- ghc-commits - Haskell.org

[Git][ghc/ghc][master] RTS: Document -K behaviour better
by Marge Bot (＠marge-bot) 19 Jan '26

19 Jan '26

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: 3fadfefe by Andreas Klebinger at 2026-01-18T21:49:01-05:00 RTS: Document -K behaviour better - - - - - 1 changed file: - docs/users_guide/runtime_control.rst Changes: ===================================== docs/users_guide/runtime_control.rst ===================================== @@ -866,6 +866,11 @@ performance. This option is there mainly to stop the program eating up all the available memory in the machine if it gets into an infinite loop. + Note that if the process is termined through a ``StackOverflow`` exception + the reported stack usage is not representative of actual stack use. It + reports stack use during error handling, rather than stack use at the time + the exception was raised initially. + .. rts-flag:: -m ⟨n⟩ :default: 3% View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/3fadfefe3fd80fb3976a55677efec9e… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/3fadfefe3fd80fb3976a55677efec9e… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][master] hadrian: enable split sections for cross stage0
by Marge Bot (＠marge-bot) 19 Jan '26

19 Jan '26

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: 323eb8f0 by Cheng Shao at 2026-01-18T21:48:19-05:00 hadrian: enable split sections for cross stage0 This patch fixes a minor issue with `splitSectionsArgs` in hadrian: previously, it's unconditionally disabled for stage0 libraries because it's not going to be shipped in the final bindists. But it's only true when not cross compiling. So for now we also need to enable it for cross stage0 as well. - - - - - 1 changed file: - hadrian/src/Settings/Builders/SplitSections.hs Changes: ===================================== hadrian/src/Settings/Builders/SplitSections.hs ===================================== @@ -4,25 +4,26 @@ module Settings.Builders.SplitSections where import Expression import Packages import Settings +import Settings.Builders.Common import Flavour.Type -import Oracles.Setting -- | Does it make sense to enable or disable split sections? splitSectionsArgs :: Args splitSectionsArgs = do pkg <- getPackage osx <- expr isOsxTarget + cross <- expr $ flag CrossCompiling notSt0 <- notStage0 flav <- expr flavour if ( ghcSplitSections flav -- Flavour enables split-sections && not osx -- OS X doesn't support split sections - && notSt0 + && (cross || notSt0) -- Disable for stage 0 because we aren't going to ship -- the resulting binaries and consequently there is no - -- reason to minimize size. + -- reason to minimize size. Unless cross compiling. && (pkg /= ghc) -- Disable section splitting for the GHC library. -- It takes too long and there is little benefit. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/323eb8f0224b0f81ffc5e7976cb5495… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/323eb8f0224b0f81ffc5e7976cb5495… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/spj-try-opt-coercion] More optCoRefl
by Simon Peyton Jones (＠simonpj) 19 Jan '26

