- ghc-commits - Haskell.org

[Git][ghc/ghc][wip/stability-test-version-nos] base-exports: update version numbers
by Teo Camarasu (＠teo) 24 Aug '25

24 Aug '25

Teo Camarasu pushed to branch wip/stability-test-version-nos at Glasgow Haskell Compiler / GHC Commits: 461c2577 by Teo Camarasu at 2025-08-24T13:27:43+01:00 base-exports: update version numbers As the version of the compiler has been bumped, a lot of the embedded version numbers will need to be updated if we ever run this test with `--test-accept` so let's just update them now, and keep future diffs clean. - - - - - 4 changed files: - testsuite/tests/interface-stability/base-exports.stdout - testsuite/tests/interface-stability/base-exports.stdout-javascript-unknown-ghcjs - testsuite/tests/interface-stability/base-exports.stdout-mingw32 - testsuite/tests/interface-stability/base-exports.stdout-ws-32 The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/461c257773cc2e3fe3262a1e4520f85… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/461c257773cc2e3fe3262a1e4520f85… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/stability-test-version-nos
by Teo Camarasu (＠teo) 24 Aug '25

24 Aug '25

Teo Camarasu pushed new branch wip/stability-test-version-nos at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/stability-test-version-nos You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T26343] base: don't expose GHC.Num.{BigNat, Integer, Natural}
by Teo Camarasu (＠teo) 24 Aug '25

24 Aug '25

Teo Camarasu pushed to branch wip/T26343 at Glasgow Haskell Compiler / GHC Commits: ad48bb4b by Teo Camarasu at 2025-08-24T13:11:48+01:00 base: don't expose GHC.Num.{BigNat, Integer, Natural} WIP - - - - - 9 changed files: - libraries/base/base.cabal.in - libraries/base/src/Data/Array/Byte.hs - libraries/base/src/GHC/Num.hs - − libraries/base/src/GHC/Num/BigNat.hs - − libraries/base/src/GHC/Num/Integer.hs - − libraries/base/src/GHC/Num/Natural.hs - libraries/base/src/System/CPUTime/Utils.hs - libraries/ghc-bignum/ghc-bignum.cabal - libraries/ghc-experimental/src/GHC/TypeNats/Experimental.hs Changes: ===================================== libraries/base/base.cabal.in ===================================== @@ -219,9 +219,6 @@ Library , GHC.MVar , GHC.Natural , GHC.Num - , GHC.Num.Integer - , GHC.Num.Natural - , GHC.Num.BigNat , GHC.OldList , GHC.OverloadedLabels , GHC.Profiling ===================================== libraries/base/src/Data/Array/Byte.hs ===================================== @@ -27,7 +27,7 @@ import qualified GHC.Internal.Data.Foldable as F import GHC.Internal.Data.Maybe (fromMaybe) import Data.Semigroup import GHC.Internal.Exts -import GHC.Num.Integer (Integer(..)) +import GHC.Internal.Bignum.Integer (Integer(..)) import GHC.Internal.Show (intToDigit) import GHC.Internal.ST (ST(..), runST) import GHC.Internal.Word (Word8(..)) ===================================== libraries/base/src/GHC/Num.hs ===================================== @@ -16,11 +16,11 @@ module GHC.Num ( Num(..) , subtract , quotRemInteger - , module GHC.Num.Integer - , module GHC.Num.Natural + , module GHC.Internal.Bignum.Integer + , module GHC.Internal.Bignum.Natural ) where import GHC.Internal.Num -import GHC.Num.Integer -import GHC.Num.Natural +import GHC.Internal.Bignum.Integer +import GHC.Internal.Bignum.Natural ===================================== libraries/base/src/GHC/Num/BigNat.hs deleted ===================================== @@ -1,6 +0,0 @@ -module GHC.Num.BigNat - ( module GHC.Internal.Bignum.BigNat - ) -where - -import GHC.Internal.Bignum.BigNat ===================================== libraries/base/src/GHC/Num/Integer.hs deleted ===================================== @@ -1,6 +0,0 @@ -module GHC.Num.Integer - ( module GHC.Internal.Bignum.Integer - ) -where - -import GHC.Internal.Bignum.Integer ===================================== libraries/base/src/GHC/Num/Natural.hs deleted ===================================== @@ -1,6 +0,0 @@ -module GHC.Num.Natural - ( module GHC.Internal.Bignum.Natural - ) -where - -import GHC.Internal.Bignum.Natural ===================================== libraries/base/src/System/CPUTime/Utils.hs ===================================== @@ -8,7 +8,7 @@ module System.CPUTime.Utils ) where import GHC.Internal.Foreign.C.Types -import GHC.Num.Integer (Integer) +import GHC.Internal.Bignum.Integer (Integer) import GHC.Internal.Real (fromIntegral) cClockToInteger :: CClock -> Integer ===================================== libraries/ghc-bignum/ghc-bignum.cabal ===================================== @@ -10,10 +10,8 @@ bug-reports: https://gitlab.haskell.org/ghc/ghc/issues/new category: Numeric, Algebra, GHC build-type: Simple description: - This package used to provide the low-level implementation of the standard + This package provides the low-level implementation of the standard 'BigNat', 'Natural' and 'Integer' types. - Use `base:GHC.Num.{Integer,Natural,BigNat}` instead or other modules from - `ghc-internal`. extra-source-files: changelog.md @@ -40,13 +38,6 @@ library , GHC.Internal.Bignum.Backend as GHC.Num.Backend , GHC.Internal.Bignum.Backend.Selected as GHC.Num.Backend.Selected , GHC.Internal.Bignum.Backend.Native as GHC.Num.Backend.Native - -- reexport from base - -- We can't reexport these modules from ghc-internal otherwise we get - -- ambiguity between: - -- ghc-bignum:GHC.Num.X - -- base:GHC.Num.X - -- we should probably just deprecate ghc-bignum and encourage users to use - -- exports from base instead. - , GHC.Num.BigNat - , GHC.Num.Natural - , GHC.Num.Integer + , GHC.Internal.Bignum.BigNat as GHC.Num.BigNat + , GHC.Internal.Bignum.Natural as GHC.Num.Natural + , GHC.Internal.Bignum.Integer as GHC.Num.Integer ===================================== libraries/ghc-experimental/src/GHC/TypeNats/Experimental.hs ===================================== @@ -12,7 +12,7 @@ module GHC.TypeNats.Experimental ( ) where import GHC.Internal.TypeNats -import GHC.Num.Natural (naturalLog2) +import GHC.Internal.Bignum.Natural (naturalLog2) plusSNat :: SNat n -> SNat m -> SNat (n + m) plusSNat (UnsafeSNat n) (UnsafeSNat m) = UnsafeSNat (n + m) View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/ad48bb4b6f76923073391d8e9f56203… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/ad48bb4b6f76923073391d8e9f56203… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/T26343
by Teo Camarasu (＠teo) 24 Aug '25

