- ghc-commits - Haskell.org

[Git][ghc/ghc][wip/mangoiv/aarch64-longjump] driver: enable -finter-module-far-jumps by default
by Magnus (＠MangoIV) 11 May '26

11 May '26

Magnus pushed to branch wip/mangoiv/aarch64-longjump at Glasgow Haskell Compiler / GHC Commits: e3df0947 by mangoiv at 2026-05-11T14:22:39+02:00 driver: enable -finter-module-far-jumps by default this fixes a compatibility bug with certain binutils/gcc versions where we were seeing jump offset overflow errors. This commit can probably reverted if we stop supporting the problematic binutils/gcc verions (2.44 and 14.2, respectively) Reolves #26994 - - - - - 3 changed files: - + changelog.d/inter-module-far-jumps-aarch64-default - compiler/GHC/Driver/DynFlags.hs - docs/users_guide/using-optimisation.rst Changes: ===================================== changelog.d/inter-module-far-jumps-aarch64-default ===================================== @@ -0,0 +1,5 @@ +section: driver +synopsis: enable -finter-module-far-jumps by default on aarch64 +issues: #26994 +mrs: !16016 +description: Some binutils and gcc versions lead to jump offset erros on aarch64 profiled/dynamic ways. To mitigate this bug, we enable inter-module-far-jumps by default on this platform. ===================================== compiler/GHC/Driver/DynFlags.hs ===================================== @@ -1233,6 +1233,11 @@ defaultFlags settings Opt_SpecialiseIncoherents, Opt_WriteSelfRecompInfo ] + {- note(mangoiv): + - on aarch64 profiled-dynamic, we are seeing jump offset overflow + - errors on certain binutils versions. This can probably be deactivated + - if we stop supporting binutils 2.44 / GCC 14.2 -} + ++ [ Opt_InterModuleFarJumps | platformArch platform == ArchAArch64 ] ++ [f | (ns,f) <- optLevelFlags, 0 `elem` ns] -- The default -O0 options @@ -1245,6 +1250,7 @@ defaultFlags settings ++ validHoleFitDefaults where platform = sTargetPlatform settings + defaultWays = ways $ defaultDynFlags settings -- | These are the default settings for the display and sorting of valid hole -- fits in typed-hole error messages. See Note [Valid hole fits include ...] ===================================== docs/users_guide/using-optimisation.rst ===================================== @@ -732,7 +732,7 @@ as such you shouldn't need to set any of them explicitly. A flag :reverse: -fno-inter-module-far-jumps :category: - :default: Off + :default: on if the target platform is AArch64, off on all others This flag forces GHC to use far jumps instead of near jumps for all jumps which cross module boundries. This removes the need for jump islands/linker View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/e3df0947028baeaaf14c419763c5570… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/e3df0947028baeaaf14c419763c5570… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/mangoiv/aarch64-longjump] driver: enable -finter-module-far-jumps by default
by Magnus (＠MangoIV) 11 May '26

11 May '26

Magnus pushed to branch wip/mangoiv/aarch64-longjump at Glasgow Haskell Compiler / GHC Commits: 49b7019d by mangoiv at 2026-05-11T14:20:33+02:00 driver: enable -finter-module-far-jumps by default this fixes a compatibility bug with certain binutils/gcc versions where we were seeing jump offset overflow errors. This commit can probably reverted if we stop supporting the problematic binutils/gcc verions (2.44 and 14.2, respectively) Reolves #26994 - - - - - 2 changed files: - + changelog.d/inter-module-far-jumps-aarch64-default - compiler/GHC/Driver/DynFlags.hs Changes: ===================================== changelog.d/inter-module-far-jumps-aarch64-default ===================================== @@ -0,0 +1,4 @@ +section: driver +synopsis: enable -finter-module-far-jumps by default on aarch64 +issues: #26994 +description: Some binutils and gcc versions lead to jump offset erros on aarch64 profiled/dynamic ways. To mitigate this bug, we enable inter-module-far-jumps by default on this platform. ===================================== compiler/GHC/Driver/DynFlags.hs ===================================== @@ -1233,6 +1233,11 @@ defaultFlags settings Opt_SpecialiseIncoherents, Opt_WriteSelfRecompInfo ] + {- note(mangoiv): + - on aarch64 profiled-dynamic, we are seeing jump offset overflow + - errors on certain binutils versions. This can probably be deactivated + - if we stop supporting binutils 2.44 / GCC 14.2 -} + ++ [ Opt_InterModuleFarJumps | platformArch platform == ArchAArch64 ] ++ [f | (ns,f) <- optLevelFlags, 0 `elem` ns] -- The default -O0 options @@ -1245,6 +1250,7 @@ defaultFlags settings ++ validHoleFitDefaults where platform = sTargetPlatform settings + defaultWays = ways $ defaultDynFlags settings -- | These are the default settings for the display and sorting of valid hole -- fits in typed-hole error messages. See Note [Valid hole fits include ...] View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/49b7019d2e3170bf1472dab2e953ef8… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/49b7019d2e3170bf1472dab2e953ef8… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/mangoiv/aarch64-longjump
by Magnus (＠MangoIV) 11 May '26

11 May '26

Magnus pushed new branch wip/mangoiv/aarch64-longjump at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/mangoiv/aarch64-longjump You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/fix-msys2-sysroot-leak
by Cheng Shao (＠TerrorJack) 11 May '26

11 May '26

Cheng Shao pushed new branch wip/fix-msys2-sysroot-leak at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/fix-msys2-sysroot-leak You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/marge_bot_batch_merge_job] 3 commits: Generalize so_inline to specify which bindings should be preserved
by Marge Bot (＠marge-bot) 11 May '26

