March 2016 - ghc-tickets

[GHC] #11687: GHC reports unimplemented/strange closure type 5937 in the ARM platform
by GHC 25 Mar '16

25 Mar '16

#11687: GHC reports unimplemented/strange closure type 5937 in the ARM platform --------------------------------+---------------------------------- Reporter: highfly | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 Keywords: | Operating System: Linux Architecture: arm | Type of failure: Runtime crash Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: --------------------------------+---------------------------------- angie: internal error: scavenge: unimplemented/strange closure type 5937 @ 0x74c01a60 (GHC version 7.10.3 for arm_unknown_linux) Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11687> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #11752: ghc-8.0 branch fails to build with ghc-7.8
by GHC 25 Mar '16

25 Mar '16

#11752: ghc-8.0 branch fails to build with ghc-7.8 -------------------------------------+------------------------------------- Reporter: erikd | Owner: Type: bug | Status: new Priority: normal | Milestone: 8.0.1 Component: Compiler | Version: 8.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Building GHC Unknown/Multiple | failed Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Problems with `getAllocationCounter` not being available in 7.8's `GHC.Conc` and ghc-7.8 not having `pure` in the prelude. Patch attached. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11752> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

Re: [GHC] #4114: Add a flag to remove/delete intermediate files generated by GHC
by GHC 25 Mar '16

25 Mar '16

#4114: Add a flag to remove/delete intermediate files generated by GHC -------------------------------------+------------------------------------- Reporter: guest | Owner: kaiha Type: feature request | Status: patch Priority: normal | Milestone: Component: Compiler | Version: 6.10.4 Resolution: | Keywords: newcomer Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: #2258 | Differential Rev(s): Phab:D2021 Wiki Page: | -------------------------------------+------------------------------------- Comment (by Phyx-): Hmm these test break on Windows since they don't properly take note of the application suffix. e.g. on Windows It needs to check for "T4114a.exe" instead of "T4114a". -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/4114#comment:26> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #11753: Type hole(?) causes compiler failure
by GHC 25 Mar '16

25 Mar '16

#11753: Type hole(?) causes compiler failure -------------------------------------+------------------------------------- Reporter: akfp | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 Keywords: | Operating System: Linux Architecture: x86_64 | Type of failure: Compile-time (amd64) | crash Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- I tried to understand soem Servant code, and added {{{#!hs bodyCheck :: _ }}} To the following code {{{#!hs -# LANGUAGE DataKinds #-} {-# LANGUAGE TypeFamilies #-} {-# LANGUAGE TypeOperators #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE TypeSynonymInstances #-} {-# LANGUAGE MultiParamTypeClasses #-} module Servant.Files ( FilesTmp , FilesMem , MultiPartData , MultiPartDataT , Tmp , Mem ) where import Control.Monad.Trans.Resource import Data.ByteString.Lazy (ByteString) import Network.Wai.Parse import Servant import Servant.Server.Internal -- Backends for file upload: in memory or in /tmp ? data Mem data Tmp class KnownBackend b where type Storage b :: * withBackend :: Proxy b -> (BackEnd (Storage b) -> IO r) -> IO r instance KnownBackend Mem where type Storage Mem = ByteString withBackend Proxy f = f lbsBackEnd instance KnownBackend Tmp where type Storage Tmp = FilePath withBackend Proxy f = runResourceT . withInternalState $ \s -> f (tempFileBackEnd s) -- * Files combinator, to get all of the uploaded files data Files b type MultiPartData b = ([Param], [File (Storage b)]) type MultiPartDataT b = ((MultiPartData b -> IO (MultiPartData b)) -> IO (MultiPartData b)) type FilesMem = Files Mem type FilesTmp = Files Tmp instance (KnownBackend b, HasServer sublayout config) => HasServer (Files b :> sublayout) config where type ServerT (Files b :> sublayout) m = MultiPartDataT b -> ServerT sublayout m route Proxy config subserver = WithRequest $ \request -> route (Proxy :: Proxy sublayout) config (addBodyCheck subserver (bodyCheck request)) where bodyCheck :: _ bodyCheck request = return $ Route (\f -> withBackend (Proxy :: Proxy b) $ \pb -> parseRequestBody pb request >>= f ) }}} Output: {{{ > :load src/Servant/Test.hs [1 of 1] Compiling Servant.Files ( src/Servant/Test.hs, interpreted ) src/Servant/Test.hs:59:59: Couldn't match type ‘r’ with ‘([Param], [File (Storage b)])’ because type variable ‘b’ would escape its scope This (rigid, skolem) type variable is bound by the instance declaration at src/Servant/Test.hs:(53,10)-(54,80) Expected type: Delayed (((([Param], [File (Storage b)]) -> IO r) -> IO r) -> Server sublayout) Actual type: Delayed (Server (Files b :> sublayout)) Relevant bindings include bodyCheck :: Network.Wai.Internal.Request -> m (RouteResult ((([Param], [File (Storage b)]) -> IO r) -> IO r)) (bound at src/Servant/Test.hs:62:7) subserver :: Delayed (Server (Files b :> sublayout)) (bound at src/Servant/Test.hs:58:22) route :: Proxy (Files b :> sublayout) -> Context config -> Delayed (Server (Files b :> sublayout)) -> Router (bound at src/Servant/Test.hs:58:3) In the first argument of ‘addBodyCheck’, namely ‘subserver’ In the third argument of ‘route’, namely ‘(addBodyCheck subserver (bodyCheck request))’ src/Servant/Test.hs:59:70: Couldn't match type ‘m’ with ‘IO’ ‘m’ is untouchable inside the constraints (KnownBackend b, HasServer sublayout config) bound by the instance declaration at src/Servant/Test.hs:(53,10)-(54,80)ghc: panic! (the 'impossible' happened) (GHC version 7.10.3 for x86_64-unknown-linux): No skolem info: m_a5An[sk] Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11753> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #11728: Core lint errors
by GHC 25 Mar '16