24 Aug '25

Teo Camarasu pushed new branch wip/T26343 at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/T26343 You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/T26318
by Ryan Scott (＠RyanGlScott) 23 Aug '25

23 Aug '25

Ryan Scott pushed new branch wip/T26318 at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/T26318 You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/14554-wasm-fix] Deleted 1 commit: ghc-toolchain: Make "Support SMP" a query on a Toolchain.Target
by Cheng Shao (＠TerrorJack) 23 Aug '25

23 Aug '25

Cheng Shao pushed to branch wip/14554-wasm-fix at Glasgow Haskell Compiler / GHC WARNING: The push did not contain any new commits, but force pushed to delete the commits and changes below. Deleted commits: da99e918 by Rodrigo Mesquita at 2025-08-23T19:21:31+02:00 ghc-toolchain: Make "Support SMP" a query on a Toolchain.Target "Support SMP" is merely a function of target, so we can represent it as such in `ghc-toolchain`. Hadrian queries the Target using this predicate to determine how to build GHC, and GHC queries the Target similarly to report under --info whether it "Support SMP" Towards #26227 - - - - - 5 changed files: - compiler/GHC/Driver/Session.hs - hadrian/bindist/Makefile - hadrian/src/Oracles/Flag.hs - hadrian/src/Rules/Generate.hs - utils/ghc-toolchain/src/GHC/Toolchain/Target.hs Changes: ===================================== compiler/GHC/Driver/Session.hs ===================================== @@ -3547,6 +3547,7 @@ compilerInfo dflags -- If true, we require that the 'id' field in installed package info -- match what is passed to the @-this-unit-id@ flag for modules -- built in it + ("Support SMP", queryBool tgtSupportsSMP), ("Requires unified installed package IDs", "YES"), -- Whether or not we support the @-this-package-key@ flag. Prefer -- "Uses unit IDs" over it. We still say yes even if @-this-package-key@ ===================================== hadrian/bindist/Makefile ===================================== @@ -88,7 +88,6 @@ lib/settings : config.mk @echo '[("unlit command", "$$topdir/../bin/$(CrossCompilePrefix)unlit")' >> $@ @echo ',("target RTS linker only supports shared libraries", "$(TargetRTSLinkerOnlySupportsSharedLibs)")' >> $@ @echo ',("Use interpreter", "$(GhcWithInterpreter)")' >> $@ - @echo ',("Support SMP", "$(GhcWithSMP)")' >> $@ @echo ',("RTS ways", "$(GhcRTSWays)")' >> $@ @echo ',("Relative Global Package DB", "package.conf.d")' >> $@ @echo ',("base unit-id", "$(BaseUnitId)")' >> $@ ===================================== hadrian/src/Oracles/Flag.hs ===================================== @@ -121,27 +121,7 @@ targetSupportsThreadedRts = do -- | Does the target support the -N RTS flag? targetSupportsSMP :: Action Bool -targetSupportsSMP = do - unreg <- queryTargetTarget tgtUnregisterised - armVer <- targetArmVersion - goodArch <- (||) <$> - anyTargetArch [ ArchX86 - , ArchX86_64 - , ArchPPC - , ArchPPC_64 ELF_V1 - , ArchPPC_64 ELF_V2 - , ArchAArch64 - , ArchS390X - , ArchRISCV64 - , ArchLoongArch64 ] <*> isArmTarget - if -- The THREADED_RTS requires `BaseReg` to be in a register and the - -- Unregisterised mode doesn't allow that. - | unreg -> return False - -- We don't support load/store barriers pre-ARMv7. See #10433. - | Just ver <- armVer - , ver < ARMv7 -> return False - | goodArch -> return True - | otherwise -> return False +targetSupportsSMP = queryTargetTarget Toolchain.tgtSupportsSMP useLibffiForAdjustors :: Action Bool useLibffiForAdjustors = queryTargetTarget tgtUseLibffiForAdjustors ===================================== hadrian/src/Rules/Generate.hs ===================================== @@ -486,7 +486,6 @@ generateSettings settingsFile = do [ ("unlit command", ("$topdir/../bin/" <>) <$> expr (programName (ctx { Context.package = unlit }))) , ("target RTS linker only supports shared libraries", expr $ yesNo <$> targetRTSLinkerOnlySupportsSharedLibs) , ("Use interpreter", expr $ yesNo <$> ghcWithInterpreter (predStage stage)) - , ("Support SMP", expr $ yesNo <$> targetSupportsSMP) , ("RTS ways", escapeArgs . map show . Set.toList <$> getRtsWays) , ("Relative Global Package DB", pure rel_pkg_db) , ("base unit-id", pure base_unit_id) ===================================== utils/ghc-toolchain/src/GHC/Toolchain/Target.hs ===================================== @@ -1,5 +1,6 @@ {-# LANGUAGE RecordWildCards #-} {-# LANGUAGE NamedFieldPuns #-} +{-# LANGUAGE MultiWayIf #-} module GHC.Toolchain.Target ( -- * A Toolchain Target @@ -7,6 +8,9 @@ module GHC.Toolchain.Target , WordSize(..), wordSize2Bytes + -- ** Queries + , tgtSupportsSMP + -- ** Lenses , _tgtCC, _tgtCxx, _tgtCpp, _tgtHsCpp @@ -151,6 +155,34 @@ instance Show Target where , "}" ] +-------------------------------------------------------------------------------- +-- Queries +-------------------------------------------------------------------------------- + +tgtSupportsSMP :: Target -> Bool +tgtSupportsSMP Target{..} = do + let goodArch = + isARM (archOS_arch tgtArchOs) + || archOS_arch tgtArchOs `elem` + [ ArchX86 + , ArchX86_64 + , ArchPPC + , ArchPPC_64 ELF_V1 + , ArchPPC_64 ELF_V2 + , ArchAArch64 + , ArchS390X + , ArchRISCV64 + , ArchLoongArch64 ] + + if -- The THREADED_RTS requires `BaseReg` to be in a register and the + -- Unregisterised mode doesn't allow that. + | tgtUnregisterised -> False + -- We don't support load/store barriers pre-ARMv7. See #10433. + | ArchARM ver _ _ <- archOS_arch tgtArchOs + , ver < ARMv7 -> False + | goodArch -> True + | otherwise -> False + -------------------------------------------------------------------------------- -- Lenses -------------------------------------------------------------------------------- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/da99e918eda2c6e7b08520c3ec65df4… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/da99e918eda2c6e7b08520c3ec65df4… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T26255] More wibbles
by Simon Peyton Jones (＠simonpj) 23 Aug '25

