- ghc-commits - Haskell.org

[Git][ghc/ghc][wip/T23162-part2] Add (CF5)
by Simon Peyton Jones (＠simonpj) 28 Dec '25

28 Dec '25

Simon Peyton Jones pushed to branch wip/T23162-part2 at Glasgow Haskell Compiler / GHC Commits: 0b6d14a9 by Simon Peyton Jones at 2025-12-28T14:47:49+00:00 Add (CF5) - - - - - 1 changed file: - compiler/GHC/Tc/Solver/FunDeps.hs Changes: ===================================== compiler/GHC/Tc/Solver/FunDeps.hs ===================================== @@ -23,6 +23,7 @@ import GHC.Tc.Types.Constraint import GHC.Core.FamInstEnv import GHC.Core.Coercion +import GHC.Core.Type( tyConAppTyCon_maybe ) import GHC.Core.Predicate( EqRel(..) ) import GHC.Core.TyCon import GHC.Core.Unify( tcUnifyTysForInjectivity, typeListsAreApart ) @@ -633,6 +634,8 @@ mkTopClosedFamEqFDs :: CoAxiom Branched -> [TcType] -> EqCt -> SolverStage [FunD -- type families without injectivity info -- See Note [Exploiting closed type families] mkTopClosedFamEqFDs ax work_args (EqCt { eq_ev = ev, eq_rhs = work_rhs }) + | Just tc <- tyConAppTyCon_maybe work_rhs + , isGenerativeTyCon tc Nominal -- See (CF5) in Note [Exploiting closed type families] = Stage $ do { let branches = fromBranches (coAxiomBranches ax) ; traceTcS "mkTopClosed" (ppr branches $$ ppr work_args $$ ppr work_rhs) @@ -640,6 +643,8 @@ mkTopClosedFamEqFDs ax work_args (EqCt { eq_ev = ev, eq_rhs = work_rhs }) [] -> insolubleFunDep True ev -- See (IFD0) in Note [Insoluble fundeps] [eqn] -> continueWith [eqn] -- If there is just one relevant equation, use it _ -> continueWith [] } + | otherwise + = Stage $ continueWith [] hasRelevantGiven :: [EqCt] -> [TcType] -> EqCt -> Bool -- See (CF1) in Note [Exploiting closed type families] @@ -1115,6 +1120,32 @@ Key point: equations that are not relevant do not need to be considered for fund I was looking at non-termination for closed type families, but it's a small improvement in general. +(CF5) Consider + type family F a where + F (Just x) = Int + + [W] F alpha ~ F (G beta) + [W] alpha ~ G beta + We can solve both Wanteds by `alpha := G beta`. But if we use fundeps on the + first Wanted, we see that there is just one relevant equation, so we'll emit + [W] alpha ~ Just gamma + [W] F (G beta) ~ Int + and try to solve them. We'll do `alpha := Just gamma`, and then it's game over; + we end up with these constraints that we can't solve + [W] Just gamma ~ F (G beta) + [W] Just gamma ~ G beta + This actually happens when compiling the libarary `type-rows`, in Data.Row.Variants. + + Even if there is only one relevant equation, we can only use it /if/ we are sure that we + cannot solve the current Wanted by reflexivity; that is, if we must do a type-family + reduction to solve it. So if the Wanted is + ``` + [W] F tys ~ rhs + ``` + we must be sure that `rhs` can't turn into `F tys`. The only way to be sure of that + is if `rhs` is headed by a generative type constructor. See the test for + `isGenerativeTyCon` in `mkTopClosedFamEqFDs`. + Note [Cache-caused loops] ~~~~~~~~~~~~~~~~~~~~~~~~~ It is very dangerous to cache a rewritten wanted family equation as 'solved' in our View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/0b6d14a97a1f0b7f6fe3d283664759c… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/0b6d14a97a1f0b7f6fe3d283664759c… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/ghci-spark] ghci: avoid arbitrary chunking in parallelized CreateBCOs serialization
by Cheng Shao (＠TerrorJack) 28 Dec '25