25 Mar '16

#11728: Core lint errors -------------------------------------+------------------------------------- Reporter: Iceland_jack | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.1 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 have an older version (8.1.20160117), does this still occur? Compiling with `ghci -dcore-lint -ignore-dot-ghci` -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11728> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #11626: type variable out of scope core lint error when compiling attoparsec
by GHC 25 Mar '16

25 Mar '16

#11626: type variable out of scope core lint error when compiling attoparsec -------------------------------------+------------------------------------- Reporter: luite | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1-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: -------------------------------------+------------------------------------- I'm getting a core lint error when compiling attoparsec with the `-dcore- lint` option {{{ *** Core Lint errors : in result of SpecConstr *** <no location info>: warning: In the type ‘Buffer -> Pos -> More -> a_aaPH -> IResult Text r_aaPI’ @ a_aaPH is out of scope *** Offending Program *** ... 57000 lines of offending program cut *** End of Offense *** }}} * attoparsec 0.13.0.2 commit 4f137347be02106765f6897059b88219c79bb86c * ghc 8.0.1-rc2 and also 892de051aaaf9a346dadc74592a9cc375b3223a4 (feb 20) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11626> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #11726: Document change in implicit quantification
by GHC 25 Mar '16

25 Mar '16

#11726: Document change in implicit quantification -------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: task | Status: new Priority: high | Milestone: 8.0.1 Component: Documentation | Version: 7.10.3 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: -------------------------------------+------------------------------------- Simon's big wildcard refactor (1e041b7382b6aa329e4ad9625439f811e0f27232) changed the way GHC does implicit quantification. I don't see documentation for this, however. (That patch doesn't touch the manual, and there are no release notes, at least.) This report was prompted by [https://mail.haskell.org/pipermail/haskell- cafe/2016-March/123487.html this Haskell-Café post], where Tom Ellis asks why we need {{{ -data PackMap a b s t = PackMap (Applicative f => (a -> f b) -> s -> f t) +data PackMap a b s t = PackMap (forall f. Applicative f => (a -> f b) -> s -> f t) }}} to update for 8.0. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11726> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #11750: 32 bit CPUs: Couldn't match expected type ‘Word#’ with actual type ‘Word64#’
by GHC 25 Mar '16

