September 2013 - ghc-tickets

[GHC] #8357: Pretty printing of kind-involving inferred types
by GHC 03 Oct '13

03 Oct '13

#8357: Pretty printing of kind-involving inferred types ------------------------------------+------------------------------------- Reporter: ksf | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- An example of inferred types (foo, bar, both), for brevity's sake omitting value-level details: {{{ {-# LANGUAGE TypeFamilies #-} {-# LANGUAGE DataKinds #-} {-# LANGUAGE KindSignatures #-} {-# LANGUAGE TypeOperators #-} import GHC.TypeLits data (:::) (sy :: Symbol) ty data Key (sy :: Symbol) data Rec (rs :: [*]) (*=) :: Key sy -> ty -> Rec '[sy ::: ty] (*=) = undefined (.*.) :: (Union xs ys ~ rs) => Rec xs -> Rec ys -> Rec rs (.*.) = undefined type family Union (xs :: [*]) (ys :: [*]) :: [*] where Union ((sy ::: t) ': xs) ys = (sy ::: t) ': Union xs ys Union '[] ys = ys fFoo :: Key "foo" fFoo = undefined fBar :: Key "bar" fBar = undefined foo :: Rec ((':) * ("foo" ::: [Char]) ('[] *)) foo = fFoo *= "foo" bar :: Rec ((':) * ("bar" ::: [Char]) ('[] *)) bar = fBar *= "bar" both :: Rec ((':) * ("foo" ::: [Char]) ((':) * ("bar" ::: [Char]) ('[] *))) both = foo .*. bar }}} There's actually two issues, here: First, the most serious one, GHC prints kind annotations (*) that it can't even parse back, and then [*] types are pretty printed with ': constructors instead of nicely with [,] syntax. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8357> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8384: -fwarn-empty-enumerations doesn't work with -fhpc
by GHC 03 Oct '13

03 Oct '13

#8384: -fwarn-empty-enumerations doesn't work with -fhpc ------------------------------------+------------------------------------- Reporter: rwbarton | Owner: Type: bug | Status: new Priority: low | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- It appears that tick insertion occurs before arithmetic sequences are desugared, so `warnAboutEmptyEnumerations` can't see the calls to `fromInteger`, because they have been wrapped in ticks already. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8384> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8334: unexpected type inference failure in main
by GHC 03 Oct '13

03 Oct '13

#8334: unexpected type inference failure in main ------------------------------------+------------------------------------- Reporter: carter | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- I'm not sure if this is a failure of me understanding the semantics of type families, or something where inference should work! Or maybe its just me failing to understand the rules / convention for main. {{{ {-# LANGUAGE MagicHash, UnboxedTuples #-} module PrefetchTest where import GHC.Prim import Data.Vector.Storable.Mutable -- import Foreign.ForeignPtr -- (unsafeForeignPtrToPtr) import Foreign.Ptr import GHC.ST import Data.Primitive.ByteArray import Control.Monad.Primitive -- newByteArray :: PrimMonad m => Int -> m (MutableByteArray (PrimState m)) -- sameMutableByteArray :: MutableByteArray s -> MutableByteArray s -> Bool --unsafeFreezeByteArray :: PrimMonad m => MutableByteArray (PrimState m) -> m ByteArray -- unsafeThawByteArray :: PrimMonad m => ByteArray -> m (MutableByteArray (PrimState m)) --main :: IO () main = do mv1 <- newByteArray 17 v1 <- unsafeFreezeByteArray mv1 return () }}} i get {{{ refetch.hs:42:5: No instance for (Monad m0) arising from a do statement The type variable ‛m0’ is ambiguous Relevant bindings include main :: m0 () (bound at prefetch.hs:41:1) Note: there are several potential instances: instance Monad (ST s) -- Defined in ‛GHC.ST’ instance Monad ((->) r) -- Defined in ‛GHC.Base’ instance Monad IO -- Defined in ‛GHC.Base’ ...plus one other In a stmt of a 'do' block: mv1 <- newByteArray 17 In the expression: do { mv1 <- newByteArray 17; v1 <- unsafeFreezeByteArray mv1; return () } In an equation for ‛main’: main = do { mv1 <- newByteArray 17; v1 <- unsafeFreezeByteArray mv1; return () } prefetch.hs:42:12: No instance for (PrimMonad m0) arising from a use of ‛newByteArray’ The type variable ‛m0’ is ambiguous Relevant bindings include main :: m0 () (bound at prefetch.hs:41:1) Note: there are several potential instances: instance PrimMonad IO -- Defined in ‛Control.Monad.Primitive’ instance PrimMonad (ST s) -- Defined in ‛Control.Monad.Primitive’ In a stmt of a 'do' block: mv1 <- newByteArray 17 In the expression: do { mv1 <- newByteArray 17; v1 <- unsafeFreezeByteArray mv1; return () } In an equation for ‛main’: main = do { mv1 <- newByteArray 17; v1 <- unsafeFreezeByt }}} this is because main seems to not be assumed to have type IO () If i add that type ascription to main, or I comment out the name, type inference succeeds. I understand and know we have special treatment for whats the main module, but this is an example of that special treatment changing the type inference! Is that a known issue? (its obvious to me now, but is it reasonable?) this is pretty much a nonissue, but i think its still worth noting. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8334> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8340: support for generating annotations from TH
by GHC 02 Oct '13