28 Dec '25

Cheng Shao pushed to branch wip/ghci-spark at Glasgow Haskell Compiler / GHC Commits: 8833cbff by Cheng Shao at 2025-12-28T06:50:55+01:00 ghci: avoid arbitrary chunking in parallelized CreateBCOs serialization Previously, when the `CreateBCOs` message is serialized, we split the `[ResolvedBCO]` list into chunks and use sparks to serialize the chunks in parallel. However, in terms of load balancing, instead of arbitrarily grouping work, it's better to spawn work on a finer granularity and rely on the work-stealing nature of the sparks mechanism to distribute work more evenly across idle Capabilities. So this patch refactors the `CreateBCOs` serialization logic to spawn one spark per `ResolvedBCO` object, both when serializing and deserializing. This is based on an added helper module `GHC.Utils.Spark` in `ghc-boot`, which might come in handy later when we want to use sparks in other parts of the compiler. - - - - - 3 changed files: - + libraries/ghc-boot/GHC/Utils/Spark.hs - libraries/ghc-boot/ghc-boot.cabal.in - libraries/ghci/GHCi/Message.hs Changes: ===================================== libraries/ghc-boot/GHC/Utils/Spark.hs ===================================== @@ -0,0 +1,25 @@ +{-# LANGUAGE MagicHash #-} + +module GHC.Utils.Spark + ( sparkST, + sparkIO, + ) +where + +import Control.Monad +import GHC.Exts +import GHC.IO +import GHC.ST +import Prelude + +-- | Same as 'sparkIO' but in the 'ST' monad, useful if no other side +-- effects are involved. +sparkST :: ST s a -> ST s a +sparkST = ST . spark# <=< unsafeInterleaveST + +-- | Spawns an 'IO' computation as a spark and returns a thunk that +-- evaluates to the computation's result. The computation is guarded +-- by 'noDuplicate' to prevent being executed by multiple threads at +-- once. +sparkIO :: IO a -> IO a +sparkIO = IO . spark# <=< unsafeInterleaveIO ===================================== libraries/ghc-boot/ghc-boot.cabal.in ===================================== @@ -55,6 +55,7 @@ Library GHC.Data.SizedSeq GHC.Utils.Encoding GHC.Utils.Encoding.UTF8 + GHC.Utils.Spark GHC.LanguageExtensions GHC.Unit.Database GHC.Serialized ===================================== libraries/ghci/GHCi/Message.hs ===================================== @@ -44,13 +44,13 @@ import qualified GHC.Exts.Heap as Heap #endif import GHC.ForeignSrcLang import GHC.Fingerprint -import GHC.Conc (pseq, par) import Control.Concurrent import Control.DeepSeq import Control.Exception #if MIN_VERSION_base(4,20,0) import Control.Exception.Context #endif +import Control.Monad.ST import Data.Binary import Data.Binary.Get import Data.Binary.Put @@ -66,6 +66,7 @@ import Data.Map (Map) import Foreign import GHC.Generics import GHC.Stack.CCS +import GHC.Utils.Spark import qualified GHC.Boot.TH.Syntax as TH import qualified GHC.Boot.TH.Monad as TH import System.Exit @@ -572,7 +573,7 @@ getMessage = do 9 -> Msg <$> RemoveLibrarySearchPath <$> get 10 -> Msg <$> return ResolveObjs 11 -> Msg <$> FindSystemLibrary <$> get - 12 -> Msg <$> (CreateBCOs . concatMap (runGet get)) <$> (get :: Get [LB.ByteString]) + 12 -> Msg <$> (CreateBCOs . \sbcos -> runST $ traverse (sparkST . pure . runGet get) sbcos) <$> (get :: Get [LB.ByteString]) -- See Note [Parallelize CreateBCOs serialization] 13 -> Msg <$> FreeHValueRefs <$> get 14 -> Msg <$> MallocData <$> get @@ -653,29 +654,17 @@ putMessage m = case m of Note [Parallelize CreateBCOs serialization] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Serializing ResolvedBCO is expensive, so we do it in parallel. -We split the list [ResolvedBCO] into chunks of length <= 100, -and serialize every chunk in parallel, getting a [LB.ByteString] -where every bytestring corresponds to a single chunk (multiple ResolvedBCOs). +For each element in the list [ResolvedBCO], we spawn a spark which +serializes it, getting a [LB.ByteString] where every fully evaluated +lazy ByteString corresponds to a single ResolvedBCO. Previously, we stored [LB.ByteString] in the Message object, but that incurs unneccessary serialization with the internal interpreter (#23919). -} serializeBCOs :: [ResolvedBCO] -> [LB.ByteString] -serializeBCOs rbcos = parMap doChunk (chunkList 100 rbcos) - where - -- make sure we force the whole lazy ByteString - doChunk c = pseq (LB.length bs) bs - where bs = runPut (put c) - - -- We don't have the parallel package, so roll our own simple parMap - parMap _ [] = [] - parMap f (x:xs) = fx `par` (fxs `pseq` (fx : fxs)) - where fx = f x; fxs = parMap f xs - - chunkList :: Int -> [a] -> [[a]] - chunkList _ [] = [] - chunkList n xs = as : chunkList n bs where (as,bs) = splitAt n xs +serializeBCOs rbcos = + runST $ traverse (sparkST . pure . force . runPut . put) rbcos -- ----------------------------------------------------------------------------- -- Reading/writing messages View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/8833cbff359bf0a2b0e6b8a0951839f… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/8833cbff359bf0a2b0e6b8a0951839f… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/ghci-spark] ghci: avoid arbitrary chunking in parallelized CreateBCOs serialization
by Cheng Shao (＠TerrorJack) 28 Dec '25

