July 2018 - ghc-tickets

[GHC] #15423: Ungrammatical error message involving a pattern binding
by GHC 31 Jul '18

31 Jul '18

#15423: Ungrammatical error message involving a pattern binding -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Compiler | Version: 8.4.3 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Take a look at this error message from GHC's `Typeable1` test case: {{{ Typeable1.hs:22:5: error: [-Winaccessible-code (in -Wdefault), -Werror =inaccessible-code] • Couldn't match type ‘ComposeK’ with ‘a3 b3’ Inaccessible code in a pattern with pattern synonym: App :: forall k2 (t :: k2). () => forall k1 (a :: k1 -> k2) (b :: k1). (t ~ a b) => TypeRep a -> TypeRep b -> TypeRep t, in a pattern binding in 'do' block }}} In particular, note the "in a pattern binding in 'do' block" part. I think that should be "in **a** 'do' block". I believe fixing this is a simple matter of applying this patch: {{{#!diff diff --git a/compiler/hsSyn/HsExpr.hs b/compiler/hsSyn/HsExpr.hs index 96d86c8..a5c65fb 100644 --- a/compiler/hsSyn/HsExpr.hs +++ b/compiler/hsSyn/HsExpr.hs @@ -2804,7 +2804,7 @@ pprMatchContextNoun PatBindGuards = text "pattern binding guards" pprMatchContextNoun LambdaExpr = text "lambda abstraction" pprMatchContextNoun ProcExpr = text "arrow abstraction" pprMatchContextNoun (StmtCtxt ctxt) = text "pattern binding in" - $$ pprStmtContext ctxt + $$ pprAStmtContext ctxt pprMatchContextNoun PatSyn = text "pattern synonym declaration" ----------------- }}} Patch incoming. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15423> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #15422: GHCi debugger doesn't see free variables when using ApplicativeDo
by GHC 31 Jul '18

31 Jul '18

#15422: GHCi debugger doesn't see free variables when using ApplicativeDo -------------------------------------+------------------------------------- Reporter: simonmar | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Compiler | Version: 8.4.3 Keywords: ApplicativeDo | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): Phab:D4991 | Wiki Page: -------------------------------------+------------------------------------- ``` > ghci break029.hs GHCi, version 8.2.2: http://www.haskell.org/ghc/ :? for help [1 of 1] Compiling Main ( break029.hs, interpreted ) Ok, one module loaded. *Main> :!cat break029.hs {-# LANGUAGE ApplicativeDo #-} f :: Int -> IO Int f x = do y <- return (x + 1) return (y * 2) *Main> :step f 3 Stopped in Main.f, break029.hs:(4,7)-(6,16) _result :: IO Int = _ x :: Int = 3 [break029.hs:(4,7)-(6,16)] [break029.hs:(4,7)-(6,16)] *Main> :list 3 f :: Int -> IO Int 4 f x = do 5 y <- return (x + 1) 6 return (y * 2) 7 [break029.hs:(4,7)-(6,16)] [break029.hs:(4,7)-(6,16)] *Main> :step Stopped in Main.f, break029.hs:5:8-21 _result :: IO Int = _ x :: Int = 3 [break029.hs:5:8-21] [break029.hs:5:8-21] *Main> :list 4 f x = do 5 y <- return (x + 1) 6 return (y * 2) [break029.hs:5:8-21] [break029.hs:5:8-21] *Main> :step Stopped in Main.f, break029.hs:6:11-15 _result :: Int = _ [break029.hs:6:11-15] [break029.hs:6:11-15] *Main> y <interactive>:7:1: error: Variable not in scope: y ``` -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15422> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #15398: GADT deriving Ord generates inaccessible code in a pattern with constructor.
by GHC 31 Jul '18

