- ghc-commits - Haskell.org

[Git][ghc/ghc][wip/T26746] 10 commits: Don't re-use stack slots for growing registers
by Simon Peyton Jones (＠simonpj) 14 Jan '26

14 Jan '26

Simon Peyton Jones pushed to branch wip/T26746 at Glasgow Haskell Compiler / GHC Commits: 023c301c by sheaf at 2026-01-13T04:57:30-05:00 Don't re-use stack slots for growing registers This commit avoids re-using a stack slot for a register that has grown but already had a stack slot. For example, suppose we have stack slot assigments %v1 :: FF64 |-> StackSlot 0 %v2 :: FF64 |-> StackSlot 1 Later, we start using %v1 at a larger format (e.g. F64x2) and we need to spill it again. Then we **must not** use StackSlot 0, as a spill at format F64x2 would clobber the data in StackSlot 1. This can cause some fragmentation of the `StackMap`, but that's probably OK. Fixes #26668 - - - - - d0966e64 by fendor at 2026-01-13T04:58:11-05:00 Remove `traceId` from ghc-pkg executable - - - - - 20d7efec by Simon Peyton Jones at 2026-01-13T12:41:22-05:00 Make SpecContr rules fire a bit later See #26615 and Note [SpecConstr rule activation] - - - - - 8bc4eb8c by Andrew Lelechenko at 2026-01-13T12:42:03-05:00 Upgrade mtl submodule to 2.3.2 Fixes #26656 - - - - - c94aaacd by Cheng Shao at 2026-01-13T12:42:44-05:00 compiler: remove iserv and only use on-demand external interpreter This patch removes `iserv` from the tree completely. Hadrian would no longer build or distribute `iserv`, and the GHC driver would use the on-demand external interpreter by default when invoked with `-fexternal-interpreter`, without needing to specify `-pgmi ""`. This has multiple benefits: - It allows cleanup of a lot of legacy hacks in the hadrian codebase. - It paves the way for running cross ghc's iserv via cross emulator (#25523), fixing TH/ghci support for cross targets other than wasm/js. - - - - - c1fe0097 by Peter Trommler at 2026-01-14T03:54:49-05:00 PPC NCG: Fix shift right MO code The shift amount in shift right [arithmetic] MOs is machine word width. Therefore remove unnecessary zero- or sign-extending of shift amount. It looks harmless to extend the shift amount argument because the shift right instruction uses only the seven lowest bits (i. e. mod 128). But now we have a conversion operation from a smaller type to word width around a memory load at word width. The types are not matching up but there is no check done in CodeGen. The necessary conversion from word width down to the smaller width would be translated into a no-op on PowerPC anyway. So all seems harmless if it was not for a small optimisation in getRegister'. In getRegister' a load instruction with the smaller width of the conversion operation was generated. This loaded the most significant bits of the word in memory on a big-endian platform. These bits were zero and hence shift right was used with shift amount zero and not one as required in test Sized. Fixes #26519 - - - - - 2dafc65a by Cheng Shao at 2026-01-14T03:55:31-05:00 Tree-wide cleanup of cygwin logic GHC has not supported cygwin for quite a few years already, and will not resume support in the forseeable future. The only supported windows toolchain is clang64/clangarm64 of the msys2 project. This patch cleans up the unused cygwin logic in the tree. Co-authored-by: Codex <codex(a)openai.com> - - - - - 66b96e2a by Teo Camarasu at 2026-01-14T03:56:13-05:00 Set default eventlog-flush-interval to 5s Resolves #26707 - - - - - d0254579 by Andrew Lelechenko at 2026-01-14T03:56:53-05:00 Document when -maxN RTS option was added - - - - - 4a01ec30 by Simon Peyton Jones at 2026-01-14T10:22:12+00:00 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] - - - - - 82 changed files: - CODEOWNERS - cabal.project-reinstall - compiler/GHC/CmmToAsm/PPC/CodeGen.hs - compiler/GHC/CmmToAsm/Reg/Linear.hs - compiler/GHC/CmmToAsm/Reg/Linear/StackMap.hs - compiler/GHC/Core/FamInstEnv.hs - compiler/GHC/Core/Opt/Simplify/Env.hs - compiler/GHC/Core/Opt/Simplify/Iteration.hs - compiler/GHC/Core/Opt/Simplify/Utils.hs - compiler/GHC/Core/Opt/SpecConstr.hs - compiler/GHC/Core/Opt/Specialise.hs - compiler/GHC/Core/Opt/WorkWrap.hs - compiler/GHC/Core/Rules.hs - compiler/GHC/Core/Unfold/Make.hs - compiler/GHC/Driver/MakeFile.hs - compiler/GHC/Runtime/Interpreter/Init.hs - compiler/GHC/Settings/IO.hs - compiler/GHC/SysTools/Terminal.hs - compiler/GHC/Tc/Instance/Family.hs - compiler/GHC/Tc/Solver/Equality.hs - compiler/GHC/Tc/Solver/Monad.hs - configure.ac - docs/users_guide/9.16.1-notes.rst - docs/users_guide/ghci.rst - docs/users_guide/packages.rst - docs/users_guide/phases.rst - docs/users_guide/using-concurrent.rst - docs/users_guide/win32-dlls.rst - driver/ghci/ghci.c - driver/utils/cwrapper.c - driver/utils/isMinTTY.c - hadrian/bindist/cwrappers/cwrapper.c - hadrian/doc/user-settings.md - hadrian/src/Packages.hs - hadrian/src/Rules/BinaryDist.hs - hadrian/src/Rules/CabalReinstall.hs - hadrian/src/Rules/Generate.hs - hadrian/src/Rules/Program.hs - hadrian/src/Rules/Test.hs - hadrian/src/Settings/Builders/Ghc.hs - hadrian/src/Settings/Default.hs - hadrian/src/Settings/Packages.hs - hadrian/src/Settings/Program.hs - libraries/base/tests/IO/T12010/cbits/initWinSock.c - libraries/ghc-internal/cbits/consUtils.c - libraries/ghc-internal/configure.ac - libraries/ghc-internal/src/GHC/Internal/ConsoleHandler.hsc - libraries/ghc-internal/src/GHC/Internal/IO/Handle.hs - libraries/mtl - m4/ghc_select_file_extensions.m4 - rts/RtsFlags.c - testsuite/driver/runtests.py - testsuite/driver/testglobals.py - testsuite/driver/testlib.py - testsuite/mk/test.mk - 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/driver/T24731.hs - testsuite/tests/driver/all.T - testsuite/tests/rts/linker/rdynamic.hs - testsuite/tests/simplCore/should_compile/T14003.stderr - testsuite/tests/simplCore/should_compile/T19672.stderr - testsuite/tests/simplCore/should_compile/T21763.stderr - testsuite/tests/simplCore/should_compile/T21763a.stderr - + testsuite/tests/simplCore/should_compile/T26615.hs - + testsuite/tests/simplCore/should_compile/T26615.stderr - + testsuite/tests/simplCore/should_compile/T26615a.hs - testsuite/tests/simplCore/should_compile/T4908.stderr - testsuite/tests/simplCore/should_compile/all.T - testsuite/tests/simplCore/should_compile/spec-inline.stderr - + 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 - utils/ghc-pkg/Main.hs - − utils/iserv/iserv.cabal.in - − utils/iserv/src/Main.hs The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/a505c4e1468274b59b744ec3053137… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/a505c4e1468274b59b744ec3053137… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/marge_bot_batch_merge_job] 9 commits: PPC NCG: Fix shift right MO code
by Marge Bot (＠marge-bot) 14 Jan '26