11 May '26

Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC Commits: 1e60023b by Facundo Domínguez at 2026-05-07T18:01:16-04:00 Generalize so_inline to specify which bindings should be preserved This commit generalizes the so_inline option of the simple optimizer so we can indicate with a predicate the specific bindings that should be kept. This feature is important for the LiquidHaskell plugin, which relies on the simple optimizer to make core programs easier to read, but needs to preserve bindings that are relevant for verification. See https://gitlab.haskell.org/ghc/ghc/-/issues/24386 for the full discussion. - - - - - ba12801b by Wolfgang Jeltsch at 2026-05-11T07:00:10-04:00 Move the `Text.Read` implementation into `base` - - - - - ede968b6 by Alice Rixte at 2026-05-11T07:00:17-04:00 Script for downloading and copying `base-exports` file - - - - - 18 changed files: - + changelog.d/so_inline_is_a_predicate - compiler/GHC/Core/SimpleOpt.hs - compiler/GHC/Driver/Config.hs - libraries/base/src/Data/Functor/Classes.hs - libraries/base/src/Data/Functor/Compose.hs - libraries/base/src/Prelude.hs - libraries/base/src/Text/Read.hs - libraries/ghc-internal/ghc-internal.cabal.in - libraries/ghc-internal/src/GHC/Internal/IO/Encoding.hs - − libraries/ghc-internal/src/GHC/Internal/Text/Read.hs - + testsuite/tests/ghc-api/T24386.hs - testsuite/tests/ghc-api/all.T - + testsuite/tests/interface-stability/.gitignore - testsuite/tests/interface-stability/README.mkd - + testsuite/tests/interface-stability/download-base-exports.sh - testsuite/tests/th/T24111.stdout - testsuite/tests/typecheck/should_compile/subsumption_sort_hole_fits.stderr - testsuite/tests/typecheck/should_fail/T21130.stderr Changes: ===================================== changelog.d/so_inline_is_a_predicate ===================================== @@ -0,0 +1,12 @@ +section: ghc-lib +synopsis: Generalize the ``so_inline`` option of the simple optimizer to a predicate + that selects the bindings to preserve. + +issues: #24386 +mrs: !15988 + +description: { + The ``so_inline`` option of the simple optimizer was a boolean and now it is a + predicate taking a binding ``Id`` and returning a boolean. ``const b`` has the + same effect as formerly setting ``b``. +} ===================================== compiler/GHC/Core/SimpleOpt.hs ===================================== @@ -108,6 +108,93 @@ unfolding-info to the scrutinee's Id.) * Bad bad bad: then the x in case x of ... may be replaced with a version that has an unfolding. See ticket #25790 + +Note [Controlling inlining in the simple optimiser] +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Sometimes, plugins that analyse Core programs may want to prevent the +inlining of certain bindings. While they could avoid running the simple +optimiser at all, that would leave plenty of generated bindings that do not +have a direct correspondence to the source code. + +For example, consider the following Haskell code: + + foo = z + where + z = z1 + z2 + z1 = 42 + z2 = 1 + +Before the simple optimizer runs, the Core programs is roughly: + + foo = + let + foo_aIb = + let + z2 + = let + z2_aHG = 1 + in + z2_aHG + in + let + z1 = + let + z1_aHR = 42 + in + z1_aHR + in + let + z = + let + z_aI5 = z1 + z2 + in + z_aI5 + in + z + in + foo_aIb + +After the simple optimizer runs, the Core program is: + + foo = 42 + 1 + +And the bindings for `z`, `z1`, and `z2` are all gone. If a plugin wanted to +analyse those bindings, it would have to deal with the unsimplified Core, but +cope with the generated bindings `z2_aHG`, `z1_aHR`, `z_aI5`, and `foo_aIb`, +all of which have no direct correspondence to the source code. + +Fortunately, a plugin can still improve the output by using the `so_inline` +field of `SimpleOpts`. The `so_inline` field is a /function/ of type +`(Id -> Bool)` that tells the simple optimiser whether or not to inline the `Id`. +The client of the GHC can thereby control precisely which bindings are inlined +and which are not. For instance, + + simplOptPgm + (defaultSimpleOpts { so_inline = (`notElem` ["z", "z1", "z2"]) }) + ... + +produces the following Core program: + + foo = + let + z2 = 1 + in + let + z1 = 42 + in + let + z = z1 + z2 + in + z + +which contains the bindings of interest and little else. + +For the specifics of how this affects a concrete plugin (Liquid Haskell), see +the discussion in https://gitlab.haskell.org/ghc/ghc/-/issues/24386 + +In addition to supporting clients of the GHC API, there is another use of +`so_inline` mentioned in 'simpleOptExprNoInline'. + -} -- | Simple optimiser options @@ -115,8 +202,11 @@ data SimpleOpts = SimpleOpts { so_uf_opts :: !UnfoldingOpts -- ^ Unfolding options , so_co_opts :: !OptCoercionOpts -- ^ Coercion optimiser options , so_eta_red :: !Bool -- ^ Eta reduction on? - , so_inline :: !Bool -- ^ False <=> do no inlining whatsoever, - -- even for trivial or used-once things + , so_inline :: !(Var -> Bool) -- ^ False <=> do no inline the given + -- binding whatsoever, even for trivial or + -- used-once things + -- + -- See Note [Controlling inlining in the simple optimiser] } -- | Default options for the Simple optimiser. @@ -125,7 +215,7 @@ defaultSimpleOpts = SimpleOpts { so_uf_opts = defaultUnfoldingOpts , so_co_opts = OptCoercionOpts { optCoercionEnabled = False } , so_eta_red = False - , so_inline = True + , so_inline = const True } simpleOptExpr :: HasDebugCallStack => SimpleOpts -> CoreExpr -> CoreExpr @@ -170,7 +260,7 @@ simpleOptExprNoInline :: HasDebugCallStack => SimpleOpts -> CoreExpr -> CoreExpr simpleOptExprNoInline opts expr = simple_opt_expr init_env expr where - init_opts = opts { so_inline = False } + init_opts = opts { so_inline = const False } init_env = (emptyEnv init_opts) { soe_subst = init_subst } init_subst = mkEmptySubst (mkInScopeSet (exprFreeVars expr)) @@ -639,12 +729,12 @@ simple_bind_pair env@(SOE { soe_inl = inl_env, soe_subst = subst, soe_opts = opt pre_inline_unconditionally :: Bool pre_inline_unconditionally - | not (so_inline opts) = False -- Not if so_inline is False - | isExportedId in_bndr = False - | stable_unf = False - | not active = False -- Note [Inline prag in simplOpt] - | not (safe_to_inline occ) = False - | otherwise = True + | not (so_inline opts in_bndr) = False -- Not if so_inline is False + | isExportedId in_bndr = False + | stable_unf = False + | not active = False -- Note [Inline prag in simplOpt] + | not (safe_to_inline occ) = False + | otherwise = True -- Unconditionally safe to inline safe_to_inline :: OccInfo -> Bool @@ -711,15 +801,15 @@ simple_out_bind_pair env@(SOE { soe_subst = subst, soe_opts = opts }) post_inline_unconditionally :: Bool post_inline_unconditionally - | not (so_inline opts) = False -- Not if so_inline is False - | isExportedId in_bndr = False -- Note [Exported Ids and trivial RHSs] - | stable_unf = False -- Note [Stable unfoldings and postInlineUnconditionally] - | not active = False -- in GHC.Core.Opt.Simplify.Utils - | is_loop_breaker = False -- If it's a loop-breaker of any kind, don't inline - -- because it might be referred to "earlier" - | exprIsTrivial out_rhs = True - | coercible_hack = True - | otherwise = False + | not (so_inline opts in_bndr) = False -- Not if so_inline is False + | isExportedId in_bndr = False -- Note [Exported Ids and trivial RHSs] + | stable_unf = False -- Note [Stable unfoldings and postInlineUnconditionally] + | not active = False -- in GHC.Core.Opt.Simplify.Utils + | is_loop_breaker = False -- If it's a loop-breaker of any kind, don't inline + -- because it might be referred to "earlier" + | exprIsTrivial out_rhs = True + | coercible_hack = True + | otherwise = False is_loop_breaker = isWeakLoopBreaker occ_info ===================================== compiler/GHC/Driver/Config.hs ===================================== @@ -26,7 +26,7 @@ initSimpleOpts dflags = SimpleOpts { so_uf_opts = unfoldingOpts dflags , so_co_opts = initOptCoercionOpts dflags , so_eta_red = gopt Opt_DoEtaReduction dflags - , so_inline = True + , so_inline = const True } -- | Instruct the interpreter evaluation to break... ===================================== libraries/base/src/Data/Functor/Classes.hs ===================================== @@ -85,7 +85,7 @@ import GHC.Internal.Read (expectP, list, paren, readField) import GHC.Internal.Show (appPrec) import GHC.Internal.Text.ParserCombinators.ReadPrec (ReadPrec, readPrec_to_S, readS_to_Prec, pfail) -import GHC.Internal.Text.Read (Read(..), parens, prec, step, reset) +import Text.Read (Read(..), parens, prec, step, reset) import GHC.Internal.Text.Read.Lex (Lexeme(..)) import GHC.Internal.Text.Show (showListWith) import Prelude ===================================== libraries/base/src/Data/Functor/Compose.hs ===================================== @@ -35,7 +35,7 @@ import GHC.Internal.Data.Foldable (Foldable(..)) import GHC.Internal.Data.Monoid (Sum(..), All(..), Any(..), Product(..)) import GHC.Internal.Data.Type.Equality (TestEquality(..), (:~:)(..)) import GHC.Generics (Generic, Generic1) -import GHC.Internal.Text.Read (Read(..), ReadPrec, readListDefault, readListPrecDefault) +import Text.Read (Read(..), ReadPrec, readListDefault, readListPrecDefault) import Prelude infixr 9 `Compose` ===================================== libraries/base/src/Prelude.hs ===================================== @@ -179,7 +179,7 @@ import GHC.Internal.Data.Tuple import GHC.Internal.Base hiding ( foldr, mapM, sequence ) import GHC.Internal.Classes import GHC.Internal.Err -import GHC.Internal.Text.Read +import Text.Read