31 Jul '18

#15398: GADT deriving Ord generates inaccessible code in a pattern with constructor. --------------------------------------+--------------------------------- Reporter: philderbeast | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Compiler | Version: 8.2.2 Keywords: | Operating System: MacOS X Architecture: x86_64 (amd64) | Type of failure: None/Unknown Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: --------------------------------------+--------------------------------- I added a second type parameter `k` to a GADT `Zone` and found that when deriving Ord with standalone deriving, the generated code has errors in ghc-8.2.2. There's a reproduction repo for this; https://github.com/BlockScope/zone-inaccessible-code-deriving-ord. I found I needed at least three constructors in `Zone` to get this error. #8128 seemed relevant for being about standalone deriving of a GADT. That being fixed, I tried with ghc-head that I built from source. With this version I found that the same code faults exist in the generated code but are treated as warnings by ghc-8.7.2 {{{ > inplace/bin/ghc-stage2 --version The Glorious Glasgow Haskell Compilation System, version 8.7.20180715 }}} {{{#!hs {-# LANGUAGE DeriveAnyClass #-} {-# LANGUAGE FlexibleContexts #-} {-# LANGUAGE GADTs #-} {-# LANGUAGE StandaloneDeriving #-} module Flight.Zone where newtype Radius a = Radius a deriving (Eq, Ord) data CourseLine data OpenDistance data EndOfSpeedSection -- TODO: Remove standalone deriving Eq & Ord for empty data after GHC 8.4.1 -- SEE: https://ghc.haskell.org/trac/ghc/ticket/7401 deriving instance Eq CourseLine deriving instance Eq OpenDistance deriving instance Eq EndOfSpeedSection deriving instance Ord CourseLine deriving instance Ord OpenDistance deriving instance Ord EndOfSpeedSection data Zone k a where Point :: (Eq a, Ord a) => Zone CourseLine a Vector :: (Eq a, Ord a) => Zone OpenDistance a Conical :: (Eq a, Ord a) => Radius a -> Zone EndOfSpeedSection a deriving instance Eq a => Eq (Zone k a) deriving instance (Eq a, Ord a) => Ord (Zone k a) }}} The error; {{{ /.../Zone.hs:25:1: error: • Couldn't match type ‘OpenDistance’ with ‘CourseLine’ Inaccessible code in a pattern with constructor: Point :: forall a. (Eq a, Ord a) => Zone CourseLine a, in a case alternative • In the pattern: Point {} In a case alternative: Point {} -> GT In the expression: case b of Point {} -> GT Vector -> EQ _ -> LT When typechecking the code for ‘compare’ in a derived instance for ‘Ord (Zone k a)’: To see the code I am typechecking, use -ddump-deriv | 25 | deriving instance (Eq a, Ord a) => Ord (Zone k a) | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ /.../Zone.hs:25:1: error: • Couldn't match type ‘OpenDistance’ with ‘CourseLine’ Inaccessible code in a pattern with constructor: Point :: forall a. (Eq a, Ord a) => Zone CourseLine a, in a case alternative • In the pattern: Point {} In a case alternative: Point {} -> False In the expression: case b of Point {} -> False Vector -> False _ -> True When typechecking the code for ‘<’ in a derived instance for ‘Ord (Zone k a)’: To see the code I am typechecking, use -ddump-deriv | 25 | deriving instance (Eq a, Ord a) => Ord (Zone k a) | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ }}} The generated code with these replacements; {{{ :%s/GHC\.Classes\.// :%s/GHC\.Types\.// :%s/Flight\.Zone\.// }}} {{{#!hs instance (Eq a, Ord a) => Ord (Zone k a) where compare a_a2hw b_a2hx = case a_a2hw of Point -> case b_a2hx of Point -> EQ _ -> LT Vector -> case b_a2hx of Point {} -> GT Vector -> EQ _ -> LT Conical a1_a2hy -> case b_a2hx of Conical b1_a2hz -> (a1_a2hy `compare` b1_a2hz) _ -> GT (<) a_a2hA b_a2hB = case a_a2hA of Point -> case b_a2hB of Point -> False _ -> True Vector -> case b_a2hB of Point {} -> False Vector -> False _ -> True Conical a1_a2hC -> case b_a2hB of Conical b1_a2hD -> (a1_a2hC < b1_a2hD) _ -> False (<=) a_a2hE b_a2hF = not ((<) b_a2hF a_a2hE) (>) a_a2hG b_a2hH = (<) b_a2hH a_a2hG (>=) a_a2hI b_a2hJ = not ((<) a_a2hI b_a2hJ) instance Eq a => Eq (Zone k a) where (==) (Point) (Point) = True (==) (Vector) (Vector) = True (==) (Conical a1_a2hK) (Conical b1_a2hL) = ((a1_a2hK == b1_a2hL)) (==) _ _ = False (/=) a_a2hM b_a2hN = not ((==) a_a2hM b_a2hN) }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15398> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 17

