May 2017 - ghc-tickets

[GHC] #13703: Internal libraries regression in GHC 8.2 (ghc-pkg handling)
by GHC 06 Jun '17

06 Jun '17

#13703: Internal libraries regression in GHC 8.2 (ghc-pkg handling) -------------------------------------+------------------------------------- Reporter: ezyang | Owner: (none) Type: bug | Status: new Priority: high | Milestone: 8.2.1 Component: ghc-pkg | Version: 8.2.1-rc2 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: -------------------------------------+------------------------------------- If I register two internal libraries from the same package, they clobber each other. They should not do this. This regression was probably introduced by 579bb7669f40ed01841dd197ee535cf26fa19580. Working on a fix. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13703> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #12710: Make some reserved Unicode symbols "specials"
by GHC 05 Jun '17

05 Jun '17

#12710: Make some reserved Unicode symbols "specials" -------------------------------------+------------------------------------- Reporter: JoshPrice247 | Owner: JoshPrice247 Type: feature | Status: new request | Priority: low | Milestone: Component: Compiler | Version: 8.0.1 (Parser) | Keywords: Unicode, | Operating System: Unknown/Multiple UnicodeSyntax | Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- The following comment is in `compiler/parser/Lexer.x`: > ToDo: ideally, → and ∷ should be "specials", so that they cannot > form part of a large operator. This would let us have a better > syntax for kinds: ɑ∷*→* would be a legal kind signature. (maybe). -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12710> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

[GHC] #10600: -fwarn-incomplete-patterns doesn't work with -fno-code
by GHC 05 Jun '17

05 Jun '17

#10600: -fwarn-incomplete-patterns doesn't work with -fno-code -------------------------------------+------------------------------------- Reporter: akio | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Incorrect Unknown/Multiple | warning at compile-time Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- `-fwarn-incomplete-patterns` doensn't seem to generate any warnings when `-fno-code` is specified. To reproduce, save this module as `foo.hs`. {{{#!hs module Foo where foo True = 4 }}} and compile with `-fwarn-incomplete-patterns`, with and without `-fno- code`. {{{ % ghc -fwarn-incomplete-patterns -fforce-recomp foo.hs [1 of 1] Compiling Foo ( foo.hs, foo.o ) foo.hs:2:1: Warning: Pattern match(es) are non-exhaustive In an equation for ‘foo’: Patterns not matched: False % ghc -fwarn-incomplete-patterns -fforce-recomp -fno-code foo.hs [1 of 1] Compiling Foo ( foo.hs, nothing ) % }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10600> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 26

[GHC] #13761: Can't create poly-kinded GADT with TypeInType enabled, but can without
by GHC 05 Jun '17

05 Jun '17

#13761: Can't create poly-kinded GADT with TypeInType enabled, but can without -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 (Type checker) | Keywords: TypeInType | Operating System: Unknown/Multiple Architecture: | Type of failure: GHC rejects Unknown/Multiple | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Surprisingly, this compiles without `TypeInType`: {{{#!hs {-# LANGUAGE DataKinds #-} {-# LANGUAGE GADTs #-} {-# LANGUAGE KindSignatures #-} {-# LANGUAGE PolyKinds #-} module Works where import Data.Kind data T :: k -> Type where MkT :: T Int }}} But once you add `TypeInType`: {{{#!hs {-# LANGUAGE DataKinds #-} {-# LANGUAGE GADTs #-} {-# LANGUAGE KindSignatures #-} {-# LANGUAGE PolyKinds #-} {-# LANGUAGE TypeInType #-} module Bug where import Data.Kind data T :: k -> Type where MkT :: T Int }}} then it stops working! {{{ GHCi, version 8.3.20170516: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/rgscott/.ghci [1 of 1] Compiling Bug ( Bug.hs, interpreted ) Bug.hs:11:12: error: • Expected kind ‘k’, but ‘Int’ has kind ‘*’ • In the first argument of ‘T’, namely ‘Int’ In the type ‘T Int’ In the definition of data constructor ‘MkT’ | 11 | MkT :: T Int | ^^^ }}} This bug is present in GHC 8.0.1, 8.0.2, 8.2.1, and HEAD. What's strange about this bug is that is requires that you write `T` with an explicit kind signature. If you write `T` like this: {{{#!hs data T (a :: k) where MkT :: T Int }}} Then it will work with `TypeInType` enabled. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13761> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #13770: HEAD: Type mentioned in error won't show up in pattern signature
by GHC 05 Jun '17