14 Jan '26

Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC Commits: c1fe0097 by Peter Trommler at 2026-01-14T03:54:49-05:00 PPC NCG: Fix shift right MO code The shift amount in shift right [arithmetic] MOs is machine word width. Therefore remove unnecessary zero- or sign-extending of shift amount. It looks harmless to extend the shift amount argument because the shift right instruction uses only the seven lowest bits (i. e. mod 128). But now we have a conversion operation from a smaller type to word width around a memory load at word width. The types are not matching up but there is no check done in CodeGen. The necessary conversion from word width down to the smaller width would be translated into a no-op on PowerPC anyway. So all seems harmless if it was not for a small optimisation in getRegister'. In getRegister' a load instruction with the smaller width of the conversion operation was generated. This loaded the most significant bits of the word in memory on a big-endian platform. These bits were zero and hence shift right was used with shift amount zero and not one as required in test Sized. Fixes #26519 - - - - - 2dafc65a by Cheng Shao at 2026-01-14T03:55:31-05:00 Tree-wide cleanup of cygwin logic GHC has not supported cygwin for quite a few years already, and will not resume support in the forseeable future. The only supported windows toolchain is clang64/clangarm64 of the msys2 project. This patch cleans up the unused cygwin logic in the tree. Co-authored-by: Codex <codex(a)openai.com> - - - - - 66b96e2a by Teo Camarasu at 2026-01-14T03:56:13-05:00 Set default eventlog-flush-interval to 5s Resolves #26707 - - - - - d0254579 by Andrew Lelechenko at 2026-01-14T03:56:53-05:00 Document when -maxN RTS option was added - - - - - 9c8c8e37 by Cheng Shao at 2026-01-14T04:29:45-05:00 testsuite: remove obsolete --ci option from the testsuite driver This patch removes the obsolete `--ci` option from the testsuite driver: neither the CI scripts nor hadrian ever invokes the testsuite driver with `--ci`, and the perf notes are always fetched to the `refs/notes/perf` local reference anyway. - - - - - 440647e0 by Julian Ospald at 2026-01-14T04:29:57-05:00 Fix fetch_cabal * download cabal if the existing one is of an older version * fix FreeBSD download url * fix unpacking on FreeBSD - - - - - 96f4fb6c by Julian Ospald at 2026-01-14T04:29:57-05:00 Bump toolchain in CI - - - - - 59f56b35 by Julian Ospald at 2026-01-14T04:29:57-05:00 Use libffi-clib Previously, we would build libffi via hadrian and bundle it manually with the GHC bindist. This now moves all that logic out of hadrian and allows us to have a clean Haskell package to build and link against and ship it without extra logic. This patch still retains the ability to link against a system libffi. The main reason of bundling libffi was that on some platforms (e.g. FreeBSD and Mac), system libffi is not visible to the C toolchain by default, so users would require settings in e.g. cabal to be able to compile anything. This adds the submodule libffi-clib to the repository. - - - - - e1320756 by Peng Fan at 2026-01-14T04:30:11-05:00 NCG/LA64: add support for la664 micro architecture Add '-mla664' flag to LA664, which has some new features: atomic instructions, dbar hints, etc. 'LA464' is the default so that unrecognized instructions are not generated. - - - - - 58 changed files: - .gitlab/ci.sh - .gitlab/generate-ci/gen_ci.hs - .gitlab/jobs.yaml - .gitmodules - compiler/GHC/CmmToAsm/Config.hs - compiler/GHC/CmmToAsm/PPC/CodeGen.hs - compiler/GHC/Driver/Config/CmmToAsm.hs - compiler/GHC/Driver/DynFlags.hs - compiler/GHC/Driver/MakeFile.hs - compiler/GHC/Driver/Session.hs - compiler/GHC/Linker/Loader.hs - compiler/GHC/Linker/Unit.hs - compiler/GHC/SysTools/Terminal.hs - compiler/GHC/Unit/State.hs - configure.ac - docs/users_guide/9.16.1-notes.rst - docs/users_guide/packages.rst - docs/users_guide/using-concurrent.rst - docs/users_guide/using.rst - docs/users_guide/win32-dlls.rst - driver/ghci/ghci.c - driver/utils/cwrapper.c - driver/utils/isMinTTY.c - hadrian/bindist/cwrappers/cwrapper.c - hadrian/hadrian.cabal - hadrian/src/Builder.hs - hadrian/src/Packages.hs - hadrian/src/Rules.hs - hadrian/src/Rules/Documentation.hs - hadrian/src/Rules/Generate.hs - − hadrian/src/Rules/Libffi.hs - hadrian/src/Rules/Register.hs - hadrian/src/Rules/Rts.hs - hadrian/src/Rules/SourceDist.hs - hadrian/src/Settings/Builders/Cabal.hs - hadrian/src/Settings/Builders/Ghc.hs - hadrian/src/Settings/Default.hs - hadrian/src/Settings/Packages.hs - hadrian/src/Settings/Program.hs - − libffi-tarballs - libraries/base/tests/IO/T12010/cbits/initWinSock.c - libraries/ghc-internal/cbits/consUtils.c - libraries/ghc-internal/configure.ac - libraries/ghc-internal/src/GHC/Internal/ConsoleHandler.hsc - libraries/ghc-internal/src/GHC/Internal/IO/Handle.hs - + libraries/libffi-clib - m4/ghc_select_file_extensions.m4 - packages - rts/RtsFlags.c - rts/include/rts/ghc_ffi.h - rts/rts.buildinfo.in - rts/rts.cabal - testsuite/driver/perf_notes.py - testsuite/driver/runtests.py - testsuite/driver/testglobals.py - testsuite/driver/testlib.py - testsuite/mk/test.mk - testsuite/tests/rts/linker/rdynamic.hs The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/6ef0ac0c8c1582a1991a6cae9a7cee… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/6ef0ac0c8c1582a1991a6cae9a7cee… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/spj-try-opt-coercion] Fuse optCoRefl and substCo
by Simon Peyton Jones (＠simonpj) 14 Jan '26