28 Dec '25

Cheng Shao pushed to branch wip/ghci-spark at Glasgow Haskell Compiler / GHC Commits: 3142ed32 by Cheng Shao at 2025-12-28T06:41:07+01:00 ghci: avoid arbitrary chunking in parallelized CreateBCOs serialization Previously, when the `CreateBCOs` message is serialized, we split the `[ResolvedBCO]` list into chunks and use sparks to serialize the chunks in parallel. However, in terms of load balancing, instead of arbitrarily grouping work, it's better to spawn work on a finer granularity and rely on the work-stealing nature of the sparks mechanism to distribute work more evenly across idle Capabilities. So this patch refactors the `CreateBCOs` serialization logic to spawn one spark per `ResolvedBCO` object. This is based on an added helper module `GHC.Utils.Spark` in `ghc-boot`, which might come in handy later when we want to use sparks in other parts of the compiler. - - - - - 3 changed files: - + libraries/ghc-boot/GHC/Utils/Spark.hs - libraries/ghc-boot/ghc-boot.cabal.in - libraries/ghci/GHCi/Message.hs Changes: ===================================== libraries/ghc-boot/GHC/Utils/Spark.hs ===================================== @@ -0,0 +1,25 @@ +{-# LANGUAGE MagicHash #-} + +module GHC.Utils.Spark + ( sparkST, + sparkIO, + ) +where + +import Control.Monad +import GHC.Exts +import GHC.IO +import GHC.ST +import Prelude + +-- | Same as 'sparkIO' but in the 'ST' monad, useful if no other side +-- effects are involved. +sparkST :: ST s a -> ST s a +sparkST = ST . spark# <=< unsafeInterleaveST + +-- | Spawns an 'IO' computation as a spark and returns a thunk that +-- evaluates to the computation's result. The computation is guarded +-- by 'noDuplicate' to prevent being executed by multiple threads at +-- once. +sparkIO :: IO a -> IO a +sparkIO = IO . spark# <=< unsafeInterleaveIO ===================================== libraries/ghc-boot/ghc-boot.cabal.in ===================================== @@ -55,6 +55,7 @@ Library GHC.Data.SizedSeq GHC.Utils.Encoding GHC.Utils.Encoding.UTF8 + GHC.Utils.Spark GHC.LanguageExtensions GHC.Unit.Database GHC.Serialized ===================================== libraries/ghci/GHCi/Message.hs ===================================== @@ -44,13 +44,13 @@ import qualified GHC.Exts.Heap as Heap #endif import GHC.ForeignSrcLang import GHC.Fingerprint -import GHC.Conc (pseq, par) import Control.Concurrent import Control.DeepSeq import Control.Exception #if MIN_VERSION_base(4,20,0) import Control.Exception.Context #endif +import Control.Monad.ST import Data.Binary import Data.Binary.Get import Data.Binary.Put @@ -66,6 +66,7 @@ import Data.Map (Map) import Foreign import GHC.Generics import GHC.Stack.CCS +import GHC.Utils.Spark import qualified GHC.Boot.TH.Syntax as TH import qualified GHC.Boot.TH.Monad as TH import System.Exit @@ -572,7 +573,7 @@ getMessage = do 9 -> Msg <$> RemoveLibrarySearchPath <$> get 10 -> Msg <$> return ResolveObjs 11 -> Msg <$> FindSystemLibrary <$> get - 12 -> Msg <$> (CreateBCOs . concatMap (runGet get)) <$> (get :: Get [LB.ByteString]) + 12 -> Msg <$> (CreateBCOs . fmap (runGet get)) <$> (get :: Get [LB.ByteString]) -- See Note [Parallelize CreateBCOs serialization] 13 -> Msg <$> FreeHValueRefs <$> get 14 -> Msg <$> MallocData <$> get @@ -653,29 +654,17 @@ putMessage m = case m of Note [Parallelize CreateBCOs serialization] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Serializing ResolvedBCO is expensive, so we do it in parallel. -We split the list [ResolvedBCO] into chunks of length <= 100, -and serialize every chunk in parallel, getting a [LB.ByteString] -where every bytestring corresponds to a single chunk (multiple ResolvedBCOs). +For each element in the list [ResolvedBCO], we spawn a spark which +serializes it, getting a [LB.ByteString] where every fully evaluated +lazy ByteString corresponds to a single ResolvedBCO. Previously, we stored [LB.ByteString] in the Message object, but that incurs unneccessary serialization with the internal interpreter (#23919). -} serializeBCOs :: [ResolvedBCO] -> [LB.ByteString] -serializeBCOs rbcos = parMap doChunk (chunkList 100 rbcos) - where - -- make sure we force the whole lazy ByteString - doChunk c = pseq (LB.length bs) bs - where bs = runPut (put c) - - -- We don't have the parallel package, so roll our own simple parMap - parMap _ [] = [] - parMap f (x:xs) = fx `par` (fxs `pseq` (fx : fxs)) - where fx = f x; fxs = parMap f xs - - chunkList :: Int -> [a] -> [[a]] - chunkList _ [] = [] - chunkList n xs = as : chunkList n bs where (as,bs) = splitAt n xs +serializeBCOs rbcos = + runST $ traverse (sparkST . pure . force . runPut . put) rbcos -- ----------------------------------------------------------------------------- -- Reading/writing messages View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/3142ed3262732173721dcd2fa8f7905… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/3142ed3262732173721dcd2fa8f7905… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new tag ghc-9.12.3-release
by Zubin (＠wz1000) 27 Dec '25

27 Dec '25

Zubin pushed new tag ghc-9.12.3-release at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/ghc-9.12.3-release You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][ghc-9.12] Finalize 9.12.3
by Zubin (＠wz1000) 27 Dec '25