05 Jun '17

#13770: HEAD: Type mentioned in error won't show up in pattern signature -------------------------------------+------------------------------------- Reporter: heisenbug | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.3 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: -------------------------------------+------------------------------------- I am encountering a stange problem: {{{ ghc-stage2 T13770.hs -c -fprint-explicit-kinds T13770.hs:29:52: error: ? Could not deduce: (r4 :: GHC.Types.RuntimeRep) ~~ ('GHC.Types.LiftedRep :: GHC.Types.RuntimeRep) from the context: (* ~ *, (f :: *) ~~ (arg -> res :: *)) bound by a pattern with pattern synonym: Fun :: forall k (fun :: k). () => forall (arg :: TYPE r1) (res :: TYPE r2). (k ~ *, (fun :: k) ~~ (arg -> res :: *)) => TypeRep (TYPE r1) arg -> TypeRep (TYPE r2) res -> TypeRep k fun, in a pattern binding in 'do' block at T13770.hs:19:19-35 or from: (* :: *) ~~ (TYPE r1 :: *) bound by a pattern with constructor: Refl :: forall k (a :: k). (:~:) k a a, in a pattern binding in 'do' block at T13770.hs:22:19-22 or from: (* :: *) ~~ (TYPE r2 :: *) bound by a pattern with constructor: Refl :: forall k (a :: k). (:~:) k a a, in a pattern binding in 'do' block at T13770.hs:23:19-22 or from: (TYPE r2 ~ *, (res :: TYPE r2) ~~ (arg1 -> res1 :: *)) bound by a pattern with pattern synonym: Fun :: forall k (fun :: k). () => forall (arg :: TYPE r1) (res :: TYPE r2). (k ~ *, (fun :: k) ~~ (arg -> res :: *)) => TypeRep (TYPE r1) arg -> TypeRep (TYPE r2) res -> TypeRep k fun, in a pattern binding in 'do' block at T13770.hs:27:37-41 ?r4? is a rigid type variable bound by a pattern with pattern synonym: Fun :: forall k (fun :: k). () => forall (arg :: TYPE r1) (res :: TYPE r2). (k ~ *, (fun :: k) ~~ (arg -> res :: *)) => TypeRep (TYPE r1) arg -> TypeRep (TYPE r2) res -> TypeRep k fun, in a pattern binding in 'do' block at T13770.hs:27:37-41 When matching the kind of ?arg1? Expected type: TypeRep * (arg1 -> res1) Actual type: TypeRep (TYPE r2) res ? In the second argument of ?postulate?, namely ?fargrep? In a stmt of a 'do' block: Refl <- postulate vrep fargrep In the expression: do A2 satur _ <- a2 retrep vrep fargrep@Fun {} <- pure retrep HRefl <- fargrep `eqTypeRep` retrep Refl <- postulate vrep fargrep .... | 29 | Refl <- postulate vrep fargrep | ^^^^^^^ }}} Notice how `r4` is mentioned in the error message but none of the signatures refer to it. This is confusing. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13770> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #13606: GHCi segfaults on Windows with D3D code
by GHC 03 Jun '17

