July 2014 - ghc-tickets

Re: [GHC] #4218: System.Random is way too lazy
by GHC 15 Jul '14

15 Jul '14

#4218: System.Random is way too lazy -------------------------------------+------------------------------------- Reporter: EyalLotem | Owner: ekmett Type: bug | Status: patch Priority: low | Milestone: 7.10.1 Component: | Version: 6.12.3 libraries/random | Keywords: random stack Resolution: | overflow Differential Revisions: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: Difficulty: Unknown | Blocking: Blocked By: | Related Tickets: | -------------------------------------+------------------------------------- Changes (by simonpj): * cc: ekmett (added) * owner: rrnewton => ekmett Comment: Making Edward K owner, since it's a Core Libraries issue. Just a question of reviewing the offered patch! Simon -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/4218#comment:12> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #5051: Typechecker behaviour change
by GHC 15 Jul '14

15 Jul '14

#5051: Typechecker behaviour change -------------------------------------+------------------------------------- Reporter: igloo | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.10.1 Component: Compiler | Version: 7.0.2 Resolution: | Keywords: Differential Revisions: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: Difficulty: Unknown | typecheck/should_compile/T5051 Blocked By: | Blocking: Related Tickets: | -------------------------------------+------------------------------------- Comment (by simonpj): As I say in comment:5, I think this is the right behaviour, so I don't propose to fix it at all. Only leaving it open in case others trip over something similar. Simon -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5051#comment:8> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #9299: GHCi type inference error
by GHC 15 Jul '14

15 Jul '14

#9299: GHCi type inference error ------------------------------------+------------------------------------- Reporter: erikd | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: GHCi | Version: 7.8.2 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: Other Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- This code (required conduit-1.1.6): {{{ {- Bolled down from https://github.com/snoyberg/conduit/blob/process/conduit- extra/Data/Conduit/Process.hs -} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE TypeFamilies #-} {-# LANGUAGE RankNTypes #-} import System.Process import Control.Monad.IO.Class (MonadIO, liftIO) import System.IO (Handle, hClose) import Data.Conduit import Data.ByteString (ByteString) import Data.Maybe (fromMaybe) import Control.Applicative ((<$>), (<*>)) data UseProvidedHandle = UseProvidedHandle data Inherited = Inherited data ClosedStream = ClosedStream class InputSource a where isStdStream :: (Maybe Handle -> IO a, Maybe StdStream) instance InputSource Handle where isStdStream = (\(Just h) -> return h, Just CreatePipe) instance InputSource ClosedStream where isStdStream = (\(Just h) -> hClose h >> return ClosedStream, Just CreatePipe) instance (r ~ (), MonadIO m, i ~ ByteString) => InputSource (ConduitM i o m r) where isStdStream = (\(Just h) -> return $ sinkHandle h, Just CreatePipe) instance (r ~ (), r' ~ (), MonadIO m, MonadIO n, i ~ ByteString) => InputSource (ConduitM i o m r, n r') where isStdStream = (\(Just h) -> return (sinkHandle h, liftIO $ hClose h), Just CreatePipe) instance InputSource Inherited where isStdStream = (\Nothing -> return Inherited, Just Inherit) instance InputSource UseProvidedHandle where isStdStream = (\Nothing -> return UseProvidedHandle, Nothing) sinkHandle :: MonadIO m => Handle -> Consumer ByteString m () sinkHandle = error "sinkHandle" conduitProcess :: (MonadIO m, InputSource stdin) => CreateProcess -> m (stdin, ProcessHandle) conduitProcess cp = liftIO $ do let (getStdin, stdinStream) = isStdStream (stdinH, _, _, ph) <- createProcess cp { std_in = fromMaybe (std_in cp) stdinStream } (,) <$> getStdin stdinH <*> return ph main :: IO () main = putStrLn "Hello" }}} Compiles fine with ghc, and runs correctly with `runghc`, but fails to load into ghci with the following error: {{{ GHCi, version 7.8.3: http://www.haskell.org/ghc/ :? for help Loading package ghc-prim ... linking ... done. Loading package integer-gmp ... linking ... done. Loading package base ... linking ... done. [1 of 1] Compiling Main ( test.hs, interpreted ) test.hs:45:9: No instance for (InputSource a0) arising from the ambiguity check for ‘stdinStream’ The type variable ‘a0’ is ambiguous When checking that ‘stdinStream’ has the inferred type ‘Maybe StdStream’ Probable cause: the inferred type is ambiguous In the second argument of ‘($)’, namely ‘do { let (getStdin, stdinStream) = isStdStream; (stdinH, _, _, ph) <- createProcess (cp {std_in = fromMaybe (std_in cp) stdinStream}); (,) <$> getStdin stdinH <*> return ph }’ In the expression: liftIO $ do { let (getStdin, stdinStream) = isStdStream; (stdinH, _, _, ph) <- createProcess (cp {std_in = fromMaybe (std_in cp) stdinStream}); (,) <$> getStdin stdinH <*> return ph } Failed, modules loaded: none. }}} Same problem in 7.6.3, 7.8.2 and 7.8.3. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9299> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #9263: Iface type variable out of scope with default associated types and polykinds
by GHC 15 Jul '14