19 Jan '26

Simon Peyton Jones pushed to branch wip/spj-try-opt-coercion at Glasgow Haskell Compiler / GHC Commits: dce958d4 by Simon Peyton Jones at 2026-01-19T00:34:26+00:00 More optCoRefl * Deal with submultiplicities * In InitialPhase run even with empty subst - - - - - 10 changed files: - compiler/GHC/Core/Coercion/Opt.hs - compiler/GHC/Core/Opt/Pipeline.hs - compiler/GHC/Core/Opt/Pipeline/Types.hs - compiler/GHC/Core/Opt/Simplify/Env.hs - compiler/GHC/Core/Opt/Simplify/Iteration.hs - compiler/GHC/Driver/Config/Core/Opt/Simplify.hs - compiler/GHC/Driver/Flags.hs - compiler/GHC/Driver/Session.hs - compiler/GHC/Utils/Outputable.hs - compiler/GHC/Utils/Trace.hs Changes: ===================================== compiler/GHC/Core/Coercion/Opt.hs ===================================== @@ -1,8 +1,7 @@ -- (c) The University of Glasgow 2006 {-# LANGUAGE CPP #-} -module GHC.Core.Coercion.Opt( optCoProgram, optCoRefl ) -where +module GHC.Core.Coercion.Opt( optCoProgram, optCoRefl ) where import GHC.Prelude @@ -28,7 +27,6 @@ import GHC.Data.Pair import GHC.Data.TrieMap import GHC.Utils.Outputable -import GHC.Utils.Constants (debugIsOn) import GHC.Utils.Misc import GHC.Utils.Panic @@ -195,44 +193,50 @@ We use the following invariants: %* * %********************************************************************* -} -optCoProgram :: CoreProgram -> CoreProgram +optCoProgram :: Bool -- True <=> do extra checks/tracking + -> CoreProgram -> CoreProgram -- Apply optCoercion to all coercions in /expressions/ -- There may also be coercions in /types/ but we `optCoProgram` doesn't -- look at them; they are typically fewer and smaller, and it doesn't seem -- worth the cost of traversing and rebuilding all the types in the program. -optCoProgram binds +optCoProgram do_checks binds = map go binds where - go (NonRec b r) = NonRec b (optCoExpr in_scope r) - go (Rec prs) = Rec (mapSnd (optCoExpr in_scope) prs) + go (NonRec b r) = NonRec b (optCoExpr (do_checks, in_scope) r) + go (Rec prs) = Rec (mapSnd (optCoExpr (do_checks, in_scope)) prs) + in_scope = mkInScopeSetList (bindersOfBinds binds) -- Put all top-level binders into scope; it is possible to have -- forward references. See Note [Glomming] in GHC.Core.Opt.OccurAnal -optCoExpr :: InScopeSet -> CoreExpr -> CoreExpr -optCoExpr _ e@(Var {}) = e -optCoExpr _ e@(Lit {}) = e -optCoExpr _ e@(Type {}) = e +optCoExpr :: (Bool, InScopeSet) -> CoreExpr -> CoreExpr +optCoExpr !_ e@(Var {}) = e +optCoExpr _ e@(Lit {}) = e +optCoExpr _ e@(Type {}) = e optCoExpr is (App e1 e2) = App (optCoExpr is e1) (optCoExpr is e2) -optCoExpr is (Lam b e) = Lam b (optCoExpr (is `extendInScopeSet` b) e) +optCoExpr is (Lam b e) = Lam b (optCoExpr (is `add_bndr` b) e) optCoExpr is (Coercion co) = Coercion (optCo is co) optCoExpr is (Cast e co) = Cast (optCoExpr is e) (optCo is co) optCoExpr is (Tick t e) = Tick t (optCoExpr is e) -optCoExpr is (Let (NonRec b r) e) = Let (NonRec b (optCoExpr is r)) - (optCoExpr (is `extendInScopeSet` b) e) -optCoExpr is (Let (Rec prs) e) = Let (Rec (mapSnd (optCoExpr is') prs)) - (optCoExpr is' e) - where - is' = is `extendInScopeSetList` map fst prs -optCoExpr is (Case e b ty alts) = Case (optCoExpr is e) b ty - (map (optCoAlt (is `extendInScopeSet` b)) alts) +optCoExpr is (Let (NonRec b r) e) = Let (NonRec b (optCoExpr is r)) + (optCoExpr (is `add_bndr` b) e) +optCoExpr is (Let (Rec prs) e) = Let (Rec (mapSnd (optCoExpr is') prs)) + (optCoExpr is' e) + where + is' = is `add_bndrs` map fst prs +optCoExpr is (Case e b ty alts) = Case (optCoExpr is e) b ty (map do_alt alts) + where + is' = is `add_bndr` b + do_alt (Alt k bs e) = Alt k bs (optCoExpr (is' `add_bndrs` bs) e) -optCo :: InScopeSet -> Coercion -> Coercion -optCo is co = optCoercion (mkEmptySubst is) co +add_bndr :: (Bool, InScopeSet) -> Var -> (Bool, InScopeSet) +add_bndr (do_checks, is) b = (do_checks, is `extendInScopeSet` b) -optCoAlt :: InScopeSet -> CoreAlt -> CoreAlt -optCoAlt is (Alt k bs e) - = Alt k bs (optCoExpr (is `extendInScopeSetList` bs) e) +add_bndrs :: (Bool, InScopeSet) -> [Var] -> (Bool, InScopeSet) +add_bndrs (do_checks, is) bs = (do_checks, is `extendInScopeSetList` bs) + +optCo :: (Bool, InScopeSet) -> Coercion -> Coercion +optCo (do_checks, is) co = optCoercionChecking do_checks (mkEmptySubst is) co {- ********************************************************************** @@ -260,8 +264,8 @@ optCoAlt is (Alt k bs e) optCoRefl :: Bool -> Subst -> Coercion -> Coercion -- See Note [optCoRefl] optCoRefl check_stuff subst in_co - | isEmptyTCvSubst subst = in_co - | not check_stuff = opt_co_refl subst in_co + | not check_stuff + = opt_co_refl subst in_co | otherwise -- Do expensive checks = let out_co = opt_co_refl subst in_co (Pair in_l in_r) = coercionKind in_co @@ -271,22 +275,29 @@ optCoRefl check_stuff subst in_co in_co' = substCo subst in_co in_sz = coercionSize in_co' out_sz = coercionSize out_co - in if not ((in_l' `eqType` out_l) && (in_r' `eqType` out_r)) - then pprTrace "Yikes: optReflCo changes type" - (vcat [ text "in_l':" <+> ppr in_l' - , text "in_r':" <+> ppr in_r' - , text "out_l:" <+> ppr out_l - , text "out_r:" <+> ppr out_r - , text "in_co:" <+> ppr in_co - , text "out_co:" <+> ppr out_co ]) $ - out_co - else if out_sz < in_sz - then pprTrace "optCoRefl: size reduction:" - (vcat [ int in_sz <+> text "-->" <+> int out_sz - , text "in_co':" <+> ppr in_co' - , text "out_co:" <+> ppr out_co ]) $ - out_co - else out_co + + details = setPprDebug False $ + vcat [ text "in_l':" <+> ppr in_l' + , text "in_r':" <+> ppr in_r' + , text "out_l:" <+> ppr out_l + , text "out_r:" <+> ppr out_r + , text "in_co:" <+> ppr in_co + , text "out_co:" <+> ppr out_co ] + + in pprTraceWhen (not ((in_l' `eqTypeIgnoringMultiplicity` out_l) && + (in_r' `eqTypeIgnoringMultiplicity` out_r))) + "Yikes: optReflCo changes type" details $ + + pprTraceWhen (out_sz > in_sz) + "Yikes: optReflCo makes coercion bigger" + (vcat [ int in_sz <+> text "-->" <+> int out_sz + , whenPprDebug details ]) $ + + pprTraceWhen (in_sz > out_sz) + "optCoRefl: size reduction:" + (vcat [ int in_sz <+> text "-->" <+> int out_sz + , whenPprDebug details ]) + out_co opt_co_refl :: Subst -> InCoercion -> OutCoercion opt_co_refl subst co = go co @@ -308,16 +319,18 @@ opt_co_refl subst co = go co go (HoleCo h) = HoleCo $!! go_hole h go (SymCo co) = mkSymCo $!! go co go (KindCo co) = mkKindCo $!! go co - go (SubCo co) = mkSubCo $!! go co + go (SubCo co) = mkSubCo $!! (let co' = go co in + if isReflexiveCo co' && not (isReflCo co') + then pprTrace "yuike" (ppr co $$ ppr co') co' + else co') go (SelCo n co) = mkSelCo n $!! go co - go (LRCo n co) = mkLRCo n $!! go co - go (AppCo co1 co2) = mkAppCo $!! go co1 $!! go co2 - go (InstCo co1 co2) = mkInstCo $!! go co1 $!! go co2 + go (LRCo n co) = mkLRCo n $!! go co + go (AppCo co1 co2) = mkAppCo $!! go co1 $!! go co2 + go (InstCo co1 co2) = mkInstCo $!! go co1 $!! go co2 go (FunCo r afl afr com coa cor) = mkFunCo2 r afl afr $!! go com $!! go coa $!! go cor go (TyConAppCo r tc cos) = mkTyConAppCo r tc $!! go_s cos - go (UnivCo p r lt rt cos) = mkUnivCo p $!! (go_s cos) $!! r - $!! (go_ty lt) $!! (go_ty rt) + go (UnivCo p r lt rt cos) = optUnivCo p $!! go_s cos $!! r $!! go_ty lt $!! go_ty rt go (AxiomCo ax cos) = mkAxiomCo ax $!! (go_s cos) go (ForAllCo v vl vr mco co) = mkForAllCo v' vl vr @@ -354,25 +367,53 @@ opt_co_refl subst co = go co data GobbleState = GS OutCoercion (TypeMap GobbleState) -- The map is keyed by OutType +optUnivCo :: UnivCoProvenance -> [Coercion] + -> Role -> Type -> Type -> Coercion +optUnivCo prov cos role lty rty + | lty `eqTypeIgnoringMultiplicity` rty + -- We only Lint multiplicities in the output of the typechecker, as + -- described in Note [Linting linearity] in GHC.Core.Lint. This means + -- we can use 'eqTypeIgnoringMultiplicity' instead of 'eqType' below. + -- + -- In particular, this gets rid of 'SubMultProv' coercions that were + -- introduced for typechecking multiplicities of data constructors, as + -- described in Note [Typechecking data constructors] in GHC.Tc.Gen.Head. + = mkReflCo role lty + + | otherwise + = UnivCo { uco_prov = prov, uco_role = role + , uco_lty = lty, uco_rty = rty + , uco_deps = cos } + {- ********************************************************************** %* * optCoercion %* * %********************************************************************* -} -optCoercion :: Subst -> Coercion -> NormalCo +optCoercionChecking :: Bool -> Subst -> Coercion -> NormalCo -- ^ optCoercion applies a substitution to a coercion, -- *and* optimises it to reduce its size -- The substitution is a vestige of an earlier era, when the coercion optimiser -- was called by the Simplifier; now it is always empty -- But I have not removed it in case we ever want it back. -optCoercion env co - | debugIsOn - = let out_co = opt_co1 lc NotSwapped co - (Pair in_ty1 in_ty2, in_role) = coercionKindRole co +optCoercionChecking do_checks subst in_co + | not do_checks + = optCoercion1 subst in_co + + | otherwise + = let out_co = optCoercion1 subst in_co + + (Pair in_ty1 in_ty2, in_role) = coercionKindRole in_co (Pair out_ty1 out_ty2, out_role) = coercionKindRole out_co - details = vcat [ text "in_co:" <+> ppr co + in_co' = substCo subst in_co + Pair in_ty1' in_ty2' = coercionKind in_co' + + in_size = coercionSize in_co' + out_size = coercionSize out_co + + details = vcat [ text "in_co:" <+> ppr in_co , text "in_ty1:" <+> ppr in_ty1 , text "in_ty2:" <+> ppr in_ty2 , text "out_co:" <+> ppr out_co @@ -382,22 +423,34 @@ optCoercion env co , text "out_role:" <+> ppr out_role ] in - warnPprTrace (not (isReflCo out_co) && isReflexiveCo out_co) - "optCoercion: reflexive but not refl" details $ + -- Check that the type isn't changed + pprTraceWhen (not ((in_ty1' `eqTypeIgnoringMultiplicity` out_ty1) && + (in_ty2' `eqTypeIgnoringMultiplicity` out_ty2))) + "optCoercion changes type!!!" details $ + -- The coercion optimiser should usually optimise -- co:ty~ty --> Refl ty -- But given a silly `newtype N = MkN N`, the axiom has type (N ~ N), -- and so that can trigger this warning (e.g. test str002). -- Maybe we should optimise that coercion to (Refl N), but it -- just doesn't seem worth the bother - out_co + pprTraceWhen (not (isReflCo out_co) && isReflexiveCo out_co) + "optCoercion: reflexive but not refl" details $ - | otherwise - = opt_co1 lc NotSwapped co - where - lc = mkSubstLiftingContext env --- ppr_one cv = ppr cv <+> dcolon <+> ppr (coVarKind cv) + -- Show a trace if the coercion shrinks + pprTraceWhen (in_size > out_size) + "optCoercion:size reduction" + (vcat [ int in_size <+> text "-->" <+> int out_size + , whenPprDebug $ -- Show details with -dppr-debug + setPprDebug False $ + details ]) $ + out_co +optCoercion1 :: Subst -> Coercion -> NormalCo +-- Starting point for the coercion optimiser: does no checking +-- but initialises the substitution and calls opt_co1 +optCoercion1 subst co + = opt_co1 (mkSubstLiftingContext subst) NotSwapped co type NormalCo = Coercion -- Invariants: @@ -791,19 +844,7 @@ opt_univ env sym prov deps role ty1 ty2 deps' = map (opt_co1 env sym) deps (ty1'', ty2'') = swapSym sym (ty1', ty2') in - -- We only Lint multiplicities in the output of the typechecker, as - -- described in Note [Linting linearity] in GHC.Core.Lint. This means - -- we can use 'eqTypeIgnoringMultiplicity' instead of 'eqType' below. - -- - -- In particular, this gets rid of 'SubMultProv' coercions that were - -- introduced for typechecking multiplicities of data constructors, as - -- described in Note [Typechecking data constructors] in GHC.Tc.Gen.Head. - if ty1'' `eqTypeIgnoringMultiplicity` ty2'' - then mkReflCo role ty2'' - else - UnivCo { uco_prov = prov, uco_role = role - , uco_lty = ty1'', uco_rty = ty2'' - , uco_deps = deps' } + optUnivCo prov deps' role ty1'' ty2'' {- opt_univ env PhantomProv cvs _r ty1 ty2 ===================================== compiler/GHC/Core/Opt/Pipeline.hs ===================================== @@ -135,7 +135,6 @@ getCoreToDo dflags hpt_rule_base extra_vars strictness = gopt Opt_Strictness dflags full_laziness = gopt Opt_FullLaziness dflags do_specialise = gopt Opt_Specialise dflags - do_co_opt = gopt Opt_OptCoercion dflags do_float_in = gopt Opt_FloatIn dflags cse = gopt Opt_CSE dflags spec_constr = gopt Opt_SpecConstr dflags @@ -147,6 +146,8 @@ getCoreToDo dflags hpt_rule_base extra_vars ww_on = gopt Opt_WorkerWrapper dflags static_ptrs = xopt LangExt.StaticPointers dflags profiling = ways dflags `hasWay` WayProf + do_co_opt = gopt Opt_OptCoercion dflags + opt_co_checks = dopt Opt_D_opt_co dflags do_simpl3 = const_fold || rules_on -- TODO: any other optimizations benefit from three-phase simplification? @@ -230,7 +231,7 @@ getCoreToDo dflags hpt_rule_base extra_vars -- Without -O, just take what the desugarer produced -- I tried running it right after desugaring, regardless of -O, -- but that was worse (longer compile times). - runWhen do_co_opt CoreOptCoercion, + runWhen do_co_opt (CoreOptCoercion opt_co_checks), if full_laziness then CoreDoFloatOutwards $ FloatOutSwitches @@ -509,8 +510,8 @@ doCorePass pass guts = do CoreCSE -> {-# SCC "CommonSubExpr" #-} updateBinds cseProgram - CoreOptCoercion -> {-# SCC "OptCoercion" #-} - updateBinds optCoProgram + CoreOptCoercion checks -> {-# SCC "OptCoercion" #-} + updateBinds (optCoProgram checks) CoreLiberateCase -> {-# SCC "LiberateCase" #-} updateBinds (liberateCase (initLiberateCaseOpts dflags)) ===================================== compiler/GHC/Core/Opt/Pipeline/Types.hs ===================================== @@ -52,7 +52,7 @@ data CoreToDo -- These are diff core-to-core passes, | CoreDoSpecialising | CoreDoSpecConstr | CoreCSE - | CoreOptCoercion -- Run the coercion optimiser + | CoreOptCoercion Bool -- Run the coercion optimiser | CoreDoRuleCheck CompilerPhase String -- Check for non-application of rules -- matching this string | CoreDoNothing -- Useful when building up @@ -82,7 +82,7 @@ instance Outputable CoreToDo where ppr CoreDoSpecialising = text "Specialise" ppr CoreDoSpecConstr = text "SpecConstr" ppr CoreCSE = text "Common sub-expression" - ppr CoreOptCoercion = text "Optimise coercions" + ppr (CoreOptCoercion {}) = text "Optimise coercions" ppr CoreDesugar = text "Desugar (before optimization)" ppr CoreDesugarOpt = text "Desugar (after optimization)" ppr CoreTidy = text "Tidy Core" ===================================== compiler/GHC/Core/Opt/Simplify/Env.hs ===================================== @@ -284,8 +284,8 @@ data SimplMode = SimplMode -- See comments in GHC.Core.Opt.Simplify.Monad , sm_rule_opts :: !RuleOpts , sm_case_folding :: !Bool , sm_case_merge :: !Bool - , sm_opt_refl_co :: !Bool - , sm_check_opt_co :: !Bool + , sm_opt_refl_co :: !Bool -- Use `optCoRefl` on each coercion + , sm_check_opt_co :: !Bool -- Do debug-checking/tracing in `optCoRefl` } -- | See Note [SimplPhase] ===================================== compiler/GHC/Core/Opt/Simplify/Iteration.hs ===================================== @@ -1390,10 +1390,18 @@ simplCoercionF env co cont simplCoercion :: SimplEnv -> InCoercion -> SimplM OutCoercion simplCoercion env co - = do { let out_co | sm_opt_refl_co mode = optCoRefl (sm_check_opt_co mode) - (getTCvSubst env) co - | otherwise = substCo env co - ; seqCo out_co `seq` return out_co } + = do { let out_co | sm_opt_refl_co mode + , not (isEmptyTCvSubst subst) || initial_phase + = optCoRefl (sm_check_opt_co mode) subst co + | otherwise + = substCo env co + subst = getTCvSubst env + initial_phase = case sePhase env of + SimplPhase InitialPhase -> True + _ -> False + + ; seqCo out_co `seq` + return out_co } where mode = seMode env ===================================== compiler/GHC/Driver/Config/Core/Opt/Simplify.hs ===================================== @@ -68,11 +68,11 @@ initSimplMode dflags phase name = SimplMode , sm_pre_inline = gopt Opt_SimplPreInlining dflags , sm_float_enable = floatEnable dflags , sm_do_eta_reduction = gopt Opt_DoEtaReduction dflags - , sm_arity_opts = initArityOpts dflags - , sm_rule_opts = initRuleOpts dflags + , sm_arity_opts = initArityOpts dflags + , sm_rule_opts = initRuleOpts dflags , sm_case_folding = gopt Opt_CaseFolding dflags - , sm_case_merge = gopt Opt_CaseMerge dflags - , sm_opt_refl_co = gopt Opt_OptReflCoercion dflags + , sm_case_merge = gopt Opt_CaseMerge dflags + , sm_opt_refl_co = gopt Opt_OptReflCoercion dflags , sm_check_opt_co = dopt Opt_D_opt_co dflags } ===================================== compiler/GHC/Driver/Flags.hs ===================================== @@ -641,8 +641,8 @@ data GeneralFlag | Opt_InlineGenerics | Opt_InlineGenericsAggressively | Opt_StaticArgumentTransformation - | Opt_OptCoercion - | Opt_OptReflCoercion + | Opt_OptCoercion -- Run the big-hammer coercion optimiser `optCoercion` + | Opt_OptReflCoercion -- Use the cheap "refl" coercion optimiser `optCoRefl` | Opt_CSE | Opt_StgCSE | Opt_StgLiftLams ===================================== compiler/GHC/Driver/Session.hs ===================================== @@ -2021,7 +2021,6 @@ dynamic_flags_deps = [ ------ Language flags ------------------------------------------------- ++ map (mkFlag turnOn "f" setExtensionFlag ) fLangFlagsDeps - ++ map (mkFlag turnOff "fno-" unSetExtensionFlag) fLangFlagsDeps ++ map (mkFlag turnOn "X" setExtensionFlag ) xFlagsDeps ++ map (mkFlag turnOff "XNo" unSetExtensionFlag) xFlagsDeps ++ map (mkFlag turnOn "X" setLanguage ) languageFlagsDeps ===================================== compiler/GHC/Utils/Outputable.hs ===================================== @@ -104,7 +104,7 @@ module GHC.Utils.Outputable ( mkUserStyle, cmdlineParserStyle, Depth(..), withUserStyle, withErrStyle, - ifPprDebug, whenPprDebug, getPprDebug, + ifPprDebug, whenPprDebug, getPprDebug, setPprDebug, bPutHDoc ) where @@ -617,6 +617,9 @@ getPprDebug :: IsOutput doc => (Bool -> doc) -> doc {-# INLINE CONLIKE getPprDebug #-} getPprDebug d = docWithContext $ \ctx -> d (sdocPprDebug ctx) +setPprDebug :: Bool -> SDoc -> SDoc +setPprDebug dbg = updSDocContext (\cxt -> cxt { sdocPprDebug = dbg }) + -- | Says what to do with and without -dppr-debug ifPprDebug :: IsOutput doc => doc -> doc -> doc {-# INLINE CONLIKE ifPprDebug #-} ===================================== compiler/GHC/Utils/Trace.hs ===================================== @@ -1,6 +1,7 @@ -- | Tracing utilities module GHC.Utils.Trace ( pprTrace + , pprTraceWhen , pprTraceM , pprTraceDebug , pprTraceIt @@ -36,6 +37,10 @@ import GHC.Stack import Debug.Trace (trace) import Control.Monad.IO.Class +pprTraceWhen :: Bool -> String -> SDoc -> a -> a +pprTraceWhen False _ _ x = x +pprTraceWhen True str doc x = pprTrace str doc x + -- | If debug output is on, show some 'SDoc' on the screen pprTrace :: String -> SDoc -> a -> a pprTrace str doc x View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/dce958d48a5a6fb5341be5f4e2d717a… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/dce958d48a5a6fb5341be5f4e2d717a… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/romes/hadrian-cross-stage2-rebase_SVEN_FINAL] 30 commits: Improve newtype unwrapping
by Sven Tennie (＠supersven) 18 Jan '26