import GHC.Internal.Enum import GHC.Internal.Num import GHC.Internal.Prim (seq) ===================================== libraries/base/src/Text/Read.hs ===================================== @@ -39,5 +39,84 @@ module Text.Read readMaybe ) where -import GHC.Internal.Text.Read +import GHC.Err (errorWithoutStackTrace) +import GHC.Read + ( + ReadS, + Read (readsPrec, readList, readPrec, readListPrec), + lex, + readParen, + readListDefault, + lexP, + parens, + readListPrecDefault + ) +import Control.Monad (return) +import Data.Function (id) +import Data.Maybe (Maybe (Nothing, Just)) +import Data.Either (Either (Left, Right), either) +import Data.String (String) +import Text.Read.Lex (Lexeme (Char, String, Punc, Ident, Symbol, Number, EOF)) +import Text.ParserCombinators.ReadP (skipSpaces) import Text.ParserCombinators.ReadPrec + +-- $setup +-- >>> import Prelude + +------------------------------------------------------------------------ +-- utility functions + +-- | equivalent to 'readsPrec' with a precedence of 0. +reads :: Read a => ReadS a +reads = readsPrec minPrec + +-- | Parse a string using the 'Read' instance. +-- Succeeds if there is exactly one valid result. +-- A 'Left' value indicates a parse error. +-- +-- >>> readEither "123" :: Either String Int +-- Right 123 +-- +-- >>> readEither "hello" :: Either String Int +-- Left "Prelude.read: no parse" +-- +-- @since base-4.6.0.0 +readEither :: Read a => String -> Either String a +readEither s = + case [ x | (x,"") <- readPrec_to_S read' minPrec s ] of + [x] -> Right x + [] -> Left "Prelude.read: no parse" + _ -> Left "Prelude.read: ambiguous parse" + where + read' = + do x <- readPrec + lift skipSpaces + return x + +-- | Parse a string using the 'Read' instance. +-- Succeeds if there is exactly one valid result. +-- +-- >>> readMaybe "123" :: Maybe Int +-- Just 123 +-- +-- >>> readMaybe "hello" :: Maybe Int +-- Nothing +-- +-- @since base-4.6.0.0 +readMaybe :: Read a => String -> Maybe a +readMaybe s = case readEither s of + Left _ -> Nothing + Right a -> Just a + +-- | The 'read' function reads input from a string, which must be +-- completely consumed by the input process. 'read' fails with an 'error' if the +-- parse is unsuccessful, and it is therefore discouraged from being used in +-- real applications. Use 'readMaybe' or 'readEither' for safe alternatives. +-- +-- >>> read "123" :: Int +-- 123 +-- +-- >>> read "hello" :: Int +-- *** Exception: Prelude.read: no parse +read :: Read a => String -> a +read s = either errorWithoutStackTrace id (readEither s) ===================================== libraries/ghc-internal/ghc-internal.cabal.in ===================================== @@ -329,7 +329,6 @@ Library GHC.Internal.System.Posix.Types GHC.Internal.Text.ParserCombinators.ReadP GHC.Internal.Text.ParserCombinators.ReadPrec - GHC.Internal.Text.Read GHC.Internal.Text.Read.Lex GHC.Internal.Text.Show GHC.Internal.Type.Reflection ===================================== libraries/ghc-internal/src/GHC/Internal/IO/Encoding.hs ===================================== @@ -46,7 +46,7 @@ import GHC.Internal.IO.Encoding.Types import qualified GHC.Internal.IO.Encoding.Iconv as Iconv #else import qualified GHC.Internal.IO.Encoding.CodePage as CodePage -import GHC.Internal.Text.Read (reads) +import GHC.Internal.Read (readsPrec) #endif import qualified GHC.Internal.IO.Encoding.Latin1 as Latin1 import qualified GHC.Internal.IO.Encoding.UTF8 as UTF8 @@ -319,7 +319,8 @@ mkTextEncoding' cfm enc = _ | isAscii -> return (Latin1.mkAscii cfm) _ | isLatin1 -> return (Latin1.mkLatin1_checked cfm) #if defined(mingw32_HOST_OS) - 'C':'P':n | [(cp,"")] <- reads n -> return $ CodePage.mkCodePageEncoding cfm cp + 'C':'P':n | [(cp,"")] <- readsPrec 0 n -> return $ CodePage.mkCodePageEncoding cfm cp + -- 'readsPrec 0' is the same as 'reads', but 'reads' is only defined in @base@. _ -> unknownEncodingErr (enc ++ codingFailureModeSuffix cfm) #else -- Otherwise, handle other encoding needs via iconv. ===================================== libraries/ghc-internal/src/GHC/Internal/Text/Read.hs deleted ===================================== @@ -1,115 +0,0 @@ -{-# LANGUAGE Trustworthy #-} -{-# LANGUAGE NoImplicitPrelude #-} - ------------------------------------------------------------------------------ --- | --- Module : GHC.Internal.Text.Read --- Copyright : (c) The University of Glasgow 2001 --- License : BSD-style (see the file libraries/base/LICENSE) --- --- Maintainer : libraries(a)haskell.org --- Stability : provisional --- Portability : non-portable (uses Text.ParserCombinators.ReadP) --- --- Converting strings to values. --- --- The "Text.Read" library is the canonical library to import for --- 'Read'-class facilities. For GHC only, it offers an extended and much --- improved 'Read' class, which constitutes a proposed alternative to the --- Haskell 2010 'Read'. In particular, writing parsers is easier, and --- the parsers are much more efficient. --- ------------------------------------------------------------------------------ - -module GHC.Internal.Text.Read ( - -- * The 'Read' class - Read(..), - ReadS, - - -- * Haskell 2010 functions - reads, - read, - readParen, - lex, - - -- * New parsing functions - module GHC.Internal.Text.ParserCombinators.ReadPrec, - L.Lexeme(..), - lexP, - parens, - readListDefault, - readListPrecDefault, - readEither, - readMaybe - - ) where - -import GHC.Internal.Base (String, id, return) -import GHC.Internal.Err (errorWithoutStackTrace) -import GHC.Internal.Maybe (Maybe(..)) -import GHC.Internal.Read -import GHC.Internal.Data.Either -import GHC.Internal.Text.ParserCombinators.ReadP as P -import GHC.Internal.Text.ParserCombinators.ReadPrec -import qualified GHC.Internal.Text.Read.Lex as L - --- $setup --- >>> import Prelude - ------------------------------------------------------------------------- --- utility functions - --- | equivalent to 'readsPrec' with a precedence of 0. -reads :: Read a => ReadS a -reads = readsPrec minPrec - --- | Parse a string using the 'Read' instance. --- Succeeds if there is exactly one valid result. --- A 'Left' value indicates a parse error. --- --- >>> readEither "123" :: Either String Int --- Right 123 --- --- >>> readEither "hello" :: Either String Int --- Left "Prelude.read: no parse" --- --- @since base-4.6.0.0 -readEither :: Read a => String -> Either String a -readEither s = - case [ x | (x,"") <- readPrec_to_S read' minPrec s ] of - [x] -> Right x - [] -> Left "Prelude.read: no parse" - _ -> Left "Prelude.read: ambiguous parse" - where - read' = - do x <- readPrec - lift P.skipSpaces - return x - --- | Parse a string using the 'Read' instance. --- Succeeds if there is exactly one valid result. --- --- >>> readMaybe "123" :: Maybe Int --- Just 123 --- --- >>> readMaybe "hello" :: Maybe Int --- Nothing --- --- @since base-4.6.0.0 -readMaybe :: Read a => String -> Maybe a -readMaybe s = case readEither s of - Left _ -> Nothing - Right a -> Just a - --- | The 'read' function reads input from a string, which must be --- completely consumed by the input process. 'read' fails with an 'error' if the --- parse is unsuccessful, and it is therefore discouraged from being used in --- real applications. Use 'readMaybe' or 'readEither' for safe alternatives. --- --- >>> read "123" :: Int --- 123 --- --- >>> read "hello" :: Int --- *** Exception: Prelude.read: no parse -read :: Read a => String -> a -read s = either errorWithoutStackTrace id (readEither s) ===================================== testsuite/tests/ghc-api/T24386.hs ===================================== @@ -0,0 +1,123 @@ + +-- This test checks that bindings are preserved when configuring the simple +-- optimizer to not inline bindings with names selected by a predicate. +-- +-- This feature is important for the LiquidHaskell plugin, which relies on the +-- simple optimizer to make core programs easier to read, but needs to preserve +-- bindings that are relevant for verification. +-- +-- See https://gitlab.haskell.org/ghc/ghc/-/issues/24386 for the full discussion. +-- + +import Control.Monad +import Data.List (find) +import Data.Time (getCurrentTime) +import GHC +import GHC.Core +import GHC.Core.SimpleOpt +import GHC.Data.StringBuffer +import GHC.Driver.Config +import GHC.Driver.DynFlags +import GHC.Driver.Env.Types +import GHC.Types.Name +import GHC.Unit.Module.ModGuts +import GHC.Unit.Types +import GHC.Utils.Error +import GHC.Utils.Outputable + +import System.Environment (getArgs) + + +main :: IO () +main = + testLocalBindingsDesugaring + +testLocalBindingsDesugaring :: IO () +testLocalBindingsDesugaring = do + let inputSource = unlines + [ "module LocalBindingsDesugaring where" + , "f :: ()" + , "f = z" + , " where" + , " z = ()" + ] + + isExpectedDesugaring p = case findExpr "f" p of + Just (Let (NonRec b _) _) + -> isIdNamed "z" b + _ -> False + + isIdNamed name v = occNameString (occName v) == name + + coreProgram <- + compileToCore + (not . isIdNamed "z") + "LocalBindingsDesugaring" + inputSource + unless (isExpectedDesugaring coreProgram) $ + fail $ unlines $ + "Unexpected desugaring: No local binding for `z` found in the Core program." + : map showPprQualified coreProgram + +-- | Find the Core expression bound to the given name. +findExpr :: String -> CoreProgram -> Maybe CoreExpr +findExpr _ [] = + Nothing +findExpr name (p:ps) = case p of + NonRec b e + | occNameString (occName b) == name + -> Just e + Rec binds + | Just (_, e) <- find (\(b, _e) -> occNameString (occName b) == name) binds + -> Just e + _ -> findExpr name ps + +showPprQualified :: Outputable a => a -> String +showPprQualified = showSDocQualified . ppr + +showSDocQualified :: SDoc -> String +showSDocQualified = renderWithContext ctx + where + ctx = defaultSDocContext { sdocStyle = cmdlineParserStyle } + + + +compileToCore :: (Id -> Bool) -> String -> String -> IO [CoreBind] +compileToCore keepBindings modName inputSource = do + [libdir] <- getArgs + now <- getCurrentTime + runGhc (Just libdir) $ do + df1 <- getSessionDynFlags + GHC.setSessionDynFlags $ df1 { GHC.backend = GHC.bytecodeBackend } + let target = Target { + targetId = TargetFile (modName ++ ".hs") Nothing + , targetUnitId = homeUnitId_ df1 + , targetAllowObjCode = False + , targetContents = Just (stringToStringBuffer inputSource, now) + } + setTargets [target] + void $ GHC.depanal [] False + + dsMod <- getModSummary + (mkModule mainUnit (mkModuleName modName)) + >>= parseModule + >>= typecheckModule NoTcMPlugins + >>= desugarModule + hsc_env <- getSession + return $ mg_binds $ simpleOptimize keepBindings hsc_env $ dm_core_module dsMod + +-- Run the simple optimizer +simpleOptimize :: (Id -> Bool) -> GHC.HscEnv -> ModGuts -> ModGuts +simpleOptimize keepBindings hsc_env guts@(ModGuts + { mg_module = mgmod + , mg_binds = binds + , mg_rules = rules + }) = + let dflags = hsc_dflags hsc_env + simpl_opts = (initSimpleOpts dflags) { so_inline = keepBindings } + (binds2, rules2, _occ_anald_binds) = + simpleOptPgm simpl_opts mgmod binds rules + in guts + { mg_binds = binds2 + , mg_rules = rules2 + } ===================================== testsuite/tests/ghc-api/all.T ===================================== @@ -81,3 +81,4 @@ test('T26910', [ extra_run_opts(f'"{config.libdir}"') test('TypeMapStringLiteral', normal, compile_and_run, ['-package ghc']) test('T25121_status', normal, compile_and_run, ['-package ghc']) +test('T24386', [extra_run_opts(f'"{config.libdir}"')], compile_and_run, ['-package ghc']) ===================================== testsuite/tests/interface-stability/.gitignore ===================================== @@ -0,0 +1 @@ +download-base-exports ===================================== testsuite/tests/interface-stability/README.mkd ===================================== @@ -1,6 +1,6 @@ # Interface stability testing -The tests in this directory verify that the interfaces of exposed by GHC's +The tests in this directory verify that the interfaces exposed by GHC's core libraries do not inadvertently change. They use the `utils/dump-decls` utility to dump all exported declarations of all exposed modules for the following packages: @@ -27,7 +27,9 @@ The `base-exports` test in particular has rather platform-dependent output. Consequently, updating its output can be a bit tricky. There are two ways by which one can do this: - * Extrapolation: The various platforms' `base-exports.stdout` files are +#### Extrapolation + +The various platforms' `base-exports.stdout` files are similar enough that one can often apply the same patch of one file to the others. For instance: ``` @@ -40,8 +42,44 @@ which one can do this: In the case of conflicts, increasing the fuzz factor (using `-F`) can be quite effective. - * Using CI: Each CI job produces a tarball, `unexpected-test-output.tar.gz`, +#### Using CI + +Each CI job produces a tarball, `unexpected-test-output.tar.gz`, which contains the output produced by the job's failing tests. Simply - download this tarball and extracting the appropriate `base-exports.stdout-*` + download this tarball and extract the appropriate `base-exports.stdout-*` files into this directory. +Doing this by hand is of course very annoying. To make things faster, use the script in this folder called `download.base-exports.sh` : + +* Running for the first time + 1. Find the URL for downloading unexpected-test-output.tar.gz. To do so + * Go to the CI job page you want to download + * Click on "Browse" + * Find unexpected-test-output.tar.gz + * Right-click the download link then "Copy link" (Firefox) + 2. The URL should look like this : + `https://gitlab.haskell.org/ghc/ghc/-/jobs/2503744/artifacts/file/unexpected-test-output.tar.gz` + * the prefix is : `https://gitlab.haskell.org/ghc/ghc/-/jobs/` + * the job ID is : `2503744` + * and the suffix : `/artifacts/file/unexpected-test-output.tar.gz` + 3. The script prompts you with URL prefix and suffix. + 4. It will save a file to remember this, so you only need to do this once. + 5. If you need to change the URL, just edit the file `download-base-exports/url-unexpected-test-output` directly. + +* Downloading the artifacts + 1. Find all the job IDs you want to download. For this, just go to the jobs + page `https://gitlab.haskell.org/<YOUR-FORK>/ghc/-/jobs` + 2. Make sure you get all the artifacts. You need 3 of them. + To get all 3 CI jobs, the label `javascript` must be on the MR. + If you don't have the rights for adding these labels, ask. + 1. The `x86` CI job for darwin or linux : `base-exports.stdout` + 2. The `windows` job : `base-exports.stdout-mingw32` + 3. The `javascript` CI job : + `base-exports.stdout-javascript-unknown-ghcjs` + 3. Run the script with all the job IDs : + `./download-base-exports.sh 2502789 2502792 2502793` + + Using a range downloads more artifacts than necessary, but is a + no-brainer: + + `./download-base-exports.sh {2502789..2502795}` ===================================== testsuite/tests/interface-stability/download-base-exports.sh ===================================== @@ -0,0 +1,55 @@ +#!/usr/bin/env bash + +# See the README file in this folder for usage + +jobIDs=("$@") + +BASE_DIR_NAME=download-base-exports +DL_DIR_NAME=dl +BASE_DIR="$(dirname "$0")/$BASE_DIR_NAME" +DL_DIR=$BASE_DIR/$DL_DIR_NAME +URL_FILE="$BASE_DIR/url-unexpected-test-output" + +DEFAULT_PREFIX="https://gitlab.haskell.org/ghc/ghc/-/jobs/" +DEFAULT_POSTFIX="/artifacts/raw/unexpected-test-output.tar.gz" + +mkdir -p "$BASE_DIR" + +# URL configuration for finding unexpected-test-output.tar.gz + +if [[ ! -f "$URL_FILE" ]]; then + echo "No URL for unexpected-test-output.tar.gz was found" + + read -p "Enter job URL prefix [${DEFAULT_PREFIX}]: " inputPrefix + read -p "Enter job URL postfix [${DEFAULT_POSTFIX}]: " inputPostfix + + urlPrefix="${inputPrefix:-$DEFAULT_PREFIX}" + urlPostfix="${inputPostfix:-$DEFAULT_POSTFIX}" + + { + echo "urlPrefix=$urlPrefix" + echo "urlPostfix=$urlPostfix" + } > "$URL_FILE" +else + source "$URL_FILE" +fi + +mkdir -p $DL_DIR + +echo "urlPrefix: $urlPrefix" +echo "jobIDs: $jobIDs" +echo "urlPostfix: $urlPostfix" +echo "" +echo "Downloading unexpected-test-output.tar.gz for each job ..." + +# Download and copy base-exports* files + +for jobID in "${jobIDs[@]}"; do + unexpectedOutputUrl="$urlPrefix$jobID$urlPostfix" + + wget -O "$DL_DIR/job$jobID.tar.gz" $unexpectedOutputUrl + + mkdir -p "$DL_DIR/job$jobID" + tar -xzf "$DL_DIR/job$jobID.tar.gz" -C "$DL_DIR/job$jobID" + cp "$DL_DIR/job$jobID"/unexpected-test-output/testsuite/tests/interface-stability/base-exports* "$BASE_DIR/.." +done ===================================== testsuite/tests/th/T24111.stdout ===================================== @@ -3,6 +3,6 @@ pattern (:+_0) :: GHC.Internal.Types.Int -> (GHC.Internal.Types.Int, GHC.Internal.Types.Int) pattern x_1 :+_0 y_2 = (x_1, y_2) pattern A_0 :: GHC.Internal.Types.Int -> GHC.Internal.Base.String -pattern A_0 n_1 <- (GHC.Internal.Text.Read.read -> n_1) where +pattern A_0 n_1 <- (Text.Read.read -> n_1) where A_0 0 = "hi" A_0 1 = "bye" ===================================== testsuite/tests/typecheck/should_compile/subsumption_sort_hole_fits.stderr ===================================== @@ -11,14 +11,13 @@ subsumption_sort_hole_fits.hs:2:5: warning: [GHC-88464] [-Wtyped-holes (in -Wdef words :: String -> [String] (imported from ‘Prelude’ (and originally defined in ‘GHC.Internal.Data.OldList’)) - read :: forall a. Read a => String -> a - with read @[String] - (imported from ‘Prelude’ - (and originally defined in ‘GHC.Internal.Text.Read’)) repeat :: forall a. a -> [a] with repeat @String (imported from ‘Prelude’ (and originally defined in ‘GHC.Internal.List’)) + read :: forall a. Read a => String -> a + with read @[String] + (imported from ‘Prelude’ (and originally defined in ‘Text.Read’)) mempty :: forall a. Monoid a => a with mempty @(String -> [String]) (imported from ‘Prelude’ ===================================== testsuite/tests/typecheck/should_fail/T21130.stderr ===================================== @@ -6,6 +6,9 @@ T21130.hs:10:6: error: [GHC-88464] In an equation for ‘x’: x = (_ f) :: Int • Relevant bindings include x :: Int (bound at T21130.hs:10:1) Valid hole fits include + read :: forall a. Read a => String -> a + with read @Int + (imported from ‘Prelude’ (and originally defined in ‘Text.Read’)) head :: forall a. GHC.Internal.Stack.Types.HasCallStack => [a] -> a with head @Int (imported from ‘Prelude’ @@ -14,10 +17,6 @@ T21130.hs:10:6: error: [GHC-88464] with last @Int (imported from ‘Prelude’ (and originally defined in ‘GHC.Internal.List’)) - read :: forall a. Read a => String -> a - with read @Int - (imported from ‘Prelude’ - (and originally defined in ‘GHC.Internal.Text.Read’)) T21130.hs:10:8: error: [GHC-39999] • Ambiguous type variable ‘t0’ arising from a use of ‘f’ View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/793302236da386b872e616859a7e78… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/793302236da386b872e616859a7e78… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/spj-reinstallable-base2] 8 commits: Better handling of krepStar etc
by Rodrigo Mesquita (＠alt-romes) 11 May '26

