June 2017 - ghc-tickets

[GHC] #13840: Pattern matching duplicated record fields
by GHC 17 Jun '17

17 Jun '17

#13840: Pattern matching duplicated record fields -------------------------------------+------------------------------------- Reporter: lapdot | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.2 Keywords: | Operating System: Linux Architecture: x86_64 | Type of failure: Compile-time (amd64) | crash or panic Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- GHC panic during compiling pattern matching with duplicated record fields. The following is a short example. Main module: {{{#!hs module Lib ( someFunc ) where import F1 import qualified F2 r :: F2.T2 r = F2.T2 { F2.foo = "ok" } someFunc :: IO () someFunc = case r of F2.T2{ foo = a } -> putStrLn a }}} The first submodule: {{{#!hs module F1 ( T1(..) ) where newtype T1 = T1 { foo :: String } }}} The second submodule: {{{#!hs module F2 ( T2(..) ) where newtype T2 = T2 { foo :: String } }}} GHC compiles the modules successfully if uses DuplicateRecordFields. The modules may be not correct without DuplicateRecordFields. But GHC should show an error message other than panic. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13840> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

[GHC] #13836: Some event log events won't be flushed if capability count decreased with setNumCapabitilies
by GHC 17 Jun '17

17 Jun '17

#13836: Some event log events won't be flushed if capability count decreased with setNumCapabitilies -------------------------------------+------------------------------------- Reporter: bgamari | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Runtime | Version: 8.0.1 System | 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: -------------------------------------+------------------------------------- Currently the eventlog code assumes that there have been at most `n_capabilities` capabilities at the time of cleanup. This may not hold in the event that the capability count was decreased using `setNumCapabilities`. This means that some `EventBuf`s won't be flushed when the program terminates. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13836> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #13804: MonoLocalBinds/RankNTypes type inference regression in GHC 8.2
by GHC 17 Jun '17

17 Jun '17

#13804: MonoLocalBinds/RankNTypes type inference regression in GHC 8.2 -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: highest | Milestone: Component: Compiler | Version: 8.2.1-rc2 (Type checker) | 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 encountered this problem when trying to compile the `ivory-bsp-stm32` library (unsuccessfully) with GHC 8.2. Here's a somewhat minimized example: {{{#!hs {-# LANGUAGE DataKinds #-} {-# LANGUAGE ExistentialQuantification #-} {-# LANGUAGE KindSignatures #-} -- Not enabling MonoLocalBinds makes it compile again {-# LANGUAGE MonoLocalBinds #-} {-# LANGUAGE RankNTypes #-} module Bug where import GHC.TypeLits i2cPeripheralDriver :: I2CPeriph -> ChanOutput ('Stored ITime) -> Monitor e () i2cPeripheralDriver periph watchdog_per = do let sendresult' :: Ivory eff () sendresult' = do -- Commenting out the line below makes it compile again clearSR1 periph return undefined sendresult = sendresult' handler watchdog_per undefined $ do callback $ \_ -> do sendresult return undefined clearSR1 :: I2CPeriph -> Ivory eff () clearSR1 = undefined ----- -- Auxiliary definitions ----- data ITime data I2CPeriph = I2CPeriph data Ivory (eff :: Effects) a instance Functor (Ivory eff) instance Applicative (Ivory eff) instance Monad (Ivory eff) data Monitor e a instance Functor (Monitor e) instance Applicative (Monitor e) instance Monad (Monitor e) data ChanOutput (a :: Area *) data RefScope = Global | forall s. Stack s type ConstRef = Pointer 'Valid 'Const -- :: RefScope -> Area * -> * data Nullability = Nullable | Valid data Constancy = Const | Mutable data Pointer (n :: Nullability) (c :: Constancy) (s :: RefScope) (a :: Area *) type AllocEffects s = 'Effects 'NoReturn 'NoBreak ('Scope s) -- :: Effects data Effects = Effects ReturnEff BreakEff AllocEff data ReturnEff = forall t. Returns t | NoReturn data BreakEff = Break | NoBreak data AllocEff = forall s. Scope s | NoAlloc callback :: (IvoryArea a, IvoryZero a) => (forall (s :: RefScope) s'. ConstRef s a -> Ivory (AllocEffects s') ()) -> Handler a e () callback _ = undefined handler :: (IvoryArea a, IvoryZero a) => ChanOutput a -> String -> Handler a e () -> Monitor e () data Handler (area :: Area *) e a handler = undefined data Area k = Struct Symbol | Array Nat (Area k) | CArray (Area k) | Stored k data Time class IvoryArea (a :: Area *) class IvoryZero (area :: Area *) where class IvoryType t instance IvoryType ITime instance IvoryType a => IvoryArea ('Stored a) class IvoryZeroVal a instance IvoryZeroVal ITime instance IvoryZeroVal a => IvoryZero ('Stored a) }}} In GHC 7.10 and 8.0, this compiles. With GHC 8.2, however, it errors: {{{#!hs GHCi, version 8.2.0.20170522: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/rgscott/.ghci [1 of 1] Compiling Bug ( Bug.hs, interpreted ) Bug.hs:25:7: error: • Couldn't match type ‘eff0’ with ‘AllocEffects s'’ because type variable ‘s'’ would escape its scope This (rigid, skolem) type variable is bound by a type expected by the context: forall (s :: RefScope) s'. ConstRef s ('Stored ITime) -> Ivory (AllocEffects s') () at Bug.hs:(24,5)-(25,16) Expected type: Ivory (AllocEffects s') () Actual type: Ivory eff0 () • In a stmt of a 'do' block: sendresult In the expression: do sendresult In the second argument of ‘($)’, namely ‘\ _ -> do sendresult’ • Relevant bindings include sendresult :: Ivory eff0 () (bound at Bug.hs:21:7) | 25 | sendresult | ^^^^^^^^^^ }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13804> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #13691: Bump time submodule
by GHC 17 Jun '17