18 Jan '26

Sven Tennie pushed to branch wip/romes/hadrian-cross-stage2-rebase_SVEN_FINAL at Glasgow Haskell Compiler / GHC Commits: 29c0aceb by Simon Peyton Jones at 2026-01-16T17:18:11-05:00 Improve newtype unwrapping Ticket #26746 describes several relatively-minor shortcomings of newtype unwrapping. This MR addresses them, while also (arguably) simplifying the code a bit. See new Note [Solving newtype equalities: overview] and Note [Decomposing newtype equalities] and Note [Eager newtype decomposition] and Note [Even more eager newtype decomposition] For some reason, on Windows only, runtime allocations decrease for test T5205 (from 52k to 48k). I have not idea why. No change at all on Linux. I'm just going to accept the change. (I saw this same effect in another MR so I think it's a fault in the baseline.) Metric Decrease: T5205 - - - - - 8b59e62c by Andreas Klebinger at 2026-01-16T17:18:52-05:00 testsuite: Widen acceptance window for T5205. Fixes #26782 - - - - - 9e5e0234 by mangoiv at 2026-01-17T06:03:03-05:00 add a new issue template for getting verified To reduce spam created by new users, we will in future not grant any rights but reporting issues to new users. That is why we will have to be able to verify them. The added issue template serves that purpose. - - - - - b18b2c42 by Cheng Shao at 2026-01-17T06:03:44-05:00 llvm: fix split sections for llvm backend This patch fixes split sections for llvm backend: - Pass missing `--data-sections`/`--function-sections` flags to llc/opt. - Use `@llvm.compiler.used` instead of `@llvm.used` to avoid sections being unnecessarily retained at link-time. Fixes #26770. ------------------------- Metric Decrease: libdir size_hello_artifact size_hello_unicode ------------------------- Co-authored-by: Codex <codex(a)openai.com> - - - - - ebf66f67 by Cheng Shao at 2026-01-17T13:16:50-05:00 Update autoconf scripts Scripts taken from autoconf a2287c3041a3f2a204eb942e09c015eab00dc7dd - - - - - 598624b9 by Andreas Klebinger at 2026-01-17T13:17:32-05:00 CString.hs: Update incorrect comment. Fixes #26322 - - - - - eea2036b by Cheng Shao at 2026-01-18T10:00:49-05:00 libraries: bump haskeline submodule to 0.8.4.1 This patch bumps the haskeline submodule to 0.8.4.1 which includes an important fix for an ANSI handling bug on Windows (https://github.com/haskell/haskeline/pull/126) - - - - - 87d8f6c2 by Cheng Shao at 2026-01-18T10:01:30-05:00 hadrian: replace default -H32m/-H64m with -O64M to improve mutator productivity Most hadrian build flavours pass `-H32m`/`-H64m` to GHC as conventional wisdom to improve mutator productivity and reduce GC overhead. They were inherited from the legacy Make build system, and there used to be make flags to instrument a build process with `-Rghc-timing` option to collect GC stats of each GHC run from stderr. It's time to revisit whether there are better defaults for `-H32m`/`-H64m`, and this patch changes it to `-O64M` which indeed improves mutator productivity based on real statistics. `-O64M` is more aggressive than `-H64m`; it allows the old generation to grow to at least 64M before triggering major GC and reduces major GC runs. The stats of a clean build with `validate` flavour and `-H64m`: ``` h64m.log matched RTS stat lines: 5499 sum MUT cpu : 2400.808 s sum GC cpu : 1378.292 s sum MUT elapsed : 2788.253 s sum GC elapsed : 1389.233 s GC/MUT cpu ratio : 0.574 (GC is 57.4% of MUT) GC/MUT elapsed ratio : 0.498 (GC is 49.8% of MUT) GC fraction of (MUT+GC) cpu : 36.5% GC fraction of (MUT+GC) elapsed : 33.3% per-line GC/MUT cpu ratio: median 0.691, p90 1.777 per-line GC/MUT elapsed ratio: median 0.519, p90 1.081 ``` The stats of a clean build with `validate` flavour and `-O64M`: ``` o64m.log matched RTS stat lines: 5499 sum MUT cpu : 2377.383 s sum GC cpu : 1127.146 s sum MUT elapsed : 2758.857 s sum GC elapsed : 1135.587 s GC/MUT cpu ratio : 0.474 (GC is 47.4% of MUT) GC/MUT elapsed ratio : 0.412 (GC is 41.2% of MUT) GC fraction of (MUT+GC) cpu : 32.2% GC fraction of (MUT+GC) elapsed : 29.2% per-line GC/MUT cpu ratio: median 0.489, p90 1.099 per-line GC/MUT elapsed ratio: median 0.367, p90 0.806 ``` Mutator time is roughly in the same ballpark, but GC CPU time has reduced by 18.22%, and mutator productivity has increased from 63.5% to 67.8%. - - - - - 8372e13d by Cheng Shao at 2026-01-18T10:02:12-05:00 rts: remove unused .def files from rts/win32 This patch removes unused .def files from `rts/win32`, given we don't build .dll files for rts/ghc-internal/ghc-prim at all. Even when we resurrect win32 dll support at some point in the future, these .def files still contain incorrect symbols anyway and won't be of any use. - - - - - f6af485d by Cheng Shao at 2026-01-18T10:03:19-05:00 .gitmodules: use gitlab mirror for the libffi-clib submodule This patch fixes .gitmodules to use the gitlab mirror for the libffi-clib submodule, to make it coherent with other submodules that allow ghc developers to experiment with wip branches in submodules for ghc patches. Fixes #26783. - - - - - 41432d25 by Cheng Shao at 2026-01-18T10:05:13-05:00 hadrian: remove the horrible i386 speedHack When hadrian builds certain rts objects for i386, there's a horrible speedHack that forces -fno-PIC even for dynamic ways of those objects. This is not compatible with newer versions of gcc/binutils as well as clang/lld, and this patch removes it. Fixes #26792. - - - - - 2ff7ba49 by Matthew Pickering at 2026-01-18T19:06:53+01:00 Add missing req_interp modifier to T18441fail3 and T18441fail19 These tests require the interpreter but they were failing in a different way with the javascript backend because the interpreter was disabled and stderr is ignored by the test. - - - - - ecc69c0c by Matthew Pickering at 2026-01-18T19:06:53+01:00 Use explicit syntax rather than pure - - - - - 990d07f0 by Matthew Pickering at 2026-01-18T19:06:53+01:00 packaging: correctly propagate build/host/target to bindist configure script At the moment the host and target which we will produce a compiler for is fixed at the initial configure time. Therefore we need to persist the choice made at this time into the installation bindist as well so we look for the right tools, with the right prefixes at install time. In the future, we want to provide a bit more control about what kind of bindist we produce so the logic about what the host/target will have to be written by hadrian rather than persisted by the configure script. In particular with cross compilers we want to either build a normal stage 2 cross bindist or a stage 3 bindist, which creates a bindist which has a native compiler for the target platform. Fixes #21970 - - - - - 1fe43aae by Matthew Pickering at 2026-01-18T19:06:53+01:00 hadrian: Fill in more of the default.host toolchain file When you are building a cross compiler this file will be used to build stage1 and it's libraries, so we need enough information here to work accurately. There is still more work to be done (see for example, word size is still fixed). - - - - - 3a92728f by Matthew Pickering at 2026-01-18T19:06:53+01:00 hadrian: Disable docs when cross compiling Before there were a variety of ad-hoc places where doc building was disabled when cross compiling. * Some CI jobs sets --docs=none in gen_ci.hs * Some CI jobs set --docs=none in .gitlab/ci.sh * There was some logic in hadrian to not need the ["docs"] target when making a bindist. Now the situation is simple: * If you are cross compiling then defaultDocsTargets is empty by default. In theory, there is no reason why we can't build documentation for cross compiler bindists, but this is left to future work to generalise the documentation building rules to allow this (#24289) - - - - - 21b4d783 by Matthew Pickering at 2026-01-18T19:16:19+01:00 hadrian: Build stage 2 cross compilers * Most of hadrian is abstracted over the stage in order to remove the assumption that the target of all stages is the same platform. This allows the RTS to be built for two different targets for example. * Abstracts the bindist creation logic to allow building either normal or cross bindists. Normal bindists use stage 1 libraries and a stage 2 compiler. Cross bindists use stage 2 libararies and a stage 2 compiler. * hadrian: Make binary-dist-dir the default build target. This allows us to have the logic in one place about which libraries/stages to build with cross compilers. Fixes #24192 New hadrian target: * `binary-dist-dir-cross`: Build a cross compiler bindist (compiler = stage 1, libraries = stage 2) This commit also contains various changes to make stage2 compilers feasible. ------------------------- Metric Decrease: ManyAlternatives MultiComponentModulesRecomp MultiLayerModulesRecomp T10421 T12425 T12707 T13035 T13379 T15703 T16577 T18698a T18698b T18923 T1969 T21839c T3294 T4801 T5030 T5321Fun T5642 T783 T9198 T9872d T9961 parsing001 T5321FD T6048 ------------------------- Co-authored-by: Sven Tennie <sven.tennie(a)gmail.com> - - - - - bc5de15c by Sven Tennie at 2026-01-18T19:16:19+01:00 Align CI scripts with master Fixup - - - - - d1491a63 by Matthew Pickering at 2026-01-18T19:16:19+01:00 ci: Test cross bindists We remove the special logic for testing in-tree cross compilers and instead test cross compiler bindists, like we do for all other platforms. - - - - - a2cbcd99 by Matthew Pickering at 2026-01-18T19:16:19+01:00 ci: Introduce CROSS_STAGE variable In preparation for building and testing stage3 bindists we introduce the CROSS_STAGE variable which is used by a CI job to determine what kind of bindist the CI job should produce. At the moment we are only using CROSS_STAGE=2 but in the future we will have some jobs which set CROSS_STAGE=3 to produce native bindists for a target, but produced by a cross compiler, which can be tested on by another CI job on the native platform. CROSS_STAGE=2: Build a normal cross compiler bindist CROSS_STAGE=3: Build a stage 3 bindist, one which is a native compiler and library for the target - - - - - 001b1d84 by Matthew Pickering at 2026-01-18T19:16:19+01:00 hadrian: Refactor system-cxx-std-lib rules0 I noticed a few things wrong with the hadrian rules for `system-cxx-std-lib` rules. * For `text` there is an ad-hoc check to depend on `system-cxx-std-lib` outside of `configurePackage`. * The `system-cxx-std-lib` dependency is not read from cabal files. * Recache is not called on the packge database after the `.conf` file is generated, a more natural place for this rule is `registerRules`. Treating this uniformly like other packages is complicated by it not having any source code or a cabal file. However we can do a bit better by reporting the dependency firstly in `PackageData` and then needing the `.conf` file in the same place as every other package in `configurePackage`. Fixes #25303 - - - - - 0c82f141 by Sven Tennie at 2026-01-18T19:16:20+01:00 ci: Increase timeout for emulators Test runs with emulators naturally take longer than on native machines. Generate jobs.yml - - - - - cef6cbd9 by Sven Tennie at 2026-01-18T19:16:20+01:00 ghc: Distinguish between having an interpreter and having an internal one Otherwise, we fail with warnings when compiling tools. Actually, these are related but different things: - ghc can run an interpreter (either internal or external) - ghc is compiled with an internal interpreter - - - - - 31b3d03f by Matthew Pickering at 2026-01-18T19:16:20+01:00 ci: Javascript don't set CROSS_EMULATOR There is no CROSS_EMULATOR needed to run javascript binaries, so we don't set the CROSS_EMULATOR to some dummy value. - - - - - 67984106 by Sven Tennie at 2026-01-18T19:16:20+01:00 Javascript skip T23697 See #22355 about how HSC2HS and the Javascript target don't play well together. - - - - - 6b4bf022 by Sven Tennie at 2026-01-18T19:16:20+01:00 Mark T24602 as fragile It was skipped before (due to CROSS_EMULATOR being set, which changed for JS), so we don't make things worse by marking it as fragile. - - - - - 2a8e33f0 by Sven Tennie at 2026-01-18T19:16:20+01:00 Windows needs NM_STAGE0 as well The stage0 always needs nm. - - - - - f0902dfe by Sven Tennie at 2026-01-18T19:16:20+01:00 T17912 sometimes works for windows-validate This seems to be timing related. However, just simply increasing the timeout (sleep) statement of this test didn't help. Maybe, it has been flaky on CI before. - - - - - f8f82145 by Sven Tennie at 2026-01-18T19:16:20+01:00 Fix rebase: settings-use-distro-mingw is now staged - - - - - 11f70549 by Sven Tennie at 2026-01-18T19:16:20+01:00 TextFile should contain no logic It's very confusing to query for one file and then get another... - - - - - 103 changed files: - .gitlab/ci.sh - .gitlab/generate-ci/gen_ci.hs - + .gitlab/issue_templates/get-verified.md - .gitlab/jobs.yaml - .gitmodules - compiler/GHC/CmmToLlvm.hs - compiler/GHC/CmmToLlvm/Base.hs - compiler/GHC/Core/FamInstEnv.hs - compiler/GHC/Driver/Pipeline/Execute.hs - compiler/GHC/Tc/Instance/Family.hs - compiler/GHC/Tc/Solver/Equality.hs - compiler/GHC/Tc/Solver/Monad.hs - config.guess - config.sub - configure.ac - distrib/configure.ac.in - ghc/GHC/Driver/Session/Mode.hs - ghc/GHCi/UI.hs - ghc/Main.hs - ghc/ghc-bin.cabal.in - hadrian/README.md - hadrian/bindist/config.mk.in - hadrian/cfg/default.host.target.in - + hadrian/cfg/system.config.host.in - hadrian/cfg/system.config.in - + hadrian/cfg/system.config.target.in - hadrian/doc/flavours.md - hadrian/hadrian.cabal - hadrian/src/Base.hs - + hadrian/src/BindistConfig.hs - hadrian/src/Builder.hs - hadrian/src/Context.hs - hadrian/src/Expression.hs - hadrian/src/Flavour.hs - hadrian/src/Flavour/Type.hs - hadrian/src/Hadrian/Builder.hs - hadrian/src/Hadrian/Haskell/Cabal/Parse.hs - hadrian/src/Hadrian/Haskell/Cabal/Type.hs - hadrian/src/Hadrian/Haskell/Hash.hs - hadrian/src/Hadrian/Oracles/TextFile.hs - hadrian/src/Oracles/Flag.hs - hadrian/src/Oracles/Flavour.hs - hadrian/src/Oracles/Setting.hs - hadrian/src/Oracles/TestSettings.hs - hadrian/src/Packages.hs - hadrian/src/Rules.hs - hadrian/src/Rules/BinaryDist.hs - hadrian/src/Rules/CabalReinstall.hs - hadrian/src/Rules/Compile.hs - hadrian/src/Rules/Documentation.hs - hadrian/src/Rules/Generate.hs - hadrian/src/Rules/Gmp.hs - + hadrian/src/Rules/Libffi.hs - hadrian/src/Rules/Library.hs - hadrian/src/Rules/Program.hs - hadrian/src/Rules/Register.hs - hadrian/src/Rules/Test.hs - hadrian/src/Settings.hs - hadrian/src/Settings/Builders/Cabal.hs - hadrian/src/Settings/Builders/Common.hs - hadrian/src/Settings/Builders/Configure.hs - hadrian/src/Settings/Builders/DeriveConstants.hs - hadrian/src/Settings/Builders/Ghc.hs - hadrian/src/Settings/Builders/Hsc2Hs.hs - hadrian/src/Settings/Builders/RunTest.hs - hadrian/src/Settings/Builders/SplitSections.hs - hadrian/src/Settings/Default.hs - hadrian/src/Settings/Flavours/Benchmark.hs - hadrian/src/Settings/Flavours/Development.hs - hadrian/src/Settings/Flavours/GhcInGhci.hs - hadrian/src/Settings/Flavours/Performance.hs - hadrian/src/Settings/Flavours/Quick.hs - hadrian/src/Settings/Flavours/QuickCross.hs - hadrian/src/Settings/Flavours/Quickest.hs - hadrian/src/Settings/Flavours/Validate.hs - hadrian/src/Settings/Packages.hs - hadrian/src/Settings/Program.hs - hadrian/src/Settings/Warnings.hs - libraries/base/tests/IO/all.T - libraries/base/tests/all.T - libraries/ghc-internal/src/GHC/Internal/CString.hs - libraries/haskeline - m4/fp_find_nm.m4 - m4/fptools_set_platform_vars.m4 - m4/prep_target_file.m4 - − rts/win32/libHSffi.def - − rts/win32/libHSghc-internal.def - − rts/win32/libHSghc-prim.def - testsuite/ghc-config/ghc-config.hs - testsuite/tests/deriving/should_fail/T8984.stderr - testsuite/tests/deriving/should_fail/deriving-via-fail.stderr - testsuite/tests/deriving/should_fail/deriving-via-fail4.stderr - testsuite/tests/deriving/should_fail/deriving-via-fail5.stderr - testsuite/tests/ghc-e/should_fail/all.T - testsuite/tests/javascript/closure/all.T - testsuite/tests/perf/should_run/all.T - + testsuite/tests/typecheck/should_compile/T26746.hs - testsuite/tests/typecheck/should_compile/all.T - testsuite/tests/typecheck/should_fail/T15801.stderr - testsuite/tests/typecheck/should_fail/T22924b.stderr - testsuite/tests/typecheck/should_fail/TcCoercibleFail.hs - testsuite/tests/typecheck/should_fail/TcCoercibleFail.stderr - testsuite/tests/typecheck/should_fail/all.T The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/3b6f72a743140a41d932ab83c238ce… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/3b6f72a743140a41d932ab83c238ce… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T24464] Wibble
by Simon Peyton Jones (＠simonpj) 18 Jan '26