15 Jul '14

#9263: Iface type variable out of scope with default associated types and polykinds ------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.9 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- I have {{{ {-# LANGUAGE DataKinds, PolyKinds, TypeFamilies #-} module A where import Data.Proxy class kproxy ~ 'KProxy => PEq (kproxy :: KProxy a) where type F (x :: a) :: Bool type F (x :: a) = False }}} {{{ {-# LANGUAGE DataKinds, KindSignatures, TypeFamilies #-} module B where import A import Data.Proxy data Void instance PEq ('KProxy :: KProxy Void) }}} {{{ module C where import B }}} Now, this happens in a terminal window: {{{ rae:21:38:44 ~/temp> ghc A.hs [1 of 1] Compiling A ( A.hs, A.o ) rae:21:38:46 ~/temp> ghc B.hs [2 of 2] Compiling B ( B.hs, B.o ) rae:21:38:47 ~/temp> ghc C.hs [3 of 3] Compiling C ( C.hs, C.o ) The interface for ‘B’ Declaration for R:FVoidx: Iface type variable out of scope: a Cannot continue after interface file error }}} If any two of the files are compiled together, the error does not occur. The emptiness of the `Void` type is irrelevant -- the error initially occurred with an inhabited type. This is a regression from 7.8.2. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9263> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #9264: Scoped kind variables do not work with default associated types
by GHC 15 Jul '14

15 Jul '14

#9264: Scoped kind variables do not work with default associated types ------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.2 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- I have {{{ {-# LANGUAGE PolyKinds, TypeFamilies, ScopedTypeVariables #-} module Bug where class C (a :: k) where type F (a :: k) type F (a :: k) = Int }}} Compiling gives me {{{ Bug.hs:7:11: The signature specified kind ‘k1’, but ‘a’ has kind ‘k2’ In the type ‘(a :: k)’ In the type instance declaration for ‘F’ In the class declaration for ‘C’ }}} The error is on the line declaring the default associated type for `F`. This happens in both 7.8.2 and HEAD. Having !ScopedTypeVariables causes no change, but I put it in the example to emphasize that adding !ScopedTypeVariables is not the solution. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9264> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #9063: Default associated type instances are too general
by GHC 15 Jul '14

15 Jul '14

#9063: Default associated type instances are too general ------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.2 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- I compile the following with `-ddump-tc` and `-fprint-explicit-kinds`: {{{ {-# LANGUAGE DataKinds, PolyKinds, TypeFamilies, TypeOperators, UndecidableInstances #-} module Bug where import Data.Type.Equality import Data.Proxy class kproxy ~ 'KProxy => PEq (kproxy :: KProxy a) where type (:==) (x :: a) (y :: a) :: Bool type x :== y = x == y }}} and I see {{{ TYPE SIGNATURES TYPE CONSTRUCTORS PEq :: forall (a :: BOX). KProxy a -> Constraint class (~) (KProxy a) kproxy (KProxy a) => PEq (a::BOX) (kproxy::KProxy a) Roles: [nominal, nominal] RecFlag NonRecursive type family (:==) (a::BOX) (x::a) (y::a) :: Bool (open) Defaults: [(k, BOX), (x, k), (y, k)] k x y = (==) k x y COERCION AXIOMS axiom Bug.NTCo:PEq :: PEq a kproxy = (~) (KProxy a) kproxy ('KProxy a) }}} The problem is the `Defaults:` line. That `k` should be an `a`. This is not a printing/tidying problem -- the default does not work when it should. For example, if I add the following {{{ instance PEq ('KProxy :: KProxy Bool) foo :: Proxy (True :== True) -> Proxy (True == True) foo = id }}} I get {{{ Couldn't match type ‘(:==) Bool 'True 'True’ with ‘'True’ Expected type: Proxy Bool ((:==) Bool 'True 'True) -> Proxy Bool ((==) Bool 'True 'True) Actual type: Proxy Bool 'True -> Proxy Bool 'True In the expression: id In an equation for ‘foo’: foo = id }}} Adding kind signatures (that is `:: a`) to the arguments of the type family default fixes the problem. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9063> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 10