17 Jun '17

#13691: Bump time submodule -------------------------------------+------------------------------------- Reporter: bgamari | Owner: (none) Type: task | Status: new Priority: highest | Milestone: 8.2.1 Component: Compiler | Version: 8.0.1 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: -------------------------------------+------------------------------------- The `time` version currently on the `ghc-8.2` branch is [[https://github.com/haskell/time/issues/74|broken]] Make sure this gets fixed before the release. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13691> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #13708: Panic! (the "impossible" happened) bug in GHC 8.2.1 rc2
by GHC 17 Jun '17

17 Jun '17

#13708: Panic! (the "impossible" happened) bug in GHC 8.2.1 rc2 -------------------------------------+------------------------------------- Reporter: deech | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | 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: -------------------------------------+------------------------------------- The bug can be reproduced but it does take some setup. Here's the error report in case it's familiar: {{{ ... [1 of 2] Compiling Valuators ( .stack- work/dist/x86_64-linux/Cabal-2.0.0.0/build/fltkhs-fluid-valuators/fltkhs- fluid-valuators-tmp/Valuators.hs, .stack- work/dist/x86_64-linux/Cabal-2.0.0.0/build/fltkhs-fluid-valuators/fltkhs- fluid-valuators-tmp/Valuators.o ) ghc: panic! (the 'impossible' happened) (GHC version 8.2.0.20170507 for x86_64-unknown-linux): idInfo a_ad4D Call stack: CallStack (from HasCallStack): prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1134:58 in ghc:Outputable callStackDoc, called at compiler/utils/Outputable.hs:1138:37 in ghc:Outputable pprPanic, called at compiler/basicTypes/Var.hs:526:34 in ghc:Var Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} The bug is a GUI binding. Reproducing is reliable but unfortunately takes some setup. To reproduce: {{{ # Install the FLTK GUI toolkit from source. > wget http://fltk.org/pub/fltk/1.3.4/fltk-1.3.4-1-source.tar.gz > tar -zxf fltk-1.3.4-1-source.tar.gz > cd fltk-1.3.4-1 > ./configure --enable-gl --enable-shared --enable-localjpeg --enable- localzlib --enable-localpng > make > sudo make install > fltk-config --version 1.3.4-1 # Install the bindings from source > git clone -b 8.2.1-rc2 https://github.com/deech/fltkhs > cd fltkhs > cabal install # Install the package causing the bug from source > git clone -b 8.2.1-rc2 https://github.com/deech/fltkhs-fluid-demos > cd fltkhs-fluid-demos > cabal install }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13708> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 14

[GHC] #13751: Runtime crash with <<loop>> after concurrent stressing of STM computatinos.
by GHC 17 Jun '17