11 May '26

Rodrigo Mesquita pushed to branch wip/spj-reinstallable-base2 at Glasgow Haskell Compiler / GHC Commits: 0bf8b3ed by Simon Peyton Jones at 2026-05-08T16:38:04+01:00 Better handling of krepStar etc See https://gitlab.haskell.org/ghc/ghc/-/merge_requests/15899#note_673935 Comments and notes Also, happens to, Fix #14480 TODO: Finish this patch and merge it first clean Mv instance get rid binary.typeable cleanup - - - - - c9ae2411 by Rodrigo Mesquita at 2026-05-08T16:44:54+01:00 fixup! Better handling of krepStar etc - - - - - 08091e50 by Rodrigo Mesquita at 2026-05-08T17:25:29+01:00 The problem with unsafeCoerce# is probably the tcLookupGlobal changed to tcLookupGlobalKnownKeyName? I am unsure - - - - - 95664bfb by Rodrigo Mesquita at 2026-05-11T09:58:45+01:00 Revert "unsafeCoercePrimName" This reverts commit 25497cb478856fc82cdc5277f85eebb66624a619. - - - - - 05a31853 by Rodrigo Mesquita at 2026-05-11T10:42:44+01:00 Merge G.I.CString into G.I.Types - - - - - a77cdf38 by Rodrigo Mesquita at 2026-05-11T10:49:04+01:00 fix duplicate name. Bytes is not exported. OK - - - - - 5e83557e by Rodrigo Mesquita at 2026-05-11T11:39:20+01:00 fixes - - - - - 1d17216c by Rodrigo Mesquita at 2026-05-11T11:59:11+01:00 progress starKindRepName - - - - - 39 changed files: - compiler/GHC/Builtin/KnownKeys.hs - compiler/GHC/Builtin/KnownOccs.hs - compiler/GHC/Builtin/Modules.hs - compiler/GHC/HsToCore.hs - compiler/GHC/Iface/Syntax.hs - compiler/GHC/Rename/Expr.hs - compiler/GHC/Tc/Instance/Typeable.hs - compiler/GHC/Tc/Module.hs - compiler/GHC/Tc/Solver/Dict.hs - compiler/GHC/Tc/Solver/Monad.hs - compiler/GHC/Tc/Utils/Concrete.hs - compiler/GHC/Utils/Binary.hs - − compiler/GHC/Utils/Binary/Typeable.hs - compiler/ghc.cabal.in - libraries/base/src/GHC/Base.hs - libraries/base/src/GHC/Essentials.hs - libraries/ghc-internal/ghc-internal.cabal.in - libraries/ghc-internal/src/GHC/Internal/Base.hs - libraries/ghc-internal/src/GHC/Internal/Bignum/Backend/Check.hs - libraries/ghc-internal/src/GHC/Internal/Bignum/Backend/FFI.hs - libraries/ghc-internal/src/GHC/Internal/Bignum/Backend/GMP.hs - libraries/ghc-internal/src/GHC/Internal/Bignum/Backend/Native.hs - − libraries/ghc-internal/src/GHC/Internal/CString.hs - libraries/ghc-internal/src/GHC/Internal/Classes.hs - libraries/ghc-internal/src/GHC/Internal/Control/Exception/Base.hs - libraries/ghc-internal/src/GHC/Internal/Data/Typeable/Internal.hs - libraries/ghc-internal/src/GHC/Internal/Encoding/UTF8.hs - libraries/ghc-internal/src/GHC/Internal/Exts.hs - libraries/ghc-internal/src/GHC/Internal/IO/Exception.hs - libraries/ghc-internal/src/GHC/Internal/JS/Prim.hs - libraries/ghc-internal/src/GHC/Internal/Pack.hs - libraries/ghc-internal/src/GHC/Internal/Show.hs - libraries/ghc-internal/src/GHC/Internal/System/Posix/Internals.hs - libraries/ghc-internal/src/GHC/Internal/TH/Lift.hs - libraries/ghc-internal/src/GHC/Internal/Types.hs - libraries/ghc-prim/ghc-prim.cabal - rts/include/rts/RtsToHsIface.h - testsuite/tests/count-deps/CountDepsAst.stdout - testsuite/tests/count-deps/CountDepsParser.stdout The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/25497cb478856fc82cdc5277f85eeb… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/25497cb478856fc82cdc5277f85eeb… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T26989] 20 commits: Use __attribute__((dllimport)) for external RTS symbol declarations
by Simon Peyton Jones (＠simonpj) 11 May '26