[GHC] #15396: -staticlib is broken on GHC 8.4.3 on Linux under certain conditions
by GHC 31 Jul '18

31 Jul '18

#15396: -staticlib is broken on GHC 8.4.3 on Linux under certain conditions -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Compiler | Version: 8.4.3 (Linking) | Keywords: | Operating System: Linux Architecture: | Type of failure: Compile-time Unknown/Multiple | crash or panic Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- When compiling this simple program: {{{#!hs main = pure () }}} Using the `-staticlib` flag on GHC 8.4.3 on Linux, it panics if using a certain toolchain: {{{ $ uname -a Linux gearloose 4.2.0-42-generic #49~14.04.1-Ubuntu SMP Wed Jun 29 20:22:11 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux $ /opt/ghc/8.4.3/bin/ghc -staticlib Bug.hs Linking Bug.a ... ghc: panic! (the 'impossible' happened) (GHC version 8.4.3 for x86_64-unknown-linux): Data.Binary.Get.runGet at position 27462987: Invalid archive header end marker CallStack (from HasCallStack): error, called at libraries/binary/src/Data/Binary/Get.hs:351:5 in binary-0.8.5.1:Data.Binary.Get }}} I say "certain toolchain" since I am able to reproduce this issue on Ubuntu 14.04, but not 18.04: {{{ $ uname -a Linux Linux-T450 4.15.0-24-generic #26-Ubuntu SMP Wed Jun 13 08:44:47 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux $ ghc -staticlib Bug.hs [1 of 1] Compiling Main ( Bug.hs, Bug.o ) Linking Bug.a ... }}} This issue does not occur on GHC 8.2 and earlier: {{{ $ /opt/ghc/8.2.2/bin/ghc -staticlib Bug.hs [1 of 1] Compiling Main ( Bug.hs, Bug.o ) Linking Bug.a ... Static archive creation only supported on Darwin/OS X/iOS }}} Commit b8f33bc6b738b0378976e42b79369f0e53b680c7 allowed the use of `-staticlib` on all platforms. angerman, do you know what might be happening here? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15396> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 13

[GHC] #15365: Argument-less infix declarations printed without parentheses in -ddump-splices
by GHC 31 Jul '18

31 Jul '18

#15365: Argument-less infix declarations printed without parentheses in -ddump- splices -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Template | Version: 8.4.3 Haskell | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Debugging Unknown/Multiple | information is incorrect Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Compile this: {{{#!hs {-# LANGUAGE TemplateHaskell #-} {-# LANGUAGE TypeOperators #-} {-# OPTIONS_GHC -ddump-splices #-} module Bug where $([d| type (|||) = Either (&&&) :: Bool -> Bool -> Bool (&&&) = (&&) data (***) |]) }}} And you'll get this: {{{ $ /opt/ghc/8.4.3/bin/ghci Bug.hs GHCi, version 8.4.3: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/rgscott/.ghci [1 of 1] Compiling Bug ( Bug.hs, interpreted ) Bug.hs:(6,3)-(12,6): Splicing declarations [d| (&&&_a1xt) :: Bool -> Bool -> Bool (&&&_a1xt) = (&&) type |||_a1xs = Either data ***_a1xr |] ======> type |||_a5bW = Either (&&&_a5bV) :: Bool -> Bool -> Bool (&&&_a5bV) = (&&) data ***_a5bX Ok, one module loaded. }}} Notice the pretty-printed declarations of `type |||_a5bW` and `data ***_a5bX`, which are lacking some necessary parentheses. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15365> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #15430: Merge llvm fix to ghc-8.6
by GHC 31 Jul '18

31 Jul '18

#15430: Merge llvm fix to ghc-8.6 -------------------------------------+------------------------------------- Reporter: monoidal | Owner: (none) Type: bug | Status: new Priority: highest | Milestone: 8.6.1 Component: Compiler | Version: 8.5 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- This ticket is a reminder to make sure that https://phabricator.haskell.org/D4969 (Fix a major copy'n'paste error in LLVM CodeGen) is merged to ghc-8.6. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15430> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #15428: GHC 8.6+ panics (piResultTys2), older GHCs don't
by GHC 31 Jul '18