14 Jan '26

Simon Peyton Jones pushed to branch wip/spj-try-opt-coercion at Glasgow Haskell Compiler / GHC Commits: e180c6dc by Simon Peyton Jones at 2026-01-14T09:07:38+00:00 Fuse optCoRefl and substCo Maybe this will be better than either - - - - - 4 changed files: - compiler/GHC/Core/Coercion/Opt.hs - compiler/GHC/Core/Opt/Simplify/Iteration.hs - compiler/GHC/Core/TyCo/Subst.hs - compiler/GHC/Utils/Misc.hs Changes: ===================================== compiler/GHC/Core/Coercion/Opt.hs ===================================== @@ -257,12 +257,17 @@ optCoAlt is (Alt k bs e) left-to-right, and won't spot (co1 ; co2 ; sym co2) -} -optCoRefl :: Coercion -> Coercion +optCoRefl :: Subst -> Coercion -> Coercion -- See Note [optCoRefl] -optCoRefl in_co -#ifdef DEBUG +optCoRefl subst in_co + | isEmptyTCvSubst subst = in_co + + | otherwise +#ifndef DEBUG + = opt_co_refl subst in_co +#else -- Debug check that optCoRefl doesn't change the type - = let out_co = go in_co + = let out_co = opt_co_refl subst in_co (Pair in_l in_r) = coercionKind in_co (Pair out_l out_r) = coercionKind out_co in if (in_l `eqType` out_l) && (in_r `eqType` out_r) @@ -274,9 +279,11 @@ optCoRefl in_co , text "in_co:" <+> ppr in_co , text "out_co:" <+> ppr out_co ]) $ out_co -#else - = go in_co #endif + + +opt_co_refl :: Subst -> Coercion -> Coercion +opt_co_refl subst co = go co where go_m MRefl = MRefl go_m (MCo co) = MCo (go co) @@ -294,12 +301,17 @@ optCoRefl in_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 (ForAllCo v vl vr mco co) = mkForAllCo v vl vr (go_m mco) (go co) 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 lt rt go (AxiomCo ax cos) = mkAxiomCo ax (go_s cos) + go (ForAllCo v vl vr mco co) = mkForAllCo v' vl vr + $!! go_m mco + $!! opt_co_refl subst' co + where + !(subst', v') = substVarBndr subst v + -- This is the main payload go (TransCo co1 co2) = gobble gs0 co1 [co2] where ===================================== compiler/GHC/Core/Opt/Simplify/Iteration.hs ===================================== @@ -1390,7 +1390,7 @@ simplCoercionF env co cont simplCoercion :: SimplEnv -> InCoercion -> SimplM OutCoercion simplCoercion env co - = do { let out_co = optCoRefl (substCo env co) + = do { let out_co = optCoRefl (getTCvSubst env) co ; seqCo out_co `seq` return out_co } ----------------------------------- ===================================== compiler/GHC/Core/TyCo/Subst.hs ===================================== @@ -860,33 +860,33 @@ subst_co subst co go_mco (MCo co) = MCo (go co) go :: Coercion -> Coercion - go (Refl ty) = mkNomReflCo $! (go_ty ty) - go (GRefl r ty mco) = (mkGReflCo r $! (go_ty ty)) $! (go_mco mco) - go (TyConAppCo r tc args)= mkTyConAppCo r tc $! go_cos args - go (AxiomCo con cos) = mkAxiomCo con $! go_cos cos - go (AppCo co arg) = (mkAppCo $! go co) $! go arg + go (Refl ty) = mkNomReflCo $!! go_ty ty + go (GRefl r ty mco) = mkGReflCo r $!! go_ty ty $!! go_mco mco + go (TyConAppCo r tc args)= mkTyConAppCo r tc $!! go_cos args + go (AxiomCo con cos) = mkAxiomCo con $!! go_cos cos + go (AppCo co arg) = mkAppCo $!! go co $!! go arg go (ForAllCo { fco_tcv = tcv, fco_visL = visL, fco_visR = visR , fco_kind = kind_co, fco_body = co }) - = ((mkForAllCo $! tcv') visL visR - $! go_mco kind_co) - $! subst_co subst' co + = (mkForAllCo $!! tcv') visL visR + $!! go_mco kind_co + $!! subst_co subst' co where !(subst', tcv') = substVarBndrUnchecked subst tcv -- Unchecked because used from substTyUnchecked - go (FunCo r afl afr w co1 co2) = ((mkFunCo2 r afl afr $! go w) $! go co1) $! go co2 + go (FunCo r afl afr w co1 co2) = mkFunCo2 r afl afr $!! go w $!! go co1 $!! go co2 go (CoVarCo cv) = substCoVar subst cv go (UnivCo { uco_prov = p, uco_role = r , uco_lty = t1, uco_rty = t2, uco_deps = deps }) - = ((((mkUnivCo $! p) $! go_cos deps) $! r) $! - (go_ty t1)) $! (go_ty t2) - go (SymCo co) = mkSymCo $! (go co) - go (TransCo co1 co2) = (mkTransCo $! (go co1)) $! (go co2) - go (SelCo d co) = mkSelCo d $! (go co) - go (LRCo lr co) = mkLRCo lr $! (go co) - go (InstCo co arg) = (mkInstCo $! (go co)) $! go arg - go (KindCo co) = mkKindCo $! (go co) - go (SubCo co) = mkSubCo $! (go co) - go (HoleCo h) = HoleCo $! go_hole h + = mkUnivCo p $!! go_cos deps $!! r + $!! go_ty t1 $!! go_ty t2 + go (SymCo co) = mkSymCo $!! go co + go (TransCo co1 co2) = mkTransCo $!! go co1 $!! go co2 + go (SelCo d co) = mkSelCo d $!! go co + go (LRCo lr co) = mkLRCo lr $!! go co + go (InstCo co arg) = mkInstCo $!! go co $!! go arg + go (KindCo co) = mkKindCo $!! go co + go (SubCo co) = mkSubCo $!! go co + go (HoleCo h) = HoleCo $!! go_hole h go_cos cos = let cos' = map go cos in cos' `seqList` cos' ===================================== compiler/GHC/Utils/Misc.hs ===================================== @@ -8,7 +8,7 @@ -- module GHC.Utils.Misc ( -- * Miscellaneous higher-order functions - applyWhen, nTimes, const2, + applyWhen, nTimes, const2, ($!!), -- * General list processing zipEqual, zipWithEqual, zipWith3Equal, zipWith4Equal, @@ -44,8 +44,7 @@ module GHC.Utils.Misc ( mergeListsBy, isSortedBy, - -- Foldable generalised functions, - + -- * Foldable generalised functions, mapMaybe', -- * Tuples @@ -153,6 +152,8 @@ import qualified Data.Set as Set import Data.Time +infixl 0 $!! -- LEFT associative + {- ************************************************************************ * * @@ -199,6 +200,14 @@ third3 f (a, b, c) = (a, b, f c) uncurry3 :: (a -> b -> c -> d) -> (a, b, c) -> d uncurry3 f (a, b, c) = f a b c +($!!) :: forall r a (b :: TYPE r). (a -> b) -> a -> b +-- | ^ ($!!) is left-associative so you can write +-- (f $!! e1 $!! e2) for a multi-argument strict application +-- In contrast ($) and ($!) are right associative +{-# INLINE ($!!) #-} +f $!! x = let !vx = x in f vx -- see #2273 + + {- ************************************************************************ * * View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/e180c6dc8f8141c31771b50d30c17c4… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/e180c6dc8f8141c31771b50d30c17c4… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][master] Document when -maxN RTS option was added
by Marge Bot (＠marge-bot) 14 Jan '26

14 Jan '26

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: d0254579 by Andrew Lelechenko at 2026-01-14T03:56:53-05:00 Document when -maxN RTS option was added - - - - - 1 changed file: - docs/users_guide/using-concurrent.rst Changes: ===================================== docs/users_guide/using-concurrent.rst ===================================== @@ -112,7 +112,7 @@ use the RTS :rts-flag:`-N ⟨x⟩` options. .. rts-flag:: -N ⟨x⟩ -N - -maxN ⟨x⟩ + -maxN ⟨x⟩ :since: 8.0 Use ⟨x⟩ simultaneous threads when running the program. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/d0254579799c8cc3885179c58b8d204… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/d0254579799c8cc3885179c58b8d204… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][master] Set default eventlog-flush-interval to 5s
by Marge Bot (＠marge-bot) 14 Jan '26

14 Jan '26

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: 66b96e2a by Teo Camarasu at 2026-01-14T03:56:13-05:00 Set default eventlog-flush-interval to 5s Resolves #26707 - - - - - 2 changed files: - docs/users_guide/9.16.1-notes.rst - rts/RtsFlags.c Changes: ===================================== docs/users_guide/9.16.1-notes.rst ===================================== @@ -88,6 +88,8 @@ Runtime system kqueue). This first one, based on `poll()`, is merely intended to be portable. +- The default for :rts-flag:`--eventlog-flush-interval=⟨seconds⟩` has changed from being disabled to being enabled with a flush inverval of 5 seconds. + Cmm ~~~ ===================================== rts/RtsFlags.c ===================================== @@ -250,7 +250,7 @@ void initRtsFlagsDefaults(void) RtsFlags.TraceFlags.ticky = false; RtsFlags.TraceFlags.trace_output = NULL; # if defined(THREADED_RTS) - RtsFlags.TraceFlags.eventlogFlushTime = 0; + RtsFlags.TraceFlags.eventlogFlushTime = SecondsToTime(5); // 5s # endif RtsFlags.TraceFlags.nullWriter = false; #endif View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/66b96e2a591d8e3d60e74af3671344d… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/66b96e2a591d8e3d60e74af3671344d… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][master] Tree-wide cleanup of cygwin logic
by Marge Bot (＠marge-bot) 14 Jan '26