17 Jun '17

#13751: Runtime crash with <<loop>> after concurrent stressing of STM computatinos. -------------------------------------+------------------------------------- Reporter: literon | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.2 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Runtime crash Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: 10414 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- For reproduction, see stack project at https://github.com/robinp/bugreports/tree/master/blackhole. Copying its readme here: ### Compiler & OS GHC 7.10.3, 8.0.1, 8.0.2 (didn't check others). Linux: 4.10.13-1-ARCH #1 SMP PREEMPT x86_64 GNU/Linux 4-core ### What does the code do? TLDR endlessly forks a batch of 8 threads, and waits for them to finish. Each thread calls `observeDuration` on an irrelevant IO action. `observeDuration` measures the time, then updates some data structures in `STM` context in a `TVar`. After a short while (usually within 1 minute), the program aborts with `<<loop>>`. See below for a more detailed investigation. ### To run stack install # Restart if doesn't terminate in a minute. loop-exe +RTS -N4 -Ds > debug 2> debug.out ; beep ### Observe `debug` shows the metering batches run for a while, then get stuck. `debug.out` contains `<<loop>>`. My debug log reading fu is poor, but in less cleaned-up versions of the program it was more trivial to see that two threads get blocked on each others blackhole. In the current log there is mentioning of blackholes, but also MVars, and I don't see what's going on. ### Tracking down When built with profiling (remove `--eventlog --debug` from cabal file, then `stack clean` then `stack build --profile`), and running with `+RTS -N4 -xc` the following stack traces appear: *** Exception (reporting due to +RTS -xc): (THUNK_STATIC), stack trace: Main.meter, called from Main.main *** Exception (reporting due to +RTS -xc): (THUNK_STATIC), stack trace: Prometheus.Metric.Summary.observe, called from Prometheus.Metric.Summary.observeDuration, called from Main.meter.action, called from Main.meter.cfork, called from Main.meter, called from Main.main *** Exception (reporting due to +RTS -xc): (THUNK_STATIC), stack trace: Prometheus.Metric.Summary.observe, called from Prometheus.Metric.Summary.observeDuration, called from Main.meter.action, called from Main.meter.cfork, called from Main.meter, called from Main.main *** Exception (reporting due to +RTS -xc): (THUNK_STATIC), stack trace: *** Exception (reporting due to +RTS -xc): (THUNK_STATIC), stack trace: Main.meter, called from Main.mainPrometheus.Metric.Summary.observe, called from Prometheus.Metric.Summary.observeDuration, called from Main.meter.action, called from Main.meter.cfork, called from Main.meter, called from Main.main *** Exception (reporting due to +RTS -xc): (IND_STATIC), stack trace: Prometheus.Metric.Summary.observe, called from Prometheus.Metric.Summary.observeDuration, called from Main.meter.action, called from Main.meter.cfork, called from Main.meter, called from Main.main loop-exe: loop-exe: <<loop>> <<loop>> loop-exe: thread blocked indefinitely in an MVar operation An other stack trace from an other run as bonus (these two are representative): *** Exception (reporting due to +RTS -xc): (THUNK_STATIC), stack trace: Prometheus.Metric.Summary.invariant, called from Prometheus.Metric.Summary.compress, called from Prometheus.Metric.Summary.observe, called from Prometheus.Metric.Summary.observeDuration, called from Main.meter.action, called from Main.meter.cfork, called from Main.meter, called from Main.main *** Exception (reporting due to +RTS -xc): (THUNK_STATIC), stack trace: Prometheus.Metric.Summary.invariant, called from Prometheus.Metric.Summary.compress, called from Prometheus.Metric.Summary.observe, called from Prometheus.Metric.Summary.observeDuration, called from Main.meter.action, called from Main.meter.cfork, called from Main.meter, called from Main.main loop-exe: <<loop>> *** Exception (reporting due to +RTS -xc): (THUNK_STATIC), stack trace: Main.meter, called from Main.main *** Exception (reporting due to +RTS -xc): (THUNK_STATIC), stack trace: Main.meter, called from Main.main loop-exe: thread blocked indefinitely in an MVar operation I tried to even simplify by removing `prometheus-client` dep, but it seems that the kind of computation done in Prometheus.Metric.Summary.observe, namely putting a lazy computation inside a data type stored in a TVar, are needed to trigger the blackholing. When I tried to simplify those operations, or bang the remaining few fields of the data type, the error didn't reproduce (at least couldn't reproduce as quick as it usually does). The relevant pieces of code from prometheus-client's P.M.Summary module. Note: I manually replaced the `MonadMonitor` constraint with plain IO in that call chain, but it didn't have much effect. {{{#!hs data Estimator = Estimator { estCount :: !Int64 , estSum :: !Double , estQuantiles :: [Quantile] , estItems :: [Item] } deriving (Show) newtype Summary = MkSummary (STM.TVar Estimator) observeDuration :: IO a -> Metric Summary -> IO a observeDuration io metric = do start <- getCurrentTime result <- io end <- getCurrentTime let dt = fromRational $ toRational $ end `diffUTCTime` start withSummary metric dt return result observe :: MonadMonitor m => Double -> Metric Summary -> m () observe v s = withSummary s (insert v) withSummary :: MonadMonitor m => Metric Summary -> (Estimator -> Estimator) -> m () withSummary (Metric {handle = MkSummary valueTVar}) f = doIO $ STM.atomically $ do STM.modifyTVar' valueTVar compress STM.modifyTVar' valueTVar f insert :: Double -> Estimator -> Estimator insert = ... compress :: Estimator -> Estimator compress = ... }}} I checked `insert` and `compress` and they don't seem to be able to loop in any edge case, so this bug is likely an RTS thing. ### Related tickets I found: https://ghc.haskell.org/trac/ghc/ticket/10218 GHC creates incorrect code which throws <<loop>> https://ghc.haskell.org/trac/ghc/ticket/10414 : Buggy behavior with threaded runtime (-N1 working, -N2 getting into <<loop>>) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13751> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #13791: Document allowed syntax in WARNING and DEPRECATED pragmas
by GHC 17 Jun '17