[GHC] #8615: All packages built with GHC / clang are failing
by GHC 14 Jul '14

14 Jul '14

#8615: All packages built with GHC / clang are failing ------------------------------------+------------------------------------- Reporter: sylvestre | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- I am trying to rebuild the Debian archive using clang instead of gcc. The biggest source of errors (622 over 2188 [1]) is related to GHC. Clang does not implement the "-u" option [2]. This is causing most (all?) of the "GHC based" packages to fail to build. with: {{{ /usr/lib/ghc/libHSrts.a(RtsStartup.o): In function `hs_exit_': (.text+0x49): undefined reference to `base_GHCziTopHandler_flushStdHandles_closure' }}} Of course, if you think it is not GHC's fault at all or there is no workaround, don't hesitate to close this bug. For example: http://clang.debian.net/logs/2013-07-14/agda_2.3.2.1-1_unstable_clang.log http://clang.debian.net/logs/2013-07-14/happy_1.18.10-2_unstable_clang.log [1] http://clang.debian.net/status.php?version=3.3&key=MISSING_OPTION_U [2] http://llvm.org/bugs/show_bug.cgi?id=18250 -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8615> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #9086: main :: IO Int does different things with runghc and when compiled
by GHC 14 Jul '14

14 Jul '14

#9086: main :: IO Int does different things with runghc and when compiled --------------------------+------------------------------------------------ Reporter: nh2 | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.10.1 Component: | Version: 7.8.2 Compiler | Operating System: Unknown/Multiple Keywords: | Type of failure: Incorrect result at runtime Architecture: | Test Case: Unknown/Multiple | Blocking: Difficulty: | Unknown | Blocked By: | Related Tickets: | --------------------------+------------------------------------------------ Consider {{{ main :: IO Int main = return 1 }}} This does different things when compiled and when run with `runghc`/`runhaskell` (it prints an extra `1` with the latter). For practical purposes, I think it is beneficial if either of the two ways to run a Haskell program have the same output. Or did somebody intend this to happen? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9086> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 10

Re: [GHC] #5108: Allow unicode sub/superscript symbols in both identifiers and operators
by GHC 14 Jul '14

14 Jul '14

#5108: Allow unicode sub/superscript symbols in both identifiers and operators ---------------------------------------+----------------------------------- Reporter: mikhail.vorozhtsov | Owner: Type: feature request | Status: infoneeded Priority: normal | Milestone: 7.10.1 Component: Compiler (Parser) | Version: 7.1 Resolution: | Keywords: lexer unicode Operating System: Unknown/Multiple | Architecture: Type of failure: None/Unknown | Unknown/Multiple Test Case: | Difficulty: Unknown Blocking: | Blocked By: | Related Tickets: ---------------------------------------+----------------------------------- Comment (by mikhail.vorozhtsov): Here is the [http://comments.gmane.org/gmane.comp.lang.haskell.glasgow.user/24760 link] to the discussion. I came up with a proposal that avoids the whole "primes on operators" thing for now. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5108#comment:13> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #8010: Add forkOSUnmasked (patch)
by GHC 14 Jul '14

14 Jul '14

#8010: Add forkOSUnmasked (patch) -------------------------------------+------------------------------------ Reporter: joeyadams | Owner: Type: feature request | Status: patch Priority: normal | Milestone: 7.10.1 Component: libraries/base | Version: 7.7 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: -------------------------------------+------------------------------------ Changes (by thoughtpolice): * cc: ekmett (added) * status: new => patch * milestone: => 7.10.1 -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8010#comment:4> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0