14 Jan '26

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: 2dafc65a by Cheng Shao at 2026-01-14T03:55:31-05:00 Tree-wide cleanup of cygwin logic GHC has not supported cygwin for quite a few years already, and will not resume support in the forseeable future. The only supported windows toolchain is clang64/clangarm64 of the msys2 project. This patch cleans up the unused cygwin logic in the tree. Co-authored-by: Codex <codex(a)openai.com> - - - - - 21 changed files: - compiler/GHC/Driver/MakeFile.hs - compiler/GHC/SysTools/Terminal.hs - configure.ac - docs/users_guide/packages.rst - docs/users_guide/win32-dlls.rst - driver/ghci/ghci.c - driver/utils/cwrapper.c - driver/utils/isMinTTY.c - hadrian/bindist/cwrappers/cwrapper.c - hadrian/src/Settings/Program.hs - libraries/base/tests/IO/T12010/cbits/initWinSock.c - libraries/ghc-internal/cbits/consUtils.c - libraries/ghc-internal/configure.ac - libraries/ghc-internal/src/GHC/Internal/ConsoleHandler.hsc - libraries/ghc-internal/src/GHC/Internal/IO/Handle.hs - m4/ghc_select_file_extensions.m4 - testsuite/driver/runtests.py - testsuite/driver/testglobals.py - testsuite/driver/testlib.py - testsuite/mk/test.mk - testsuite/tests/rts/linker/rdynamic.hs Changes: ===================================== compiler/GHC/Driver/MakeFile.hs ===================================== @@ -331,7 +331,7 @@ writeDependency :: FilePath -> Handle -> [FilePath] -> FilePath -> IO () writeDependency root hdl targets dep = do let -- We need to avoid making deps on -- c:/foo/... - -- on cygwin as make gets confused by the : + -- on Windows as make gets confused by the : -- Making relative deps avoids some instances of this. dep' = makeRelative root dep forOutput = escapeSpaces . reslash Forwards . normalise ===================================== compiler/GHC/SysTools/Terminal.hs ===================================== @@ -41,7 +41,7 @@ stderrSupportsAnsiColors' = do case eMode of Left (_ :: IOError) -> Win32.isMinTTYHandle h -- Check if the we're in a MinTTY terminal - -- (e.g., Cygwin or MSYS2) + -- (e.g., MSYS2) Right mode | modeHasVTP mode -> pure True | otherwise -> enableVTP h mode ===================================== configure.ac ===================================== @@ -469,7 +469,7 @@ dnl -------------------------------------------------------------- dnl Note: we may not have objdump on OS X, and we only need it on dnl Windows (for DLL checks), OpenBSD, and AIX case $HostOS_CPP in - cygwin32|mingw32|openbsd|aix) + mingw32|openbsd|aix) AC_CHECK_TARGET_TOOL([OBJDUMP], [objdump]) ;; esac ===================================== docs/users_guide/packages.rst ===================================== @@ -1139,7 +1139,7 @@ read and write this syntax respectively. Here's an example of the Operating System Interface for Computing Environments - IEEE Std. 1003.1). . - The package is not supported under Windows (except under Cygwin). + The package is not supported under Windows. category: System author: exposed: True ===================================== docs/users_guide/win32-dlls.rst ===================================== @@ -33,23 +33,16 @@ We recommend running GHCi in a standard Windows console: select the use ``Start->Run->cmd`` to get a Windows console and invoke ``ghci`` from there (as long as it's in your ``PATH``). -If you run GHCi in a Cygwin or MSYS shell, then the Control-C behaviour -is adversely affected. In one of these environments you should use the +If you run GHCi in an MSYS shell, then the Control-C behaviour +is adversely affected. In this environment you should use the ``ghcii.sh`` script to start GHCi, otherwise when you hit Control-C you'll be returned to the shell prompt but the GHCi process will still be running. However, even using the ``ghcii.sh`` script, if you hit Control-C then the GHCi process will be killed immediately, rather than letting you interrupt a running program inside GHCi as it should. This -problem is caused by the fact that the Cygwin and MSYS shell -environments don't pass Control-C events to non-Cygwin child processes, -because in order to do that there needs to be a Windows console. - -There's an exception: you can use a Cygwin shell if the ``CYGWIN`` -environment variable does *not* contain ``tty``. In this mode, the -Cygwin shell behaves like a Windows console shell and console events are -propagated to child processes. Note that the ``CYGWIN`` environment -variable must be set *before* starting the Cygwin shell; changing it -afterwards has no effect on the shell. +problem is caused by the fact that the MSYS shell environment doesn't +pass Control-C events to non-console child processes, because in order +to do that there needs to be a Windows console. This problem doesn't just affect GHCi, it affects any GHC-compiled program that wants to catch console events. See the @@ -138,78 +131,6 @@ See the for more details. -.. _ghci-cygwin: - -Using GHC (and other GHC-compiled executables) with Cygwin ----------------------------------------------------------- - -Background -~~~~~~~~~~ - -The Cygwin tools aim to provide a Unix-style API on top of the windows -libraries, to facilitate ports of Unix software to windows. To this end, -they introduce a Unix-style directory hierarchy under some root -directory (typically ``/`` is ``C:\cygwin\``). Moreover, everything -built against the Cygwin API (including the Cygwin tools and programs -compiled with Cygwin's GHC) will see ``/`` as the root of their file system, -happily pretending to work in a typical unix environment, and finding -things like ``/bin`` and ``/usr/include`` without ever explicitly -bothering with their actual location on the windows system (probably -``C:\cygwin\bin`` and ``C:\cygwin\usr\include``). - -The problem -~~~~~~~~~~~ - -GHC, by default, no longer depends on cygwin, but is a native Windows -program. It is built using mingw, and it uses mingw's GHC while -compiling your Haskell sources (even if you call it from cygwin's bash), -but what matters here is that - just like any other normal windows -program - neither GHC nor the executables it produces are aware of -Cygwin's pretended unix hierarchy. GHC will happily accept either ``/`` or -``\`` as path separators, but it won't know where to find ``/home/joe/Main.hs`` -or ``/bin/bash`` or the like. This causes all kinds of fun when GHC is used from -within Cygwin's bash, or in make-sessions running under Cygwin. - -Things to do -~~~~~~~~~~~~ - -- Don't use absolute paths in ``make``, ``configure`` & co if there is any - chance that those might be passed to GHC (or to GHC-compiled - programs). Relative paths are fine because cygwin tools are happy - with them and GHC accepts ``/`` as path-separator. And relative paths - don't depend on where Cygwin's root directory is located, or on which - partition or network drive your source tree happens to reside, as - long as you ``cd`` there first. - -- If you have to use absolute paths (beware of the innocent-looking - ``ROOT=$(pwd)`` in makefile hierarchies or configure scripts), Cygwin - provides a tool called ``cygpath`` that can convert Cygwin's - Unix-style paths to their actual Windows-style counterparts. Many - Cygwin tools actually accept absolute Windows-style paths (remember, - though, that you either need to escape ``\`` or convert ``\`` to ``/``), - so you should be fine just using those everywhere. If you need to use - tools that do some kind of path-mangling that depends on unix-style - paths (one fun example is trying to interpret ``:`` as a separator in - path lists), you can still try to convert paths using ``cygpath`` - just before they are passed to GHC and friends. - -- If you don't have ``cygpath``, you probably don't have cygwin and - hence no problems with it... unless you want to write one build - process for several platforms. Again, relative paths are your friend, - but if you have to use absolute paths, and don't want to use - different tools on different platforms, you can simply write a short - Haskell program to print the current directory (thanks to George - Russell for this idea): compiled with GHC, this will give you the - view of the file system that GHC depends on (which will differ - depending on whether GHC is compiled with cygwin's gcc or mingw's gcc - or on a real Unix system..) - that little program can also deal with - escaping ``\`` in paths. Apart from the banner and the startup time, - something like this would also do: - - .. code-block:: none - - $ echo "Directory.getCurrentDirectory >>= putStrLn . init . tail . show " | ghci - .. _win32-dlls: Building and using Win32 DLLs ===================================== driver/ghci/ghci.c ===================================== @@ -27,7 +27,7 @@ int main(int argc, char** argv) { char *preArgv[1]; if (isMinTTY()) { - printf("WARNING: GHCi invoked via 'ghci.exe' in MinTTY consoles (e.g., Cygwin or MSYS)\n"); + printf("WARNING: GHCi invoked via 'ghci.exe' in MinTTY consoles (e.g., MSYS2)\n"); printf(" doesn't handle Ctrl-C well; use the 'ghcii.sh' shell wrapper instead\n"); fflush(stdout); } ===================================== driver/utils/cwrapper.c ===================================== @@ -116,7 +116,7 @@ __attribute__((noreturn)) int run (char *exePath, didn't do the trick either. Irrespective of this issue, using CreateProcess() is preferable, - as it makes this wrapper work on both mingw and cygwin. + as it makes this wrapper work reliably on Windows. */ #if 0 fprintf(stderr, "Invoking %s\n", new_cmdline); fflush(stderr); ===================================== driver/utils/isMinTTY.c ===================================== @@ -28,6 +28,6 @@ bool isMinTTY() { PWSTR fn = pfni->FileName; fn[pfni->FileNameLength] = L'\0'; - return ((wcsstr(fn, L"\\cygwin-") || wcsstr(fn, L"\\msys-")) && - wcsstr(fn, L"-pty") && wcsstr(fn, L"-master")); + return (wcsstr(fn, L"\\msys-") && + wcsstr(fn, L"-pty") && wcsstr(fn, L"-master")); } ===================================== hadrian/bindist/cwrappers/cwrapper.c ===================================== @@ -116,7 +116,7 @@ __attribute__((noreturn)) int run (char *exePath, didn't do the trick either. Irrespective of this issue, using CreateProcess() is preferable, - as it makes this wrapper work on both mingw and cygwin. + as it makes this wrapper work reliably on Windows. */ #if 0 fprintf(stderr, "Invoking %s\n", new_cmdline); fflush(stderr); ===================================== hadrian/src/Settings/Program.hs ===================================== @@ -40,7 +40,7 @@ programContext stage pkg = do ghcWithInterpreter :: Stage -> Action Bool ghcWithInterpreter stage = do is_cross <- flag CrossCompiling - goodOs <- anyTargetOs [ OSMinGW32, OSLinux, OSSolaris2 -- TODO "cygwin32"?, + goodOs <- anyTargetOs [ OSMinGW32, OSLinux, OSSolaris2 , OSFreeBSD, OSDragonFly, OSNetBSD, OSOpenBSD , OSDarwin, OSKFreeBSD , OSWasi ] ===================================== libraries/base/tests/IO/T12010/cbits/initWinSock.c ===================================== @@ -1,7 +1,7 @@ #include "HsBase.h" #include "HsFFI.h" -#if defined(HAVE_WINSOCK_H) && !defined(__CYGWIN__) +#if defined(HAVE_WINSOCK_H) static int winsock_inited = 0; ===================================== libraries/ghc-internal/cbits/consUtils.c ===================================== @@ -3,17 +3,13 @@ * * Win32 Console API support */ -#if defined(_WIN32) || defined(__CYGWIN__) +#if defined(_WIN32) /* to the end */ #include "consUtils.h" #include <windows.h> #include <io.h> -#if defined(__CYGWIN__) -#define _get_osfhandle get_osfhandle -#endif - int is_console__(int fd) { DWORD st; HANDLE h; ===================================== libraries/ghc-internal/configure.ac ===================================== @@ -12,7 +12,7 @@ AC_USE_SYSTEM_EXTENSIONS AC_MSG_CHECKING(for WINDOWS platform) case $host_alias in - *mingw32*|*mingw64*|*cygwin*|*msys*|*windows*) + *mingw32*|*mingw64*|*msys*|*windows*) WINDOWS=YES;; *) WINDOWS=NO;; ===================================== libraries/ghc-internal/src/GHC/Internal/ConsoleHandler.hsc ===================================== @@ -67,9 +67,7 @@ data Handler -- running in a Windows console. Certain environments that look like consoles -- do not support console events, these include: -- --- * Cygwin shells with @CYGWIN=tty@ set (if you don't set @CYGWIN=tty@, --- then a Cygwin shell behaves like a Windows console). --- * Cygwin xterm and rxvt windows +-- * MSYS MinTTY shells -- * MSYS rxvt windows -- -- In order for your application to receive console events, avoid running ===================================== libraries/ghc-internal/src/GHC/Internal/IO/Handle.hs ===================================== @@ -575,7 +575,7 @@ hGetEcho handle = do -- | Is the handle connected to a terminal? -- -- On Windows the result of 'hIsTerminalDevide' might be misleading, --- because non-native terminals, such as MinTTY used in MSYS and Cygwin environments, +-- because non-native terminals, such as MinTTY used in MSYS environments, -- are implemented via redirection. -- Use @System.Win32.Types.withHandleToHANDLE System.Win32.MinTTY.isMinTTYHandle@ -- to recognise it. Also consider @ansi-terminal@ package for crossplatform terminal ===================================== m4/ghc_select_file_extensions.m4 ===================================== @@ -3,11 +3,6 @@ AC_DEFUN([GHC_SELECT_FILE_EXTENSIONS], $2='' $3='.so' case $1 in - *-unknown-cygwin32) - AC_MSG_WARN([GHC does not support the Cygwin target at the moment]) - AC_MSG_WARN([I'm assuming you wanted to build for x86_64-w64-mingw32]) - exit 1 - ;; # examples: x86_64-w64-mingw32 *-mingw32) windows=YES ===================================== testsuite/driver/runtests.py ===================================== @@ -21,8 +21,8 @@ from pathlib import Path # We don't actually need subprocess in runtests.py, but: # * We do need it in testlibs.py # * We can't import testlibs.py until after we have imported ctypes -# * If we import ctypes before subprocess on cygwin, then sys.exit(0) -# says "Aborted" and we fail with exit code 134. +# * If we import ctypes before subprocess on some Windows environments, +# then sys.exit(0) says "Aborted" and we fail with exit code 134. # So we import it here first, so that the testsuite doesn't appear to fail. import subprocess @@ -182,16 +182,13 @@ elif args.ignore_perf_failures == 'decreases': if args.test_env: config.test_env = args.test_env -config.cygwin = False config.msys = False if windows: h = os.popen('uname -s', 'r') v = h.read() h.close() - if v.startswith("CYGWIN"): - config.cygwin = True - elif v.startswith("MINGW") or v.startswith("MSYS"): + if v.startswith("MINGW") or v.startswith("MSYS"): # msys gives "MINGW32" # msys2 gives "MINGW_NT-6.2" or "MSYS_NT-6.3" config.msys = True @@ -274,12 +271,6 @@ def format_path(path): # letter representation. Otherwise it thinks we're adding two env # variables E and /Foo when we add E:/Foo. path = re.sub('([a-zA-Z]):', '/\\1', path) - if config.cygwin: - # On cygwin we can't put "c:\foo" in $PATH, as : is a - # field separator. So convert to /cygdrive/c/foo instead. - # Other pythons use ; as the separator, so no problem. - path = re.sub('([a-zA-Z]):', '/cygdrive/\\1', path) - path = re.sub('\\\\', '/', path) return path # On Windows we need to set $PATH to include the paths to all the DLLs ===================================== testsuite/driver/testglobals.py ===================================== @@ -80,7 +80,6 @@ class TestConfig: self.os = '' self.arch = '' self.msys = False - self.cygwin = False # What is the wordsize (in bits) of this platform? self.wordsize = '' ===================================== testsuite/driver/testlib.py ===================================== @@ -1023,9 +1023,6 @@ def wordsize( ws: int ) -> bool: def msys( ) -> bool: return config.msys -def cygwin( ) -> bool: - return config.cygwin - def js_arch() -> bool: return arch("javascript"); ===================================== testsuite/mk/test.mk ===================================== @@ -65,13 +65,13 @@ TEST_HC_OPTS_INTERACTIVE = $(TEST_HC_OPTS) --interactive -v0 -ignore-dot-ghci -f RUNTEST_OPTS = -ifeq "$(filter $(TargetOS_CPP), cygwin32 mingw32)" "" +ifeq "$(filter $(TargetOS_CPP), mingw32)" "" exeext = else exeext = .exe endif -ifneq "$(filter $(TargetOS_CPP),cygwin32 mingw32)" "" +ifneq "$(filter $(TargetOS_CPP),mingw32)" "" dllext = .dll else ifeq "$(TargetOS_CPP)" "darwin" dllext = .dylib ===================================== testsuite/tests/rts/linker/rdynamic.hs ===================================== @@ -39,10 +39,8 @@ loadFunction mpkg m valsym = do where prefixUnderscore = case (os, arch) of ("mingw32", "x86_64") -> "" - ("cygwin" , "x86_64") -> "" ("mingw32", _ ) -> "_" ("darwin" , _ ) -> "_" - ("cygwin" , _ ) -> "_" _ -> "" foreign import ccall safe "lookupSymbol" c_lookupSymbol :: CString -> IO (Ptr a) View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/2dafc65a48f7d0d959030b801ee45fb… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/2dafc65a48f7d0d959030b801ee45fb… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][master] PPC NCG: Fix shift right MO code
by Marge Bot (＠marge-bot) 14 Jan '26

