internal.Control.Monad.Fix` keeps its instance as it is one of the very last modules compiled in `ghc-internal` and so inverting the relationship here would risk adding GHC.Internal.Generics back onto the critical path. We also don't change modules that are re-exported from the `template-haskell` or `ghc-heap`. This is done to make it easy to eventually move `Generics` to `base` once something like #26657 is implemented. Resolves #26930 Metric Decrease: T21839c - - - - - 25 changed files: - libraries/ghc-internal/src/GHC/Internal/ByteOrder.hs - − libraries/ghc-internal/src/GHC/Internal/ByteOrder.hs-boot - libraries/ghc-internal/src/GHC/Internal/Conc/IO.hs - libraries/ghc-internal/src/GHC/Internal/Data/Foldable.hs - libraries/ghc-internal/src/GHC/Internal/Data/Functor/Const.hs - libraries/ghc-internal/src/GHC/Internal/Data/Functor/Identity.hs - libraries/ghc-internal/src/GHC/Internal/Data/Monoid.hs - libraries/ghc-internal/src/GHC/Internal/Data/Semigroup/Internal.hs - libraries/ghc-internal/src/GHC/Internal/Data/Traversable.hs - libraries/ghc-internal/src/GHC/Internal/Data/Version.hs - libraries/ghc-internal/src/GHC/Internal/Event/Control.hs - libraries/ghc-internal/src/GHC/Internal/Functor/ZipList.hs - libraries/ghc-internal/src/GHC/Internal/Generics.hs - libraries/ghc-internal/src/GHC/Internal/IO/Exception.hs - libraries/ghc-internal/src/GHC/Internal/RTS/Flags.hsc - libraries/ghc-internal/src/GHC/Internal/RTS/Flags/Test.hsc - libraries/ghc-internal/src/GHC/Internal/Read.hs - libraries/ghc-internal/src/GHC/Internal/Unicode/Bits.hs - testsuite/tests/ghci/scripts/ListTuplePunsPpr.stdout - testsuite/tests/ghci/scripts/T10963.stderr - testsuite/tests/ghci/scripts/ghci064.stdout - testsuite/tests/interface-stability/base-exports.stdout - testsuite/tests/interface-stability/base-exports.stdout-javascript-unknown-ghcjs - testsuite/tests/interface-stability/base-exports.stdout-mingw32 - testsuite/tests/interface-stability/base-exports.stdout-ws-32 The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/9cfe99bc5acef54792480ffee69e4c6… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/9cfe99bc5acef54792480ffee69e4c6… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/abstract-q] try without runInIo
by Teo Camarasu (＠teo) 19 Mar '26

19 Mar '26

Teo Camarasu pushed to branch wip/abstract-q at Glasgow Haskell Compiler / GHC Commits: 6047c446 by Teo Camarasu at 2026-03-19T23:27:24+00:00 try without runInIo - - - - - 3 changed files: - compiler/GHC/Tc/Gen/Splice.hs - libraries/ghc-internal/src/GHC/Internal/TH/Monad.hs - libraries/ghci/GHCi/TH.hs Changes: ===================================== compiler/GHC/Tc/Gen/Splice.hs ===================================== @@ -1140,7 +1140,7 @@ convertAnnotationWrapper fhv = do -} runQuasi :: TH.Q a -> TcM a -runQuasi (TH.Q act) = unliftIOEnv $ \runInIO -> liftIO $ act runInIO metaHandlersTcM +runQuasi (TH.Q act) = unliftIOEnv $ \runInIO -> liftIO $ act (metaHandlersTcM runInIO) runRemoteModFinalizers :: ThModFinalizers -> TcM () runRemoteModFinalizers (ThModFinalizers finRefs) = do @@ -1557,71 +1557,71 @@ location = do { m <- getModule , TH.loc_start = (srcSpanStartLine r, srcSpanStartCol r) , TH.loc_end = (srcSpanEndLine r, srcSpanEndCol r) }) } -metaHandlersTcM :: TH.MetaHandlers TcM -metaHandlersTcM = TH.MetaHandlers { - mFail = fail - , mNewName = \s -> do { u <- newUnique +metaHandlersTcM :: (forall x. TcM x -> IO x) -> TH.MetaHandlers IO +metaHandlersTcM runInIO = TH.MetaHandlers { + mFail = \s -> runInIO $ fail s + , mNewName = \s -> runInIO $ do { u <- newUnique ; let i = toInteger (getKey u) ; return (TH.mkNameU s i) } -- 'msg' is forced to ensure exceptions don't escape, -- see Note [Exceptions in TH] - , mReport = report - - , mLocation = location - - , mLookupName = lookupName - , mReify = reify - , mReifyFixity = \nm -> lookupThName nm >>= reifyFixity - , mReifyType = reifyTypeOfThing - , mReifyInstances = reifyInstances - , mReifyRoles = reifyRoles - , mReifyAnnotations = reifyAnnotations - , mReifyModule = reifyModule - , mReifyConStrictness = \nm -> do { nm' <- lookupThName nm + , mReport = fmap runInIO . report + + , mLocation = runInIO location + + , mLookupName = fmap runInIO . lookupName + , mReify = runInIO . reify + , mReifyFixity = \nm -> runInIO $ lookupThName nm >>= reifyFixity + , mReifyType = runInIO . reifyTypeOfThing + , mReifyInstances = fmap runInIO . reifyInstances + , mReifyRoles = runInIO . reifyRoles + , mReifyAnnotations = runInIO . reifyAnnotations + , mReifyModule = runInIO . reifyModule + , mReifyConStrictness = \nm -> runInIO $ do { nm' <- lookupThName nm ; dc <- tcLookupDataCon nm' ; let bangs = dataConImplBangs dc ; return (map reifyDecidedStrictness bangs) } - -- For qRecover, discard error messages if - -- the recovery action is chosen. Otherwise - -- we'll only fail higher up. - -- NB: extremely subtle!!! TODO: write up note - -- tryTcDiscardingErrs manipulates the reader env so we need to be careful we don't sneak in the outside env - , mRecover = \recover main -> tryTcDiscardingErrs (runQuasi recover) (runQuasi main) + -- -- For qRecover, discard error messages if + -- -- the recovery action is chosen. Otherwise + -- -- we'll only fail higher up. + -- -- NB: extremely subtle!!! TODO: write up note + -- -- tryTcDiscardingErrs manipulates the reader env so we need to be careful we don't sneak in the outside env + , mRecover = \recover main -> runInIO $ tryTcDiscardingErrs (runQuasi recover) (runQuasi main) - , mGetPackageRoot = do + , mGetPackageRoot = runInIO $ do dflags <- getDynFlags return $ fromMaybe "." (workingDirectory dflags) - , mAddDependentFile = \fp -> do + , mAddDependentFile = \fp -> runInIO $ do ref <- fmap tcg_dependent_files getGblEnv dep_files <- readTcRef ref writeTcRef ref (fp:dep_files) - , mAddDependentDirectory = \dp -> do + , mAddDependentDirectory = \dp -> runInIO $ do ref <- fmap tcg_dependent_dirs getGblEnv dep_dirs <- readTcRef ref writeTcRef ref (dp:dep_dirs) - , mAddTempFile = \suffix -> do + , mAddTempFile = \suffix -> runInIO $ do dflags <- getDynFlags logger <- getLogger tmpfs <- hsc_tmpfs <$> getTopEnv liftIO $ newTempName logger tmpfs (tmpDir dflags) TFL_GhcSession suffix - , mAddTopDecls = addTopDecls + , mAddTopDecls = runInIO . addTopDecls - , mAddForeignFilePath = \lang fp -> do + , mAddForeignFilePath = \lang fp -> runInIO $ do var <- fmap tcg_th_foreign_files getGblEnv updTcRef var ((lang, fp) :) - , mAddModFinalizer = \fin -> do + , mAddModFinalizer = \fin -> runInIO $ do r <- liftIO $ mkRemoteRef fin fref <- liftIO $ mkForeignRef r (freeRemoteRef r) addModFinalizerRef fref - , mAddCorePlugin = \plugin -> do + , mAddCorePlugin = \plugin -> runInIO $ do hsc_env <- getTopEnv let fc = hsc_FC hsc_env let home_unit = hsc_home_unit hsc_env @@ -1636,20 +1636,20 @@ metaHandlersTcM = TH.MetaHandlers { th_coreplugins_var <- tcg_th_coreplugins <$> getGblEnv updTcRef th_coreplugins_var (plugin:) - , mGetQ = getQ + , mGetQ = runInIO getQ - , mPutQ = \x -> do + , mPutQ = \x -> runInIO $ do th_state_var <- fmap tcg_th_state getGblEnv updTcRef th_state_var (\m -> Map.insert (typeOf x) (toDyn x) m) - , mIsExtEnabled = xoptM + , mIsExtEnabled = runInIO . xoptM - , mExtsEnabled = + , mExtsEnabled = runInIO $ EnumSet.toList . extensionFlags . hsc_dflags <$> getTopEnv - , mPutDoc = putDoc + , mPutDoc = fmap runInIO . putDoc - , mGetDoc = getDoc + , mGetDoc = runInIO . getDoc } -- | Looks up documentation for a declaration in first the current module, ===================================== libraries/ghc-internal/src/GHC/Internal/TH/Monad.hs ===================================== @@ -150,7 +150,7 @@ class (MonadIO m, MonadFail m) => Quasi m where -- type environment, so reification isn't going to -- work. instance Quasi IO where - qRunQ (Q m) = m id metaHandlersIO + qRunQ (Q m) = m metaHandlersIO qNewName = newNameIO qReport True msg = hPutStrLn stderr ("Template Haskell error: " ++ msg) @@ -332,7 +332,7 @@ counter = unsafePerformIO (newIORef 0) -- inversion](https://en.wikipedia.org/wiki/Dependency_inversion_principle), -- providing an abstract interface for the user which is later concretely -- fufilled by an concrete 'Quasi' instance, internal to GHC. -newtype Q a = Q { unQ :: forall m. (forall x. m x -> IO x) -> MetaHandlers m -> IO a } +newtype Q a = Q { unQ :: MetaHandlers IO -> IO a } -- | \"Runs\" the 'Q' monad. Normal users of Template Haskell -- should not need this function, as the splice brackets @$( ... )@ @@ -349,19 +349,19 @@ runQ :: Quasi m => Q a -> m a runQ = qRunQ instance Monad Q where - Q m >>= k = Q $ \r h -> (m r h >>= \x -> unQ (k x) r h) + Q m >>= k = Q $ \h -> (m h >>= \x -> unQ (k x) h) (>>) = (*>) instance MonadFail Q where - fail s = report True s >> Q (\r h -> r $ mFail h "Q monad failure") + fail s = report True s >> Q (\h -> mFail h "Q monad failure") instance Functor Q where - fmap f (Q x) = Q $ \r h -> fmap f (x r h) + fmap f (Q x) = Q $ \h -> fmap f (x h) instance Applicative Q where - pure x = Q $ \_ _ -> pure x - Q f <*> Q x = Q $ \r h -> (f r h <*> x r h) - Q m *> Q n = Q $ \r h -> (m r h *> n r h) + pure x = Q $ \_ -> pure x + Q f <*> Q x = Q $ \h -> (f h <*> x h) + Q m *> Q n = Q $ \h -> (m h *> n h) -- | @since 2.17.0.0 instance Semigroup a => Semigroup (Q a) where @@ -431,13 +431,13 @@ class Monad m => Quote m where newName :: String -> m Name runHandler :: (forall m. MetaHandlers m -> m a) -> Q a -runHandler op = Q $ \r h -> r (op h) +runHandler op = Q $ \h -> (op h) runHandler1 :: (forall m. MetaHandlers m -> a -> m b) -> a -> Q b -runHandler1 op = \x -> Q $ \r h -> r (op h x) +runHandler1 op = \x -> Q $ \h -> (op h x) runHandler2 :: (forall m. MetaHandlers m -> a -> b -> m c) -> a -> b -> Q c -runHandler2 op = \x y -> Q $ \r h -> r (op h x y) +runHandler2 op = \x y -> Q $ \h -> (op h x y) instance Quote Q where newName = runHandler1 mNewName @@ -653,7 +653,7 @@ reportWarning = report False recover :: Q a -- ^ handler to invoke on failure -> Q a -- ^ computation to run -> Q a -recover rec main = Q $ \r h -> r $ mRecover h rec main +recover rec main = Q $ \h -> mRecover h rec main -- We don't export lookupName; the Bool isn't a great API -- Instead we export lookupTypeName, lookupValueName @@ -920,7 +920,7 @@ location = runHandler mLocation -- necessarily flushed when the compiler finishes running, so you should -- flush them yourself. runIO :: IO a -> Q a -runIO m = Q $ \_ _ -> m +runIO m = Q $ \_ -> m -- | Get the package root for the current package which is being compiled. -- This can be set explicitly with the -package-root flag but is normally ===================================== libraries/ghci/GHCi/TH.hs ===================================== @@ -1,5 +1,5 @@ {-# LANGUAGE ScopedTypeVariables, StandaloneDeriving, DeriveGeneric, - TupleSections, RecordWildCards, InstanceSigs, CPP #-} + TupleSections, RecordWildCards, InstanceSigs, CPP, RankNTypes #-} {-# OPTIONS_GHC -fno-warn-name-shadowing #-} -- | @@ -174,18 +174,19 @@ reifyAnnotations lookup = where typerep = typeOf (undefined :: a) runQinGHCiQ :: TH.Q a -> GHCiQ a -runQinGHCiQ (TH.Q m) = GHCiQ $ \sRef -> m (runInIO sRef) metaHandlersGHCiQ +runQinGHCiQ (TH.Q m) = GHCiQ $ \sRef -> m (metaHandlersGHCiQ (runInIO sRef)) where runInIO :: IORef QState -> GHCiQ a -> IO a runInIO sRef (GHCiQ m) = m sRef -metaHandlersGHCiQ = TH.MetaHandlers { - mFail = fail - , mNewName = \str -> ghcCmd (NewName str) - , mReport = \isError msg -> ghcCmd (Report isError msg) +metaHandlersGHCiQ :: (forall x. GHCiQ x -> IO x) -> TH.MetaHandlers IO +metaHandlersGHCiQ runInIO = TH.MetaHandlers { + mFail = runInIO . fail + , mNewName = \str -> runInIO $ ghcCmd (NewName str) + , mReport = \isError msg -> runInIO $ ghcCmd (Report isError msg) -- See Note [TH recover with -fexternal-interpreter] in GHC.Tc.Gen.Splice - , mRecover = \h a -> GHCiQ $ \sRef -> mask $ \unmask -> do + , mRecover = \h a -> runInIO $ GHCiQ $ \sRef -> mask $ \unmask -> do s <- readIORef sRef remoteTHCall (qsPipe s) StartRecover e <- try $ unmask $ runGHCiQ (runQinGHCiQ a <* ghcCmd FailIfErrs) sRef @@ -195,37 +196,37 @@ metaHandlersGHCiQ = TH.MetaHandlers { -- in case of error, restore the state to the start of the `recover` block. newIORef s >>= runGHCiQ (runQinGHCiQ h) Right r -> return r - , mLookupName = \isType occ -> ghcCmd (LookupName isType occ) - , mReify = \name -> ghcCmd (Reify name) - , mReifyFixity = \name -> ghcCmd (ReifyFixity name) - , mReifyType = \name -> ghcCmd (ReifyType name) - , mReifyInstances = \name tys -> ghcCmd (ReifyInstances name tys) - , mReifyRoles = \name -> ghcCmd (ReifyRoles name) - - , mReifyAnnotations = reifyAnnotations - , mReifyModule = \m -> ghcCmd (ReifyModule m) - , mReifyConStrictness = \name -> ghcCmd (ReifyConStrictness name) - , mLocation = fromMaybe noLoc . qsLocation <$> getState - , mGetPackageRoot = ghcCmd GetPackageRoot - , mAddDependentFile = \file -> ghcCmd (AddDependentFile file) - , mAddDependentDirectory = \dir -> ghcCmd (AddDependentDirectory dir) - , mAddTempFile = \suffix -> ghcCmd (AddTempFile suffix) - , mAddTopDecls = \decls -> ghcCmd (AddTopDecls decls) - , mAddForeignFilePath = \lang fp -> ghcCmd (AddForeignFilePath lang fp) - , mAddModFinalizer = \fin -> GHCiQ (\s -> mkRemoteRef fin) >>= + , mLookupName = \isType occ -> runInIO $ ghcCmd (LookupName isType occ) + , mReify = \name ->runInIO $ ghcCmd (Reify name) + , mReifyFixity = \name ->runInIO $ ghcCmd (ReifyFixity name) + , mReifyType = \name -> runInIO $ ghcCmd (ReifyType name) + , mReifyInstances = \name tys -> runInIO $ghcCmd (ReifyInstances name tys) + , mReifyRoles = \name -> runInIO $ ghcCmd (ReifyRoles name) + + , mReifyAnnotations = runInIO . reifyAnnotations + , mReifyModule = \m -> runInIO $ ghcCmd (ReifyModule m) + , mReifyConStrictness = \name -> runInIO $ ghcCmd (ReifyConStrictness name) + , mLocation = runInIO $ fromMaybe noLoc . qsLocation <$> getState + , mGetPackageRoot = runInIO $ ghcCmd GetPackageRoot + , mAddDependentFile = \file -> runInIO $ ghcCmd (AddDependentFile file) + , mAddDependentDirectory = \dir -> runInIO $ ghcCmd (AddDependentDirectory dir) + , mAddTempFile = \suffix -> runInIO $ ghcCmd (AddTempFile suffix) + , mAddTopDecls = \decls -> runInIO $ ghcCmd (AddTopDecls decls) + , mAddForeignFilePath = \lang fp -> runInIO $ ghcCmd (AddForeignFilePath lang fp) + , mAddModFinalizer = \fin -> runInIO $ GHCiQ (\s -> mkRemoteRef fin) >>= ghcCmd . AddModFinalizer - , mAddCorePlugin = \str -> ghcCmd (AddCorePlugin str) - , mGetQ = do + , mAddCorePlugin = \str -> runInIO $ ghcCmd (AddCorePlugin str) + , mGetQ = runInIO $ do s <- getState let lookup :: forall a. Typeable a => Map TypeRep Dynamic -> Maybe a lookup m = fromDynamic =<< M.lookup (typeOf (undefined::a)) m return $ lookup (qsMap s) - , mPutQ = \k -> GHCiQ $ \sRef -> + , mPutQ = \k -> runInIO $ GHCiQ $ \sRef -> modifyIORef' sRef (\s -> s { qsMap = M.insert (typeOf k) (toDyn k) (qsMap s) }) - , mIsExtEnabled = \x -> ghcCmd (IsExtEnabled x) - , mExtsEnabled = ghcCmd ExtsEnabled - , mPutDoc = \l s -> ghcCmd (PutDoc l s) - , mGetDoc = \l -> ghcCmd (GetDoc l) + , mIsExtEnabled = \x -> runInIO $ ghcCmd (IsExtEnabled x) + , mExtsEnabled = runInIO $ ghcCmd ExtsEnabled + , mPutDoc = \l s -> runInIO $ ghcCmd (PutDoc l s) + , mGetDoc = \l -> runInIO $ ghcCmd (GetDoc l) } -- | The implementation of the 'StartTH' message: create View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/6047c446732ba2243b69c39d4ade28b… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/6047c446732ba2243b69c39d4ade28b… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/marge_bot_batch_merge_job] 17 commits: ghci: Mention active language edition in startup banner
by Marge Bot (＠marge-bot) 19 Mar '26

19 Mar '26

Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC Commits: 2441cb61 by Adam Gundry at 2026-03-19T18:15:22-04:00 ghci: Mention active language edition in startup banner Per GHC proposal 632, this makes the GHCi startup banner include the active language edition, plus an indication of whether this was the default (as opposed to being explicitly selected via an option such as `-XGHC2024`). For example: ``` $ ghci GHCi, version 9.14.1: https://www.haskell.org/ghc/ :? for help Using default language edition: GHC2024 ghci> ``` Fixes #26037. - - - - - 1ef695d5 by sheaf at 2026-03-19T18:15:29-04:00 Improve incomplete record selector warnings This commit stops GHC from emitting spurious incomplete record selector warnings for bare selectors/projections such as .fld There are two places we currently emit incomplete record selector warnings: 1. In the desugarer, when we see a record selector or an occurrence of 'getField'. Here, we can use pattern matching information to ensure we don't give false positives. 2. In the typechecker, which might sometimes give false positives but can emit warnings in cases that the pattern match checker would otherwise miss. This is explained in Note [Detecting incomplete record selectors] in GHC.HsToCore.Pmc. Now, we obviously don't want to emit the same error twice, and generally we prefer (1), as those messages contain fewer false positives. So we suppress (2) when we are sure we are going to emit (1); the logic for doing so is in GHC.Tc.Instance.Class.warnIncompleteRecSel, and works by looking at the CtOrigin. Now, the issue was that this logic handled explicit record selectors as well as overloaded record field selectors such as "x.r" (which turns into a simple GetFieldOrigin CtOrigin), but it didn't properly handle record projectors like ".fld" or ".fld1.fld2" (which result in other CtOrigins such as 'RecordFieldProjectionOrigin'). To solve this problem, we re-use the 'isHasFieldOrigin' introduced in fbdc623a (slightly adjusted). On the way, we also had to update the desugarer with special handling for the 'ExpandedThingTc' case in 'ds_app', to make sure that 'ds_app_var' sees all the type arguments to 'getField' in order for it to indeed emit warnings like in (1). Fixes #26686 - - - - - cbe28216 by Cheng Shao at 2026-03-19T18:15:31-04:00 rts: opportunistically grow the MutableByteArray# in-place in resizeMutableByteArray# Following !15234, this patch improves `resizeMutableByteArray#` memory efficiency by growing the `MutableByteArray#` in-place if possible, addressing an old todo comment here. Also adds a new test case `resizeMutableByteArrayInPlace` that stresses this behavior. - - - - - fc1f1569 by Matthew Craven at 2026-03-19T18:15:36-04:00 Change representation of floating point literals This commit changes the representation of floating point literals throughough the compiler, in particular in Core and Cmm. The Rational type is deficient for this purpose, dealing poorly with NaN, +/-Infinity, and negative zero. Instead, the new module GHC.Types.Literal.Floating uses the host Float/Double type to represent NaNs, infinities and negative zero. It also contains a Rational constructor, for the benefit of -fexcess-precision. Other changes: - Remove Note [negative zero] and related code This also removes the restrictions on constant-folding of division by zero, and should make any problems with NaN/Infinity more obvious. - Use -0.0 as the additive identity for Core constant folding rules for floating-point addition, fixing #21227. - Manual worker-wrapper for GHC.Float.rationalToDouble. This is intended to prevent the compiler's WW on this function from interfering with constant-folding. This change means that we now avoid allocating a box for the result of a 'realToFrac' call in T10359. - Combine floatDecodeOp and doubleDecodeOp. This change also fixes a bug in doubleDecodeOp wherein it would incorrectly produce an Int# instead of an Int64# literal for the mantissa component with 64-bit targets. - Use Float/Double for assembly immediates, and update the X86 and PowerPC backends to properly handle special values such as NaN and infinity. - Allow 'rational_to' to handle zero denominators, fixing a TODO in GHC.Core.Opt.ConstantFold. Fixes #8364 #9811 #18897 #21227 Progress towards #26919 Metric Decrease: T10359 Co-authored-by: sheaf <sam.derbyshire(a)gmail.com> ------------------------- Metric Decrease: T1969 T5321FD ------------------------- - - - - - 3b1f5a18 by Zubin Duggal at 2026-03-19T18:15:38-04:00 compiler/ffi: Collapse void pointer chains in capi