18 Jan '26

Simon Peyton Jones pushed to branch wip/T24464 at Glasgow Haskell Compiler / GHC Commits: 7cf2b4e3 by Simon Peyton Jones at 2026-01-18T18:15:37+00:00 Wibble - - - - - 2 changed files: - compiler/GHC/Tc/Gen/Sig.hs - compiler/GHC/Tc/Utils/Env.hs Changes: ===================================== compiler/GHC/Tc/Gen/Sig.hs ===================================== @@ -163,6 +163,9 @@ errors were dealt with by the renamer. ********************************************************************* -} tcTySigs :: [LSig GhcRn] -> TcM ([TcId], TcSigFun) +-- The returned [TcId] are the ones for which we have +-- a /complete/ type signature. +-- See Note [Complete and partial type signatures] tcTySigs hs_sigs = checkNoErrs $ do { -- Fail if any of the signatures is duff @@ -172,9 +175,6 @@ tcTySigs hs_sigs ; let ty_sigs = concat ty_sigs_s poly_ids = mapMaybe completeSigPolyId_maybe ty_sigs - -- The returned [TcId] are the ones for which we have - -- a complete type signature. - -- See Note [Complete and partial type signatures] env = mkNameEnv [(tcSigInfoName sig, sig) | sig <- ty_sigs] ; return (poly_ids, lookupNameEnv env) } ===================================== compiler/GHC/Tc/Utils/Env.hs ===================================== @@ -723,7 +723,7 @@ tcExtendLetEnv top_lvl sig_fn closed ids thing_inside [ (id_nm, ATcId { tct_id = id, tct_info = LetBound closed }) | Scaled _ id <- ids , let id_nm = idName id - , isNothing (sig_fn id_nm) -- See (ELE) above + , not (hasCompleteSig sig_fn id_nm) -- See (ELE) above ] $ foldr check_one_usg thing_inside ids where View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/7cf2b4e35e7db5fde1b9d5bb0e21cb5… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/7cf2b4e35e7db5fde1b9d5bb0e21cb5… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/romes/hadrian-cross-stage2-rebase_SVEN_FINAL] 4 commits: Windows needs NM_STAGE0 as well
by Sven Tennie (＠supersven) 18 Jan '26

18 Jan '26

Sven Tennie pushed to branch wip/romes/hadrian-cross-stage2-rebase_SVEN_FINAL at Glasgow Haskell Compiler / GHC Commits: b788c97c by Sven Tennie at 2026-01-18T17:52:31+01:00 Windows needs NM_STAGE0 as well The stage0 always needs nm. - - - - - e88a442d by Sven Tennie at 2026-01-18T17:52:31+01:00 T17912 sometimes works for windows-validate This seems to be timing related. However, just simply increasing the timeout (sleep) statement of this test didn't help. Maybe, it has been flaky on CI before. - - - - - 408de4a2 by Sven Tennie at 2026-01-18T17:52:31+01:00 Fix rebase: settings-use-distro-mingw is now staged - - - - - 3b6f72a7 by Sven Tennie at 2026-01-18T18:19:29+01:00 TextFile should contain no logic It's very confusing to query for one file and then get another... - - - - - 5 changed files: - hadrian/src/Builder.hs - hadrian/src/Hadrian/Oracles/TextFile.hs - hadrian/src/Oracles/Setting.hs - libraries/base/tests/IO/all.T - m4/fp_find_nm.m4 Changes: ===================================== hadrian/src/Builder.hs ===================================== @@ -228,10 +228,10 @@ instance H.Builder Builder where -- changes (#18001). _bootGhcVersion <- setting GhcVersion pure [] - Ghc _ _ -> do + Ghc _ stage -> do root <- buildRoot unlitPath <- builderPath Unlit - distro_mingw <- lookupSystemConfig "settings-use-distro-mingw" + distro_mingw <- lookupStageBuildConfig "settings-use-distro-mingw" stage return $ [ unlitPath ] ++ [ root -/- mingwStamp | windowsHost, distro_mingw == "NO" ] ===================================== hadrian/src/Hadrian/Oracles/TextFile.hs ===================================== @@ -125,18 +125,7 @@ getBuildTarget = getTargetConfig buildTargetFile -- | Get the host target configuration through 'getTargetConfig' getHostTarget :: Action Toolchain.Target -getHostTarget = do - -- MP: If we are not cross compiling then we should use the target file in order to - -- build things for the host, in particular we want to use the configured values for the - -- target for building the RTS (ie are we using Libffi for adjustors, and the wordsize) - -- TODO: Use "flag CrossCompiling" - ht <- getTargetConfig hostTargetFile - tt <- getTargetConfig targetTargetFile - if (Toolchain.targetPlatformTriple ht) == (Toolchain.targetPlatformTriple tt) - then return tt - else return ht - -- where - -- msg = "The host's target configuration file " ++ quote hostTargetFile ++ " does not exist! ghc-toolchain might have failed to generate it." +getHostTarget = getTargetConfig hostTargetFile -- | Get the target target configuration through 'getTargetConfig' getTargetTarget :: Action Toolchain.Target ===================================== hadrian/src/Oracles/Setting.hs ===================================== @@ -236,7 +236,17 @@ libsuf st way -- For example, we want to build RTS with stage1 for the host target as we -- produce a host executable with stage1 (which cross-compiles to stage2). targetStage :: Stage -> Action Target -targetStage stage | isHostStage stage = getHostTarget +targetStage Stage0 {} = getHostTarget +targetStage stage | isHostStage stage = do + -- MP: If we are not cross compiling then we should use the target file in order to + -- build things for the host, in particular we want to use the configured values for the + -- target for building the RTS (ie are we using Libffi for adjustors, and the wordsize) + -- TODO: Use "flag CrossCompiling" + ht <- getHostTarget + tt <- getTargetTarget + if targetPlatformTriple ht == targetPlatformTriple tt + then return tt + else return ht targetStage _ = getTargetTarget isHostStage :: Stage -> Bool ===================================== libraries/base/tests/IO/all.T ===================================== @@ -182,7 +182,7 @@ test('T17414', compile_and_run, ['']) test('T17510', expect_broken(17510), compile_and_run, ['']) test('bytestringread001', extra_run_opts('test.data'), compile_and_run, ['']) -test('T17912', [only_ways(['threaded1']), when(opsys('mingw32'),expect_broken(1))], compile_and_run, ['']) +test('T17912', [only_ways(['threaded1']), when(opsys('mingw32'),fragile(1))], compile_and_run, ['']) test('T18832', only_ways(['threaded1']), compile_and_run, ['']) test('mkdirExists', [exit_code(1), when(opsys('mingw32'), ignore_stderr)], compile_and_run, ['']) ===================================== m4/fp_find_nm.m4 ===================================== @@ -12,6 +12,7 @@ AC_DEFUN([FP_FIND_NM], if test "$HostOS" = "mingw32" then NmCmd=$(cygpath -m "$NM") + NM_STAGE0=$NmCmd else NmCmd="$NM" fi View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/565e3e7c82a265704335dd6ac355d6… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/565e3e7c82a265704335dd6ac355d6… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/26737] 10 commits: add a new issue template for getting verified
by Cheng Shao (＠TerrorJack) 18 Jan '26