25 Mar '16

#11750: 32 bit CPUs: Couldn't match expected type ‘Word#’ with actual type ‘Word64#’ -------------------------------------+------------------------------------- Reporter: erikd | Owner: erikd Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Building GHC Unknown/Multiple | failed Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Getting this on PowerPC and Arm: {{{ libraries/base/GHC/Word.hs:605:39: error: • Couldn't match expected type ‘Word#’ with actual type ‘Word64#’ • In the first argument of ‘eqWord#’, namely ‘x’ In the first argument of ‘isTrue#’, namely ‘(x `eqWord#` y)’ In the expression: isTrue# (x `eqWord#` y) libraries/base/GHC/Word.hs:605:51: error: • Couldn't match expected type ‘Word#’ with actual type ‘Word64#’ • In the second argument of ‘eqWord#’, namely ‘y’ In the first argument of ‘isTrue#’, namely ‘(x `eqWord#` y)’ In the expression: isTrue# (x `eqWord#` y) libraries/base/GHC/Word.hs:606:39: error: • Couldn't match expected type ‘Word#’ with actual type ‘Word64#’ • In the first argument of ‘neWord#’, namely ‘x’ In the first argument of ‘isTrue#’, namely ‘(x `neWord#` y)’ In the expression: isTrue# (x `neWord#` y) libraries/base/GHC/Word.hs:606:51: error: • Couldn't match expected type ‘Word#’ with actual type ‘Word64#’ • In the second argument of ‘neWord#’, namely ‘y’ In the first argument of ‘isTrue#’, namely ‘(x `neWord#` y)’ In the expression: isTrue# (x `neWord#` y) }}} Same happens for `Int64` in `GHC/Int.hs`. Patch on the way. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11750> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #11667: Incorrect pattern synonym types in error messages
by GHC 24 Mar '16

24 Mar '16

#11667: Incorrect pattern synonym types in error messages -------------------------------------+------------------------------------- Reporter: rdragon | Owner: rdragon Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1-rc2 Keywords: | Operating System: Unknown/Multiple PatternSynonyms | Architecture: | Type of failure: Incorrect Unknown/Multiple | warning at compile-time Test Case: | Blocked By: Blocking: | Related Tickets: #11213 Differential Rev(s): Phab:D1967 | Wiki Page: -------------------------------------+------------------------------------- Here are some examples of incorrect pattern synonym types in error messages. **Example 1** {{{#!hs {-# LANGUAGE PatternSynonyms #-} module Mod1 where pattern Pat :: Eq a => Maybe a pattern Pat <- Just 42 }}} This gives the error (note the missing `Eq a =>`): {{{ Mod1.hs:4:21: error: * Could not deduce (Num a) arising from the literal `42' from the context: Eq a bound by the type signature for pattern synonym `Pat': Maybe a at Mod1.hs:4:9-11 Possible fix: add (Num a) to the context of the type signature for pattern synonym `Pat': Maybe a * In the pattern: 42 In the pattern: Just 42 In the declaration for pattern synonym `Pat' }}} **Example 2** {{{#!hs {-# LANGUAGE PatternSynonyms, GADTs #-} module Mod1 where data T where MkT :: a -> T pattern Pat :: () => b ~ Bool => a -> b -> (b, T) pattern Pat x y = (y, MkT x) }}} We get the error (note the missing contexts and the incorrect quantification `forall b`, missing `a`): {{{ Mod1.hs:6:27: error: * Couldn't match type `b' with `Bool' arising from the "provided" constraints claimed by the signature of `Pat' `b' is a rigid type variable bound by the type signature for pattern synonym `Pat': forall b. a -> b -> (b, T) at Mod1.hs:5:16 * In the declaration for pattern synonym `Pat' * Relevant bindings include y :: b (bound at Mod1.hs:6:20) }}} **Example 3** {{{#!hs {-# LANGUAGE PatternSynonyms #-} module Mod1 where pattern Pat :: Eq a => Show a => a -> Maybe a pattern Pat x <- Just x }}} We get the error: {{{ Mod1.hs:4:23: error: * Could not deduce (Show a) arising from the "provided" constraints claimed by the signature of `Pat' from the context: Eq a bound by the type signature for pattern synonym `Pat': a -> Maybe a at Mod1.hs:4:9-11 * In the declaration for pattern synonym `Pat' }}} Here we could perhaps remove the whole "from the context ... at Mod1.hs:4:9-11" part, as it doesn't make much sense to tell the user that we could not deduce a "provided" constraint from the "required" context. **Example 4** {{{#!hs {-# LANGUAGE PatternSynonyms, GADTs #-} module Mod1 where data T a where MkT :: (Num a, Show a) => a -> T a pattern Pat :: (Eq a) => (Show a) => T a pattern Pat = MkT 42 }}} This gives the error (note that the signature in the error is of the pattern synonym `Pat` when used in an expression context): {{{ Mod1.hs:6:15: error: * Could not deduce (Num a) arising from a use of `MkT' from the context: (Eq a, Show a) bound by the type signature for pattern synonym `Pat': (Eq a, Show a) => T a at Mod1.hs:6:1-20 Possible fix: add (Num a) to the context of the type signature for pattern synonym `Pat': (Eq a, Show a) => T a * In the expression: MkT 42 In an equation for `$bPat': $bPat = MkT 42 }}} The problems in these four examples are caused by the use of `SigSkol (PatSynCtxt n) (Check ty)` for representing pattern synonym types (and the types of bidirectional pattern synonyms when used in an expression context). A possible solution can be found in Phab:D1967. Simon notes the following: > Example 3 raises an interesting point we can debate once we have a ticket. Namely, is this ok? > {{{#!hs > pattern Pat :: Ix a => Ix a => a -> Maybe a > pattern Pat x <- Just x > }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11667> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #11644: Core lint error in result of Specialise for TEST=T3220 WAY=profasm
by GHC 24 Mar '16