14 Jan '26

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: c1fe0097 by Peter Trommler at 2026-01-14T03:54:49-05:00 PPC NCG: Fix shift right MO code The shift amount in shift right [arithmetic] MOs is machine word width. Therefore remove unnecessary zero- or sign-extending of shift amount. It looks harmless to extend the shift amount argument because the shift right instruction uses only the seven lowest bits (i. e. mod 128). But now we have a conversion operation from a smaller type to word width around a memory load at word width. The types are not matching up but there is no check done in CodeGen. The necessary conversion from word width down to the smaller width would be translated into a no-op on PowerPC anyway. So all seems harmless if it was not for a small optimisation in getRegister'. In getRegister' a load instruction with the smaller width of the conversion operation was generated. This loaded the most significant bits of the word in memory on a big-endian platform. These bits were zero and hence shift right was used with shift amount zero and not one as required in test Sized. Fixes #26519 - - - - - 1 changed file: - compiler/GHC/CmmToAsm/PPC/CodeGen.hs Changes: ===================================== compiler/GHC/CmmToAsm/PPC/CodeGen.hs ===================================== @@ -467,48 +467,26 @@ getRegister' _ platform (CmmLoad mem pk _) return (Any II64 code) -- catch simple cases of zero- or sign-extended load -getRegister' _ _ (CmmMachOp (MO_UU_Conv W8 W32) [CmmLoad mem _ _]) = do - Amode addr addr_code <- getAmode D mem - return (Any II32 (\dst -> addr_code `snocOL` LD II8 dst addr)) - -getRegister' _ _ (CmmMachOp (MO_XX_Conv W8 W32) [CmmLoad mem _ _]) = do - Amode addr addr_code <- getAmode D mem - return (Any II32 (\dst -> addr_code `snocOL` LD II8 dst addr)) - -getRegister' _ _ (CmmMachOp (MO_UU_Conv W8 W64) [CmmLoad mem _ _]) = do - Amode addr addr_code <- getAmode D mem - return (Any II64 (\dst -> addr_code `snocOL` LD II8 dst addr)) - -getRegister' _ _ (CmmMachOp (MO_XX_Conv W8 W64) [CmmLoad mem _ _]) = do - Amode addr addr_code <- getAmode D mem - return (Any II64 (\dst -> addr_code `snocOL` LD II8 dst addr)) - --- Note: there is no Load Byte Arithmetic instruction, so no signed case here - -getRegister' _ _ (CmmMachOp (MO_UU_Conv W16 W32) [CmmLoad mem _ _]) = do - Amode addr addr_code <- getAmode D mem - return (Any II32 (\dst -> addr_code `snocOL` LD II16 dst addr)) - -getRegister' _ _ (CmmMachOp (MO_SS_Conv W16 W32) [CmmLoad mem _ _]) = do - Amode addr addr_code <- getAmode D mem - return (Any II32 (\dst -> addr_code `snocOL` LA II16 dst addr)) - -getRegister' _ _ (CmmMachOp (MO_UU_Conv W16 W64) [CmmLoad mem _ _]) = do - Amode addr addr_code <- getAmode D mem - return (Any II64 (\dst -> addr_code `snocOL` LD II16 dst addr)) - -getRegister' _ _ (CmmMachOp (MO_SS_Conv W16 W64) [CmmLoad mem _ _]) = do - Amode addr addr_code <- getAmode D mem - return (Any II64 (\dst -> addr_code `snocOL` LA II16 dst addr)) - -getRegister' _ _ (CmmMachOp (MO_UU_Conv W32 W64) [CmmLoad mem _ _]) = do - Amode addr addr_code <- getAmode D mem - return (Any II64 (\dst -> addr_code `snocOL` LD II32 dst addr)) - -getRegister' _ _ (CmmMachOp (MO_SS_Conv W32 W64) [CmmLoad mem _ _]) = do - -- lwa is DS-form. See Note [Power instruction format] - Amode addr addr_code <- getAmode DS mem - return (Any II64 (\dst -> addr_code `snocOL` LA II32 dst addr)) +getRegister' _ _ (CmmMachOp (MO_UU_Conv src tgt) [CmmLoad mem pk _]) + | src < tgt + , cmmTypeFormat pk == intFormat src = loadZeroExpand mem pk tgt + +getRegister' _ _ (CmmMachOp (MO_XX_Conv src tgt) [CmmLoad mem pk _]) + | src < tgt + , cmmTypeFormat pk == intFormat src = loadZeroExpand mem pk tgt + + -- XXX: This is ugly, refactor +getRegister' _ _ (CmmMachOp (MO_SS_Conv src tgt) [CmmLoad mem pk _]) + -- Note: there is no Load Byte Arithmetic instruction + | cmmTypeFormat pk /= II8 + , src < tgt = do + let format = cmmTypeFormat pk + -- lwa is DS-form. See Note [Power instruction format] + let form = if format >= II32 then DS else D + Amode addr addr_code <- getAmode form mem + let code dst = assert (format == intFormat src) + $ addr_code `snocOL` LA format dst addr + return (Any (intFormat tgt) code) getRegister' config platform (CmmMachOp (MO_RelaxedRead w) [e]) = getRegister' config platform (CmmLoad e (cmmBits w) NaturallyAligned) @@ -789,6 +767,12 @@ extendSExpr from to x = CmmMachOp (MO_SS_Conv from to) [x] extendUExpr :: Width -> Width -> CmmExpr -> CmmExpr extendUExpr from to x = CmmMachOp (MO_UU_Conv from to) [x] +loadZeroExpand :: CmmExpr -> CmmType -> Width -> NatM Register +loadZeroExpand mem pk tgt = do + Amode addr addr_code <- getAmode D mem + let code dst = addr_code `snocOL` LD (cmmTypeFormat pk) dst addr + return (Any (intFormat tgt) code) + -- ----------------------------------------------------------------------------- -- The 'Amode' type: Memory addressing modes passed up the tree. @@ -2448,8 +2432,8 @@ srCode width sgn instr x y = do let op_len = max W32 width extend = if sgn then extendSExpr else extendUExpr (src1, code1) <- getSomeReg (extend width op_len x) - (src2, code2) <- getSomeReg (extendUExpr width op_len y) - -- Note: Shift amount `y` is unsigned + (src2, code2) <- getSomeReg y + let code dst = code1 `appOL` code2 `snocOL` instr (intFormat op_len) dst src1 (RIReg src2) return (Any (intFormat width) code) View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/c1fe009739da8b368e59b87b00d7c56… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/c1fe009739da8b368e59b87b00d7c56… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T26746] Wibble
by Simon Peyton Jones (＠simonpj) 14 Jan '26