31 Jul '18

#15428: GHC 8.6+ panics (piResultTys2), older GHCs don't -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: high | Milestone: 8.6.1 Component: Compiler | Version: 8.5 Keywords: TypeInType, | Operating System: Unknown/Multiple TypeFamilies | Architecture: | Type of failure: Compile-time Unknown/Multiple | crash or panic Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- The following program panics when compiled with GHC 8.6.1 or HEAD: {{{#!hs {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE TypeFamilies #-} {-# LANGUAGE TypeInType #-} {-# LANGUAGE TypeOperators #-} module Bug where import Data.Kind import Data.Proxy import Data.Type.Equality type family Flurmp :: k type family Pure (x :: a) :: f a wat :: forall (f :: Type -> Type) (p :: Type). Proxy (f p) -> () wat _ = let s :: (Flurmp :: f p) :~: Pure (Flurmp :: p -> p) (Flurmp :: p) s = undefined in () }}} {{{ $ /opt/ghc/8.6.1/bin/ghc Bug.hs [1 of 1] Compiling Bug ( Bug.hs, Bug.o ) ghc: panic! (the 'impossible' happened) (GHC version 8.6.0.20180714 for x86_64-unknown-linux): piResultTys2 f_aAT a_aAU [(->) p_a1uI[sk:1], p_a1uI[sk:1] -> p_a1uI[sk:1], Flurmp, Flurmp] [Flurmp] Call stack: CallStack (from HasCallStack): callStackDoc, called at compiler/utils/Outputable.hs:1164:37 in ghc:Outputable pprPanic, called at compiler/types/Type.hs:1041:9 in ghc:Type }}} On GHC 8.4 and earlier, this simply gives an error message: {{{ $ /opt/ghc/8.4.3/bin/ghc Bug.hs [1 of 1] Compiling Bug ( Bug.hs, Bug.o ) Bug.hs:18:16: error: • Expecting one more argument to ‘Pure (Flurmp :: p -> p)’ Expected kind ‘p -> f p’, but ‘Pure (Flurmp :: p -> p)’ has kind ‘p -> p -> p’ • In the second argument of ‘(:~:)’, namely ‘Pure (Flurmp :: p -> p) (Flurmp :: p)’ In the type signature: s :: (Flurmp :: f p) :~: Pure (Flurmp :: p -> p) (Flurmp :: p) In the expression: let s :: (Flurmp :: f p) :~: Pure (Flurmp :: p -> p) (Flurmp :: p) s = undefined in () • Relevant bindings include wat :: Proxy (f p) -> () (bound at Bug.hs:16:1) | 18 | :~: Pure (Flurmp :: p -> p) (Flurmp :: p) | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15428> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #15388: GHC reports missing INLINABLE pragmas in vector and ghc-prim
by GHC 31 Jul '18