wrappers New gcc/clang treat -Wincompatible-pointer-types as an error by default. Since C only allows implicit conversion from void*, not void**, capi wrappers for functions taking e.g. abstract** would fail to compile when the Haskell type Ptr (Ptr Abstract) was naively translated to void**. Collapse nested void pointers to a single void* when the pointee type has no known C representation. Fixes #26852 - - - - - e74a3e68 by Luite Stegeman at 2026-03-19T18:15:43-04:00 Move some functions related to pointer tagging to a separate module - - - - - cfda1289 by Luite Stegeman at 2026-03-19T18:15:43-04:00 Branchless unpacking for enumeration types Change unpacking for enumeration types to go to Word8#/Word16#/Word# directly instead of going through an intermediate unboxed sum. This allows us to do a branchless conversion using DataToTag and TagToEnum. Fixes #26970 - - - - - a9b86f37 by Luite Stegeman at 2026-03-19T18:15:48-04:00 bytecode: Carefully SLIDE off the end of a stack chunk The SLIDE bytecode instruction was not checking for stack chunk boundaries and could corrupt the stack underflow frame, leading to crashes. We add a check to use safe writes if we cross the chunk boundary and also handle stack underflow if Sp is advanced past the underflow frame. fix #27001 - - - - - ccda4658 by Cheng Shao at 2026-03-19T18:15:50-04:00 ghci: serialize BCOByteArray buffer directly when possible This patch changes the `Binary` instances of `BCOByteArray` to directly serialize the underlying buffer when possible, while also taking into account the issue of host-dependent `Word` width. See added comments and amended `Note [BCOByteArray serialization]` for detailed explanation. Closes #27020. - - - - - c3d8df5f by Sylvain Henry at 2026-03-19T18:15:57-04:00 JS: replace BigInt with Number arithmetic for 32/64-bit quot/rem (#23597) Replace BigInt-based implementations of quotWord32, remWord32, quotRemWord32, quotRem2Word32, quotWord64, remWord64, quotInt64, and remInt64 with pure Number (double/integer) arithmetic to avoid the overhead of BigInt promotion. - - - - - 3800c431 by Sylvain Henry at 2026-03-19T18:16:02-04:00 Core: add constant-folding rules for Addr# eq/ne (#18032) - - - - - ddba914b by Matthew Pickering at 2026-03-19T18:16:04-04:00 Use OsPath rather than FilePath in Downsweep cache This gets us one step closure to uniformly using `OsPath` in the compiler. - - - - - 9dbf72a3 by Cheng Shao at 2026-03-19T18:16:05-04:00 hadrian: fix ghc-in-ghci flavour stage0 shared libraries This patch fixes missing stage0 shared libraries in hadrian ghc-in-ghci flavour, which was accidentally dropped in 669d09f950a6e88b903d9fd8a7571531774d4d5d and resulted in a regression in HLS support on linux/macos. Fixes #27057. - - - - - 1bc10aab by Sylvain Henry at 2026-03-19T18:16:10-04:00 JS: install rts/Types.h header file (#27033) It was an omission, making HsFFI.h not usable with GHC using the JS backend. - - - - - d76b4dc5 by Cheng Shao at 2026-03-19T18:16:11-04:00 hadrian: don't compile RTS with -Winline This patch removes `-Winline` from cflags when compiling the RTS, given that: 1. It generates a huge pile of spam and hurts developer experience 2. Whether inlining happens is highly dependent on toolchains, flavours, etc, and it's not really an issue to fix if inlining doesn't happen; it's a hint to the C compiler anyway. Fixes #27060. - - - - - fc78df8a by Cheng Shao at 2026-03-19T18:16:11-04:00 hadrian: compile libffi-clib with -Wno-deprecated-declarations This patch adds `-Wno-deprecated-declarations` to cflags of `libffi-clib`, given that it produces noise at compile-time that aren't really our issue to fix anyway, it's from vendored libffi source code. - - - - - 2f21fbdb by Rodrigo Mesquita at 2026-03-19T18:16:13-04:00 Expose decodeStackWithIpe from ghc-experimental This decoding is useful to the debugger and it wasn't originally exported as an oversight. - - - - - 101 changed files: - compiler/GHC/Builtin/Names.hs - compiler/GHC/ByteCode/Asm.hs - compiler/GHC/Cmm/CommonBlockElim.hs - compiler/GHC/Cmm/Expr.hs - compiler/GHC/Cmm/Opt.hs - compiler/GHC/Cmm/Parser.y - compiler/GHC/Cmm/Type.hs - compiler/GHC/Cmm/Utils.hs - compiler/GHC/CmmToAsm/AArch64/CodeGen.hs - compiler/GHC/CmmToAsm/AArch64/Ppr.hs - compiler/GHC/CmmToAsm/AArch64/Regs.hs - compiler/GHC/CmmToAsm/LA64/CodeGen.hs - compiler/GHC/CmmToAsm/LA64/Ppr.hs - compiler/GHC/CmmToAsm/LA64/Regs.hs - compiler/GHC/CmmToAsm/PPC/CodeGen.hs - compiler/GHC/CmmToAsm/PPC/Ppr.hs - compiler/GHC/CmmToAsm/PPC/Regs.hs - compiler/GHC/CmmToAsm/RV64/CodeGen.hs - compiler/GHC/CmmToAsm/RV64/Ppr.hs - compiler/GHC/CmmToAsm/RV64/Regs.hs - compiler/GHC/CmmToAsm/Wasm/FromCmm.hs - compiler/GHC/CmmToAsm/X86/CodeGen.hs - compiler/GHC/CmmToAsm/X86/Ppr.hs - compiler/GHC/CmmToAsm/X86/Regs.hs - compiler/GHC/CmmToC.hs - compiler/GHC/CmmToLlvm/CodeGen.hs - compiler/GHC/CmmToLlvm/Data.hs - compiler/GHC/Core.hs - compiler/GHC/Core/Make.hs - compiler/GHC/Core/Opt/ConstantFold.hs - compiler/GHC/Driver/Downsweep.hs - compiler/GHC/HsToCore/Expr.hs - compiler/GHC/HsToCore/Foreign/C.hs - compiler/GHC/HsToCore/Match/Literal.hs - compiler/GHC/HsToCore/Pmc.hs - compiler/GHC/HsToCore/Pmc/Solver/Types.hs - + compiler/GHC/Platform/Tag.hs - compiler/GHC/Stg/Unarise.hs - compiler/GHC/StgToByteCode.hs - compiler/GHC/StgToCmm/Closure.hs - compiler/GHC/StgToCmm/Expr.hs - compiler/GHC/StgToCmm/Lit.hs - compiler/GHC/StgToCmm/Prim.hs - compiler/GHC/StgToJS/Literal.hs - compiler/GHC/Tc/Errors.hs - compiler/GHC/Tc/Instance/Class.hs - compiler/GHC/Tc/TyCl/Build.hs - compiler/GHC/Tc/Types/Origin.hs - compiler/GHC/Types/Id/Make.hs - compiler/GHC/Types/Literal.hs - + compiler/GHC/Types/Literal/Floating.hs - compiler/GHC/Types/RepType.hs - compiler/GHC/Utils/Binary.hs - compiler/ghc.cabal.in - docs/users_guide/9.16.1-notes.rst - ghc/GHCi/UI.hs - ghc/Main.hs - hadrian/src/Settings/Flavours/GhcInGhci.hs - hadrian/src/Settings/Packages.hs - libraries/ghc-experimental/ghc-experimental.cabal.in - + libraries/ghc-experimental/src/GHC/Stack/Decode/Experimental.hs - libraries/ghc-internal/src/GHC/Internal/Float.hs - libraries/ghci/GHCi/ResolvedBCO.hs - rts/Interpreter.c - rts/PrimOps.cmm - rts/js/arith.js - rts/rts.cabal - + testsuite/tests/bytecode/T27001.hs - + testsuite/tests/bytecode/T27001.stdout - testsuite/tests/bytecode/all.T - + testsuite/tests/codeGen/should_run/T21227.hs - + testsuite/tests/codeGen/should_run/T21227.stdout - + testsuite/tests/codeGen/should_run/T9811.hs - + testsuite/tests/codeGen/should_run/T9811.stdout - testsuite/tests/codeGen/should_run/all.T - testsuite/tests/count-deps/CountDepsAst.stdout - testsuite/tests/count-deps/CountDepsParser.stdout - + testsuite/tests/ffi/should_compile/T26852.h - + testsuite/tests/ffi/should_compile/T26852.hs - + testsuite/tests/ffi/should_compile/T26852.stderr - testsuite/tests/ffi/should_compile/all.T - testsuite/tests/ghci/should_run/BinaryArray.hs - testsuite/tests/interface-stability/ghc-experimental-exports.stdout - testsuite/tests/interface-stability/ghc-experimental-exports.stdout-mingw32 - + testsuite/tests/javascript/js-c-sources/T27033.hs - + testsuite/tests/javascript/js-c-sources/T27033.stdout - + testsuite/tests/javascript/js-c-sources/T27033_c.c - + testsuite/tests/javascript/js-c-sources/T27033_js.js - testsuite/tests/javascript/js-c-sources/all.T - testsuite/tests/numeric/should_run/T7014.hs - + testsuite/tests/overloadedrecflds/should_compile/T26686.hs - + testsuite/tests/overloadedrecflds/should_compile/T26686.stderr - testsuite/tests/overloadedrecflds/should_compile/all.T - testsuite/tests/rts/all.T - + testsuite/tests/rts/resizeMutableByteArrayInPlace.hs - + testsuite/tests/simplCore/should_compile/T18032.hs - + testsuite/tests/simplCore/should_compile/T18032.stderr - testsuite/tests/simplCore/should_compile/all.T - testsuite/tests/simplStg/should_run/all.T - + testsuite/tests/simplStg/should_run/unpack_enum.hs - + testsuite/tests/simplStg/should_run/unpack_enum.stdout The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/e5cf0804542bd1c063ed295a5248d5… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/e5cf0804542bd1c063ed295a5248d5… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/dcoutts/idle-gc
by Duncan Coutts (＠dcoutts) 19 Mar '26

19 Mar '26

Duncan Coutts pushed new branch wip/dcoutts/idle-gc at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/dcoutts/idle-gc You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/spj-apporv-Oct24] remove srcCodeOrigin references, add some WIP flag
by Apoorv Ingle (＠ani) 19 Mar '26

19 Mar '26

Apoorv Ingle pushed to branch wip/spj-apporv-Oct24 at Glasgow Haskell Compiler / GHC Commits: 7d2a15c1 by Apoorv Ingle at 2026-03-19T16:09:36-05:00 remove srcCodeOrigin references, add some WIP flag - - - - - 7 changed files: - compiler/GHC/Tc/Gen/App.hs - compiler/GHC/Tc/Gen/Do.hs - compiler/GHC/Tc/Gen/Expr.hs - compiler/GHC/Tc/Gen/Head.hs - compiler/GHC/Tc/Types/LclEnv.hs - compiler/GHC/Tc/Types/Origin.hs - compiler/GHC/Tc/Utils/Monad.hs Changes: ===================================== compiler/GHC/Tc/Gen/App.hs ===================================== @@ -1943,7 +1943,7 @@ quickLookArg1 pos app_lspan (fun, fun_lspan) larg@(L _ arg) sc_arg_ty@(Scaled _ do { ((rn_fun_arg, fun_lspan_arg), rn_args) <- splitHsApps arg -- Step 1: get the type of the head of the argument - ; (fun_ue, mb_fun_ty) <- (tcCollectingUsage $ tcInferAppHead_maybe rn_fun_arg) + ; (fun_ue, mb_fun_ty) <- tcCollectingUsage $ tcInferAppHead_maybe rn_fun_arg -- tcCollectingUsage: the use of an Id at the head generates usage-info -- See the call to `tcEmitBindingUsage` in `check_local_id`. So we must -- capture and save it in the `EValArgQL`. See (QLA6) in @@ -2020,9 +2020,9 @@ mk_origin fun_lspan rn_fun = return $ exprCtOrigin rn_fun | otherwise -- if the location is generated, -- the best we can do is to approximate by looking on top of the error message stack - = do { code_orig <- getSrcCodeOrigin + = do { code_orig <- getHsCtxt ; traceTc "mk_origin" (pprHsCtxt code_orig) - ; return $ srcCodeOriginCtOrigin code_orig + ; return $ hsCtxtCtOrigin code_orig } ===================================== compiler/GHC/Tc/Gen/Do.hs ===================================== @@ -448,7 +448,7 @@ It stores the original statement (with location) and the expanded expression of the error context stack which contains the error message for the previous statement: eg. "In the stmt of a do block: e1". This popping is implicitly done when we push the error context message for the next statment. - See Note [ErrCtxtStack Manipulation] and `LclEnv.setLclCtxtSrcCodeOrigin` + See Note [ErrCtxtStack Manipulation] and `LclEnv.setLclCtxtHsCtxt` Sans the popping business for error context stack, if there were to be a type error in `e2`, we would get a spurious and confusing error message ===================================== compiler/GHC/Tc/Gen/Expr.hs ===================================== @@ -677,6 +677,7 @@ tcExpr expr@(RecordCon { rcon_con = L loc qcon@(WithUserRdr _ con_name) -- in the renamer. See Note [Overview of record dot syntax] in -- GHC.Hs.Expr. This is why we match on 'rupd_flds = Left rbnds' here -- and panic otherwise. +-- WIP: To be fixed soon expandRecordUpd needs to return HsExpansion and not a separate ds_res_ty tcExpr expr@(RecordUpd { rupd_expr = record_expr , rupd_flds = RegularRecUpdFields @@ -1862,14 +1863,3 @@ checkMissingFields con_like rbinds arg_tys field_strs = conLikeImplBangs con_like fl `elemField` flds = any (\ fl' -> flSelector fl == fl') flds - - --- Expands the expression on the fly --- See Note [Handling overloaded and rebindable constructs] --- See Note [Typechecking by expansion: overview] -tcExpandExpr :: HsExpr GhcRn -> TcM (HsExpr GhcRn) -tcExpandExpr orig_expr@(HsDo _ flav (L _ stmts)) - = do { expanded_expr <- expandDoStmts flav stmts - ; return (mkExpandedLExpr orig_expr expanded_expr) } - -tcExpandExpr e = return e ===================================== compiler/GHC/Tc/Gen/Head.hs ===================================== @@ -467,7 +467,7 @@ tcInferAppHead_maybe fun = case fun of -- visible type applications in the argument. -- c.f. T19167 (\ (e, ds_flag, ty) -> (mkExpandedTc o (L loc e), ds_flag, ty)) <$> - tcExprSigma False (errCtxtCtOrigin o) e + tcExprSigma False (hsCtxtCtOrigin o) e ) _ -> return Nothing ===================================== compiler/GHC/Tc/Types/LclEnv.hs ===================================== @@ -21,9 +21,9 @@ module GHC.Tc.Types.LclEnv ( , setLclEnvTypeEnv , modifyLclEnvTcLevel - , getLclEnvSrcCodeOrigin - , setLclEnvSrcCodeOrigin - , setLclCtxtSrcCodeOrigin + , getLclEnvHsCtxt + , setLclEnvHsCtxt + , setLclCtxtHsCtxt , lclEnvInGeneratedCode , addLclEnvErrCtxt @@ -110,7 +110,7 @@ This data structure keeps track of two things: In the current design, if the top of the ErrCtxtStack is an ExpansionCodeCtxt i.e. we are currently typechecking a compiler generated expression, and we encounter an XExpr, then we _replace_ the top of the stack with the new XExpr. Otherwise, we -push the new expression error message on top of the stack. cf. `LclEnv.setLclCtxtSrcCodeOrigin` +push the new expression error message on top of the stack. cf. `LclEnv.setLclCtxtHsCtxt` -} @@ -119,9 +119,9 @@ push the new expression error message on top of the stack. cf. `LclEnv.setLclCtx type ErrCtxtStack = [ErrCtxt] -- | Get the top of the error message stack -get_src_code_origin :: ErrCtxtStack -> HsCtxt -get_src_code_origin (e : _) = e -get_src_code_origin _ = error "get_src_code_origin: oops! Empty error message