27 Dec '25

Zubin pushed to branch ghc-9.12 at Glasgow Haskell Compiler / GHC Commits: 8b63dfe0 by Zubin Duggal at 2025-12-26T02:18:47+05:30 Finalize 9.12.3 - - - - - 5 changed files: - .gitlab/rel_eng/upload.sh - configure.ac - docs/users_guide/9.12.1-notes.rst - docs/users_guide/9.12.2-notes.rst - docs/users_guide/9.12.3-notes.rst Changes: ===================================== .gitlab/rel_eng/upload.sh ===================================== @@ -1,4 +1,4 @@ -#!/usr/bin/env nix-shell +#!/usr/bin/bash #! nix-shell -i bash -p moreutils lzip zip lftp gnupg set -Eeuo pipefail @@ -41,7 +41,7 @@ set -Eeuo pipefail : ${ver:=$(ls ghc-*.tar.* | sed -ne 's/ghc-$[0-9]\+\.[0-9]\+\.[0-9]\+\(\.[0-9]\+$\?\).\+/\1/p' | head -n1)} if [ -z "$ver" ]; then echo "Failed to infer \$ver"; exit 1; fi -host="gitlab-storage.haskell.org" +host="gitlab.haskell.org:2222" usage() { echo "Usage: [rel_name=<name>] ver=7.10.3-rc2 $0 <action>" ===================================== configure.ac ===================================== @@ -13,7 +13,7 @@ dnl # see what flags are available. (Better yet, read the documentation!) # -AC_INIT([The Glorious Glasgow Haskell Compilation System], [9.12.2], [glasgow-haskell-bugs(a)haskell.org] [ghc-AC_PACKAGE_VERSION]) +AC_INIT([The Glorious Glasgow Haskell Compilation System], [9.12.3], [glasgow-haskell-bugs(a)haskell.org] [ghc-AC_PACKAGE_VERSION]) # Version on master must be X.Y (not X.Y.Z) for ProjectVersionMunged variable # to be useful (cf #19058). However, the version must have three components # (X.Y.Z) on stable branches (e.g. ghc-9.2) to ensure that pre-releases are @@ -22,7 +22,7 @@ AC_INIT([The Glorious Glasgow Haskell Compilation System], [9.12.2], [glasgow-ha AC_CONFIG_MACRO_DIRS([m4]) # Set this to YES for a released version, otherwise NO -: ${RELEASE=NO} +: ${RELEASE=YES} # The primary version (e.g. 7.5, 7.4.1) is set in the AC_INIT line # above. If this is not a released version, then we will append the ===================================== docs/users_guide/9.12.1-notes.rst ===================================== @@ -290,55 +290,3 @@ j~~~~~~~~~~~~~~~~~~~~~~~~~~~~ <https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0281-v…>`_. - ``template-haskell`` is no longer wired-in. All wired-in identifiers have been moved to ``ghc-internal``. - ``Lift`` instances were added for the ``template-haskell`` AST. - -Included libraries -~~~~~~~~~~~~~~~~~~ - -The package database provided with this distribution also contains a number of -packages other than GHC itself. See the changelogs provided with these packages -for further change information. - -.. ghc-package-list:: - - compiler/ghc.cabal: The compiler itself - libraries/array/array.cabal: Dependency of ``ghc`` library - libraries/base/base.cabal: Core library - libraries/binary/binary.cabal: Dependency of ``ghc`` library - libraries/bytestring/bytestring.cabal: Dependency of ``ghc`` library - libraries/Cabal/Cabal/Cabal.cabal: Dependency of ``ghc-pkg`` utility - libraries/Cabal/Cabal-syntax/Cabal-syntax.cabal: Dependency of ``ghc-pkg`` utility - libraries/containers/containers/containers.cabal: Dependency of ``ghc`` library - libraries/deepseq/deepseq.cabal: Dependency of ``ghc`` library - libraries/directory/directory.cabal: Dependency of ``ghc`` library - libraries/exceptions/exceptions.cabal: Dependency of ``ghc`` and ``haskeline`` library - libraries/file-io/file-io.cabal: Dependency of ``directory`` library - libraries/filepath/filepath.cabal: Dependency of ``ghc`` library - libraries/ghc-boot/ghc-boot.cabal: Internal compiler library - libraries/ghc-boot-th/ghc-boot-th.cabal: Internal compiler library - libraries/ghc-compact/ghc-compact.cabal: Core library - libraries/ghc-experimental/ghc-experimental.cabal: Core library - libraries/ghc-heap/ghc-heap.cabal: GHC heap-walking library - libraries/ghci/ghci.cabal: The REPL interface - libraries/ghc-internal/ghc-internal.cabal: Core library - libraries/ghc-platform/ghc-platform.cabal: Internal library - libraries/ghc-prim/ghc-prim.cabal: Core library - libraries/haskeline/haskeline.cabal: Dependency of ``ghci`` executable - libraries/hpc/hpc.cabal: Dependency of ``hpc`` executable - libraries/integer-gmp/integer-gmp.cabal: Core library - libraries/mtl/mtl.cabal: Dependency of ``Cabal`` library - libraries/os-string/os-string.cabal: Dependency of ``filepath`` library - libraries/parsec/parsec.cabal: Dependency of ``Cabal`` library - libraries/pretty/pretty.cabal: Dependency of ``ghc`` library - libraries/process/process.cabal: Dependency of ``ghc`` library - libraries/semaphore-compat/semaphore-compat.cabal: Dependency of ``ghc`` library - libraries/stm/stm.cabal: Dependency of ``haskeline`` library - libraries/template-haskell/template-haskell.cabal: Core library - libraries/terminfo/terminfo.cabal: Dependency of ``haskeline`` library - libraries/text/text.cabal: Dependency of ``Cabal`` library - libraries/time/time.cabal: Dependency of ``ghc`` library - libraries/transformers/transformers.cabal: Dependency of ``ghc`` library - libraries/unix/unix.cabal: Dependency of ``ghc`` library - libraries/Win32/Win32.cabal: Dependency of ``ghc`` library - libraries/xhtml/xhtml.cabal: Dependency of ``haddock`` executable - utils/haddock/haddock-api/haddock-api.cabal: Dependency of ``haddock`` executable - utils/haddock/haddock-library/haddock-library.cabal: Dependency of ``haddock`` executable ===================================== docs/users_guide/9.12.2-notes.rst ===================================== @@ -12,56 +12,3 @@ Compiler ~~~~~~~~ - Fixed miscompilation of certain division operations (:ghc-ticket:`25653`) - -Included libraries -~~~~~~~~~~~~~~~~~~ - -The package database provided with this distribution also contains a number of -packages other than GHC itself. See the changelogs provided with these packages -for further change information. - -.. ghc-package-list:: - - compiler/ghc.cabal: The compiler itself - libraries/array/array.cabal: Dependency of ``ghc`` library - libraries/base/base.cabal: Core library - libraries/binary/binary.cabal: Dependency of ``ghc`` library - libraries/bytestring/bytestring.cabal: Dependency of ``ghc`` library - libraries/Cabal/Cabal/Cabal.cabal: Dependency of ``ghc-pkg`` utility - libraries/Cabal/Cabal-syntax/Cabal-syntax.cabal: Dependency of ``ghc-pkg`` utility - libraries/containers/containers/containers.cabal: Dependency of ``ghc`` library - libraries/deepseq/deepseq.cabal: Dependency of ``ghc`` library - libraries/directory/directory.cabal: Dependency of ``ghc`` library - libraries/exceptions/exceptions.cabal: Dependency of ``ghc`` and ``haskeline`` library - libraries/file-io/file-io.cabal: Dependency of ``directory`` library - libraries/filepath/filepath.cabal: Dependency of ``ghc`` library - libraries/ghc-boot/ghc-boot.cabal: Internal compiler library - libraries/ghc-boot-th/ghc-boot-th.cabal: Internal compiler library - libraries/ghc-compact/ghc-compact.cabal: Core library - libraries/ghc-experimental/ghc-experimental.cabal: Core library - libraries/ghc-heap/ghc-heap.cabal: GHC heap-walking library - libraries/ghci/ghci.cabal: The REPL interface - libraries/ghc-internal/ghc-internal.cabal: Core library - libraries/ghc-platform/ghc-platform.cabal: Internal library - libraries/ghc-prim/ghc-prim.cabal: Core library - libraries/haskeline/haskeline.cabal: Dependency of ``ghci`` executable - libraries/hpc/hpc.cabal: Dependency of ``hpc`` executable - libraries/integer-gmp/integer-gmp.cabal: Core library - libraries/mtl/mtl.cabal: Dependency of ``Cabal`` library - libraries/os-string/os-string.cabal: Dependency of ``filepath`` library - libraries/parsec/parsec.cabal: Dependency of ``Cabal`` library - libraries/pretty/pretty.cabal: Dependency of ``ghc`` library - libraries/process/process.cabal: Dependency of ``ghc`` library - libraries/semaphore-compat/semaphore-compat.cabal: Dependency of ``ghc`` library - libraries/stm/stm.cabal: Dependency of ``haskeline`` library - libraries/template-haskell/template-haskell.cabal: Core library - libraries/terminfo/terminfo.cabal: Dependency of ``haskeline`` library - libraries/text/text.cabal: Dependency of ``Cabal`` library - libraries/time/time.cabal: Dependency of ``ghc`` library - libraries/transformers/transformers.cabal: Dependency of ``ghc`` library - libraries/unix/unix.cabal: Dependency of ``ghc`` library - libraries/Win32/Win32.cabal: Dependency of ``ghc`` library - libraries/xhtml/xhtml.cabal: Dependency of ``haddock`` executable - utils/haddock/haddock-api/haddock-api.cabal: Dependency of ``haddock`` executable - utils/haddock/haddock-library/haddock-library.cabal: Dependency of ``haddock`` executable - ===================================== docs/users_guide/9.12.3-notes.rst ===================================== @@ -17,11 +17,14 @@ Compiler - Fixed CprAnal to detect recursive newtypes (:ghc-ticket:`25944`) - Fixed specialisation of incoherent instances (:ghc-ticket:`25883`) - Fixed bytecode generation for ``tagToEnum# <LITERAL>`` (:ghc-ticket:`25975`) +- Fixed bytecode generation for unsaturated data constructor workers (:ghc-ticket:`23210`) - Fixed panic with EmptyCase and RequiredTypeArguments (:ghc-ticket:`25004`) +- Fixed EmptyCase panic in tcMatches with invisible forall (:ghc-ticket:`25960`) - Fixed ``tyConStupidTheta`` to handle ``PromotedDataCon`` (:ghc-ticket:`25739`) - Fixed unused import warnings for duplicate record fields (:ghc-ticket:`24035`) - Fixed lexing of ``"\^\"`` (:ghc-ticket:`25937`) - Fixed string gap collapsing (:ghc-ticket:`25784`) +- Fixed tabs in string gaps (:ghc-ticket:`26415`) - Fixed lexing of comments in multiline strings (:ghc-ticket:`25609`) - Made unexpected LLVM versions a warning rather than an error (:ghc-ticket:`25915`) - Disabled ``-fprof-late-overloaded-calls`` for join points to avoid invalid transformations @@ -38,6 +41,7 @@ Runtime System - Fixed ``MessageBlackHole.link`` to always be a valid closure - Fixed handling of WHITEHOLE in ``messageBlackHole`` (:ghc-ticket:`26205`) +- Fixed handling of WHITEHOLEs in ``scavenge_one`` (:ghc-ticket:`26204`) - Fixed ``rts_clearMemory`` logic when sanity checks are enabled (:ghc-ticket:`26011`) - Fixed underflow frame lookups in the bytecode interpreter (:ghc-ticket:`25750`) - Fixed overflows and reentrancy in interpreter statistics calculation (:ghc-ticket:`25756`) @@ -76,6 +80,8 @@ Linker - Fixed incorrect use of ``break`` in nested for loop (:ghc-ticket:`26052`) - Fixed linker to not fail due to ``RTLD_NOW`` (:ghc-ticket:`25943`) - Dropped obsolete Windows XP compatibility checks +- Fix the Windows runtime linker to copy DLL path strings before inserting + them into the cache, preventing use-after-free issues (:ghc-ticket:`26613`) GHCi ~~~~ View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/8b63dfe0a81d08fef5b4acba6e60d80… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/8b63dfe0a81d08fef5b4acba6e60d80… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/rts-sparkpool-1m
by Cheng Shao (＠TerrorJack) 27 Dec '25