03 Jun '17

#13606: GHCi segfaults on Windows with D3D code ----------------------------------------+------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: GHCi | Version: 8.2.1-rc2 Keywords: | Operating System: Windows Architecture: Unknown/Multiple | Type of failure: GHCi crash Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: ----------------------------------------+------------------------------- Warning: this is about as Windows-specific of a bug as it possibly gets, since it requires the use of D3D. I noticed this when trying to run [https://github.com/jwvg0425/d3d11binding examples] from the `d3d11binding` library in GHCi, as they all failed. First, to conjure up the code needed for this: * `Main.hs`: {{{#!hs {-# LANGUAGE CPP #-} {-# LANGUAGE ScopedTypeVariables #-} module Main (main) where import Data.Bits (Bits(..)) import Data.Int (Int32) import Data.Word (Word32) import Foreign.C.String (CString, peekCString, withCString, withCStringLen) import Foreign.Marshal.Alloc (alloca) import Foreign.Ptr (Ptr, castPtr, nullPtr) import Foreign.Storable (Storable(..)) import System.IO (IOMode(..), hGetContents, withFile) #if defined(i386_HOST_ARCH) # define WINDOWS_CCONV stdcall #elif defined(x86_64_HOST_ARCH) # define WINDOWS_CCONV ccall #else # error Unknown mingw32 arch #endif foreign import WINDOWS_CCONV "D3DCompile" c_d3dCompile :: Ptr () -> Word32 -> CString -> Ptr D3DShaderMacro -> Ptr ID3DInclude -> CString -> CString -> D3DCompileFlag -> D3DCompileEffectFlag -> Ptr (Ptr ID3DBlob) -> Ptr (Ptr ID3DBlob) -> IO HRESULT maybePoke :: (Storable a) => Maybe a -> (Ptr a -> IO b) -> IO b maybePoke Nothing proc = proc nullPtr maybePoke (Just m) proc = alloca $ \ptr -> do poke ptr m proc ptr maybeWithCString :: Maybe String -> (CString -> IO a) -> IO a maybeWithCString Nothing proc = proc nullPtr maybeWithCString (Just m) proc = withCString m proc type HRESULT = LONG data ID3DBlob = ID3DBlob data ID3DInclude = ID3DInclue type LONG = Int32 data D3DShaderMacro = D3DShaderMacro { _name :: String , _definition :: String } instance Storable D3DShaderMacro where sizeOf _ = 8 alignment _ = 8 peek ptr = do n <- peekByteOff ptr 0 d <- peekByteOff ptr 4 n' <- peekCString n d' <- peekCString d return $ D3DShaderMacro n' d' poke ptr (D3DShaderMacro n d) = do withCString n $ \n' -> withCString d $ \d' -> do pokeByteOff ptr 0 n' pokeByteOff ptr 4 d' type D3DCompileFlag = Word32 type D3DCompileEffectFlag = Word32 d3dCompileEnableStrictness :: D3DCompileFlag d3dCompileEnableStrictness = shift 1 11 d3dCompile :: String -> Maybe String -> Maybe D3DShaderMacro -> Ptr ID3DInclude -> Maybe String -> String -> [D3DCompileFlag] -> [D3DCompileEffectFlag] -> IO (Either (HRESULT, Ptr ID3DBlob) (Ptr ID3DBlob)) d3dCompile source sourceName defines pInclude entryPoint target compileFlags effectFlags = do withCStringLen source $ \(csource, len) -> withCString target $ \pTarget -> maybeWithCString sourceName $ \pSourceName -> maybePoke defines $ \pDefines -> maybeWithCString entryPoint $ \pEntryPoint -> alloca $ \ppCode -> alloca $ \ppErrorMsgs -> do let sFlag = foldl (.|.) 0 compileFlags let eFlag = foldl (.|.) 0 effectFlags putStrLn "Before d3dCompile" hr <- c_d3dCompile (castPtr csource) (fromIntegral len) pSourceName pDefines pInclude pEntryPoint pTarget sFlag eFlag ppCode ppErrorMsgs putStrLn "After d3dCompile" if hr < 0 then do pErrorMsgs <- peek ppErrorMsgs return $ Left (hr, pErrorMsgs) else do pCode <- peek ppCode return $ Right pCode d3dCompileFromFile :: String -> Maybe String -> Maybe D3DShaderMacro -> Ptr ID3DInclude -> Maybe String -> String -> [D3DCompileFlag] -> [D3DCompileEffectFlag] -> IO (Either (HRESULT, Ptr ID3DBlob) (Ptr ID3DBlob)) d3dCompileFromFile fileName sourceName defines pInclude entryPoint target compileFlags effectFlags = withFile fileName ReadMode $ \handle -> do contents <- hGetContents handle d3dCompile contents sourceName defines pInclude entryPoint target compileFlags effectFlags main :: IO () main = do _vb <- compileShaderFromFile "Triangle.fx" "VS" "vs_4_0" return () compileShaderFromFile :: String -> String -> String -> IO (Ptr ID3DBlob) compileShaderFromFile fileName entryPoint shaderModel = do Right res <- d3dCompileFromFile fileName Nothing Nothing nullPtr (Just entryPoint) shaderModel [d3dCompileEnableStrictness] [] return res }}} * `Triangle.fx` {{{ float4 VS( float4 Pos : POSITION ) : SV_POSITION { return Pos; } float4 PS( float4 Pos : SV_POSITION ) : SV_Target { return float4( 1.0f, 1.0f, 0.0f, 1.0f ); // Yellow, with Alpha = 1 } }}} Make sure that `Triangle.fx` is in the same directory as `Main.hs` when running this program. When compiled, this program works OK: {{{ $ C:\Users\RyanGlScott\Software\ghc-8.2.0.20170404\bin\ghc -lD3DCompiler Main.hs -fforce-recomp [1 of 1] Compiling Main ( Main.hs, Main.o ) Linking Main.exe ... $ .\Main.exe Before d3dCompile After d3dCompile }}} But with GHCi, it crashes: {{{ $ C:\Users\RyanGlScott\Software\ghc-8.2.0.20170404\bin\runghc -lD3DCompiler Main.hs Before d3dCompile Access violation in generated code when writing 0000000000000000 }}} I ran these tests on GHC 8.2.1, but I've also reproduced this bug in the past on GHC 8.0.2, so I don't think this is a new bug by any means. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13606> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #13757: Are you for or against writing "otherwise" as a keyword?
by GHC 03 Jun '17