stack" +get_err_ctxt_stack_head :: ErrCtxtStack -> HsCtxt +get_err_ctxt_stack_head (e : _) = e +get_err_ctxt_stack_head _ = error "get_err_ctxt_stack_head: oops! Empty error message stack" data TcLclCtxt = TcLclCtxt { @@ -199,17 +199,17 @@ setLclEnvErrCtxt ctxt = modifyLclCtxt (\env -> env { tcl_err_ctxt = ctxt }) -- See Note [ErrCtxtStack Manipulation] addLclEnvErrCtxt :: ErrCtxt -> TcLclEnv -> TcLclEnv -addLclEnvErrCtxt ec = setLclEnvSrcCodeOrigin ec +addLclEnvErrCtxt ec = setLclEnvHsCtxt ec -getLclEnvSrcCodeOrigin :: TcLclEnv -> HsCtxt -getLclEnvSrcCodeOrigin = get_src_code_origin . tcl_err_ctxt . tcl_lcl_ctxt +getLclEnvHsCtxt :: TcLclEnv -> HsCtxt +getLclEnvHsCtxt = get_err_ctxt_stack_head . tcl_err_ctxt . tcl_lcl_ctxt -setLclEnvSrcCodeOrigin :: ErrCtxt -> TcLclEnv -> TcLclEnv -setLclEnvSrcCodeOrigin ec = modifyLclCtxt (setLclCtxtSrcCodeOrigin ec) +setLclEnvHsCtxt :: ErrCtxt -> TcLclEnv -> TcLclEnv +setLclEnvHsCtxt ec = modifyLclCtxt (setLclCtxtHsCtxt ec) -- See Note [ErrCtxtStack Manipulation] -setLclCtxtSrcCodeOrigin :: ErrCtxt -> TcLclCtxt -> TcLclCtxt -setLclCtxtSrcCodeOrigin ec lclCtxt +setLclCtxtHsCtxt :: ErrCtxt -> TcLclCtxt -> TcLclCtxt +setLclCtxtHsCtxt ec lclCtxt -- never stack 2 statement error contexts on top of each other | StmtErrCtxt{} : ecs <- tcl_err_ctxt lclCtxt , StmtErrCtxt{} <- ec ===================================== compiler/GHC/Tc/Types/Origin.hs ===================================== @@ -9,7 +9,7 @@ module GHC.Tc.Types.Origin ( -- * CtOrigin CtOrigin(..), exprCtOrigin, lexprCtOrigin, matchesCtOrigin, grhssCtOrigin, - srcCodeOriginCtOrigin, errCtxtCtOrigin, + hsCtxtCtOrigin, invisibleOrigin_maybe, isVisibleOrigin, toInvisibleOrigin, pprCtOrigin, pprCtOriginBriefly, isGivenOrigin, defaultReprEqOrigins, isWantedSuperclassOrigin, @@ -625,20 +625,16 @@ exprCtOrigin (HsIf {}) = IfThenElseOrigin exprCtOrigin (HsProjection _ p) = RecordFieldProjectionOrigin (FieldLabelStrings $ fmap noLocA p) exprCtOrigin (RecordUpd{}) = RecordUpdOrigin exprCtOrigin (HsGetField _ _ f) = GetFieldOrigin (fmap field_label $ dfoLabel (unLoc f)) -exprCtOrigin (XExpr (ExpandedThingRn (HSE o _))) = errCtxtCtOrigin o +exprCtOrigin (XExpr (ExpandedThingRn (HSE o _))) = hsCtxtCtOrigin o exprCtOrigin (XExpr (HsRecSelRn f)) = OccurrenceOfRecSel $ L (getLoc $ foLabel f) (foExt f) -srcCodeOriginCtOrigin :: HsCtxt -> CtOrigin -srcCodeOriginCtOrigin = errCtxtCtOrigin - -errCtxtCtOrigin :: HsCtxt -> CtOrigin -errCtxtCtOrigin (ExprCtxt e) = exprCtOrigin e -errCtxtCtOrigin (FunAppCtxt (FunAppCtxtExpr _ e) _) = exprCtOrigin e -errCtxtCtOrigin (StmtErrCtxt{}) = DoStmtOrigin -errCtxtCtOrigin (StmtErrCtxtPat p) = DoPatOrigin p -errCtxtCtOrigin (RecordUpdCtxt{}) = RecordUpdOrigin -errCtxtCtOrigin _ = Shouldn'tHappenOrigin "errCtxtCtOrigin" - +hsCtxtCtOrigin :: HsCtxt -> CtOrigin +hsCtxtCtOrigin (ExprCtxt e) = exprCtOrigin e +hsCtxtCtOrigin (FunAppCtxt (FunAppCtxtExpr _ e) _) = exprCtOrigin e +hsCtxtCtOrigin (StmtErrCtxt{}) = DoStmtOrigin +hsCtxtCtOrigin (StmtErrCtxtPat p) = DoPatOrigin p +hsCtxtCtOrigin (RecordUpdCtxt{}) = RecordUpdOrigin +hsCtxtCtOrigin _ = Shouldn'tHappenOrigin "hsCtxtCtOrigin" -- | Extract a suitable CtOrigin from a MatchGroup matchesCtOrigin :: MatchGroup GhcRn (LHsExpr GhcRn) -> CtOrigin ===================================== compiler/GHC/Tc/Utils/Monad.hs ===================================== @@ -61,7 +61,7 @@ module GHC.Tc.Utils.Monad( addDependentFiles, addDependentDirectories, -- * Error management - getSrcCodeOrigin, + getHsCtxt, getSrcSpanM, getRealSrcSpanM, setSrcSpan, setSrcSpanA, addLocM, inGeneratedCode, wrapLocM, wrapLocFstM, wrapLocFstMA, wrapLocSndM, wrapLocSndMA, wrapLocM_, @@ -1094,8 +1094,8 @@ setSrcSpan (GeneratedSrcSpan{}) thing_inside setSrcSpan _ thing_inside = thing_inside -getSrcCodeOrigin :: TcRn HsCtxt -getSrcCodeOrigin = getLclEnvSrcCodeOrigin <$> getLclEnv +getHsCtxt :: TcRn HsCtxt +getHsCtxt = getLclEnvHsCtxt <$> getLclEnv setSrcSpanA :: EpAnn ann -> TcRn a -> TcRn a setSrcSpanA l = setSrcSpan (locA l) View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/7d2a15c1226cfa12c0df3b011d91d73… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/7d2a15c1226cfa12c0df3b011d91d73… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T26930] ghc-internal: Float Generics to near top of module graph
by Teo Camarasu (＠teo) 19 Mar '26

19 Mar '26

Teo Camarasu pushed to branch wip/T26930 at Glasgow Haskell Compiler / GHC Commits: b169da6c by Teo Camarasu at 2026-03-19T20:26:05+00:00 ghc-internal: Float Generics to near top of module graph We remove GHC.Internal.Generics from the critical path of the `ghc-internal` module graph. GHC.Internal.Generics used to be in the middle of the module graph, but now it is nearer the top (built later). This change thins out the module graph and allows us to get rid of the ByteOrder hs-boot file. We implement this by moving Generics instances from the module where the datatype is defined to the GHC.Internal.Generics module. This trades off increasing the compiled size of GHC.Internal.Generics with reducing the dependency footprint of datatype modules. Not all instances are moved to GHC.Internal.Generics. For instance, `GHC.Internal.Control.Monad.Fix` keeps its instance as it is one of the very last modules compiled in `ghc-internal` and so inverting the relationship here would risk adding GHC.Internal.Generics back onto the critical path. We also don't change modules that are re-exported from the `template-haskell` or `ghc-heap`. This is done to make it easy to eventually move `Generics` to `base` once something like #26657 is implemented. Resolves #26930 Metric Decrease: T21839c - - - - - 25 changed files: - libraries/ghc-internal/src/GHC/Internal/ByteOrder.hs - − libraries/ghc-internal/src/GHC/Internal/ByteOrder.hs-boot - libraries/ghc-internal/src/GHC/Internal/Conc/IO.hs - libraries/ghc-internal/src/GHC/Internal/Data/Foldable.hs - libraries/ghc-internal/src/GHC/Internal/Data/Functor/Const.hs - libraries/ghc-internal/src/GHC/Internal/Data/Functor/Identity.hs - libraries/ghc-internal/src/GHC/Internal/Data/Monoid.hs - libraries/ghc-internal/src/GHC/Internal/Data/Semigroup/Internal.hs - libraries/ghc-internal/src/GHC/Internal/Data/Traversable.hs - libraries/ghc-internal/src/GHC/Internal/Data/Version.hs - libraries/ghc-internal/src/GHC/Internal/Event/Control.hs - libraries/ghc-internal/src/GHC/Internal/Functor/ZipList.hs - libraries/ghc-internal/src/GHC/Internal/Generics.hs - libraries/ghc-internal/src/GHC/Internal/IO/Exception.hs - libraries/ghc-internal/src/GHC/Internal/RTS/Flags.hsc - libraries/ghc-internal/src/GHC/Internal/RTS/Flags/Test.hsc - libraries/ghc-internal/src/GHC/Internal/Read.hs - libraries/ghc-internal/src/GHC/Internal/Unicode/Bits.hs - testsuite/tests/ghci/scripts/ListTuplePunsPpr.stdout - testsuite/tests/ghci/scripts/T10963.stderr - testsuite/tests/ghci/scripts/ghci064.stdout - testsuite/tests/interface-stability/base-exports.stdout - testsuite/tests/interface-stability/base-exports.stdout-javascript-unknown-ghcjs - testsuite/tests/interface-stability/base-exports.stdout-mingw32 - testsuite/tests/interface-stability/base-exports.stdout-ws-32 The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/b169da6c2f6e2bc86e3a0a1c1239927… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/b169da6c2f6e2bc86e3a0a1c1239927… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/marge_bot_batch_merge_job] 17 commits: ghci: Mention active language edition in startup banner