23 Aug '25

Simon Peyton Jones pushed to branch wip/T26255 at Glasgow Haskell Compiler / GHC Commits: 745cae95 by Simon Peyton Jones at 2025-08-23T23:31:48+01:00 More wibbles - - - - - 3 changed files: - compiler/GHC/Tc/Errors.hs - testsuite/tests/typecheck/should_fail/T18851.hs - testsuite/tests/typecheck/should_fail/UnliftedNewtypesFamilyKindFail2.stderr Changes: ===================================== compiler/GHC/Tc/Errors.hs ===================================== @@ -443,7 +443,6 @@ ignoreConstraint :: Ct -> Bool ignoreConstraint ct = case ctOrigin ct of AssocFamPatOrigin -> True -- See (CIG1) - WantedSuperclassOrigin {} -> True -- See (CIG2) _ -> False -- | Makes an error item from a constraint, calculating whether or not @@ -540,7 +539,7 @@ reportWanteds ctxt tc_lvl wc@(WC { wc_simple = simples, wc_impl = implics ; when (null simples) $ reportMultiplicityCoercionErrs ctxt_for_insols mult_co_errs -- See Note [Suppressing confusing errors] - ; let (suppressed_items, items0) = partition suppress tidy_items + ; let (suppressed_items, items0) = partition suppressItem tidy_items ; traceTc "reportWanteds suppressed:" (ppr suppressed_items) ; (ctxt1, items1) <- tryReporters ctxt_for_insols report1 items0 @@ -548,7 +547,7 @@ reportWanteds ctxt tc_lvl wc@(WC { wc_simple = simples, wc_impl = implics -- any of the first batch failed, or if the enclosing context -- says to suppress ; let ctxt2 = ctxt1 { cec_suppress = cec_suppress ctxt || cec_suppress ctxt1 } - ; (ctxt3, leftovers) <- tryReporters ctxt2 report2 items1 + ; (_, leftovers) <- tryReporters ctxt2 report2 items1 ; massertPpr (null leftovers) (text "The following unsolved Wanted constraints \ \have not been reported to the user:" @@ -559,12 +558,15 @@ reportWanteds ctxt tc_lvl wc@(WC { wc_simple = simples, wc_impl = implics -- wanted insoluble here; but do suppress inner insolubles -- if there's a *given* insoluble here (= inaccessible code) - -- Only now, if there are no errors, do we report suppressed ones - -- See Note [Suppressing confusing errors] - -- We don't need to update the context further because of the - -- whenNoErrs guard + -- Only now, if there are no errors, do we report suppressed ones + -- See Note [Suppressing confusing errors]. We don't need to update + -- the context further because of the whenNoErrs guard ; whenNoErrs $ - do { (_, more_leftovers) <- tryReporters ctxt3 report3 suppressed_items + do { (_, more_leftovers) <- tryReporters ctxt_for_insols (report1++report2) + suppressed_items + -- ctxt_for_insols: the suppressed errors can be Int~Bool, which + -- will have made the incoming `ctxt` be True; don't make that suppress + -- the Int~Bool error! ; massertPpr (null more_leftovers) (ppr more_leftovers) } } where env = cec_tidy ctxt @@ -587,14 +589,6 @@ reportWanteds ctxt tc_lvl wc@(WC { wc_simple = simples, wc_impl = implics DE_Multiplicity mult_co loc -> (es1, es2, es3, (mult_co, loc):es4) - -- See Note [Suppressing confusing errors] - suppress :: ErrorItem -> Bool - suppress item - | Wanted <- ei_flavour item - = is_ww_fundep_item item - | otherwise - = False - -- report1: ones that should *not* be suppressed by -- an insoluble somewhere else in the tree -- It's crucial that anything that is considered insoluble @@ -656,11 +650,6 @@ reportWanteds ctxt tc_lvl wc@(WC { wc_simple = simples, wc_impl = implics , ("Dicts", is_dict, False, mkGroupReporter mkDictErr) , ("Quantified", is_qc, False, mkGroupReporter mkQCErr) ] - -- report3: suppressed errors should be reported as categorized by either report1 - -- or report2. Keep this in sync with the suppress function above - report3 = [ ("wanted/wanted fundeps", is_ww_fundep, True, mkGroupReporter mkEqErr) - ] - -- rigid_nom_eq, rigid_nom_tv_eq, is_dict, is_equality, is_ip, is_FRR, is_irred :: ErrorItem -> Pred -> Bool @@ -728,10 +717,6 @@ reportWanteds ctxt tc_lvl wc@(WC { wc_simple = simples, wc_impl = implics is_qc _ (ForAllPred {}) = True is_qc _ _ = False - -- See situation (1) of Note [Suppressing confusing errors] - is_ww_fundep item _ = is_ww_fundep_item item - is_ww_fundep_item = isWantedWantedFunDepOrigin . errorItemOrigin - given_eq_spec -- See Note [Given errors] | has_gadt_match_here = ("insoluble1a", is_given_eq, True, mkGivenErrorReporter) @@ -757,6 +742,16 @@ reportWanteds ctxt tc_lvl wc@(WC { wc_simple = simples, wc_impl = implics = has_gadt_match implics --------------- +suppressItem :: ErrorItem -> Bool + -- See Note [Suppressing confusing errors] +suppressItem item + | Wanted <- ei_flavour item + , let orig = errorItemOrigin item + = isWantedSuperclassOrigin orig -- See (SCE1) + || isWantedWantedFunDepOrigin orig -- See (SCE2) + | otherwise + = False + isSkolemTy :: TcLevel -> Type -> Bool -- The type is a skolem tyvar isSkolemTy tc_lvl ty @@ -781,7 +776,23 @@ If there are any other errors to report, at all, we want to suppress these. Which errors (only 1 case right now): -1) Errors which arise from the interaction of two Wanted fun-dep constraints. +(SCE1) Superclasses of Wanteds. These are generated on in case they trigger functional + dependencies. If such a constraint is unsolved, then its "parent" constraint must + also be unsolved, and is much more informative to the user. Example (#26255): + class (MinVersion <= F era) => Era era where { ... } + f :: forall era. EraFamily era -> IO () + f = ..blah... -- [W] Era era + Here we have simply omitted "Era era =>" from f's type. But we'll end up with + /two/ Wanted constraints: + [W] d1 : Era era + [W] d2 : MinVersion <= F era -- Superclass of d1 + We definitely want to report d1 and not d2! Happily it's easy to filter out those + superclass-Wanteds, becuase their Origin betrays them. + + See test T18851 for an example of how it is (just, barely) possible for the /only/ + errors to be superclass-of-Wanted constraints. + +(SCE2) Errors which arise from the interaction of two Wanted fun-dep constraints. Example: class C a b | a -> b where @@ -852,18 +863,8 @@ Currently, the constraints to ignore are: If there is any trouble, checkValidFamInst bleats, aborting compilation. -(CIG2) Superclasses of Wanteds. These are generated on in case they trigger functional - dependencies. If such a constraint is unsolved, then its "parent" constraint must - also be unsolved, and is much more informative to the user. Example (#26255): - class (MinVersion <= F era) => Era era where { ... } - f :: forall era. EraFamily era -> IO () - f = ..blah... -- [W] Era era - Here we have simply omitted "Era era =>" from f's type. But we'll end up with - /two/ Wanted constraints: - [W] d1 : Era era - [W] d2 : MinVersion <= F era -- Superclass of d1 - We definitely want to report d1 and not d2! Happily it's easy to filter out those - superclass-Wanteds, becuase their Origin betrays them. +(Note: Aug 25: this seems a rather tricky corner; + c.f. Note [Suppressing confusing errors]) Note [Implementation of Unsatisfiable constraints] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ===================================== testsuite/tests/typecheck/should_fail/T18851.hs ===================================== @@ -33,3 +33,14 @@ f = show (undefined :: c) -- blows up at run time once type checks g :: String g = f @A @B + +{- +[W] Show c, Num int, C int A, C int B, C int c +Superclasses + C_FD int ~ A + C_FD int ~ B + C_FD int ~ c +--> + C_FD int ~ B + B ~ A +-} ===================================== testsuite/tests/typecheck/should_fail/UnliftedNewtypesFamilyKindFail2.stderr ===================================== @@ -1,3 +1,7 @@ +UnliftedNewtypesFamilyKindFail2.hs:12:1: error: [GHC-83865] + • Expected a type, but ‘F 5’ has kind ‘5’ + • In the newtype family instance declaration for ‘F’ + UnliftedNewtypesFamilyKindFail2.hs:12:20: error: [GHC-83865] • Expected a type, but ‘5’ has kind View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/745cae95484f32660086d8abeb88bf4… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/745cae95484f32660086d8abeb88bf4… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/14554-wasm-fix] Deleted 1 commit: ghc-toolchain: Make "tgt rts linker only supports shared libs" function on Target
by Cheng Shao (＠TerrorJack) 23 Aug '25