11 May '26

Simon Peyton Jones pushed to branch wip/T26989 at Glasgow Haskell Compiler / GHC Commits: 9a9ae4df by Duncan Coutts at 2026-05-05T14:44:37-04:00 Use __attribute__((dllimport)) for external RTS symbol declarations This is needed to be hygenic about DLL symbol imports and exports. The attribute is ignored on platforms other than Windows. Use of the attribute however means that external data symbols do not have a compile-time constant address (they are loaded using an indirection). This means we have to adjust the rtsSyms initial linker table so that it is a local constant in a function, rather than a global constant. We now define it within a function that pre-populates the symbol table with the RTS symbols. - - - - - 2ad3e01e by Duncan Coutts at 2026-05-05T14:44:37-04:00 Fix the rts linker declarations for a few data symbols and ensure that the (windows only) rts_IOManagerIsWin32Native data symbol is marked as externally visible. - - - - - 8ff4fdb5 by David Eichmann at 2026-05-05T14:44:37-04:00 Hadrian: Disable runtime pseudo relocations for RTS on windows hosts - - - - - 96974723 by Teo Camarasu at 2026-05-05T14:45:20-04:00 ghci/TH: refactor to use IORef QState This is a pure refactor and shouldn't modify semantics at all - - - - - eff6bfaf by Teo Camarasu at 2026-05-05T14:45:20-04:00 iserv: recover/getQ/putQ should behave same as internal interpreter The internal and external interpreter should behave the same when handling `recover`, the exeception recovery method of Q. In practice, they diverge. In case of failure, the internal interpreter only restores error message state to before the computation, wheras the external interperter restores error message state *and* the state of putQ/getQ. As far as I can tell this is a simple mistake in the implementation. Note [TH recover with -fexternal-interpreter] describes the correct behaviour but the implementation doesn't mirror this. This change restores the correct behaviour by keeping the effects of putQ in the erroring computation. This is a breaking change since it modifies the behaviour of programs that rely on recover ignoring putQ from failling computations when used with the external interpreter. Although I highly doubt anyone relies on this behaviour. This divergence was first introduced in d00c308633fe7d216d31a1087e00e63532d87d6d. As far as I can tell this was unintentional and tha commit was trying to solve a different bug. Resolves #27022 - - - - - 1cb1d672 by Wen Kokke at 2026-05-06T09:53:40-04:00 rts: Add dynamic trace flags API This commit adds an API to the RTS (exposed via Rts.h) that allows users to dynamically change the trace flags. Prior to this commit, users were able to stop and start the profiling and heap profiling timers (via startProfTimer/stopProfTimer and startHeapProfTimer/stopHeapProfTimer). This extends that functionality to also cover the core event types. The getTraceFlag/setTraceFlag functions read and write the values of the trace flag cache, which is allocated by Trace.c, rather than modifying the members of RtsFlags.TraceFlags. This is done under the assumption that the members of RtsFlags should not be modified after RTS initialisation. Consequently, if the user modifies the trace flags using setTraceFlag, the object returned by getTraceFlags (from base) will not reflect these changes. The trace flags are not protected by locks of any sort. Hence, these functions are not thread-safe. However, the trace flags are not modified by the RTS after initialisation, only read, so the race conditions introduced by one user modifying them are most likely benign. This PR also puts the trace flag cache in a single global struct, as opposed to a collection of global variables, and changes the types of the individual flags from uint8_t to bool, as these have the same size on both Clang and GCC and are a better semantic match. Prior to the change to uint8_t, they had type int, see 42c47cd6. Even with its deprecation in C23, I don't think there should be any issue depending on stdbool.h. The TRACE_X macros are redefined to access the global struct, with values cast to const bool to ensure they are read-only. - - - - - 9d54dc94 by Wen Kokke at 2026-05-06T09:53:40-04:00 rts: Ensure TRACE_X values are used in place of RtsFlags.TraceFlags.X - - - - - 418d737b by Wen Kokke at 2026-05-06T09:53:40-04:00 rts: Fix nonmoving-GC tracing The current nonmoving-GC tracing functions were written in a different style from the other tracing functions. They were directly implemented as, e.g., a traceConcMarkEnd function that called postConcMarkEnd. The other tracing functions are implemented as, e.g., traceThreadLabel_, a function that posts the thread label event, and traceThreadLabel, a macro that checks whether TRACE_scheduler is set. This commit fixes that implementation, and ensures that the nonmoving-GC tracing functions only emit events if nonmoving-GC tracing is enabled. - - - - - 99f4afa4 by Wen Kokke at 2026-05-06T09:53:40-04:00 rts: Add SymI_HasProto for get/setTraceFlag - - - - - 7e9eb8b9 by Wen Kokke at 2026-05-06T09:53:40-04:00 rts: Add SymI_HasProto for start/endEventLogging - - - - - 3a3045fb by Wen Kokke at 2026-05-06T09:53:41-04:00 rts: Add changelog entry - - - - - a3b339a4 by Teo Camarasu at 2026-05-06T09:54:25-04:00 interface-stability/base: don't distinguish ws-32 The interface of base is identical when the Word size is 32bits. Therefore, there is no need to have another file for this case. So, we delete it. Step towards: #26752 - - - - - eb922183 by Duncan Coutts at 2026-05-07T14:28:50+01:00 Add a rts posix FdWakup utility module This will be used to implement wakeupIOManager for in-RTS I/O managers. It provides a notification/wakeup mechanism using FDs, suitable for situations when a thread is blocked on a set of fds anyway. It uses the classic self-pipe trick, or equivalently eventfd on supported platforms. This will initially be used to implement prompt interrupt or shutdown of the posix ticker thread. - - - - - 01b0e233 by Duncan Coutts at 2026-05-07T14:28:50+01:00 Add prompt shutdown to the pthread ticker implementation. The Linux timerfd ticker monitors a pipe which is used by exitTicker to ensure a prompt wakeup and shutdown. The pthread ticker lacked this and so would only exit at the next ticker wakeup (10ms by default). This patch adds the same mechanism to the pthread ticker. This changes the pthread ticker from waiting by using nanosleep() to waiting using either ppoll() or select(), so that it can wait on both a time and a file descriptor. On Linux at least, a test program to compare the timing jitter of these APIs shows that using nanpsleep, ppoll or select makes no statistical difference to the maximum or average jitter. This is a step towards unifying the posix ticker implementations, so that we can have just one portable one (albeit with some limited cpp). It is also a step towards using the ticker as part of a more general implementation of wakeUpRts, since this will require a method to wake the rts from a signal handler context (ctl-c handler). - - - - - bc41d646 by Duncan Coutts at 2026-05-07T14:28:50+01:00 Update ticker header commentary It was antique and didn't apply even to the previous implementation, and certainly not to the updated one. - - - - - 4ed9a386 by Duncan Coutts at 2026-05-07T14:28:50+01:00 Remove the timerfd-based ticker implementation There does not appear to be any remaining advantage on Linux to using the timerfd ticker implementation over the portable one (using ppoll on Linux for precise timing). The eventfd implementation was originally added at a time when Linux was still using a signal based implementation. So it made sense at the time. See (closed) issue #10840. - - - - - 97504fa6 by Duncan Coutts at 2026-05-07T14:28:50+01:00 Consolidate to a single posix ticker implementation Previously we had four implementations, two using signals and two using threads. Having just one should make behaviour more consistent between platforms, and should make maintenance easier. - - - - - 1e60023b by Facundo Domínguez at 2026-05-07T18:01:16-04:00 Generalize so_inline to specify which bindings should be preserved This commit generalizes the so_inline option of the simple optimizer so we can indicate with a predicate the specific bindings that should be kept. This feature is important for the LiquidHaskell plugin, which relies on the simple optimizer to make core programs easier to read, but needs to preserve bindings that are relevant for verification. See https://gitlab.haskell.org/ghc/ghc/-/issues/24386 for the full discussion. - - - - - d61c2bbe by Simon Peyton Jones at 2026-05-11T11:41:05+01:00 Do not use mkCast during typechecking This commit fixes #27219. The problem was that the typechecker was using `mkCast`, whose assertion checks legitimately fail when applied to types that contain unification variables. - - - - - 4093ea81 by Simon Peyton Jones at 2026-05-11T11:41:05+01:00 Major refactor of the Simplifier The main payload of this patch is to refactor the Simplifer to avoid repeated simplification when using Plan (AFTER) for rule rewrites. The need for this was shown up by #26989. See Note [Avoid repeated simplification] in GHC.Core.Opt.Simplify.Iteration. Related refactoring: * Refactor the two fields `sc_dup` and `sc_env` in `ApplyToVal` into one, `sc_env`. Reason: the envt is irrelevant in the "simplified" case, so the data type describes the possiblitiies much more accurately now. * Some refactoring in `knownCon` to split off `wrapDataConFloats`. * Refactor `lookupRule` and its auxiliary functions to return `RuleMatch`, a new data type. See Note [data RuleMatch] in GHC.Core. Ditto for BuiltinRule. This RuleMatch returns fragments of the target in rm_args and rm_floats, leaving `rm_rhs` to be the stuff from the RULE itself. Doing this has routine consequences in GHC.Core.Opt.ConstantFold. Many changes there but all routine. * When doing occurrence analysis on RULEs, make the occ-info on the rule binders relate just to the RHS, not the LHS. See (OUR1) in Note Note [OccInfo in unfoldings and rules] This means that Lint must not complain about the fact that the patterns in the RULE mentions binders that are marked dead. See Note [Dead occurrences] in GHC.Core.Lint. I changed the Core pretty-printer so that it didn't suppress dead binders, else I can't see those binders in RULEs. That led to quite a lot of testsuite wibbles. * Refactor FloatBinds, so that it is used both by `exprIsConApp_mabye` and by `lookupRule` * Move the definition of FloatBinds out of GHc.Core.Make, into GHC.Core. * Add FloatTick as an extra constructor. * Refactor `lookupRule` to use `FloatBinds` instead of `BindWrapper`. This refactor just shares more code. (Rename GHC.Core.Opt.FloatOut.FloatBinds to FloatLets, to avoid gratuitious name clash with GHC.Core.FloatBinds.) Corecion optimisation * In simpleOpt, when composing coercions, call new function `optTransCo`. This is much lighter weight than full blown coercion optimisation. * Make `GHC.Core.Opt.Arity.pushCoValArg` and `pushCoTyArg` return the coercionLKind of the coercion. This saves recomputing that coercionLKind at the key call sites in GHC.Core.Opt.Simplify.Iteration.pushCast. * Rename `addCoerce` in GHC.Core.Simplify.Iteration to become `pushCast`. * In the `ApplyToVal` case of `pushCast` we had a very unsavoury call to `simplArg`. I eliminated it by adding a field `sc_cast` to `ApplyToVal` that records any pending casts. Much nicer now. See Note [The sc_cast field of ApplyToVal]. * Don't optimise coercions if the type-substitution is empty. See Note [Optimising coercions] in GHC.Core.Opt.Simplify.Iteration. The fix for #26838 is dramatic. For the test in perf/compiler/T26839 we have Compiler allocs: Before: 7,363M After: 688M Compile time goes down generally. Here are compiler-alloc changes over 0.5%: CoOpt_Read(normal) 729,184,920 -0.7% CoOpt_Singletons(normal) 666,916,960 -4.6% GOOD LargeRecord(normal) 1,227,056,876 +1.1% T12227(normal) 256,827,604 -4.6% GOOD T12425(optasm) 76,879,410 -0.8% T12545(normal) 787,826,918 -10.8% GOOD T12707(normal) 775,186,464 -0.9% T13253(normal) 318,599,596 -0.8% T14766(normal) 685,857,320 -1.0% T15304(normal) 1,123,333,422 -2.2% T15630(normal) 123,142,330 -2.6% T15630a(normal) 123,092,100 -2.6% T15703(normal) 299,751,682 -2.9% GOOD T17516(normal) 964,072,280 +1.0% T18223(normal) 367,016,820 -6.2% GOOD T18730(optasm) 130,643,770 -3.3% GOOD T20261(normal) 535,608,584 -0.7% T21839c(normal) 340,340,436 -0.9% T24984(normal) 85,568,392 -1.9% T3064(normal) 174,631,992 -1.2% T3294(normal) 1,215,886,432 -0.7% T5030(normal) 141,449,704 -17.2% GOOD T5321Fun(normal) 258,484,744 -1.9% T8095(normal) 770,532,232 -2.7% T9630(normal) 858,423,408 -14.5% GOOD T9872c(normal) 1,591,709,448 +0.7% info_table_map_perf(normal) 19,700,614,458 -1.3% geo. mean -0.7% minimum -17.2% maximum +1.1% Metric Decrease: CoOpt_Singletons T12227 T12545 T15703 T18223 T18730 T21839c T5030 T9630 - - - - - 83 changed files: - + changelog.d/T27022 - + changelog.d/dynamic-trace-flags - + changelog.d/so_inline_is_a_predicate - compiler/GHC/Core.hs - compiler/GHC/Core/Coercion.hs - compiler/GHC/Core/Coercion/Opt.hs - compiler/GHC/Core/Lint.hs - compiler/GHC/Core/Make.hs - compiler/GHC/Core/Opt/Arity.hs - compiler/GHC/Core/Opt/ConstantFold.hs - compiler/GHC/Core/Opt/FloatIn.hs - compiler/GHC/Core/Opt/FloatOut.hs - compiler/GHC/Core/Opt/OccurAnal.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/Ppr.hs - compiler/GHC/Core/Rules.hs - compiler/GHC/Core/SimpleOpt.hs - compiler/GHC/Core/TyCo/Subst.hs - compiler/GHC/CoreToStg/Prep.hs - compiler/GHC/Data/List/SetOps.hs - compiler/GHC/Driver/Config.hs - compiler/GHC/Driver/Config/Core/Lint.hs - compiler/GHC/HsToCore/Pmc/Solver.hs - compiler/GHC/IfaceToCore.hs - compiler/GHC/Tc/Types/Evidence.hs - compiler/GHC/Types/Id/Make.hs - hadrian/src/Settings/Packages.hs - libraries/ghci/GHCi/TH.hs - rts/IOManager.h - rts/Linker.c - rts/LinkerInternals.h - rts/RtsSymbols.c - rts/RtsSymbols.h - rts/Trace.c - rts/Trace.h - rts/include/rts/EventLogWriter.h - rts/linker/Elf.c - + rts/posix/FdWakeup.c - + rts/posix/FdWakeup.h - rts/posix/Ticker.c - − rts/posix/ticker/Pthread.c - − rts/posix/ticker/TimerFd.c - rts/rts.cabal - rts/sm/NonMoving.c - testsuite/tests/codeGen/should_compile/T25177.stderr - testsuite/tests/deSugar/should_compile/T13208.stdout - + testsuite/tests/ghc-api/T24386.hs - testsuite/tests/ghc-api/all.T - − testsuite/tests/interface-stability/base-exports.stdout-ws-32 - testsuite/tests/linters/notes.stdout - testsuite/tests/numeric/should_compile/T15547.stderr - testsuite/tests/numeric/should_compile/T20347.stderr - testsuite/tests/numeric/should_compile/T20374.stderr - testsuite/tests/numeric/should_compile/T20376.stderr - + testsuite/tests/perf/compiler/T26989.hs - + testsuite/tests/perf/compiler/T26989a.hs - testsuite/tests/perf/compiler/all.T - testsuite/tests/printer/T18052a.stderr - testsuite/tests/simplCore/should_compile/DsSpecPragmas.stderr - testsuite/tests/simplCore/should_compile/RewriteHigherOrderPatterns.stderr - testsuite/tests/simplCore/should_compile/T15205.stderr - testsuite/tests/simplCore/should_compile/T18668.stderr - testsuite/tests/simplCore/should_compile/T19246.stderr - testsuite/tests/simplCore/should_compile/T19599.stderr - testsuite/tests/simplCore/should_compile/T19599a.stderr - testsuite/tests/simplCore/should_compile/T21917.stderr - testsuite/tests/simplCore/should_compile/T23074.stderr - testsuite/tests/simplCore/should_compile/T24359a.stderr - testsuite/tests/simplCore/should_compile/T25160.stderr - testsuite/tests/simplCore/should_compile/T25718c.stderr-ws-64 - testsuite/tests/simplCore/should_compile/T26051.stderr - testsuite/tests/simplCore/should_compile/T26116.stderr - testsuite/tests/simplCore/should_compile/T8331.stderr - testsuite/tests/simplCore/should_compile/T8848a.stderr - testsuite/tests/simplCore/should_compile/spec004.stderr - + testsuite/tests/th/T27022.hs - + testsuite/tests/th/T27022.stdout - testsuite/tests/th/all.T - testsuite/tests/typecheck/should_compile/T13032.stderr The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/a99ce92f5779f57403372f631cc5cb… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/a99ce92f5779f57403372f631cc5cb… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][ghc-9.12] 3 commits: rts: Eliminate uses of implicit constant arrays
by Magnus (＠MangoIV) 11 May '26