by Marge Bot (＠marge-bot) 19 Mar '26

19 Mar '26

Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC Commits: f2374966 by Adam Gundry at 2026-03-19T15:40:43-04:00 ghci: Mention active language edition in startup banner Per GHC proposal 632, this makes the GHCi startup banner include the active language edition, plus an indication of whether this was the default (as opposed to being explicitly selected via an option such as `-XGHC2024`). For example: ``` $ ghci GHCi, version 9.14.1: https://www.haskell.org/ghc/ :? for help Using default language edition: GHC2024 ghci> ``` Fixes #26037. - - - - - b0ea6b92 by sheaf at 2026-03-19T15:40:49-04:00 Improve incomplete record selector warnings This commit stops GHC from emitting spurious incomplete record selector warnings for bare selectors/projections such as .fld There are two places we currently emit incomplete record selector warnings: 1. In the desugarer, when we see a record selector or an occurrence of 'getField'. Here, we can use pattern matching information to ensure we don't give false positives. 2. In the typechecker, which might sometimes give false positives but can emit warnings in cases that the pattern match checker would otherwise miss. This is explained in Note [Detecting incomplete record selectors] in GHC.HsToCore.Pmc. Now, we obviously don't want to emit the same error twice, and generally we prefer (1), as those messages contain fewer false positives. So we suppress (2) when we are sure we are going to emit (1); the logic for doing so is in GHC.Tc.Instance.Class.warnIncompleteRecSel, and works by looking at the CtOrigin. Now, the issue was that this logic handled explicit record selectors as well as overloaded record field selectors such as "x.r" (which turns into a simple GetFieldOrigin CtOrigin), but it didn't properly handle record projectors like ".fld" or ".fld1.fld2" (which result in other CtOrigins such as 'RecordFieldProjectionOrigin'). To solve this problem, we re-use the 'isHasFieldOrigin' introduced in fbdc623a (slightly adjusted). On the way, we also had to update the desugarer with special handling for the 'ExpandedThingTc' case in 'ds_app', to make sure that 'ds_app_var' sees all the type arguments to 'getField' in order for it to indeed emit warnings like in (1). Fixes #26686 - - - - - 06cec669 by Cheng Shao at 2026-03-19T15:40:51-04:00 rts: opportunistically grow the MutableByteArray# in-place in resizeMutableByteArray# Following !15234, this patch improves `resizeMutableByteArray#` memory efficiency by growing the `MutableByteArray#` in-place if possible, addressing an old todo comment here. Also adds a new test case `resizeMutableByteArrayInPlace` that stresses this behavior. - - - - - a564f77c by Matthew Craven at 2026-03-19T15:40:56-04:00 Change representation of floating point literals This commit changes the representation of floating point literals throughough the compiler, in particular in Core and Cmm. The Rational type is deficient for this purpose, dealing poorly with NaN, +/-Infinity, and negative zero. Instead, the new module GHC.Types.Literal.Floating uses the host Float/Double type to represent NaNs, infinities and negative zero. It also contains a Rational constructor, for the benefit of -fexcess-precision. Other changes: - Remove Note [negative zero] and related code This also removes the restrictions on constant-folding of division by zero, and should make any problems with NaN/Infinity more obvious. - Use -0.0 as the additive identity for Core constant folding rules for floating-point addition, fixing #21227. - Manual worker-wrapper for GHC.Float.rationalToDouble. This is intended to prevent the compiler's WW on this function from interfering with constant-folding. This change means that we now avoid allocating a box for the result of a 'realToFrac' call in T10359. - Combine floatDecodeOp and doubleDecodeOp. This change also fixes a bug in doubleDecodeOp wherein it would incorrectly produce an Int# instead of an Int64# literal for the mantissa component with 64-bit targets. - Use Float/Double for assembly immediates, and update the X86 and PowerPC backends to properly handle special values such as NaN and infinity. - Allow 'rational_to' to handle zero denominators, fixing a TODO in GHC.Core.Opt.ConstantFold. Fixes #8364 #9811 #18897 #21227 Progress towards #26919 Metric Decrease: T10359 Co-authored-by: sheaf <sam.derbyshire(a)gmail.com> ------------------------- Metric Decrease: T1969 T5321FD ------------------------- - - - - - 60a9d8c0 by Zubin Duggal at 2026-03-19T15:40:58-04:00 compiler/ffi: Collapse void pointer chains in capi wrappers New gcc/clang treat -Wincompatible-pointer-types as an error by default. Since C only allows implicit conversion from void*, not void**, capi wrappers for functions taking e.g. abstract** would fail to compile when the Haskell type Ptr (Ptr Abstract) was naively translated to void**. Collapse nested void pointers to a single void* when the pointee type has no known C representation. Fixes #26852 - - - - - 6d64deec by Luite Stegeman at 2026-03-19T15:41:02-04:00 Move some functions related to pointer tagging to a separate module - - - - - b83f6ae2 by Luite Stegeman at 2026-03-19T15:41:02-04:00 Branchless unpacking for enumeration types Change unpacking for enumeration types to go to Word8#/Word16#/Word# directly instead of going through an intermediate unboxed sum. This allows us to do a branchless conversion using DataToTag and TagToEnum. Fixes #26970 - - - - - 0454b140 by Luite Stegeman at 2026-03-19T15:41:07-04:00 bytecode: Carefully SLIDE off the end of a stack chunk The SLIDE bytecode instruction was not checking for stack chunk boundaries and could corrupt the stack underflow frame, leading to crashes. We add a check to use safe writes if we cross the chunk boundary and also handle stack underflow if Sp is advanced past the underflow frame. fix #27001 - - - - - 68f15429 by Cheng Shao at 2026-03-19T15:41:08-04:00 ghci: serialize BCOByteArray buffer directly when possible This patch changes the `Binary` instances