18 Jan '26

Cheng Shao pushed to branch wip/26737 at Glasgow Haskell Compiler / GHC Commits: 9e5e0234 by mangoiv at 2026-01-17T06:03:03-05:00 add a new issue template for getting verified To reduce spam created by new users, we will in future not grant any rights but reporting issues to new users. That is why we will have to be able to verify them. The added issue template serves that purpose. - - - - - b18b2c42 by Cheng Shao at 2026-01-17T06:03:44-05:00 llvm: fix split sections for llvm backend This patch fixes split sections for llvm backend: - Pass missing `--data-sections`/`--function-sections` flags to llc/opt. - Use `@llvm.compiler.used` instead of `@llvm.used` to avoid sections being unnecessarily retained at link-time. Fixes #26770. ------------------------- Metric Decrease: libdir size_hello_artifact size_hello_unicode ------------------------- Co-authored-by: Codex <codex(a)openai.com> - - - - - ebf66f67 by Cheng Shao at 2026-01-17T13:16:50-05:00 Update autoconf scripts Scripts taken from autoconf a2287c3041a3f2a204eb942e09c015eab00dc7dd - - - - - 598624b9 by Andreas Klebinger at 2026-01-17T13:17:32-05:00 CString.hs: Update incorrect comment. Fixes #26322 - - - - - eea2036b by Cheng Shao at 2026-01-18T10:00:49-05:00 libraries: bump haskeline submodule to 0.8.4.1 This patch bumps the haskeline submodule to 0.8.4.1 which includes an important fix for an ANSI handling bug on Windows (https://github.com/haskell/haskeline/pull/126) - - - - - 87d8f6c2 by Cheng Shao at 2026-01-18T10:01:30-05:00 hadrian: replace default -H32m/-H64m with -O64M to improve mutator productivity Most hadrian build flavours pass `-H32m`/`-H64m` to GHC as conventional wisdom to improve mutator productivity and reduce GC overhead. They were inherited from the legacy Make build system, and there used to be make flags to instrument a build process with `-Rghc-timing` option to collect GC stats of each GHC run from stderr. It's time to revisit whether there are better defaults for `-H32m`/`-H64m`, and this patch changes it to `-O64M` which indeed improves mutator productivity based on real statistics. `-O64M` is more aggressive than `-H64m`; it allows the old generation to grow to at least 64M before triggering major GC and reduces major GC runs. The stats of a clean build with `validate` flavour and `-H64m`: ``` h64m.log matched RTS stat lines: 5499 sum MUT cpu : 2400.808 s sum GC cpu : 1378.292 s sum MUT elapsed : 2788.253 s sum GC elapsed : 1389.233 s GC/MUT cpu ratio : 0.574 (GC is 57.4% of MUT) GC/MUT elapsed ratio : 0.498 (GC is 49.8% of MUT) GC fraction of (MUT+GC) cpu : 36.5% GC fraction of (MUT+GC) elapsed : 33.3% per-line GC/MUT cpu ratio: median 0.691, p90 1.777 per-line GC/MUT elapsed ratio: median 0.519, p90 1.081 ``` The stats of a clean build with `validate` flavour and `-O64M`: ``` o64m.log matched RTS stat lines: 5499 sum MUT cpu : 2377.383 s sum GC cpu : 1127.146 s sum MUT elapsed : 2758.857 s sum GC elapsed : 1135.587 s GC/MUT cpu ratio : 0.474 (GC is 47.4% of MUT) GC/MUT elapsed ratio : 0.412 (GC is 41.2% of MUT) GC fraction of (MUT+GC) cpu : 32.2% GC fraction of (MUT+GC) elapsed : 29.2% per-line GC/MUT cpu ratio: median 0.489, p90 1.099 per-line GC/MUT elapsed ratio: median 0.367, p90 0.806 ``` Mutator time is roughly in the same ballpark, but GC CPU time has reduced by 18.22%, and mutator productivity has increased from 63.5% to 67.8%. - - - - - 8372e13d by Cheng Shao at 2026-01-18T10:02:12-05:00 rts: remove unused .def files from rts/win32 This patch removes unused .def files from `rts/win32`, given we don't build .dll files for rts/ghc-internal/ghc-prim at all. Even when we resurrect win32 dll support at some point in the future, these .def files still contain incorrect symbols anyway and won't be of any use. - - - - - f6af485d by Cheng Shao at 2026-01-18T10:03:19-05:00 .gitmodules: use gitlab mirror for the libffi-clib submodule This patch fixes .gitmodules to use the gitlab mirror for the libffi-clib submodule, to make it coherent with other submodules that allow ghc developers to experiment with wip branches in submodules for ghc patches. Fixes #26783. - - - - - 41432d25 by Cheng Shao at 2026-01-18T10:05:13-05:00 hadrian: remove the horrible i386 speedHack When hadrian builds certain rts objects for i386, there's a horrible speedHack that forces -fno-PIC even for dynamic ways of those objects. This is not compatible with newer versions of gcc/binutils as well as clang/lld, and this patch removes it. Fixes #26792. - - - - - 708ad5ca by Simon Peyton Jones at 2026-01-18T17:49:11+01:00 Make the implicit-parameter class have representational role This MR addresses #26737, by making the built-in class IP have a representational role for its second parameter. See Note [IP: implicit parameter class] in ghc-internal:GHC.Internal.Classes.IP In fact, IP is (unfortunately, currently) exposed by base:GHC.Base, so we ran a quick CLC proposal to agree the change: https://github.com/haskell/core-libraries-committee/issues/385 Some (small) compilations get faster because they only need to load (small) interface file GHC.Internal.Classes.IP.hi, rather than (large) GHC.Internal.Classes.hi. Metric Decrease: T10421 T12150 T12425 T24582 T5837 - - - - - 37 changed files: - + .gitlab/issue_templates/get-verified.md - .gitmodules - compiler/GHC/Builtin/Names.hs - compiler/GHC/CmmToLlvm.hs - compiler/GHC/CmmToLlvm/Base.hs - compiler/GHC/Driver/Pipeline/Execute.hs - config.guess - config.sub - docs/users_guide/9.16.1-notes.rst - hadrian/doc/flavours.md - hadrian/src/Settings/Default.hs - hadrian/src/Settings/Flavours/Benchmark.hs - hadrian/src/Settings/Flavours/Development.hs - hadrian/src/Settings/Flavours/GhcInGhci.hs - hadrian/src/Settings/Flavours/Performance.hs - hadrian/src/Settings/Flavours/Quick.hs - hadrian/src/Settings/Flavours/QuickCross.hs - hadrian/src/Settings/Flavours/Quickest.hs - hadrian/src/Settings/Flavours/Validate.hs - hadrian/src/Settings/Packages.hs - libraries/ghc-internal/ghc-internal.cabal.in - libraries/ghc-internal/src/GHC/Internal/CString.hs - libraries/ghc-internal/src/GHC/Internal/Classes.hs - + libraries/ghc-internal/src/GHC/Internal/Classes/IP.hs - libraries/haskeline - − rts/win32/libHSffi.def - − rts/win32/libHSghc-internal.def - − rts/win32/libHSghc-prim.def - 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 - testsuite/tests/interface-stability/ghc-prim-exports.stdout - testsuite/tests/interface-stability/ghc-prim-exports.stdout-mingw32 - testsuite/tests/th/TH_implicitParams.stdout - + testsuite/tests/typecheck/should_compile/T26737.hs - testsuite/tests/typecheck/should_compile/all.T The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/22fb0ced78a19b3e23ab6eb9f1192f… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/22fb0ced78a19b3e23ab6eb9f1192f… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/marge_bot_batch_merge_job] 7 commits: libraries: bump haskeline submodule to 0.8.4.1
by Marge Bot (＠marge-bot) 18 Jan '26