11 May '26

Magnus pushed to branch ghc-9.12 at Glasgow Haskell Compiler / GHC Commits: 89df7715 by Ben Gamari at 2026-05-05T11:49:13+02:00 rts: Eliminate uses of implicit constant arrays Folding of `const`-sized variable-length arrays to a constant-length array is a gnu extension which clang complains about. Closes #26502. (cherry picked from commit 0c00c9c3b4e9b8515d4839f2c1d7d771781dc6f4) - - - - - 6062c663 by mangoiv at 2026-05-05T11:49:23+02:00 testsuite: filter stderr for static001 on darwin This reactivates the test on x86_64 darwin as this should have been done long ago and ignores warnings emitted by ranlib on newer version of the darwin toolchain since they are benign. (no symbols for stub libraries) Fixes #27116 (cherry picked from commit b822c30aa6c0bf008c06c5bd4ee86313c40f652d) - - - - - 97981d8b by Cheng Shao at 2026-05-05T11:49:33+02:00 testsuite: fix testdir cleanup logic on Windows testdir cleanup is unreliable on Windows (#13162) and despite existing hacks in the driver, new failure mode has occurred. This patch makes it print the warning and carry on when failed to clean up a testdir, instead of reporting a spurious framework failure. See added comment for detailed explanation. (cherry picked from commit ed2c65707aacdf442edb8098a7e5cea5fee5d2b0) - - - - - 4 changed files: - rts/Printer.c - rts/posix/OSMem.c - testsuite/driver/testlib.py - testsuite/tests/driver/all.T Changes: ===================================== rts/Printer.c ===================================== @@ -1009,8 +1009,8 @@ findPtr(P_ p, int follow) { uint32_t g, n; bdescr *bd; - const int arr_size = 1024; - StgPtr arr[arr_size]; +#define ARR_SIZE 1024 + StgPtr arr[ARR_SIZE]; int i = 0; searched = 0; @@ -1020,24 +1020,24 @@ findPtr(P_ p, int follow) // just before a block is used. for (n = 0; n < getNumCapabilities(); n++) { bd = nurseries[i].blocks; - i = findPtrBlocks(p,bd,arr,arr_size,i); - if (i >= arr_size) return; + i = findPtrBlocks(p,bd,arr,ARR_SIZE,i); + if (i >= ARR_SIZE) return; } #endif for (g = 0; g < RtsFlags.GcFlags.generations; g++) { bd = generations[g].blocks; - i = findPtrBlocks(p,bd,arr,arr_size,i); + i = findPtrBlocks(p,bd,arr,ARR_SIZE,i); bd = generations[g].large_objects; - i = findPtrBlocks(p,bd,arr,arr_size,i); - if (i >= arr_size) return; + i = findPtrBlocks(p,bd,arr,ARR_SIZE,i); + if (i >= ARR_SIZE) return; for (n = 0; n < getNumCapabilities(); n++) { i = findPtrBlocks(p, gc_threads[n]->gens[g].part_list, - arr, arr_size, i); + arr, ARR_SIZE, i); i = findPtrBlocks(p, gc_threads[n]->gens[g].todo_bd, - arr, arr_size, i); + arr, ARR_SIZE, i); } - if (i >= arr_size) return; + if (i >= ARR_SIZE) return; } if (follow && i == 1) { debugBelch("-->\n"); ===================================== rts/posix/OSMem.c ===================================== @@ -585,7 +585,7 @@ void *osReserveHeapMemory(void *startAddressPtr, W_ *len) } #endif - const int MAX_ATTEMPTS = 256; +#define MAX_ATTEMPTS 256 void *bad_allocs[MAX_ATTEMPTS]; size_t bad_alloc_lens[MAX_ATTEMPTS]; memset(bad_allocs, 0, sizeof(void*) * MAX_ATTEMPTS); ===================================== testsuite/driver/testlib.py ===================================== @@ -25,7 +25,7 @@ from testglobals import config, ghc_env, default_testopts, brokens, t, \ from testutil import strip_quotes, lndir, link_or_copy_file, passed, \ failBecause, testing_metrics, residency_testing_metrics, \ stable_perf_counters, \ - PassFail, badResult, memoize, str_removeprefix + PassFail, badResult, str_warn, str_removeprefix from term_color import Color, colored import testutil from cpu_features import have_cpu_feature @@ -2992,6 +2992,12 @@ def normalise_errmsg(s: str) -> str: # Old emcc warns when we export HEAP8 but new one requires it (see #26290) s = s.replace('warning: invalid item in EXPORTED_RUNTIME_METHODS: HEAP8\nwarning: invalid item in EXPORTED_RUNTIME_METHODS: HEAPU8\nemcc: warning: warnings in JS library compilation [-Wjs-compiler]\n','') + # on newer versions of MacOS X, the shipped ranlib warns about object files with no symbols, + # however, these are completely benign stubs. + # See https://gitlab.haskell.org/ghc/ghc/-/issues/27116 + if opsys('darwin'): + s = modify_lines(s, lambda l: re.sub(r'.*ranlib:.*has no symbols', '', l)) + return s # normalise a .prof file, so that we can reasonably compare it against @@ -3407,9 +3413,20 @@ if config.msys: exception = e retries -= 1 + # Don't fail as framework error if cleanup fails here, just + # print the warning and proceed. I've seen new failure mode + # here on Windows Server 2025 and recent msys2 installation: + # fifo.lnk is created as read-only and the on_error trick + # above somehow doesn't work. + # + # For a local testsuite run, it's in %TEMP% that will be + # periodically cleaned up anyway; for CI, there's post-job + # cleanup and runner level cleanup. It's better to report + # actual job pass/failure than to waste CPU cycles to spurious + # Windows misery. if retries == 0 and testdir.exists(): - raise Exception("Unable to remove folder '%s': %s\nUnable to start current test." - % (testdir, exception)) + print(str_warn("Unable to remove folder '%s': %s\nUnable to start current test." + % (testdir, exception))) else: def cleanup() -> None: testdir = getTestOpts().testdir_raw ===================================== testsuite/tests/driver/all.T ===================================== @@ -120,9 +120,7 @@ if config.os == 'darwin': else: only_darwin = skip -test('static001', [extra_files(['Static001.hs']), - only_darwin, - when(arch('x86_64'), expect_broken(8127))], +test('static001', [extra_files(['Static001.hs']), only_darwin], makefile_test, ['static001']) test('dynHelloWorld', View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/44c6796f535d60861fa2510025ec79… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/44c6796f535d60861fa2510025ec79… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/dcoutts/remove-signal-tickers-docs] 2 commits: Document removal of the signal-based interval timer
by Duncan Coutts (＠dcoutts) 10 May '26