27 Dec '25

Cheng Shao pushed new branch wip/rts-sparkpool-1m at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/rts-sparkpool-1m You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/ghci-spark
by Cheng Shao (＠TerrorJack) 26 Dec '25

26 Dec '25

Cheng Shao pushed new branch wip/ghci-spark at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/ghci-spark You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/cleanup-split-objs
by Cheng Shao (＠TerrorJack) 26 Dec '25

26 Dec '25

Cheng Shao pushed new branch wip/cleanup-split-objs at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/cleanup-split-objs You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/9.12.3-final
by Zubin (＠wz1000) 26 Dec '25

26 Dec '25

Zubin pushed new branch wip/9.12.3-final at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/9.12.3-final You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][master] Check for rubbish literals in Lint
by Marge Bot (＠marge-bot) 26 Dec '25

26 Dec '25

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: a3afae0c by Simon Peyton Jones at 2025-12-25T15:26:36-05:00 Check for rubbish literals in Lint Addresses #26607. See new Note [Checking for rubbish literals] in GHC.Core.Lint - - - - - 5 changed files: - compiler/GHC/Core/Lint.hs - compiler/GHC/Core/Utils.hs - compiler/GHC/Driver/Config/Core/Lint.hs - compiler/GHC/Driver/Config/Core/Lint/Interactive.hs - compiler/GHC/IfaceToCore.hs Changes: ===================================== compiler/GHC/Core/Lint.hs ===================================== @@ -405,7 +405,6 @@ data LintPassResultConfig = LintPassResultConfig { lpr_diagOpts :: !DiagOpts , lpr_platform :: !Platform , lpr_makeLintFlags :: !LintFlags - , lpr_showLintWarnings :: !Bool , lpr_passPpr :: !SDoc , lpr_localsInScope :: ![Var] } @@ -425,18 +424,16 @@ lintPassResult logger cfg binds "Core Linted result of " ++ renderWithContext defaultSDocContext (lpr_passPpr cfg) ; displayLintResults logger - (lpr_showLintWarnings cfg) (lpr_passPpr cfg) + (lpr_passPpr cfg) (pprCoreBindings binds) warns_and_errs } displayLintResults :: Logger - -> Bool -- ^ If 'True', display linter warnings. - -- If 'False', ignore linter warnings. -> SDoc -- ^ The source of the linted program -> SDoc -- ^ The linted program, pretty-printed -> WarnsAndErrs -> IO () -displayLintResults logger display_warnings pp_what pp_pgm (warns, errs) +displayLintResults logger pp_what pp_pgm (warns, errs) | not (isEmptyBag errs) = do { lintMessage logger (vcat [ lint_banner "errors" pp_what, Err.pprMessageBag errs @@ -447,7 +444,6 @@ displayLintResults logger display_warnings pp_what pp_pgm (warns, errs) | not (isEmptyBag warns) , log_enable_debug (logFlags logger) - , display_warnings = lintMessage logger (lint_banner "warnings" pp_what $$ Err.pprMessageBag (mapBag ($$ blankLine) warns)) @@ -858,6 +854,31 @@ remember the details, but could probably recover it if we want to revisit. So Lint current accepts (Coercion co) in arbitrary places. There is no harm in that: it really is a value, albeit a zero-bit value. +Note [Checking for rubbish literals] +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +GHC uses "rubbish literals" (see Note [Rubbish literals] in GHC.Types.Literal) +to fill for absent arguments, in worker/wrapper. See Note [Absent fillers] in +GHC.Core.Opt.WorkWrap.Utils. + +These rubbish literals are almost always discarded as dead code, in post-w/w +optimisation. If they are ever used at runtime, something is wrong. For lifted +absent fillers, instead of using a `LitRubbish` we use an error thunk. So if, +through some catastrophe, it is used at runtime after all, we get a civilised +runtime error + +But for /unlifted/ ones we can't use a thunk, so we use a random "rubbish" value +of the right type. That could lead to bizarre behaviour at runtime. + +Worse, for dictionaries, for reasons explained in Note [Absent fillers], we also +don't use an error thunk. Now if we use it at runtime after all, a seg-fault will +happen (e.g. #26416). Yikes! + +So Lint warns about the presence of rubbish literals, in