14 Jan '26

Simon Peyton Jones pushed to branch wip/T26746 at Glasgow Haskell Compiler / GHC Commits: a505c4e1 by Simon Peyton Jones at 2026-01-14T08:13:54+00:00 Wibble - - - - - 1 changed file: - compiler/GHC/Tc/Solver/Equality.hs Changes: ===================================== compiler/GHC/Tc/Solver/Equality.hs ===================================== @@ -357,9 +357,9 @@ can_eq_nc _rewritten _rdr_env _envs ev eq_rel ty1 _ ty2 _ -- You might think that representational role would also be OK, but -- see Note [Even more eager newtype decomposition] ok (ty1:tys1) (ty2:tys2) rs - | Phantom : rs <- r = ok tys1 tys2 rs - | ty1 `tcEqType` ty2 = ok tys1 tys2 (drop 1 rs) - | otherwise = False + | Phantom : rs' <- rs = ok tys1 tys2 rs' + | ty1 `tcEqType` ty2 = ok tys1 tys2 (drop 1 rs) + | otherwise = False ok [] [] _ = True ok _ _ _ = False -- Mis-matched lengths, just about possible because of -- kind polymorphism. Anyway False is a safe result! View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/a505c4e1468274b59b744ec30531377… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/a505c4e1468274b59b744ec30531377… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/fix-llvm-split-sections] llvm: fix split sections for llvm backend
by Cheng Shao (＠TerrorJack) 14 Jan '26

14 Jan '26

Cheng Shao pushed to branch wip/fix-llvm-split-sections at Glasgow Haskell Compiler / GHC Commits: 41074dd7 by Cheng Shao at 2026-01-14T03:28:25+01:00 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 `(a)llvm.compiler.used` instead of `(a)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> - - - - - 3 changed files: - compiler/GHC/CmmToLlvm.hs - compiler/GHC/CmmToLlvm/Base.hs - compiler/GHC/Driver/Pipeline/Execute.hs Changes: ===================================== compiler/GHC/CmmToLlvm.hs ===================================== @@ -271,7 +271,7 @@ cmmUsedLlvmGens = do -- used if we didn't provide these hints. This will generate a -- definition of the form -- - -- @llvm.used = appending global [42 x i8*] [i8* bitcast <var> to i8*, ...] + -- @llvm.compiler.used = appending global [42 x i8*] [i8* bitcast <var> to i8*, ...] -- -- Which is the LLVM way of protecting them against getting removed. ivars <- getUsedVars @@ -279,7 +279,7 @@ cmmUsedLlvmGens = do ty = LMArray (length ivars) i8Ptr usedArray = LMStaticArray (map cast ivars) ty sectName = Just $ fsLit "llvm.metadata" - lmUsedVar = LMGlobalVar (fsLit "llvm.used") ty Appending sectName Nothing Constant + lmUsedVar = LMGlobalVar (fsLit "llvm.compiler.used") ty Appending sectName Nothing Constant lmUsed = LMGlobal lmUsedVar (Just usedArray) if null ivars then return () ===================================== compiler/GHC/CmmToLlvm/Base.hs ===================================== @@ -286,7 +286,7 @@ data LlvmEnv = LlvmEnv , envUniqMeta :: UniqFM Unique MetaId -- ^ Global metadata nodes , envFunMap :: LlvmEnvMap -- ^ Global functions so far, with type , envAliases :: UniqSet LMString -- ^ Globals that we had to alias, see [Llvm Forward References] - , envUsedVars :: [LlvmVar] -- ^ Pointers to be added to llvm.used (see @cmmUsedLlvmGens@) + , envUsedVars :: [LlvmVar] -- ^ Pointers to be added to llvm.compiler.used (see @cmmUsedLlvmGens@) -- the following get cleared for every function (see @withClearVars@) , envVarMap :: LlvmEnvMap -- ^ Local variables so far, with type ===================================== compiler/GHC/Driver/Pipeline/Execute.hs ===================================== @@ -934,6 +934,14 @@ llvmOptions llvm_config llvm_version dflags = [("-relocation-model=" ++ rmodel ,"-relocation-model=" ++ rmodel) | not (null rmodel)] + -- Both llc/opt need these flags for split sections + ++ [ ("--data-sections", "--data-sections") + | gopt Opt_SplitSections dflags + ] + ++ [ ("--function-sections", "--function-sections") + | gopt Opt_SplitSections dflags + ] + -- Additional llc flags ++ [("", "-mcpu=" ++ mcpu) | not (null mcpu) , not (any (isInfixOf "-mcpu") (getOpts dflags opt_lc)) ] View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/41074dd7f7695979bb90180c19fe78c… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/41074dd7f7695979bb90180c19fe78c… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/26737] 21 commits: ghc-internal: Move STM utilities out of GHC.Internal.Conc.Sync
by Simon Peyton Jones (＠simonpj) 13 Jan '26