10 May '26

Duncan Coutts pushed to branch wip/dcoutts/remove-signal-tickers-docs at Glasgow Haskell Compiler / GHC Commits: 8948af81 by Duncan Coutts at 2026-05-10T21:50:02+01:00 Document removal of the signal-based interval timer Update mentions within the RTS section of the users guide. Add a changelog entry. - - - - - baf136c5 by Duncan Coutts at 2026-05-10T21:50:02+01:00 Fix section for an recent changelog entry - - - - - 4 changed files: - changelog.d/dynamic-trace-flags - + changelog.d/no-more-timer-signal - docs/users_guide/profiling.rst - docs/users_guide/runtime_control.rst Changes: ===================================== changelog.d/dynamic-trace-flags ===================================== @@ -1,4 +1,4 @@ -section: compiler +section: rts synopsis: Support dynamic trace flags in RTS issues: #27186 mrs: !15936 ===================================== changelog.d/no-more-timer-signal ===================================== @@ -0,0 +1,20 @@ +section: rts +synopsis: The RTS no longer uses signals for its interval timer +issues: #27073 +mrs: !15757 !16014 +description: { + Historically, on unix platforms, the RTS used a signal for its interval + timer. The interval timer is used for thread context switching amongst other + things. This would use either SIGVTALRM or SIGALRM. It was always somewhat + problematic: signals can cause system calls to be interrupted, the signal + must be disabled before forking a new process or the child process will also + receive the signals, and code executed within a signal handler is very + limited in what it can do. + + Most unix platforms were already using other interval timer implementations. + In particular OSX and recent Linux were using thread-based timers. As of + this release, no platforms use a signal based interval timer. So it is no + longer necessary to mitigate the issues that arise. In particular the + `stopTimer` and `startTimer` RTS C functions (used by some libraries) are + now no-ops and are deprecated. +} ===================================== docs/users_guide/profiling.rst ===================================== @@ -678,11 +678,11 @@ enclosed between ``+RTS ... -RTS`` as usual): Sets the interval that the RTS clock ticks at, which is also the sampling interval of the time and allocation profile. The default is 0.001 seconds - when profiling, and 0.01 otherwise. The runtime uses a single timer signal - to count ticks; this timer signal is used to control the context switch + when profiling, and 0.01 otherwise. The runtime uses a single timer event + to count ticks; this timer event is used to control the context switch timer (:ref:`using-concurrent`) and the heap profiling timer :ref:`rts-options-heap-prof`. Also, the time profiler uses the RTS timer - signal directly to record time profiling samples. + event directly to record time profiling samples. Normally, setting the :rts-flag:`-V ⟨secs⟩` option directly is not necessary: the resolution of the RTS timer is adjusted automatically if a ===================================== docs/users_guide/runtime_control.rst ===================================== @@ -259,12 +259,6 @@ Miscellaneous RTS options using the Haskell code as a DLL, and want to set your own signal handlers. - Note that even with ``--install-signal-handlers=no``, the RTS - interval timer signal is still enabled. The timer signal is either - SIGVTALRM or SIGALRM, depending on the RTS configuration and OS - capabilities. To disable the timer signal, use the ``-V0`` RTS - option (see :rts-flag:`-V ⟨secs⟩`). - .. rts-flag:: --install-seh-handlers=⟨yes|no⟩ If yes (the default), the RTS on Windows installs exception handlers to View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/a2bee34936fcf0ac43b5c835e7e9be… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/a2bee34936fcf0ac43b5c835e7e9be… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/dcoutts/remove-signal-tickers-docs] 2 commits: Document removal of the signal-based interval timer
by Duncan Coutts (＠dcoutts) 10 May '26

10 May '26

Duncan Coutts pushed to branch wip/dcoutts/remove-signal-tickers-docs at Glasgow Haskell Compiler / GHC Commits: 410748f3 by Duncan Coutts at 2026-05-10T16:47:32+01:00 Document removal of the signal-based interval timer Update mentions within the RTS section of the users guide. Add a changelog entry. - - - - - a2bee349 by Duncan Coutts at 2026-05-10T16:47:41+01:00 Fix section for an recent changelog entry - - - - - 4 changed files: - changelog.d/dynamic-trace-flags - + changelog.d/no-more-timer-signal - docs/users_guide/profiling.rst - docs/users_guide/runtime_control.rst Changes: ===================================== changelog.d/dynamic-trace-flags ===================================== @@ -1,4 +1,4 @@ -section: compiler +section: rts synopsis: Support dynamic trace flags in RTS issues: #27186 mrs: !15936 ===================================== changelog.d/no-more-timer-signal ===================================== @@ -0,0 +1,20 @@ +section: rts +synopsis: The RTS no longer uses signals for its interval timer +issues: #27073 +mrs: !15757 +description: { + Historically, on unix platforms, the RTS used a signal for its interval + timer. The interval timer is used for thread context switching amongst other + things. This would use either SIGVTALRM or SIGALRM. It was always somewhat + problematic: signals can cause system calls to be interrupted, the signal + must be disabled before forking a new process or the child process will also + receive the signals, and code executed within a signal handler is very + limited in what it can do. + + Most unix platforms were already using other interval timer implementations. + In particular OSX and recent Linux were using thread-based timers. As of + this release, no platforms use a signal based interval timer. So it is no + longer necessary to mitigate the issues that arise. In particular the + `stopTimer` and `startTimer` RTS C functions (used by some libraries) are + now no-ops and are deprecated. +} ===================================== docs/users_guide/profiling.rst ===================================== @@ -678,11 +678,11 @@ enclosed between ``+RTS ... -RTS`` as usual): Sets the interval that the RTS clock ticks at, which is also the sampling interval of the time and allocation profile. The default is 0.001 seconds - when profiling, and 0.01 otherwise. The runtime uses a single timer signal - to count ticks; this timer signal is used to control the context switch + when profiling, and 0.01 otherwise. The runtime uses a single timer event + to count ticks; this timer event is used to control the context switch timer (:ref:`using-concurrent`) and the heap profiling timer :ref:`rts-options-heap-prof`. Also, the time profiler uses the RTS timer - signal directly to record time profiling samples. + event directly to record time profiling samples. Normally, setting the :rts-flag:`-V ⟨secs⟩` option directly is not necessary: the resolution of the RTS timer is adjusted automatically if a ===================================== docs/users_guide/runtime_control.rst ===================================== @@ -259,12 +259,6 @@ Miscellaneous RTS options using the Haskell code as a DLL, and want to set your own signal handlers. - Note that even with ``--install-signal-handlers=no``, the RTS - interval timer signal is still enabled. The timer signal is either - SIGVTALRM or SIGALRM, depending on the RTS configuration and OS - capabilities. To disable the timer signal, use the ``-V0`` RTS - option (see :rts-flag:`-V ⟨secs⟩`). - .. rts-flag:: --install-seh-handlers=⟨yes|no⟩ If yes (the default), the RTS on Windows installs exception handlers to View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/80b5312a5256b4b898a57488b02663… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/80b5312a5256b4b898a57488b02663… You're receiving this email because of your account on gitlab.haskell.org.

1 0