24 Mar '16

#11644: Core lint error in result of Specialise for TEST=T3220 WAY=profasm -------------------------------------+------------------------------------- Reporter: thomie | Owner: Type: bug | Status: new Priority: highest | Milestone: 8.0.1 Component: Compiler | Version: 8.0.1-rc2 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: indexed- | Blocked By: types/should_compile/T3220 | Blocking: | Related Tickets: #11371, #11643 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- This is the code: {{{ {-# LANGUAGE TypeFamilies, ScopedTypeVariables#-} module T3220 where class Foo m where type Bar m :: * action :: m -> Bar m -> m right x m = action m (Right x) right' :: (Either a b ~ Bar m, Foo m) => b -> m -> m right' x m = action m (Right x) instance Foo Int where type Bar Int = Either Int Int action m a = either (*) (+) a m instance Foo Float where type Bar Float = Either Float Float action m a = either (*) (+) a m foo = print $ right (1::Int) (3 :: Int) bar = print $ right (1::Float) (3 :: Float) }}} Invocation: {{{ $ ghc-8.0.1 T3220.hs -prof -O -dcore-lint }}} {{{ [1 of 1] Compiling T3220 ( T3220.hs, T3220.o ) *** Core Lint errors : in result of Specialise *** <no location info>: warning: In the expression: action @ Float $fFooFloat m_ayb ((Right @ Float @ Float x_aya) `cast` (Sub (Sym cobox_aUe) :: Either a_aU3 b_aU4 ~R# Bar m_aU1)) cobox_aUe :: Bar m_aU1 ~# Either a_aU3 b_aU4 [LclId[CoVarId], Str=DmdType] is out of scope }}} Interestingly, with HEAD (2aee41960aa00fe09a2cd1983e02c15e06013037), it hits an assert from #11371. {{{ =====> T3220(profasm) 12 of 21 [0, 11, 0] cd ./indexed-types/should_compile && "/home/thomas/ghc- validate/inplace/test spaces/ghc-stage2" -c T3220.hs -fforce-recomp -dcore-lint -dcmm-lint -dno-debug-output -no-user-package-db -rtsopts -fno-warn-tabs -fno-warn-missed-specialisations -fno-ghci-history -O -prof -static -auto-all > T3220.comp.stderr 2>&1 Compile failed (status 256) errors were: ghc-stage2: panic! (the 'impossible' happened) (GHC version 8.1.20160221 for x86_64-unknown-linux): ASSERT failed! CallStack (from HasCallStack): assertPprPanic, called at compiler/types/TyCoRep.hs:1942:51 in ghc:TyCoRep checkValidSubst, called at compiler/types/TyCoRep.hs:2076:17 in ghc:TyCoRep substCo, called at compiler/coreSyn/CoreSubst.hs:605:20 in ghc:CoreSubst in_scope InScope {x_axS m_axT $caction_a1dp $caction_a1dR right right' foo bar $tcFoo $tc'C:Foo $fFooFloat $fFooInt $trModule a_s1fZ a_s1g0 a_s1g1 a_s1g2} tenv [aTT :-> Float, aTV :-> Float, aTW :-> Float] cenv [] tys [] cos [Sub (Sym cobox_aU6)] needInScope [aU6 :-> cobox_aU6] Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} This is a regression from 7.10.3. Setting priority=highest. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11644> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 10