the output of CorePrep +only (see GHC.Driver.Config.Core.Lint.perPassFlags). It's a warning, not an +error, because it's not /necessarily/ wrong to see a rubbish literal. But it's +suspicious and worth investigating if you have a seg-fault or bizarre behaviour. + + ************************************************************************ * * \subsection[lintCoreExpr]{lintCoreExpr} @@ -883,7 +904,14 @@ lintCoreExpr (Var var) ; return var_pair } lintCoreExpr (Lit lit) - = return (literalType lit, zeroUE) + = do { flags <- getLintFlags + + ; -- See Note [Checking for rubbish literals] + when (lf_check_rubbish_lits flags) $ + checkWarnL (not (isLitRubbish lit)) $ + hang (text "Unexpected rubbish literal:") 2 (ppr lit) + + ; return (literalType lit, zeroUE) } lintCoreExpr (Cast expr co) = do { (expr_ty, ue) <- markAllJoinsBad (lintCoreExpr expr) @@ -2785,7 +2813,7 @@ lintAxioms :: Logger -> [CoAxiom Branched] -> IO () lintAxioms logger cfg what axioms = - displayLintResults logger True what (vcat $ map pprCoAxiom axioms) $ + displayLintResults logger what (vcat $ map pprCoAxiom axioms) $ initL cfg $ do { mapM_ lint_axiom axioms ; let axiom_groups = groupWith coAxiomTyCon axioms @@ -2968,9 +2996,10 @@ data LintFlags = LF { lf_check_global_ids :: Bool -- See Note [Checking for global Ids] , lf_check_inline_loop_breakers :: Bool -- See Note [Checking for INLINE loop breakers] , lf_check_static_ptrs :: StaticPtrCheck -- ^ See Note [Checking StaticPtrs] - , lf_report_unsat_syns :: Bool -- ^ See Note [Linting type synonym applications] - , lf_check_linearity :: Bool -- ^ See Note [Linting linearity] - , lf_check_fixed_rep :: Bool -- See Note [Checking for representation polymorphism] + , lf_report_unsat_syns :: Bool -- ^ See Note [Linting type synonym applications] + , lf_check_linearity :: Bool -- ^ See Note [Linting linearity] + , lf_check_fixed_rep :: Bool -- ^ See Note [Checking for representation polymorphism] + , lf_check_rubbish_lits :: Bool -- ^ See Note [Checking for rubbish literals] } -- See Note [Checking StaticPtrs] ===================================== compiler/GHC/Core/Utils.hs ===================================== @@ -2168,20 +2168,24 @@ it doesn't have the trickiness of the let-can-float invariant to worry about. -- Suppose @f x@ diverges; then @C (f x)@ is not a value. -- We check for this using needsCaseBinding below exprIsHNF :: CoreExpr -> Bool -- True => Value-lambda, constructor, PAP -exprIsHNF = exprIsHNFlike isDataConWorkId isEvaldUnfolding +exprIsHNF = exprIsHNFlike isDataConWorkId isEvaldUnfolding True -- | Similar to 'exprIsHNF' but includes CONLIKE functions as well as -- data constructors. Conlike arguments are considered interesting by the -- inliner. exprIsConLike :: CoreExpr -> Bool -- True => lambda, conlike, PAP -exprIsConLike = exprIsHNFlike isConLikeId isConLikeUnfolding +exprIsConLike = exprIsHNFlike isConLikeId isConLikeUnfolding False -- | Returns true for values or value-like expressions. These are lambdas, -- constructors / CONLIKE functions (as determined by the function argument) -- or PAPs. -- -exprIsHNFlike :: HasDebugCallStack => (Var -> Bool) -> (Unfolding -> Bool) -> CoreExpr -> Bool -exprIsHNFlike is_con is_con_unf e +exprIsHNFlike :: HasDebugCallStack => (Var -> Bool) + -> (Unfolding -> Bool) + -> Bool + -> CoreExpr -> Bool +{-# INLINE exprIsHNFlike #-} -- Specialise at its two call sites +exprIsHNFlike is_con is_con_unf rubbish_lit_result e = -- pprTraceWith "hnf" (\r -> ppr r <+> ppr e) $ is_hnf_like e where @@ -2200,9 +2204,12 @@ exprIsHNFlike is_con is_con_unf e || definitelyUnliftedType (idType v) -- Unlifted binders are always evaluated (#20140) - is_hnf_like (Lit l) = not (isLitRubbish l) + is_hnf_like (Lit lit) + | isLitRubbish lit = rubbish_lit_result + | otherwise = True -- Regarding a LitRubbish as ConLike leads to unproductive inlining in -- WWRec, see #20035 + is_hnf_like (Type _) = True -- Types are honorary Values; -- we don't mind copying them is_hnf_like (Coercion _) = True -- Same for coercions @@ -2283,7 +2290,6 @@ exprIsHNFlike is_con is_con_unf