31 Jul '18

#15388: GHC reports missing INLINABLE pragmas in vector and ghc-prim -------------------------------------+------------------------------------- Reporter: ChaiTRex | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Compiler | Version: 8.4.3 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- When compiling a program with GHC 8.4.3, vector 0.12.0.1 (with a dependency on primitive 0.6.4.0) and `-Weverything`, I get several warnings that `INLINABLE` pragmas should be added to the vector and ghc- prim packages. It would be nicer if these warnings didn't happen, even with `-Weverything`. Also, perhaps there are optimization opportunities that are lost. A somewhat reduced example (can be reduced further by removing the function definitions and unused imports that creates): {{{#!hs lineno=1 {-# OPTIONS_GHC -Wall-missed-specialisations #-} {-# LANGUAGE MultiParamTypeClasses, TypeFamilies #-} module Warnings2 where import GHC.Base (liftM) import qualified Data.Vector.Generic.Base as G (Vector, basicUnsafeFreeze, basicUnsafeThaw, basicLength, basicUnsafeSlice, basicUnsafeIndexM, basicUnsafeCopy, elemseq) import qualified Data.Vector.Generic.Mutable.Base as GM (MVector, basicLength, basicUnsafeSlice, basicOverlaps, basicUnsafeNew, basicInitialize, basicUnsafeReplicate, basicUnsafeRead, basicUnsafeWrite, basicClear, basicSet, basicUnsafeCopy, basicUnsafeMove, basicUnsafeGrow) import qualified Data.Vector.Unboxed as U (Unbox, Vector) import qualified Data.Vector.Unboxed.Mutable as U (MVector) newtype T = T () deriving (Show, Eq, Ord) newtype C = C (T, U.Vector T) deriving (Show, Eq, Ord) instance U.Unbox T newtype instance U.MVector s T = MV_T (U.MVector s ()) instance GM.MVector U.MVector T where basicLength (MV_T mv) = GM.basicLength mv basicUnsafeSlice i l (MV_T mv) = MV_T (GM.basicUnsafeSlice i l mv) basicOverlaps (MV_T mv1) (MV_T mv2) = GM.basicOverlaps mv1 mv2 basicUnsafeNew l = MV_T `liftM` GM.basicUnsafeNew l basicInitialize (MV_T mv) = GM.basicInitialize mv basicUnsafeReplicate l _ = MV_T `liftM` GM.basicUnsafeReplicate l () basicUnsafeRead (MV_T mv) i = const (T ()) `liftM` GM.basicUnsafeRead mv i basicUnsafeWrite (MV_T mv) i _ = GM.basicUnsafeWrite mv i () basicClear (MV_T mv) = GM.basicClear mv basicSet (MV_T mv) x = GM.basicSet mv () basicUnsafeCopy (MV_T mv1) (MV_T mv2) = GM.basicUnsafeCopy mv1 mv2 basicUnsafeMove (MV_T mv1) (MV_T mv2) = GM.basicUnsafeMove mv1 mv2 basicUnsafeGrow (MV_T mv) l = MV_T `liftM` GM.basicUnsafeGrow mv l newtype instance U.Vector T = V_T (U.Vector ()) instance G.Vector U.Vector T where basicUnsafeFreeze (MV_T mv) = V_T `liftM` G.basicUnsafeFreeze mv basicUnsafeThaw (V_T v) = MV_T `liftM` G.basicUnsafeThaw v basicLength (V_T v) = G.basicLength v basicUnsafeSlice i l (V_T v) = V_T (G.basicUnsafeSlice i l v) basicUnsafeIndexM (V_T v) i = const (T ()) `liftM` G.basicUnsafeIndexM v i basicUnsafeCopy (MV_T mv) (V_T v) = G.basicUnsafeCopy mv v elemseq (V_T v) _ = G.elemseq v () }}} The warnings: {{{#!default lineno=1 marks=16,31,46 src/Warnings2.hs: warning: Could not specialise imported function ‘Data.Vector.Unboxed.$w$cshowsPrec’ when specialising ‘Data.Vector.Unboxed.$fShowVector_$cshowsPrec’ when specialising ‘Data.Vector.Unboxed.$fShowVector’ Probable fix: add INLINABLE pragma on ‘Data.Vector.Unboxed.$w$cshowsPrec’ src/Warnings2.hs: warning: Could not specialise imported function ‘Data.Vector.Unboxed.$fShowVector_$cshow’ when specialising ‘Data.Vector.Unboxed.$fShowVector’ Probable fix: add INLINABLE pragma on ‘Data.Vector.Unboxed.$fShowVector_$cshow’ src/Warnings2.hs: warning: Could not specialise imported function ‘Data.Vector.Unboxed.$fShowVector_$cshowList’ when specialising ‘Data.Vector.Unboxed.$fShowVector’ Probable fix: add INLINABLE pragma on ‘Data.Vector.Unboxed.$fShowVector_$cshowList’ src/Warnings2.hs: warning: Could not specialise imported function ‘ghc- prim-0.5.2.0:GHC.Classes.$w$c==’ when specialising ‘ghc-prim-0.5.2.0:GHC.Classes.$fEq(,)_$c==’ Probable fix: add INLINABLE pragma on ‘ghc- prim-0.5.2.0:GHC.Classes.$w$c==’ src/Warnings2.hs: warning: Could not specialise imported function ‘Data.Vector.Unboxed.$fOrdVector_$cmax’ when specialising ‘Data.Vector.Unboxed.$fOrdVector’ Probable fix: add INLINABLE pragma on ‘Data.Vector.Unboxed.$fOrdVector_$cmax’ src/Warnings2.hs: warning: Could not specialise imported function ‘Data.Vector.Unboxed.$fOrdVector_$cmin’ when specialising ‘Data.Vector.Unboxed.$fOrdVector’ Probable fix: add INLINABLE pragma on ‘Data.Vector.Unboxed.$fOrdVector_$cmin’ src/Warnings2.hs: warning: Could not specialise imported function ‘ghc- prim-0.5.2.0:GHC.Classes.$w$c>=’ when specialising ‘ghc-prim-0.5.2.0:GHC.Classes.$fOrd(,)_$c>=’ Probable fix: add INLINABLE pragma on ‘ghc- prim-0.5.2.0:GHC.Classes.$w$c>=’ src/Warnings2.hs: warning: Could not specialise imported function ‘ghc- prim-0.5.2.0:GHC.Classes.$w$c<’ when specialising ‘ghc-prim-0.5.2.0:GHC.Classes.$fOrd(,)_$c<’ Probable fix: add INLINABLE pragma on ‘ghc- prim-0.5.2.0:GHC.Classes.$w$c<’ src/Warnings2.hs: warning: Could not specialise imported function ‘ghc- prim-0.5.2.0:GHC.Classes.$w$c<=’ when specialising ‘ghc-prim-0.5.2.0:GHC.Classes.$fOrd(,)_$c<=’ Probable fix: add INLINABLE pragma on ‘ghc- prim-0.5.2.0:GHC.Classes.$w$c<=’ src/Warnings2.hs: warning: Could not specialise imported function ‘ghc- prim-0.5.2.0:GHC.Classes.$w$ccompare’ when specialising ‘ghc-prim-0.5.2.0:GHC.Classes.$fOrd(,)_$ccompare’ Probable fix: add INLINABLE pragma on ‘ghc- prim-0.5.2.0:GHC.Classes.$w$ccompare’ }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15388> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #14928: TH eats 50 GB memory when creating ADT with multiple constructors
by GHC 31 Jul '18