23 Aug '25

Cheng Shao pushed to branch wip/14554-wasm-fix at Glasgow Haskell Compiler / GHC WARNING: The push did not contain any new commits, but force pushed to delete the commits and changes below. Deleted commits: f6d63a45 by Rodrigo Mesquita at 2025-08-23T19:21:31+02:00 ghc-toolchain: Make "tgt rts linker only supports shared libs" function on Target Just like with "Support SMP", "target RTS linker only supports shared libraries" is a predicate on a `Target` so we can just compute it when necessary from the given `Target`. Towards #26227 - - - - - 6 changed files: - compiler/GHC/Driver/Session.hs - compiler/GHC/Settings/IO.hs - hadrian/bindist/Makefile - hadrian/src/Oracles/Flag.hs - hadrian/src/Rules/Generate.hs - utils/ghc-toolchain/src/GHC/Toolchain/Target.hs Changes: ===================================== compiler/GHC/Driver/Session.hs ===================================== @@ -3502,6 +3502,7 @@ compilerInfo dflags ("target has libm", queryBool tgtHasLibm), ("target has .ident directive", queryBool tgtSupportsIdentDirective), ("target has subsections via symbols", queryBool tgtSupportsSubsectionsViaSymbols), + ("target RTS linker only supports shared libraries", queryBool tgtRTSLinkerOnlySupportsSharedLibs), ("Unregisterised", queryBool tgtUnregisterised), ("LLVM target", query tgtLlvmTarget), ("LLVM llc command", queryCmdMaybe id tgtLlc), ===================================== compiler/GHC/Settings/IO.hs ===================================== @@ -146,7 +146,6 @@ initSettings top_dir = do pure (ld_r_path, map Option ld_r_args) iserv_prog = libexec "ghc-iserv" - targetRTSLinkerOnlySupportsSharedLibs <- getBooleanSetting "target RTS linker only supports shared libraries" ghcWithInterpreter <- getBooleanSetting "Use interpreter" baseUnitId <- getSetting_raw "base unit-id" @@ -231,7 +230,7 @@ initSettings top_dir = do , platformMisc_ghcWithInterpreter = ghcWithInterpreter , platformMisc_libFFI = tgtUseLibffiForAdjustors target , platformMisc_llvmTarget = tgtLlvmTarget target - , platformMisc_targetRTSLinkerOnlySupportsSharedLibs = targetRTSLinkerOnlySupportsSharedLibs + , platformMisc_targetRTSLinkerOnlySupportsSharedLibs = tgtRTSLinkerOnlySupportsSharedLibs target } , sRawSettings = settingsList ===================================== hadrian/bindist/Makefile ===================================== @@ -86,7 +86,6 @@ WrapperBinsDir=${bindir} lib/settings : config.mk @rm -f $@ @echo '[("unlit command", "$$topdir/../bin/$(CrossCompilePrefix)unlit")' >> $@ - @echo ',("target RTS linker only supports shared libraries", "$(TargetRTSLinkerOnlySupportsSharedLibs)")' >> $@ @echo ',("Use interpreter", "$(GhcWithInterpreter)")' >> $@ @echo ',("RTS ways", "$(GhcRTSWays)")' >> $@ @echo ',("Relative Global Package DB", "package.conf.d")' >> $@ ===================================== hadrian/src/Oracles/Flag.hs ===================================== @@ -80,23 +80,8 @@ platformSupportsGhciObjects = do only_shared_libs <- targetRTSLinkerOnlySupportsSharedLibs pure $ has_merge_objs && not only_shared_libs --- | Does the target RTS linker only support loading shared libraries? --- If true, this has several implications: --- 1. The GHC driver must not do loadArchive/loadObj etc and must --- always do loadDLL, regardless of whether host GHC is dynamic or --- not. --- 2. The GHC driver will always enable -dynamic-too when compiling --- vanilla way with TH codegen requirement. --- 3. ghci will always enforce dynamic ways even if -dynamic or --- -dynamic-too