03 Jun '17

#13757: Are you for or against writing "otherwise" as a keyword? -------------------------------------+------------------------------------- Reporter: vanto | Owner: (none) Type: feature | Status: new request | Priority: normal | Milestone: Component: Compiler | Version: 8.0.2 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Poor/confusing Unknown/Multiple | error message Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- {{{otherwise}}} is frequently used and therefore should have been written as a keyword.\\ {{{otherwise}}} was coded in prelude like this {{{otherwise = True}}}\\ So what is the reason that {{{otherwise}}} was not coded as a keyword?\\ \\ To understand you have to go back. It seems that this has been done to minimize the number of keywords.\\ Today it seems surprising. And I wonder why we wrote the conditional expression with {{{if}}}{{{then}}}{{{else}}}? That's three additional keywords.\\ \\ It would have been better to write the conditional expression as in Sasl {{{condition -> exp1 ; exp22}}} (By adapting the operators of course.)\\ Let's go back to {{{otherwise}}} . And take a look at this code.\\ {{{ otherwise :: Bool otherwise = Prelude.otherwise sig :: Integer -> Integer sig n | a = -1 | b = 0 | Main.otherwise = 1 where { a = n < 0; b = n == 0 } z :: Bool z = if Main.otherwise == Prelude.otherwise then True else False main :: IO () main = do { print (sig 0); print (z) } }}} This code works but the compiler result is different by adding or removing the first two lines of code.\\ For me it is urgent that {{{otherwise}}} become a keyword as in Miranda. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13757> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 12

[GHC] #13669: Identifier "Otherwise" in guarded equation can crash a program
by GHC 03 Jun '17