02 Oct '13

#8340: support for generating annotations from TH -------------------------------------------+------------------------------- Reporter: errge | Owner: Type: feature request | Status: new Priority: normal | Milestone: 7.8.1 Component: Compiler | Version: 7.7 Keywords: | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Easy (less than 1 hour) | Type of failure: Blocked By: | None/Unknown Related Tickets: | Test Case: | Blocking: 7867 -------------------------------------------+------------------------------- This patch adds support for template haskell generation of annotations. This is currently only possible directly via ```{-# ANN function "whatever annotation #-}```. The attached patch makes it possible to do the same from TH. Module and type annotations are supported too. This is part of a larger effort to make it possible to cross-module communicate between template haskell runs. See #7867 for a longer discussion. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8340> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #8283: utils/fingerprint/fingerprint.py doesn't work on Windows
by GHC 02 Oct '13

02 Oct '13

#8283: utils/fingerprint/fingerprint.py doesn't work on Windows -------------------------------------------+------------------------------- Reporter: schyler | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Build System | Version: 7.7 Keywords: | Operating System: Windows Architecture: Unknown/Multiple | Type of failure: Difficulty: Easy (less than 1 hour) | None/Unknown Blocked By: | Test Case: Related Tickets: | Blocking: -------------------------------------------+------------------------------- Invocation: {{{ utils/fingerprint/fingerprint.py create }}} Produces: {{{ WindowsError: [Error 193] %1 is not a valid Win32 application }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8283> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #8343: Interrupting readProcess doesn't work
by GHC 02 Oct '13

02 Oct '13

#8343: Interrupting readProcess doesn't work --------------------------+------------------------------------------------ Reporter: akio | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: | Version: 7.6.3 libraries/base | Operating System: Linux Keywords: | Type of failure: Incorrect result at runtime Architecture: x86_64 | Test Case: (amd64) | Blocking: Difficulty: | Unknown | Blocked By: | Related Tickets: #5766 | --------------------------+------------------------------------------------ The following program should terminate in 1 second, but it does not: {{{ import System.Process import System.Timeout main = timeout 1000000 $ print =<< readProcess "sleep" ["2h"] "" }}} The problem is in the exception handler in {{{readProcess}}}. When it receives an asynchronous exception, it tries to clean up by closing the pipes. However the attempt to close {{{outh}}} blocks because the reader thread (reading with {{{hGetContents}}}) is blocking on the handle. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8343> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #8359: ConstraintKinds require UndecidableInstances when it doesn't need it
by GHC 01 Oct '13