is not explicitly passed. --- 4. Cabal must not build ghci objects since it's not supported by --- the target. --- 5. The testsuite driver will use dyn way for TH/ghci tests even --- when host GHC is static. --- 6. TH/ghci doesn't work if stage1 is built without shared libraries --- (e.g. quickest/fully_static). targetRTSLinkerOnlySupportsSharedLibs :: Action Bool -targetRTSLinkerOnlySupportsSharedLibs = anyTargetArch [ ArchWasm32 ] +targetRTSLinkerOnlySupportsSharedLibs = queryTargetTarget Toolchain.tgtRTSLinkerOnlySupportsSharedLibs arSupportsDashL :: Stage -> Action Bool arSupportsDashL stage = Toolchain.arSupportsDashL . tgtAr <$> targetStage stage ===================================== hadrian/src/Rules/Generate.hs ===================================== @@ -484,7 +484,6 @@ generateSettings settingsFile = do settings <- traverse sequence $ [ ("unlit command", ("$topdir/../bin/" <>) <$> expr (programName (ctx { Context.package = unlit }))) - , ("target RTS linker only supports shared libraries", expr $ yesNo <$> targetRTSLinkerOnlySupportsSharedLibs) , ("Use interpreter", expr $ yesNo <$> ghcWithInterpreter (predStage stage)) , ("RTS ways", escapeArgs . map show . Set.toList <$> getRtsWays) , ("Relative Global Package DB", pure rel_pkg_db) ===================================== utils/ghc-toolchain/src/GHC/Toolchain/Target.hs ===================================== @@ -9,7 +9,7 @@ module GHC.Toolchain.Target , WordSize(..), wordSize2Bytes -- ** Queries - , tgtSupportsSMP + , tgtSupportsSMP, tgtRTSLinkerOnlySupportsSharedLibs -- ** Lenses , _tgtCC, _tgtCxx, _tgtCpp, _tgtHsCpp @@ -183,6 +183,26 @@ tgtSupportsSMP Target{..} = do | goodArch -> True | otherwise -> False +-- | Does the target RTS linker only support loading shared libraries? +-- If true, this has several implications: +-- 1. The GHC driver must not do loadArchive/loadObj etc and must +-- always do loadDLL, regardless of whether host GHC is dynamic or +-- not. +-- 2. The GHC driver will always enable -dynamic-too when compiling +-- vanilla way with TH codegen requirement. +-- 3. ghci will always enforce dynamic ways even if -dynamic or +-- -dynamic-too is not explicitly passed. +-- 4. Cabal must not build ghci objects since it's not supported by +-- the target. +-- 5. The testsuite driver will use dyn way for TH/ghci tests even +-- when host GHC is static. +-- 6. TH/ghci doesn't work if stage1 is built without shared libraries +-- (e.g. quickest/fully_static). +tgtRTSLinkerOnlySupportsSharedLibs :: Target -> Bool +tgtRTSLinkerOnlySupportsSharedLibs Target{tgtArchOs} = + archOS_arch tgtArchOs `elem` + [ ArchWasm32 ] + -------------------------------------------------------------------------------- -- Lenses -------------------------------------------------------------------------------- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/f6d63a45a27479cf08774e99faba23b… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/f6d63a45a27479cf08774e99faba23b… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/14554-wasm-fix
by Cheng Shao (＠TerrorJack) 23 Aug '25

23 Aug '25

Cheng Shao pushed new branch wip/14554-wasm-fix at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/14554-wasm-fix You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Deleted branch wip/14554-5
by Cheng Shao (＠TerrorJack) 23 Aug '25

23 Aug '25

Cheng Shao deleted branch wip/14554-5 at Glasgow Haskell Compiler / GHC -- You're receiving this email because of your account on gitlab.haskell.org.

1 0