03 Jun '17

#13669: Identifier "Otherwise" in guarded equation can crash a program -------------------------------------+------------------------------------- Reporter: vanto | Owner: (none) Type: feature | Status: new request | Priority: normal | Milestone: Component: Compiler | Version: 8.0.2 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Other Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- In GHC.Base the identifier {{{otherwise}}} is defined as the value "True" and it helps to make guards more readable.\\ The identifier {{{otherwise}}} is not secure.\\ We can change his property easily and trick the compiler.\\ As written in GHC Base, {{{otherwise = True}}} is perfectible.\\ The problem here is to protect the "otherwise" identifier from uncontrolled access.\\ See the code below and the result of the compilation as well as the output.\\ {{{ import qualified Prelude as P otherwise :: P.Bool otherwise = P.False fun :: (P.Num a, P.Ord a) => a -> [P.Char] fun x | x P.< 0 = "down" | otherwise = "up" main :: P.IO () main = P.print (fun 7) c:\Sourcehs>ghc -Wall otherwise.hs [1 of 1] Compiling Main ( otherwise.hs, otherwise.o ) otherwise.hs:7:1: warning: [-Wincomplete-patterns] Pattern match(es) are non-exhaustive In an equation for `fun': Patterns not matched: _ otherwise.hs:11:17: warning: [-Wtype-defaults] * Defaulting the following constraints to type `P.Integer' (P.Num a0) arising from a use of `fun' at otherwise.hs:11:17-21 (P.Ord a0) arising from a use of `fun' at otherwise.hs:11:17-21 * In the first argument of `P.print', namely `(fun 7)' In the expression: P.print (fun 7) In an equation for `main': main = P.print (fun 7) Linking otherwise.exe ... c:\Sourcehs>otherwise otherwise: otherwise.hs:(7,1)-(8,27): Non-exhaustive patterns in function fun }}} I propose three solutions: 1- Delete {{{otherwise}}} from GHC.Base and use only TRUE.\\ 2- Add {{{otherwise}}} as a reserved word or keyword.\\ 3- Create protected objects in Haskell. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13669> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #13764: Creating a ticket named "New proposal" in GHC Trac.
by GHC 03 Jun '17

03 Jun '17

#13764: Creating a ticket named "New proposal" in GHC Trac. -------------------------------------+------------------------------------- Reporter: vanto | Owner: hvr Type: feature | Status: new request | Priority: normal | Milestone: Component: Trac & Git | Version: 8.0.2 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: -------------------------------------+------------------------------------- Instead of going to GitHub to make a proposal, I suggest instead to do it in GHC TRAC by creating a new ticket named {{{New Proposal}}} I think it's easier.\\ Of course all the boxes of the ticket will not be for information.\\ Either you write directly in the ticket or you write in a file attached to the ticket.\\ The attached file would be a template.\\ Of course if the text file is too long it can be shared in several parts.\\ In Trac one can managed the time spent since the opening of the proposal.\\ They can also be archived.\\ Making simple is my creed.\\ To be honest I don't understand anything about GitHub and making a proposal in GitHub is complicated.\\ So I created a specimen of proposal attached to ticket number #13757 to show you that it is feasible.\\ And I invite you already to criticize this specimen of proposal in the legal time and to give a conclusive answer in the attached ticket. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13764> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #11032: Missing result type handling for State# s in foreign import prim.
by GHC 02 Jun '17

02 Jun '17

#11032: Missing result type handling for State# s in foreign import prim. -------------------------------------+------------------------------------- Reporter: ekmett | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.2 (FFI) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: GHC rejects Unknown/Multiple | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- I've been using `foreign import prim` for a while now to craft custom primops on an as-needed basis. Attempting to write {{{#!hs {-# LANGUAGE MagicHash #-} {-# LANGUAGE UnboxedTuples #-} {-# LANGUAGE UnliftedFFITypes #-} {-# LANGUAGE GHCForeignImportPrim #-} {-# LANGUAGE ForeignFunctionInterface #-} foreign import prim "pinThreadzh" pinThread# :: State# s -> (# State# s, Int#, Int# #) foreign import prim "unpinThreadzh" unpinThread# :: State# s -> State# s }}} threw me for a loop when it didn't work. The former works fine. The latter complains that `State# s` isn't a valid result type! {{{ src/Concurrent/Internal/Thread.hs:23:1: Unacceptable result type in foreign declaration: ‘State# s’ cannot be marshalled in a foreign call When checking declaration: foreign import prim safe "static unpinThreadzh" unpinThread# :: State# s -> State# s }}} Builtin primitives work this way all the time, so I'm guessing we just missed it in a case statement somewhere inside the compiler when dealing with `foreign import prim` support. I can likely hack around it by giving back something like `(# State# s #)` for now but that is a rather unsatisfying solution. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11032> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4