31 Jul '18

#14928: TH eats 50 GB memory when creating ADT with multiple constructors -------------------------------------+------------------------------------- Reporter: YitzGale | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Template | Version: 8.2.2 Haskell | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- When TH creates a data type with multiple constructors, GHC consumes huge amounts of memory in what appears to be a highly superlinear manner. A common use case: in the Yesod web framework, localized strings are represented by constructors of a Messages data type, created by a TH splice. There is one constructor for each localized string on the site, possibly hundreds. The splice also creates a class instance for the data type whose method matches against all the constructors for each language for which localizations are provided; this may or may not play a role in the memory leak. This Trac ticket corresponds to this Yesod issue: https://github.com/yesodweb/yesod/issues/1487 Here are two reproductions, and one NON-reproduction: 1. A blank "hello world" Yesod web site, with 500 messages defined for about 30 languages. The single page displays the messages in the user's language. Compiling this program in GHC 8.2.2 (stackage lts-10.5) on Ubuntu 16.04 eats over 50 GB of memory. https://gitub.com/ygale/yesod-bug1487 2. @snoyberg has cut down this reproduction to avoid using any libraries not included with GHC. It is in the same repo, on the {{{snoyberg- master}}} branch. NON-reproduction: The code in this gist, which is similar to what is generated by the TH in the above reproductions, is compiled by GHC without the bat of an eyelash. This demonstrates that the bug requires TH to reproduce. https://gist.github.com/92347aa93d226e31f977a0b62b443aa7 -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14928> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 7

[GHC] #15387: Unable to set testsuite verbosity to zero from command line
by GHC 31 Jul '18

31 Jul '18

#15387: Unable to set testsuite verbosity to zero from command line -------------------------------------+------------------------------------- Reporter: lantti | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Test Suite | Version: 8.4.3 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Other Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Setting testsuite verbosity to zero from commandline using 'make test VERBOSE=0' does not work because runtests.py confuses the value 0 with value None. Solution: check explicitly for None instead of any value considered equal to False. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15387> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5