18 Jan '26

Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC Commits: eea2036b by Cheng Shao at 2026-01-18T10:00:49-05:00 libraries: bump haskeline submodule to 0.8.4.1 This patch bumps the haskeline submodule to 0.8.4.1 which includes an important fix for an ANSI handling bug on Windows (https://github.com/haskell/haskeline/pull/126) - - - - - 87d8f6c2 by Cheng Shao at 2026-01-18T10:01:30-05:00 hadrian: replace default -H32m/-H64m with -O64M to improve mutator productivity Most hadrian build flavours pass `-H32m`/`-H64m` to GHC as conventional wisdom to improve mutator productivity and reduce GC overhead. They were inherited from the legacy Make build system, and there used to be make flags to instrument a build process with `-Rghc-timing` option to collect GC stats of each GHC run from stderr. It's time to revisit whether there are better defaults for `-H32m`/`-H64m`, and this patch changes it to `-O64M` which indeed improves mutator productivity based on real statistics. `-O64M` is more aggressive than `-H64m`; it allows the old generation to grow to at least 64M before triggering major GC and reduces major GC runs. The stats of a clean build with `validate` flavour and `-H64m`: ``` h64m.log matched RTS stat lines: 5499 sum MUT cpu : 2400.808 s sum GC cpu : 1378.292 s sum MUT elapsed : 2788.253 s sum GC elapsed : 1389.233 s GC/MUT cpu ratio : 0.574 (GC is 57.4% of MUT) GC/MUT elapsed ratio : 0.498 (GC is 49.8% of MUT) GC fraction of (MUT+GC) cpu : 36.5% GC fraction of (MUT+GC) elapsed : 33.3% per-line GC/MUT cpu ratio: median 0.691, p90 1.777 per-line GC/MUT elapsed ratio: median 0.519, p90 1.081 ``` The stats of a clean build with `validate` flavour and `-O64M`: ``` o64m.log matched RTS stat lines: 5499 sum MUT cpu : 2377.383 s sum GC cpu : 1127.146 s sum MUT elapsed : 2758.857 s sum GC elapsed : 1135.587 s GC/MUT cpu ratio : 0.474 (GC is 47.4% of MUT) GC/MUT elapsed ratio : 0.412 (GC is 41.2% of MUT) GC fraction of (MUT+GC) cpu : 32.2% GC fraction of (MUT+GC) elapsed : 29.2% per-line GC/MUT cpu ratio: median 0.489, p90 1.099 per-line GC/MUT elapsed ratio: median 0.367, p90 0.806 ``` Mutator time is roughly in the same ballpark, but GC CPU time has reduced by 18.22%, and mutator productivity has increased from 63.5% to 67.8%. - - - - - 8372e13d by Cheng Shao at 2026-01-18T10:02:12-05:00 rts: remove unused .def files from rts/win32 This patch removes unused .def files from `rts/win32`, given we don't build .dll files for rts/ghc-internal/ghc-prim at all. Even when we resurrect win32 dll support at some point in the future, these .def files still contain incorrect symbols anyway and won't be of any use. - - - - - f6af485d by Cheng Shao at 2026-01-18T10:03:19-05:00 .gitmodules: use gitlab mirror for the libffi-clib submodule This patch fixes .gitmodules to use the gitlab mirror for the libffi-clib submodule, to make it coherent with other submodules that allow ghc developers to experiment with wip branches in submodules for ghc patches. Fixes #26783. - - - - - 41432d25 by Cheng Shao at 2026-01-18T10:05:13-05:00 hadrian: remove the horrible i386 speedHack When hadrian builds certain rts objects for i386, there's a horrible speedHack that forces -fno-PIC even for dynamic ways of those objects. This is not compatible with newer versions of gcc/binutils as well as clang/lld, and this patch removes it. Fixes #26792. - - - - - 06a24790 by Cheng Shao at 2026-01-18T10:37:45-05:00 hadrian: enable split sections for cross stage0 This patch fixes a minor issue with `splitSectionsArgs` in hadrian: previously, it's unconditionally disabled for stage0 libraries because it's not going to be shipped in the final bindists. But it's only true when not cross compiling. So for now we also need to enable it for cross stage0 as well. - - - - - 1a25b097 by Andreas Klebinger at 2026-01-18T10:37:46-05:00 RTS: Document -K behaviour better - - - - - 18 changed files: - .gitmodules - docs/users_guide/runtime_control.rst - hadrian/doc/flavours.md - hadrian/src/Settings/Builders/SplitSections.hs - hadrian/src/Settings/Default.hs - hadrian/src/Settings/Flavours/Benchmark.hs - hadrian/src/Settings/Flavours/Development.hs - hadrian/src/Settings/Flavours/GhcInGhci.hs - hadrian/src/Settings/Flavours/Performance.hs - hadrian/src/Settings/Flavours/Quick.hs - hadrian/src/Settings/Flavours/QuickCross.hs - hadrian/src/Settings/Flavours/Quickest.hs - hadrian/src/Settings/Flavours/Validate.hs - hadrian/src/Settings/Packages.hs - libraries/haskeline - − rts/win32/libHSffi.def - − rts/win32/libHSghc-internal.def - − rts/win32/libHSghc-prim.def Changes: ===================================== .gitmodules ===================================== @@ -122,4 +122,4 @@ url = https://gitlab.haskell.org/ghc/template-haskell-quasiquoter.git [submodule "libraries/libffi-clib"] path = libraries/libffi-clib - url = https://github.com/stable-haskell/libffi-clib.git + url = https://gitlab.haskell.org/ghc/libffi-clib.git ===================================== docs/users_guide/runtime_control.rst ===================================== @@ -866,6 +866,11 @@ performance. This option is there mainly to stop the program eating up all the available memory in the machine if it gets into an infinite loop. + Note that if the process is termined through a ``StackOverflow`` exception + the reported stack usage is not representative of actual stack use. It + reports stack use during error handling, rather than stack use at the time + the exception was raised initially. + .. rts-flag:: -m ⟨n⟩ :default: 3% ===================================== hadrian/doc/flavours.md ===================================== @@ -37,8 +37,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>default </td> <td> </td> - <td>-O -H32m </td> - <td>-O2 -H32m</td> + <td>-O +RTS -O64M -RTS </td> + <td>-O2 +RTS -O64M -RTS</td> <td></td> <td>-haddock</td> <td></td> @@ -49,8 +49,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>quick</td> <td> </td> - <td>-O0 -H64m</td> - <td>-O0 -H64m</td> + <td>-O0 +RTS -O64M -RTS</td> + <td>-O0 +RTS -O64M -RTS</td> <td></td> <td>-O</td> <td>-O2</td> @@ -61,8 +61,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>quick-validate</td> <td></td> - <td>-O0 -H64m -Werror</td> - <td>-O0 -H64m -Werror</td> + <td>-O0 +RTS -O64M -RTS -Werror</td> + <td>-O0 +RTS -O64M -RTS -Werror</td> <td></td> <td>-O</td> <td>-O2</td> @@ -73,8 +73,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>quick-debug</td> <td></td> - <td>-O0 -H64m</td> - <td>-O0 -H64m</td> + <td>-O0 +RTS -O64M -RTS</td> + <td>-O0 +RTS -O64M -RTS</td> <td></td> <td>-O</td> <td>-O2</td> @@ -85,8 +85,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>quickest</td> <td></td> - <td>-O0 -H64m</td> - <td>-O0 -H64m</td> + <td>-O0 +RTS -O64M -RTS</td> + <td>-O0 +RTS -O64M -RTS</td> <td></td> <td></td> <td>-O</td> @@ -97,8 +97,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>perf</td> <td> Yes (on supported platforms) </td> - <td>-O -H64m</td> - <td>-O -H64m</td> + <td>-O +RTS -O64M -RTS</td> + <td>-O +RTS -O64M -RTS</td> <td></td> <td>-O2</td> <td>-O2</td> @@ -109,8 +109,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>release (same as perf with -haddock and +no_self_recomp+hash_unit_ids)</td> <td></td> - <td>-O -H64m</td> - <td>-O -H64m</td> + <td>-O +RTS -O64M -RTS</td> + <td>-O +RTS -O64M -RTS</td> <td></td> <td>-O2</td> <td>-O2</td> @@ -121,8 +121,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>bench</td> <td></td> - <td>-O -H64m</td> - <td>-O -H64m</td> + <td>-O +RTS -O64M -RTS</td> + <td>-O +RTS -O64M -RTS</td> <td></td> <td>-O2</td> <td>-O2</td> @@ -133,8 +133,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>devel1</td> <td></td> - <td>-O -H64m</td> - <td>-O -H64m</td> + <td>-O +RTS -O64M -RTS</td> + <td>-O +RTS -O64M -RTS</td> <td></td> <td>-dcore-lint</td> <td>-O0 -DDEBUG</td> @@ -145,8 +145,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>devel2</td> <td></td> - <td>-O -H64m</td> - <td>-O -H64m</td> + <td>-O +RTS -O64M -RTS</td> + <td>-O +RTS -O64M -RTS</td> <td></td> <td>-dcore-lint</td> <td>-O2</td> @@ -157,7 +157,7 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>validate</td> <td></td> - <td>-O0 -H64m</td> + <td>-O0 +RTS -O64M -RTS</td> <td>-fllvm-fill-undef-with-garbage -fcheck-prim-bounds</td> <td></td> <td>-O -dcore-lint -dno-debug-output</td> @@ -169,7 +169,7 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>slow-validate</td> <td></td> - <td>-O0 -H64m</td> + <td>-O0 +RTS -O64M -RTS</td> <td>-fllvm-fill-undef-with-garbage</td> <td></td> <td>-O -dcore-lint -dno-debug-output</td> @@ -181,8 +181,8 @@ when compiling the `compiler` library, and `hsGhc` when compiling/linking the GH <tr> <th>static</td> <td></td> - <td>-O -H64m -fPIC -static</td> - <td>-O -H64m -fPIC -static</td> + <td>-O +RTS -O64M -RTS -fPIC -static</td> + <td>-O +RTS -O64M -RTS -fPIC -static</td> <td></td> <td>-O2</td> <td>-O2</td> ===================================== hadrian/src/Settings/Builders/SplitSections.hs ===================================== @@ -4,25 +4,26 @@ module Settings.Builders.SplitSections where import Expression import Packages import Settings +import Settings.Builders.Common import Flavour.Type -import Oracles.Setting -- | Does it make sense to enable or disable split sections? splitSectionsArgs :: Args splitSectionsArgs = do pkg <- getPackage osx <- expr isOsxTarget + cross <- expr $ flag CrossCompiling notSt0 <- notStage0 flav <- expr flavour if ( ghcSplitSections flav -- Flavour enables split-sections && not osx -- OS X doesn't support split sections - && notSt0 + && (cross || notSt0) -- Disable for stage 0 because we aren't going to ship -- the resulting binaries and consequently there is no - -- reason to minimize size. + -- reason to minimize size. Unless cross compiling. && (pkg /= ghc) -- Disable section splitting for the GHC library. -- It takes too long and there is little benefit. ===================================== hadrian/src/Settings/Default.hs ===================================== @@ -274,7 +274,7 @@ defaultSourceArgs :: SourceArgs defaultSourceArgs = SourceArgs { hsDefault = mconcat [ stage0 ? arg "-O" , notStage0 ? arg "-O2" - , arg "-H32m" ] + , pure ["+RTS", "-O64M", "-RTS"] ] , hsLibrary = notStage0 ? arg "-haddock" , hsCompiler = mempty , hsGhc = mempty } ===================================== hadrian/src/Settings/Flavours/Benchmark.hs ===================================== @@ -16,7 +16,7 @@ benchmarkFlavour = defaultFlavour benchmarkArgs :: Args