of `BCOByteArray` to directly serialize the underlying buffer when possible, while also taking into account the issue of host-dependent `Word` width. See added comments and amended `Note [BCOByteArray serialization]` for detailed explanation. Closes #27020. - - - - - cfdfde8d by Sylvain Henry at 2026-03-19T15:41:15-04:00 JS: replace BigInt with Number arithmetic for 32/64-bit quot/rem (#23597) Replace BigInt-based implementations of quotWord32, remWord32, quotRemWord32, quotRem2Word32, quotWord64, remWord64, quotInt64, and remInt64 with pure Number (double/integer) arithmetic to avoid the overhead of BigInt promotion. - - - - - fe2949c0 by Sylvain Henry at 2026-03-19T15:41:21-04:00 Core: add constant-folding rules for Addr# eq/ne (#18032) - - - - - eaaf9930 by Matthew Pickering at 2026-03-19T15:41:23-04:00 Use OsPath rather than FilePath in Downsweep cache This gets us one step closure to uniformly using `OsPath` in the compiler. - - - - - b5475f85 by Cheng Shao at 2026-03-19T15:41:24-04:00 hadrian: fix ghc-in-ghci flavour stage0 shared libraries This patch fixes missing stage0 shared libraries in hadrian ghc-in-ghci flavour, which was accidentally dropped in 669d09f950a6e88b903d9fd8a7571531774d4d5d and resulted in a regression in HLS support on linux/macos. Fixes #27057. - - - - - 9c39c6b4 by Cheng Shao at 2026-03-19T15:41:25-04:00 ghc-internal: fix unused imports again Fixes #27059. - - - - - a97dc10f by Sylvain Henry at 2026-03-19T15:41:31-04:00 JS: install rts/Types.h header file (#27033) It was an omission, making HsFFI.h not usable with GHC using the JS backend. - - - - - 178179d0 by Cheng Shao at 2026-03-19T15:41:32-04:00 hadrian: don't compile RTS with -Winline This patch removes `-Winline` from cflags when compiling the RTS, given that: 1. It generates a huge pile of spam and hurts developer experience 2. Whether inlining happens is highly dependent on toolchains, flavours, etc, and it's not really an issue to fix if inlining doesn't happen; it's a hint to the C compiler anyway. Fixes #27060. - - - - - e5cf0804 by Cheng Shao at 2026-03-19T15:41:33-04:00 hadrian: compile libffi-clib with -Wno-deprecated-declarations This patch adds `-Wno-deprecated-declarations` to cflags of `libffi-clib`, given that it produces noise at compile-time that aren't really our issue to fix anyway, it's from vendored libffi source code. - - - - - 103 changed files: - compiler/GHC/Builtin/Names.hs - compiler/GHC/ByteCode/Asm.hs - compiler/GHC/Cmm/CommonBlockElim.hs - compiler/GHC/Cmm/Expr.hs - compiler/GHC/Cmm/Opt.hs - compiler/GHC/Cmm/Parser.y - compiler/GHC/Cmm/Type.hs - compiler/GHC/Cmm/Utils.hs - compiler/GHC/CmmToAsm/AArch64/CodeGen.hs - compiler/GHC/CmmToAsm/AArch64/Ppr.hs - compiler/GHC/CmmToAsm/AArch64/Regs.hs - compiler/GHC/CmmToAsm/LA64/CodeGen.hs - compiler/GHC/CmmToAsm/LA64/Ppr.hs - compiler/GHC/CmmToAsm/LA64/Regs.hs - compiler/GHC/CmmToAsm/PPC/CodeGen.hs - compiler/GHC/CmmToAsm/PPC/Ppr.hs - compiler/GHC/CmmToAsm/PPC/Regs.hs - compiler/GHC/CmmToAsm/RV64/CodeGen.hs - compiler/GHC/CmmToAsm/RV64/Ppr.hs - compiler/GHC/CmmToAsm/RV64/Regs.hs - compiler/GHC/CmmToAsm/Wasm/FromCmm.hs - compiler/GHC/CmmToAsm/X86/CodeGen.hs - compiler/GHC/CmmToAsm/X86/Ppr.hs - compiler/GHC/CmmToAsm/X86/Regs.hs - compiler/GHC/CmmToC.hs - compiler/GHC/CmmToLlvm/CodeGen.hs - compiler/GHC/CmmToLlvm/Data.hs - compiler/GHC/Core.hs - compiler/GHC/Core/Make.hs - compiler/GHC/Core/Opt/ConstantFold.hs - compiler/GHC/Driver/Downsweep.hs - compiler/GHC/HsToCore/Expr.hs - compiler/GHC/HsToCore/Foreign/C.hs - compiler/GHC/HsToCore/Match/Literal.hs - compiler/GHC/HsToCore/Pmc.hs - compiler/GHC/HsToCore/Pmc/Solver/Types.hs - + compiler/GHC/Platform/Tag.hs - compiler/GHC/Stg/Unarise.hs - compiler/GHC/StgToByteCode.hs - compiler/GHC/StgToCmm/Closure.hs - compiler/GHC/StgToCmm/Expr.hs - compiler/GHC/StgToCmm/Lit.hs - compiler/GHC/StgToCmm/Prim.hs - compiler/GHC/StgToJS/Literal.hs - compiler/GHC/Tc/Errors.hs - compiler/GHC/Tc/Instance/Class.hs - compiler/GHC/Tc/TyCl/Build.hs - compiler/GHC/Tc/Types/Origin.hs - compiler/GHC/Types/Id/Make.hs - compiler/GHC/Types/Literal.hs - + compiler/GHC/Types/Literal/Floating.hs - compiler/GHC/Types/RepType.hs - compiler/GHC/Utils/Binary.hs - compiler/ghc.cabal.in - docs/users_guide/9.16.1-notes.rst - ghc/GHCi/UI.hs - ghc/Main.hs - hadrian/src/Settings/Flavours/GhcInGhci.hs - hadrian/src/Settings/Packages.hs - libraries/ghc-internal/src/GHC/Internal/Conc/IO.hs - libraries/ghc-internal/src/GHC/Internal/Event/Control.hs - libraries/ghc-internal/src/GHC/Internal/Float.hs - libraries/ghc-internal/src/GHC/Internal/RTS/Flags.hsc - libraries/ghc-internal/src/GHC/Internal/RTS/Flags/Test.hsc - libraries/ghc-internal/src/GHC/Internal/System/Posix/Internals.hs - libraries/ghc-internal/src/GHC/Internal/TopHandler.hs - libraries/ghci/GHCi/ResolvedBCO.hs - rts/Interpreter.c - rts/PrimOps.cmm - rts/js/arith.js - rts/rts.cabal - + testsuite/tests/bytecode/T27001.hs - + testsuite/tests/bytecode/T27001.stdout - testsuite/tests/bytecode/all.T - + testsuite/tests/codeGen/should_run/T21227.hs - + testsuite/tests/codeGen/should_run/T21227.stdout - + testsuite/tests/codeGen/should_run/T9811.hs - + testsuite/tests/codeGen/should_run/T9811.stdout - testsuite/tests/codeGen/should_run/all.T - testsuite/tests/count-deps/CountDepsAst.stdout - testsuite/tests/count-deps/CountDepsParser.stdout - + testsuite/tests/ffi/should_compile/T26852.h - + testsuite/tests/ffi/should_compile/T26852.hs - + testsuite/tests/ffi/should_compile/T26852.stderr - testsuite/tests/ffi/should_compile/all.T - testsuite/tests/ghci/should_run/BinaryArray.hs - + testsuite/tests/javascript/js-c-sources/T27033.hs - + testsuite/tests/javascript/js-c-sources/T27033.stdout - + testsuite/tests/javascript/js-c-sources/T27033_c.c - + testsuite/tests/javascript/js-c-sources/T27033_js.js - testsuite/tests/javascript/js-c-sources/all.T - testsuite/tests/numeric/should_run/T7014.hs - + testsuite/tests/overloadedrecflds/should_compile/T26686.hs - + testsuite/tests/overloadedrecflds/should_compile/T26686.stderr - testsuite/tests/overloadedrecflds/should_compile/all.T - testsuite/tests/rts/all.T - + testsuite/tests/rts/resizeMutableByteArrayInPlace.hs - + testsuite/tests/simplCore/should_compile/T18032.hs - + testsuite/tests/simplCore/should_compile/T18032.stderr - testsuite/tests/simplCore/should_compile/all.T - testsuite/tests/simplStg/should_run/all.T - + testsuite/tests/simplStg/should_run/unpack_enum.hs - + testsuite/tests/simplStg/should_run/unpack_enum.stdout The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/3f18ff78e6b177bb3220e37044ec13… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/3f18ff78e6b177bb3220e37044ec13… You're receiving this email because of your account on gitlab.haskell.org.

1 0