17 Jun '17

#13791: Document allowed syntax in WARNING and DEPRECATED pragmas -------------------------------------+------------------------------------- Reporter: phischu | Owner: (none) Type: bug | Status: new Priority: lowest | Milestone: Component: Compiler | Version: 8.0.2 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Documentation Unknown/Multiple | bug Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- GHC accepts the following deprecation pragma: {{{#!hs {-# DEPRECATED f ["f is deprecated","use g instead"] #-} f :: a -> a f x = x }}} The deprecation message here is a list of strings. I didn't expect this to be allowed because [https://downloads.haskell.org/~ghc/latest/docs/html/users_guide/glasgow_ext… #warning-and-deprecated-pragmas] only shows examples where the deprecation message is a single string. Are there other forms of deprecation message allowed by GHC? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13791> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #13807: GHC 8.2 nondeterministic with foreign imports
by GHC 17 Jun '17

17 Jun '17

#13807: GHC 8.2 nondeterministic with foreign imports -------------------------------------+------------------------------------- Reporter: niteria | Owner: (none) Type: bug | Status: new Priority: high | Milestone: Component: Compiler | 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: -------------------------------------+------------------------------------- See the attached Repro.hs. Reproduction steps: {{{ $ rm Repro.{o,hi}; ghc -O -dunique-increment=-1 Repro.hs; md5sum Repro.hi; rm Repro.{o,hi}; ghc -O -dunique-increment=1 Repro.hs; md5sum Repro.hi; [1 of 1] Compiling Repro ( Repro.hs, Repro.o ) 97f005e3959b657bfac761aa3e8a9447 Repro.hi [1 of 1] Compiling Repro ( Repro.hs, Repro.o ) 691f891fb404eb874e8bedd7bda1c8b7 Repro.hi }}} The crucial difference comes from: {{{ 08fffd1551f79b5aca3cafdceaf865b9 mkStringWriter1 :: Int -> State# RealWorld -> (# State# RealWorld, Ptr Int #) {- Arity: 2, HasNoCafRefs, Strictness: <L,U><S,U>, Unfolding: InlineRule (2, True, False) (\ (ds :: Int) (s :: State# RealWorld) -> case makeStablePtr# @ Int ds s of ds1 { (#,#) ipv ipv1 -> case {__pkg_ccall Int# -> StablePtr# Int -> Addr# -> Addr# -> State# RealWorld -> (# State# RealWorld, Addr# #)} 1# ipv1 __label "Repro_d105" (function) mkStringWriter2 ipv of wild { (#,#) ds2 ds3 -> (# ds2, Ptr @ Int ds3 #) } }) -} }}} vs {{{ a766a748e60cec63a74534e03db3bf30 mkStringWriter1 :: Int -> State# RealWorld -> (# State# RealWorld, Ptr Int #) {- Arity: 2, HasNoCafRefs, Strictness: <L,U><S,U>, Unfolding: InlineRule (2, True, False) (\ (ds :: Int) (s :: State# RealWorld) -> case makeStablePtr# @ Int ds s of ds1 { (#,#) ipv ipv1 -> case {__pkg_ccall Int# -> StablePtr# Int -> Addr# -> Addr# -> State# RealWorld -> (# State# RealWorld, Addr# #)} 1# ipv1 __label "Repro_d5k1wlNFGaX" (function) mkStringWriter2 ipv of wild { (#,#) ds2 ds3 -> (# ds2, Ptr @ Int ds3 #) } }) -} }}} Notice that the labels are different. This doesn't reproduce under GHC 8.0.2, but that may be only because that version doesn't optimize it this way. This is what GHC 8.0.2 produces: {{{ ccf55489bd76322a9404b365b7be7628 mkStringWriter :: Int -> IO (Ptr Int) {- Arity: 2, HasNoCafRefs, Strictness: <L,U><S,U>, Inline: [NEVER] -} }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13807> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 7