benchmarkArgs = sourceArgs SourceArgs - { hsDefault = pure ["-O", "-H64m"] + { hsDefault = pure ["-O", "+RTS", "-O64M", "-RTS"] , hsLibrary = pure ["-O2"] -- TODO: We should really pass -O2 when notStage0. Otherwise, we aren't -- really measuring the overhead of a potential new optimisation we want ===================================== hadrian/src/Settings/Flavours/Development.hs ===================================== @@ -26,7 +26,7 @@ developmentFlavour ghcStage = defaultFlavour developmentArgs :: Stage -> Args developmentArgs ghcStage = sourceArgs SourceArgs - { hsDefault = mconcat [ pure ["-O", "-H64m"], + { hsDefault = mconcat [ pure ["-O", "+RTS", "-O64M", "-RTS"], -- Disable optimization when building Cabal; -- this saves many minutes of build time. package cabal ? pure ["-O0"]] ===================================== hadrian/src/Settings/Flavours/GhcInGhci.hs ===================================== @@ -22,7 +22,7 @@ ghcInGhciFlavour = defaultFlavour ghciArgs :: Args ghciArgs = sourceArgs SourceArgs { hsDefault = mconcat $ - [ pure ["-O0", "-H64m"] + [ pure ["-O0", "+RTS", "-O64M", "-RTS"] ] , hsLibrary = mempty , hsCompiler = mempty ===================================== hadrian/src/Settings/Flavours/Performance.hs ===================================== @@ -12,7 +12,7 @@ performanceFlavour = splitSections $ enableLateCCS $ defaultFlavour performanceArgs :: Args performanceArgs = sourceArgs SourceArgs - { hsDefault = pure ["-O", "-H64m"] + { hsDefault = pure ["-O", "+RTS", "-O64M", "-RTS"] , hsLibrary = orM [notStage0, cross] ? arg "-O2" , hsCompiler = pure ["-O2"] , hsGhc = mconcat ===================================== hadrian/src/Settings/Flavours/Quick.hs ===================================== @@ -34,7 +34,7 @@ quickFlavour = defaultFlavour quickArgs :: Args quickArgs = sourceArgs SourceArgs - { hsDefault = mconcat [ pure ["-O0", "-H64m"] ] + { hsDefault = mconcat [ pure ["-O0", "+RTS", "-O64M", "-RTS"] ] , hsLibrary = notStage0 ? arg "-O" , hsCompiler = stage0 ? arg "-O2" , hsGhc = stage0 ? arg "-O" } ===================================== hadrian/src/Settings/Flavours/QuickCross.hs ===================================== @@ -32,7 +32,7 @@ quickCrossFlavour = defaultFlavour quickCrossArgs :: Args quickCrossArgs = sourceArgs SourceArgs { hsDefault = mconcat $ - [ pure ["-O0", "-H64m"] + [ pure ["-O0", "+RTS", "-O64M", "-RTS"] ] , hsLibrary = notStage0 ? mconcat [ arg "-O", arg "-fllvm" ] , hsCompiler = stage0 ? arg "-O2" ===================================== hadrian/src/Settings/Flavours/Quickest.hs ===================================== @@ -19,7 +19,7 @@ quickestFlavour = defaultFlavour quickestArgs :: Args quickestArgs = sourceArgs SourceArgs { hsDefault = mconcat $ - [ pure ["-O0", "-H64m"] + [ pure ["-O0", "+RTS", "-O64M", "-RTS"] ] , hsLibrary = mempty , hsCompiler = stage0 ? arg "-O" ===================================== hadrian/src/Settings/Flavours/Validate.hs ===================================== @@ -16,11 +16,12 @@ validateFlavour = enableLinting $ quickValidateFlavour validateArgs :: Args validateArgs = sourceArgs SourceArgs - { hsDefault = mconcat [ stage0 ? pure ["-O0", "-H64m"] + { hsDefault = mconcat [ stage0 ? pure ["-O0"] -- See #11487 , notStage0 ? arg "-fllvm-fill-undef-with-garbage" , notStage0 ? arg "-dno-debug-output" , notStage0 ? arg "-fcheck-prim-bounds" + , pure ["+RTS", "-O64M", "-RTS"] ] , hsLibrary = pure ["-O"] , hsCompiler = mconcat [ stage0 ? pure ["-O2"] @@ -37,7 +38,7 @@ slowValidateFlavour = validateFlavour quickValidateArgs :: Args quickValidateArgs = sourceArgs SourceArgs - { hsDefault = mempty + { hsDefault = pure ["+RTS", "-O64M", "-RTS"] , hsLibrary = pure [ "-O" ] , hsCompiler = mconcat [ stage0 ? arg "-O2", notStage0 ? arg "-O"] , hsGhc = pure [ "-O", "-hide-all-packages" ] ===================================== hadrian/src/Settings/Packages.hs ===================================== @@ -370,21 +370,6 @@ rtsPackageArgs = package rts ? do , inputs ["**/Evac.c", "**/Evac_thr.c"] ? arg "-funroll-loops" - , speedHack ? - inputs [ "**/Evac.c", "**/Evac_thr.c" - , "**/Scav.c", "**/Scav_thr.c" - , "**/Compact.c", "**/GC.c" ] ? arg "-fno-PIC" - -- @-static@ is necessary for these bits, as otherwise the NCG - -- generates dynamic references. - , speedHack ? - inputs [ "**/Updates.c", "**/StgMiscClosures.c" - , "**/Jumps_D.c", "**/Jumps_V16.c", "**/Jumps_V32.c", "**/Jumps_V64.c" - , "**/PrimOps.c", "**/Apply.c" - , "**/AutoApply.c" - , "**/AutoApply_V16.c" - , "**/AutoApply_V32.c" - , "**/AutoApply_V64.c" ] ? pure ["-fno-PIC", "-static"] - -- See Note [AutoApply.cmm for vectors] in genapply/Main.hs , inputs ["**/AutoApply_V32.c"] ? pure [ "-mavx2" | x86 ] , inputs ["**/AutoApply_V64.c"] ? pure [ "-mavx512f" | x86 ] @@ -452,42 +437,6 @@ rtsPackageArgs = package rts ? do , builder HsCpp ? useLibdw ? arg "-DUSE_LIBDW" ] --- Compile various performance-critical pieces *without* -fPIC -dynamic --- even when building a shared library. If we don't do this, then the --- GC runs about 50% slower on x86 due to the overheads of PIC. The --- cost of doing this is a little runtime linking and less sharing, but --- not much. --- --- On x86_64 this doesn't work, because all objects in a shared library --- must be compiled with -fPIC (since the 32-bit relocations generated --- by the default small memory can't be resolved at runtime). So we --- only do this on i386. --- --- This apparently doesn't work on OS X (Darwin) nor on Solaris. --- On Darwin we get errors of the form --- --- ld: absolute addressing (perhaps -mdynamic-no-pic) used in _stg_ap_0_fast --- from rts/dist-install/build/Apply.dyn_o not allowed in slidable image --- --- and lots of these warnings: --- --- ld: warning codegen in _stg_ap_pppv_fast (offset 0x0000005E) prevents image --- from loading in dyld shared cache --- --- On Solaris we get errors like: --- --- Text relocation remains referenced --- against symbol offset in file --- .rodata (section) 0x11 rts/dist-install/build/Apply.dyn_o --- ... --- ld: fatal: relocations remain against allocatable but non-writable sections --- collect2: ld returned 1 exit status -speedHack :: Action Bool -speedHack = do - i386 <- anyTargetArch [ArchX86] - goodOS <- not <$> anyTargetOs [OSSolaris2] - return $ i386 && goodOS - -- See @rts/ghc.mk@. rtsWarnings :: Args rtsWarnings = mconcat ===================================== libraries/haskeline ===================================== @@ -1 +1 @@ -Subproject commit 991953cd5d3bb9e8057de4a0d8f2cae3455865d8 +Subproject commit d0d8902f88fde549f76771feeb40f16941a22a9f ===================================== rts/win32/libHSffi.def deleted ===================================== @@ -1,19 +0,0 @@ - -LIBRARY "libHSffi-ghc@ProjectVersion@.dll" - -EXPORTS - - ffi_prep_cif - ffi_call - ffi_type_void - ffi_type_float - ffi_type_double - ffi_type_sint64 - ffi_type_uint64 - ffi_type_sint32 - ffi_type_uint32 - ffi_type_sint16 - ffi_type_uint16 - ffi_type_sint8 - ffi_type_uint8 - ffi_type_pointer ===================================== rts/win32/libHSghc-internal.def deleted ===================================== @@ -1,49 +0,0 @@ - -LIBRARY "libHSghc-internal-@LibVersion@-ghc@ProjectVersion@.dll" - -EXPORTS - ghczminternal_GHCziInternalziInt_I8zh_con_info - ghczminternal_GHCziInternalziInt_I16zh_con_info - ghczminternal_GHCziInternalziInt_I32zh_con_info - ghczminternal_GHCziInternalziInt_I64zh_con_info - - ghczminternal_GHCziInternalziWord_W8zh_con_info - ghczminternal_GHCziInternalziWord_W16zh_con_info - ghczminternal_GHCziInternalziWord_W32zh_con_info - ghczminternal_GHCziInternalziWord_W64zh_con_info - - ghczminternal_GHCziInternalziStable_StablePtr_con_info - - ghczminternal_GHCziInternalziPack_unpackCString_closure - - ghczminternal_GHCziInternalziTopHandler_runIO_closure - ghczminternal_GHCziInternalziTopHandler_runNonIO_closure - - ghczminternal_GHCziInternalziIOziException_stackOverflow_closure - ghczminternal_GHCziInternalziIOziException_heapOverflow_closure - - ghczminternal_GHCziInternalziPtr_Ptr_con_info - ghczminternal_GHCziInternalziPtr_FunPtr_con_info - - ghczminternal_GHCziInternalziConcziIO_ensureIOManagerIsRunning_closure - ghczminternal_GHCziInternalziConcziIO_interruptIOManager_closure - ghczminternal_GHCziInternalziConcziIO_ioManagerCapabilitiesChanged_closure - ghczminternal_GHCziInternalziConcziSync_runSparks_closure - ghczminternal_GHCziInternalziEventziWindows_processRemoteCompletion_closure - - ghczminternal_GHCziInternalziTopHandler_flushStdHandles_closure - - ghczminternal_GHCziInternalziWeakziFinalizze_runFinalizzerBatch_closure - ghczminternal_GHCziInternalziPack_unpackCString_closure - ghczminternal_GHCziInternalziIOziException_blockedIndefinitelyOnMVar_closure - ghczminternal_GHCziInternalziIOziException_blockedIndefinitelyOnSTM_closure - ghczminternal_GHCziInternalziIOziException_allocationLimitExceeded_closure - ghczminternal_GHCziInternalziIOziException_stackOverflow_closure - ghczminternal_GHCziInternalziIOziException_cannotCompactFunction_closure - ghczminternal_GHCziInternalziIOziException_cannotCompactPinned_closure - ghczminternal_GHCziInternalziIOziException_cannotCompactMutable_closure - ghczminternal_GHCziInternalziControlziExceptionziBase_nonTermination_closure - ghczminternal_GHCziInternalziControlziExceptionziBase_nestedAtomically_closure - ghczminternal_GHCziInternalziExceptionziType_divZZeroException_closure - ghczminternal_GHCziInternalziExceptionziType_underflowException_closure - ghczminternal_GHCziInternalziExceptionziType_overflowException_closure ===================================== rts/win32/libHSghc-prim.def deleted ===================================== @@ -1,14 +0,0 @@ - -LIBRARY "libHSghc-internal-@LibVersion@-ghc@ProjectVersion@.dll" - -EXPORTS - - ghczminternal_GHCziInternalziTypes_True_closure - ghczminternal_GHCziInternalziTypes_False_closure - ghczminternal_GHCziInternalziTypes_Czh_con_info - ghczminternal_GHCziInternalziTypes_Izh_con_info - ghczminternal_GHCziInternalziTypes_Fzh_con_info - ghczminternal_GHCziInternalziTypes_Dzh_con_info - ghczminternal_GHCziInternalziTypes_Wzh_con_info - ghczminternal_GHCziInternalziTypes_Czh_static_info - ghczminternal_GHCziInternalziTypes_Izh_static_info View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/e052213f4c06b8c1fbb7815aa4ad1f… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/e052213f4c06b8c1fbb7815aa4ad1f… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][master] hadrian: remove the horrible i386 speedHack
by Marge Bot (＠marge-bot) 18 Jan '26