01 Oct '13

#8359: ConstraintKinds require UndecidableInstances when it doesn't need it ----------------------------+---------------------------------------------- Reporter: | Owner: thomaseding | Status: new Type: bug | Milestone: Priority: normal | Version: 7.4.2 Component: Compiler | Operating System: Unknown/Multiple Keywords: | Type of failure: GHC rejects valid program Architecture: | Test Case: Unknown/Multiple | Blocking: Difficulty: Unknown | Blocked By: | Related Tickets: | ----------------------------+---------------------------------------------- Using ConstraintKinds to alias a bunch of class constraints fails to compile without UndecidableInstances. The same code that manually spells out class constraints without ConstraintKinds compiles just fine though. Test case below ------------------------------------- {-# LANGUAGE ConstraintKinds #-} {-# LANGUAGE MultiParamTypeClasses #-} class DifferentTypes a b type DifferentTypes3 a b c = (DifferentTypes a b, DifferentTypes b c, DifferentTypes a c) class Foo a class Bar a -- Buggy instance requires UndecidableInstances to compile instance (DifferentTypes3 a b c, Bar a, Bar b, Bar c) => Foo (a, b, c) -- Equivalent instance compiles when manually expanding constraint type instance (DifferentTypes a b, DifferentTypes b c, DifferentTypes a c, Bar a, Bar b, Bar c) => Foo (a, b, c) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8359> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #8329: dmdTransformDictSelSig panic
by GHC 01 Oct '13

01 Oct '13

#8329: dmdTransformDictSelSig panic -----------------------------------+--------------------------------------- Reporter: monoidal | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Compile-time crash Unknown/Multiple | Test Case: Difficulty: Unknown | Blocking: Blocked By: | Related Tickets: | -----------------------------------+--------------------------------------- Installing `xmlhtml` with HEAD panicks, while it works in 7.6.3. {{{ ghc: panic! (the 'impossible' happened) (GHC version 7.7.20130918 for x86_64-unknown-linux): dmdTransformDictSelSig: split failed }}} I reduced this panic to four files: Common.hs {{{ module Common where import Data.String (fromString) encodingName :: String encodingName = fromString "" }}} Meta.hs {{{ {-# OPTIONS_GHC -O0 #-} module Meta where import Data.String }}} Parse.hs {{{ module Parse where import Meta }}} XmlHtml.hs {{{ module XmlHtml where import Parse import Common }}} Remove *.hi and *.o files and compile `ghc -O XmlHtml` to get the panic. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8329> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #8247: Dependency tracking (--make) broken for re-exported modules
by GHC 01 Oct '13

01 Oct '13

#8247: Dependency tracking (--make) broken for re-exported modules ------------------------------------+------------------------------------- Reporter: heisenbug | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- Say, I re-export a module, from which I hide some bindings: {{{ module B (module A) where import A hiding (a1) }}} where `A` is defined thus: {{{ module A where a1 = 5 a2 = 42 a3 = 113 }}} I use `B` from `C`: {{{ module C where import B a2 = 142 }}} Build `C`, with `ghc --make C.hs`, observing redefinition error (of `a2`). Now hide also `a2` in `B` like this (by editing `B.hs`): {{{ module B (module A) where import A hiding (a1, a2) }}} Build `C` again, and observe that the error persists, because `B` has not been rebuilt, and its interface file not regenerated. Deleting `B.hi` and recompiling resolves the situation. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8247> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #8377: GHCi crashes with unboxed values on 32-bit platforms
by GHC 01 Oct '13

01 Oct '13

#8377: GHCi crashes with unboxed values on 32-bit platforms ---------------------------+------------------------------------- Reporter: akio | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: GHCi | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: x86 | Type of failure: GHCi crash Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ---------------------------+------------------------------------- The following program causes GHCi to crash, on x86 Linux and x86 Windows. {{{ {-# LANGUAGE MagicHash #-} import System.Mem import GHC.Base main = do let list = concatMap buildThunk [0..10000] length list `seq` performGC print $ last list buildThunk :: Int -> [Int] buildThunk (I# k) = [f k] f :: Int# -> Int f x = I# x }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8377> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6