e | otherwise = Nothing -{-# INLINE exprIsHNFlike #-} {- Note [exprIsHNF Tick] ===================================== compiler/GHC/Driver/Config/Core/Lint.hs ===================================== @@ -106,26 +106,18 @@ initLintPassResultConfig dflags extra_vars pass = LintPassResultConfig { lpr_diagOpts = initDiagOpts dflags , lpr_platform = targetPlatform dflags , lpr_makeLintFlags = perPassFlags dflags pass - , lpr_showLintWarnings = showLintWarnings pass , lpr_passPpr = ppr pass , lpr_localsInScope = extra_vars } -showLintWarnings :: CoreToDo -> Bool --- Disable Lint warnings on the first simplifier pass, because --- there may be some INLINE knots still tied, which is tiresomely noisy -showLintWarnings (CoreDoSimplify cfg) - | SimplPhase InitialPhase <- sm_phase (so_mode cfg) - = False -showLintWarnings _ = True - perPassFlags :: DynFlags -> CoreToDo -> LintFlags perPassFlags dflags pass = (defaultLintFlags dflags) - { lf_check_global_ids = check_globals + { lf_check_global_ids = check_globals , lf_check_inline_loop_breakers = check_lbs - , lf_check_static_ptrs = check_static_ptrs - , lf_check_linearity = check_linearity } + , lf_check_static_ptrs = check_static_ptrs + , lf_check_linearity = check_linearity + , lf_check_rubbish_lits = check_rubbish } where -- See Note [Checking for global Ids] check_globals = case pass of @@ -137,6 +129,14 @@ perPassFlags dflags pass check_lbs = case pass of CoreDesugar -> False CoreDesugarOpt -> False + + -- Disable Lint warnings on the first simplifier pass, because + -- there may be some INLINE knots still tied, which is tiresomely noisy + CoreDoSimplify cfg + | SimplPhase InitialPhase <- sm_phase (so_mode cfg) + -> True + | otherwise + -> False _ -> True -- See Note [Checking StaticPtrs] @@ -153,6 +153,11 @@ perPassFlags dflags pass CoreDesugar -> True _ -> False) + -- See Note [Checking for rubbish literals] in GHC.Core.Lint + check_rubbish = case pass of + CorePrep -> True + _ -> False + initLintConfig :: DynFlags -> [Var] -> LintConfig initLintConfig dflags vars =LintConfig { l_diagOpts = initDiagOpts dflags @@ -168,4 +173,5 @@ defaultLintFlags dflags = LF { lf_check_global_ids = False , lf_check_linearity = gopt Opt_DoLinearCoreLinting dflags , lf_report_unsat_syns = True , lf_check_fixed_rep = True + , lf_check_rubbish_lits = True } ===================================== compiler/GHC/Driver/Config/Core/Lint/Interactive.hs ===================================== @@ -10,12 +10,9 @@ import GHC.Driver.Config.Core.Lint import GHC.Core import GHC.Core.Ppr - import GHC.Core.Lint import GHC.Core.Lint.Interactive ---import GHC.Runtime.Context - import GHC.Data.Bag import GHC.Utils.Outputable as Outputable @@ -27,7 +24,7 @@ lintInteractiveExpr what hsc_env expr | not (gopt Opt_DoCoreLinting dflags) = return () | Just err <- lintExpr (initLintConfig dflags $ interactiveInScope $ hsc_IC hsc_env) expr - = displayLintResults logger False what (pprCoreExpr expr) (emptyBag, err) + = displayLintResults logger what (pprCoreExpr expr) (emptyBag, err) | otherwise = return () where ===================================== compiler/GHC/IfaceToCore.hs ===================================== @@ -1413,7 +1413,7 @@ tcIfaceRule (IfaceRule {ifRuleName = name, ifActivation = act, ifRuleBndrs = bnd Nothing -> return () Just errs -> do logger <- getLogger - liftIO $ displayLintResults logger False doc + liftIO $ displayLintResults logger doc (pprCoreExpr rhs') (emptyBag, errs) } ; return (bndrs', args', rhs') } @@ -2006,7 +2006,7 @@ tcUnfoldingRhs is_compulsory toplvl name expr case lintUnfolding is_compulsory (initLintConfig dflags in_scope) noSrcLoc core_expr' of Nothing -> return () Just errs -> liftIO $ - displayLintResults logger False doc + displayLintResults logger doc (pprCoreExpr core_expr') (emptyBag, errs) return core_expr' where View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/a3afae0c85fa57479378a5df5bbde2e… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/a3afae0c85fa57479378a5df5bbde2e… You're receiving this email because of your account on gitlab.haskell.org.

1 0