18 Jan '26

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: 41432d25 by Cheng Shao at 2026-01-18T10:05:13-05:00 hadrian: remove the horrible i386 speedHack When hadrian builds certain rts objects for i386, there's a horrible speedHack that forces -fno-PIC even for dynamic ways of those objects. This is not compatible with newer versions of gcc/binutils as well as clang/lld, and this patch removes it. Fixes #26792. - - - - - 1 changed file: - hadrian/src/Settings/Packages.hs Changes: ===================================== hadrian/src/Settings/Packages.hs ===================================== @@ -370,21 +370,6 @@ rtsPackageArgs = package rts ? do , inputs ["**/Evac.c", "**/Evac_thr.c"] ? arg "-funroll-loops" - , speedHack ? - inputs [ "**/Evac.c", "**/Evac_thr.c" - , "**/Scav.c", "**/Scav_thr.c" - , "**/Compact.c", "**/GC.c" ] ? arg "-fno-PIC" - -- @-static@ is necessary for these bits, as otherwise the NCG - -- generates dynamic references. - , speedHack ? - inputs [ "**/Updates.c", "**/StgMiscClosures.c" - , "**/Jumps_D.c", "**/Jumps_V16.c", "**/Jumps_V32.c", "**/Jumps_V64.c" - , "**/PrimOps.c", "**/Apply.c" - , "**/AutoApply.c" - , "**/AutoApply_V16.c" - , "**/AutoApply_V32.c" - , "**/AutoApply_V64.c" ] ? pure ["-fno-PIC", "-static"] - -- See Note [AutoApply.cmm for vectors] in genapply/Main.hs , inputs ["**/AutoApply_V32.c"] ? pure [ "-mavx2" | x86 ] , inputs ["**/AutoApply_V64.c"] ? pure [ "-mavx512f" | x86 ] @@ -452,42 +437,6 @@ rtsPackageArgs = package rts ? do , builder HsCpp ? useLibdw ? arg "-DUSE_LIBDW" ] --- Compile various performance-critical pieces *without* -fPIC -dynamic --- even when building a shared library. If we don't do this, then the --- GC runs about 50% slower on x86 due to the overheads of PIC. The --- cost of doing this is a little runtime linking and less sharing, but --- not much. --- --- On x86_64 this doesn't work, because all objects in a shared library --- must be compiled with -fPIC (since the 32-bit relocations generated --- by the default small memory can't be resolved at runtime). So we --- only do this on i386. --- --- This apparently doesn't work on OS X (Darwin) nor on Solaris. --- On Darwin we get errors of the form --- --- ld: absolute addressing (perhaps -mdynamic-no-pic) used in _stg_ap_0_fast --- from rts/dist-install/build/Apply.dyn_o not allowed in slidable image --- --- and lots of these warnings: --- --- ld: warning codegen in _stg_ap_pppv_fast (offset 0x0000005E) prevents image --- from loading in dyld shared cache --- --- On Solaris we get errors like: --- --- Text relocation remains referenced --- against symbol offset in file --- .rodata (section) 0x11 rts/dist-install/build/Apply.dyn_o --- ... --- ld: fatal: relocations remain against allocatable but non-writable sections --- collect2: ld returned 1 exit status -speedHack :: Action Bool -speedHack = do - i386 <- anyTargetArch [ArchX86] - goodOS <- not <$> anyTargetOs [OSSolaris2] - return $ i386 && goodOS - -- See @rts/ghc.mk@. rtsWarnings :: Args rtsWarnings = mconcat View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/41432d25b7a66fa36a0e3def429fd98… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/41432d25b7a66fa36a0e3def429fd98… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][master] .gitmodules: use gitlab mirror for the libffi-clib submodule
by Marge Bot (＠marge-bot) 18 Jan '26

18 Jan '26

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: f6af485d by Cheng Shao at 2026-01-18T10:03:19-05:00 .gitmodules: use gitlab mirror for the libffi-clib submodule This patch fixes .gitmodules to use the gitlab mirror for the libffi-clib submodule, to make it coherent with other submodules that allow ghc developers to experiment with wip branches in submodules for ghc patches. Fixes #26783. - - - - - 1 changed file: - .gitmodules Changes: ===================================== .gitmodules ===================================== @@ -122,4 +122,4 @@ url = https://gitlab.haskell.org/ghc/template-haskell-quasiquoter.git [submodule "libraries/libffi-clib"] path = libraries/libffi-clib - url = https://github.com/stable-haskell/libffi-clib.git + url = https://gitlab.haskell.org/ghc/libffi-clib.git View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/f6af485d2583c4e629d8516c32e73bc… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/f6af485d2583c4e629d8516c32e73bc… You're receiving this email because of your account on gitlab.haskell.org.

1 0