13 Jan '26

Simon Peyton Jones pushed to branch wip/26737 at Glasgow Haskell Compiler / GHC Commits: 039bac4c by Ben Gamari at 2026-01-09T20:22:16-05:00 ghc-internal: Move STM utilities out of GHC.Internal.Conc.Sync This is necessary to avoid an import cycle on Windows when importing `GHC.Internal.Exception.Context` in `GHC.Internal.Conc.Sync`. On the road to address #25365. - - - - - 8c389e8c by Ben Gamari at 2026-01-09T20:22:16-05:00 base: Capture backtrace from throwSTM Implements core-libraries-committee#297. Fixes #25365. - - - - - e1ce1fc3 by Ben Gamari at 2026-01-09T20:22:16-05:00 base: Annotate rethrown exceptions in catchSTM with WhileHandling Implements core-libraries-committee#298 - - - - - c4ebdbdf by Cheng Shao at 2026-01-09T20:23:06-05:00 compiler: make getPrim eagerly evaluate its result This commit makes `GHC.Utils.Binary.getPrim` eagerly evaluate its result, to avoid accidental laziness when future patches build other binary parsers using `getPrim`. - - - - - 66a0c4f7 by Cheng Shao at 2026-01-09T20:23:06-05:00 compiler: implement fast get/put for Word16/Word32/Word64 Previously, `GHC.Utils.Binary` contains `get`/`put` functions for `Word16`/`Word32`/`Word64` which always loads and stores them as big-endian words at a potentially unaligned address. The previous implementation is based on loads/stores of individual bytes and concatenating bytes with bitwise operations, which currently cannot be fused to a single load/store operation by GHC. This patch implements fast `get`/`put` functions for `Word16`/`Word32`/`Word64` based on a single memory load/store, with an additional `byteSwap` operation on little-endian hosts. It is based on unaligned load/store primops added since GHC 9.10, and we already require booting with at least 9.10, so it's about time to switch to this faster path. - - - - - 641ec3f0 by Simon Peyton Jones at 2026-01-09T20:23:55-05:00 Fix scoping errors in specialisation Using -fspecialise-aggressively in #26682 showed up a couple of subtle errors in the type-class specialiser. * dumpBindUDs failed to call `deleteCallsMentioning`, resulting in a call that mentioned a dictionary that was not in scope. This call has been missing since 2009! commit c43c981705ec33da92a9ce91eb90f2ecf00be9fe Author: Simon Peyton Jones <simonpj(a)microsoft.com> Date: Fri Oct 23 16:15:51 2009 +0000 Fixed by re-combining `dumpBindUDs` and `dumpUDs`. * I think there was another bug involving the quantified type variables in polymorphic specialisation. In any case I refactored `specHeader` and `spec_call` so that the former looks for the extra quantified type variables rather than the latter. This is quite a worthwhile simplification: less code, easier to grok. Test case in simplCore/should_compile/T26682, brilliantly minimised by @sheaf. - - - - - 2433e91d by Cheng Shao at 2026-01-09T20:24:43-05:00 compiler: change sectionProtection to take SectionType argument This commit changes `sectionProtection` to only take `SectionType` argument instead of whole `Section`, since it doesn't need the Cmm section content anyway, and it can then be called in parts of NCG where we only have a `SectionType` in scope. - - - - - e5926fbe by Cheng Shao at 2026-01-09T20:24:43-05:00 compiler: change isInitOrFiniSection to take SectionType argument This commit changes `isInitOrFiniSection` to only take `SectionType` argument instead of whole `Section`, since it doesn't need the Cmm section content anyway, and it can then be called in parts of NCG where we only have a `SectionType` in scope. Also marks it as exported. - - - - - 244d57d7 by Cheng Shao at 2026-01-09T20:24:43-05:00 compiler: fix split sections on windows This patch fixes split sections on windows by emitting the right COMDAT section header in NCG, see added comment for more explanation. Fix #26696 #26494. ------------------------- Metric Decrease: LargeRecord T9675 size_hello_artifact size_hello_artifact_gzip size_hello_unicode size_hello_unicode_gzip Metric Increase: T13035 ------------------------- Co-authored-by: Codex <codex(a)openai.com> - - - - - 182f3d0f by Cheng Shao at 2026-01-09T20:25:28-05:00 iserv: add comment about -fkeep-cafs - - - - - 49675b69 by Matthew Craven at 2026-01-09T20:26:14-05:00 Account for "stupid theta" in demand sig for DataCon wrappers Fixes #26748. - - - - - f3c18890 by Samuel Thibault at 2026-01-10T15:48:22+01:00 hurd: Fix getExecutablePath build 3939a8bf93e27 ("GNU/Hurd: Add getExecutablePath support") added using /proc/self/exe for GNU/Hurd but missed adding the required imports for the corresponding code. - - - - - 7f15bd15 by Samuel Thibault at 2026-01-12T07:16:25-05:00 Fix the OS string encoding for GNU/Hurd Following https://github.com/haskell/cabal/pull/9434/files , and as seen in the various gnu_HOST_OS usages in the source code, it is expected that GNU/Hurd is advertised as "gnu", like the autotools do. - - - - - 1db2f240 by Andrew Lelechenko at 2026-01-12T07:17:06-05:00 Add since annotation for Data.Bifoldable1 Fixes #26432 - - - - - e038a383 by Sven Tennie at 2026-01-12T07:17:49-05:00 Ignore Windows CI tool directories in Git Otherwise, we see thousands of changes in `git status` which is very confusing to work with. - - - - - 023c301c by sheaf at 2026-01-13T04:57:30-05:00 Don't re-use stack slots for growing registers This commit avoids re-using a stack slot for a register that has grown but already had a stack slot. For example, suppose we have stack slot assigments %v1 :: FF64 |-> StackSlot 0 %v2 :: FF64 |-> StackSlot 1 Later, we start using %v1 at a larger format (e.g. F64x2) and we need to spill it again. Then we **must not** use StackSlot 0, as a spill at format F64x2 would clobber the data in StackSlot 1. This can cause some fragmentation of the `StackMap`, but that's probably OK. Fixes #26668 - - - - - d0966e64 by fendor at 2026-01-13T04:58:11-05:00 Remove `traceId` from ghc-pkg executable - - - - - 20d7efec by Simon Peyton Jones at 2026-01-13T12:41:22-05:00 Make SpecContr rules fire a bit later See #26615 and Note [SpecConstr rule activation] - - - - - 8bc4eb8c by Andrew Lelechenko at 2026-01-13T12:42:03-05:00 Upgrade mtl submodule to 2.3.2 Fixes #26656 - - - - - c94aaacd by Cheng Shao at 2026-01-13T12:42:44-05:00 compiler: remove iserv and only use on-demand external interpreter This patch removes `iserv` from the tree completely. Hadrian would no longer build or distribute `iserv`, and the GHC driver would use the on-demand external interpreter by default when invoked with `-fexternal-interpreter`, without needing to specify `-pgmi ""`. This has multiple benefits: - It allows cleanup of a lot of legacy hacks in the hadrian codebase. - It paves the way for running cross ghc's iserv via cross emulator (#25523), fixing TH/ghci support for cross targets other than wasm/js. - - - - - ecc0f11c by Simon Peyton Jones at 2026-01-13T23:56:46+00:00 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 - - - - - 82 changed files: - .gitignore - CODEOWNERS - cabal.project-reinstall - compiler/GHC/Builtin/Names.hs - compiler/GHC/Cmm.hs - compiler/GHC/Cmm/InitFini.hs - compiler/GHC/CmmToAsm/AArch64/Ppr.hs - compiler/GHC/CmmToAsm/Ppr.hs - compiler/GHC/CmmToAsm/Reg/Linear.hs - compiler/GHC/CmmToAsm/Reg/Linear/StackMap.hs - compiler/GHC/CmmToAsm/Wasm/FromCmm.hs - compiler/GHC/CmmToAsm/X86/Ppr.hs - compiler/GHC/CmmToC.hs - compiler/GHC/CmmToLlvm/Data.hs - compiler/GHC/Core/Opt/Simplify/Env.hs - compiler/GHC/Core/Opt/Simplify/Iteration.hs - compiler/GHC/Core/Opt/Simplify/Utils.hs - compiler/GHC/Core/Opt/SpecConstr.hs - compiler/GHC/Core/Opt/Specialise.hs - compiler/GHC/Core/Opt/WorkWrap.hs - compiler/GHC/Core/Rules.hs - compiler/GHC/Core/Unfold/Make.hs - compiler/GHC/Runtime/Interpreter/Init.hs - compiler/GHC/Settings/IO.hs - compiler/GHC/Types/Id/Make.hs - compiler/GHC/Utils/Binary.hs - docs/users_guide/ghci.rst - docs/users_guide/phases.rst - hadrian/doc/user-settings.md - hadrian/src/Packages.hs - hadrian/src/Rules/BinaryDist.hs - hadrian/src/Rules/CabalReinstall.hs - hadrian/src/Rules/Generate.hs - hadrian/src/Rules/Program.hs - hadrian/src/Rules/Test.hs - hadrian/src/Settings/Builders/Ghc.hs - hadrian/src/Settings/Default.hs - hadrian/src/Settings/Packages.hs - libraries/base/changelog.md - libraries/base/src/Data/Bifoldable1.hs - libraries/base/src/GHC/Conc.hs - libraries/base/src/GHC/Conc/Sync.hs - libraries/ghc-internal/ghc-internal.cabal.in - libraries/ghc-internal/src/GHC/Internal/Classes.hs - + libraries/ghc-internal/src/GHC/Internal/Classes/IP.hs - libraries/ghc-internal/src/GHC/Internal/Conc/IO.hs - libraries/ghc-internal/src/GHC/Internal/Conc/POSIX.hs - libraries/ghc-internal/src/GHC/Internal/Conc/Sync.hs - libraries/ghc-internal/src/GHC/Internal/Conc/Sync.hs-boot - libraries/ghc-internal/src/GHC/Internal/Conc/Windows.hs - libraries/ghc-internal/src/GHC/Internal/Event/Thread.hs - libraries/ghc-internal/src/GHC/Internal/Event/Windows/Thread.hs - + libraries/ghc-internal/src/GHC/Internal/STM.hs - libraries/ghc-internal/src/GHC/Internal/System/Environment/ExecutablePath.hsc - libraries/ghc-platform/src/GHC/Platform/ArchOS.hs - libraries/mtl - + testsuite/tests/dmdanal/should_run/T26748.hs - + testsuite/tests/dmdanal/should_run/T26748.stdout - testsuite/tests/dmdanal/should_run/all.T - − testsuite/tests/driver/T24731.hs - testsuite/tests/driver/all.T - 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/simplCore/should_compile/T14003.stderr - testsuite/tests/simplCore/should_compile/T19672.stderr - testsuite/tests/simplCore/should_compile/T21763.stderr - testsuite/tests/simplCore/should_compile/T21763a.stderr - + testsuite/tests/simplCore/should_compile/T26615.hs - + testsuite/tests/simplCore/should_compile/T26615.stderr - + testsuite/tests/simplCore/should_compile/T26615a.hs - + testsuite/tests/simplCore/should_compile/T26682.hs - + testsuite/tests/simplCore/should_compile/T26682a.hs - testsuite/tests/simplCore/should_compile/T4908.stderr - testsuite/tests/simplCore/should_compile/all.T - testsuite/tests/simplCore/should_compile/spec-inline.stderr - + testsuite/tests/typecheck/should_compile/T26737.hs - testsuite/tests/typecheck/should_compile/all.T - utils/ghc-pkg/Main.hs - − utils/iserv/iserv.cabal.in - − utils/iserv/src/Main.hs The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/86cd91adca0c4f269382c525859f70… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/86cd91adca0c4f269382c525859f70… You're receiving this email because of your account on gitlab.haskell.org.

1 0