[GHC] #8697: Type rationals
by GHC 16 Jun '17

16 Jun '17

#8697: Type rationals ------------------------------------+------------------------------------- Reporter: MikeIzbicki | Owner: Type: feature request | 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've used GHC's Type Nats to implement my own type level rationals (code below). This works fine, but the syntax is a little funky because you have to write out the number as a fraction. For example, you must write 13/10 instead of 1.3; or even worse, 13/1 instead of 13. It would be nice to just write the decimal notation. I only see one potential problem with this feature: the dot in the decimal notation could interfere with the dot in forall statements. I guess this can be parsed unambiguously, but maybe not. {{{ data Frac = Frac Nat Nat data instance Sing (n::Frac) = SFrac Integer Integer instance (SingI a, SingI b) => SingI ('Frac a b) where sing = SFrac (fromSing (sing :: Sing a)) (fromSing (sing :: Sing b)) instance Fractional r => SingE (Kind :: Frac) r where fromSing (SFrac a b) = fromIntegral a/fromIntegral b type (/) a b = 'Frac a b }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8697> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 14

[GHC] #13267: Constraint synonym instances
by GHC 16 Jun '17

16 Jun '17

#13267: Constraint synonym instances -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.1 (Type checker) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: GHC accepts Unknown/Multiple | invalid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Something funny happens when you try to declare an instance of a constraint synonym: {{{ {-# LANGUAGE ConstraintKinds #-} module F where type ShowF a = Show (a -> Bool) instance ShowF Int where show _ = "Fun" }}} I get: {{{ F.hs:8:5: error: ‘show’ is not a (visible) method of class ‘ShowF’ | 8 | show _ = "Fun" | ^^^^ }}} OK, but it gets weirder. Look at: {{{ {-# LANGUAGE ConstraintKinds, FlexibleContexts, FlexibleInstances #-} module F where type ShowF a = (Show (a -> Bool)) instance ShowF Int where }}} This is accepted (with a complaint that `show` is not implemented.) It gets even more awful: {{{ {-# LANGUAGE ConstraintKinds, FlexibleContexts, FlexibleInstances, MultiParamTypeClasses #-} module F where type ShowF a = (Show Bool, Show Int) instance ShowF Int where }}} This is awful: GHC treats `Show Bool` and `Show Int` as if they were constraints, and then emits the following DFun: {{{ df9d1b4635f2a752f29ff327ab66d1cb $f(%,%)ShowShow :: (Show Bool, Show Int) DFunId {- Strictness: m, Inline: CONLIKE, Unfolding: DFun: @ a @ b. @ (Show Bool) @ (Show Int) $fShowBool $fShowInt -} }}} I don't even know what this is supposed to mean. OK, so what should we do? I think there are a few possibilities: 1. Completely outlaw instance declarations on constraint synonyms. 2. Allow instance declarations on constraint synonyms, but only if after desugaring the synonym, you end up with a single class head. I would find this useful in a few cases, for example, if you are writing `instance MonadSample (Impl t) MyMonad`, if you had `type MonadSample2 t a = MonadSample (Impl t) a` you might prefer writing `instance MonadSample t MyMonad` instead 3. Figure out what instance declarations with multiple class heads, and proceed accordingly. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13267> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 13