February 2017 - ghc-tickets

[GHC] #12983: Loading temp shared object failed: TemplateHaskell and recompilation
by GHC 27 Mar '17

27 Mar '17

#12983: Loading temp shared object failed: TemplateHaskell and recompilation -------------------------------------+------------------------------------- Reporter: StefanWehr | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Keywords: template | Operating System: Unknown/Multiple haskell, recompilation | Architecture: | Type of failure: Compile-time Unknown/Multiple | crash or panic Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- In our project, we get from time to time an error like this: {{{ ghc: panic! (the 'impossible' happened) (GHC version 8.0.1 for x86_64-apple-darwin): Loading temp shared object failed: dlopen(/var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40483_0/libghc_1.dylib, 5): Symbol not found: _ShortText_ShortText_con_info Referenced from: /var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40483_0/libghc_1.dylib Expected in: flat namespace in /var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40483_0/libghc_1.dylib }}} I was able to create a somewhat minimal setup for reproducing the problem. I've reproduced the problem on Mac and Linux with GHC 8.0.1, didn't try on Windows. The root of the problem seems to be an interaction with template haskell and the recompilation manager. I'll now explain the setup. I've attached an archive which contains all code and a simple build script. Just unpack the archive and execute ./build.sh to reproduce the problem. The setup consists of four files. (Don't be confused by the module names. They are derived from our project). The problem arises after a change to one of these files and a recompilation. Here are the initial versions of the four files. {{{#!hs -- src/ShortText.hs module ShortText where data ShortText = ShortText String toString :: ShortText -> String toString (ShortText s) = s }}} {{{#!hs -- src/Hospital.hs module Hospital where foo :: Int -> String foo i = show i }}} {{{#!hs -- src/Types.hs {-# LANGUAGE TemplateHaskell #-} module Types where import qualified Hospital import Language.Haskell.TH genCode :: Q [Dec] genCode = let s = Hospital.foo 5 in [d|string = s|] }}} {{{#!hs -- src/MetaHandler.hs {-# LANGUAGE TemplateHaskell #-} module Main where import qualified Types -- splices in "string :: String" $(Types.genCode) main :: IO () main = putStrLn string }}} Compiling the project for the first time works without problems: {{{ $ /Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/bin/ghc -O0 -isrc -ibuild -dynamic -outputdir build -package template-haskell --make src/MetaHandler.hs [1 of 3] Compiling Hospital ( src/Hospital.hs, build/Hospital.o ) [2 of 3] Compiling Types ( src/Types.hs, build/Types.o ) [3 of 3] Compiling Main ( src/MetaHandler.hs, build/Main.o ) Linking src/MetaHandler ... }}} I then changed the content of src/Hospital.hs to use the ShortText module. Here is the new content: {{{#!hs -- src/Hospital.hs module Hospital where import ShortText foo :: Int -> String foo i = toString (ShortText (show i)) }}} Then recompilation fails: {{{ [1 of 4] Compiling ShortText ( src/ShortText.hs, build/ShortText.o ) [2 of 4] Compiling Hospital ( src/Hospital.hs, build/Hospital.o ) [4 of 4] Compiling Main ( src/MetaHandler.hs, build/Main.o ) [TH] ghc: panic! (the 'impossible' happened) (GHC version 8.0.1 for x86_64-apple-darwin): Loading temp shared object failed: dlopen(/var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40827_0/libghc_7.dylib, 5): Symbol not found: _ShortText_ShortText_con_info Referenced from: /var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40827_0/libghc_7.dylib Expected in: flat namespace in /var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40827_0/libghc_7.dylib Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} Looking at the output when compiling with -v3 gives a clue why: {{{ ... *** Linker: /usr/bin/gcc -m64 -fno-stack-protector -DTABLES_NEXT_TO_CODE -m64 -dynamiclib -o /var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40913_0/libghc_7.dylib build/Hospital.o build/Types.o -undefined dynamic_lookup -single_module -install_name '@rpath/libghc_7.dylib' -L/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1 /template-haskell-2.11.0.0 -Wl,-rpath -Wl,/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1 /template-haskell-2.11.0.0 -L/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/pretty-1.1.3.3 -Wl,-rpath -Wl,/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/pretty-1.1.3.3 -L/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/deepseq-1.4.2.0 -Wl,-rpath -Wl,/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/deepseq-1.4.2.0 -L/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/array-0.5.1.1 -Wl,-rpath -Wl,/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/array-0.5.1.1 -L/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/ghc- boot-th-8.0.1 -Wl,-rpath -Wl,/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/ghc- boot-th-8.0.1 -L/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/base-4.9.0.0 -Wl,-rpath -Wl,/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/base-4.9.0.0 -L/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/integer- gmp-1.0.0.1 -Wl,-rpath -Wl,/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1 /integer-gmp-1.0.0.1 -L/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/ghc- prim-0.5.0.0 -Wl,-rpath -Wl,/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/ghc- prim-0.5.0.0 -L/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/rts -Wl,-rpath -Wl,/Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/lib/ghc-8.0.1/rts -lHStemplate-haskell-2.11.0.0-ghc8.0.1 -lHSpretty-1.1.3.3-ghc8.0.1 -lHSdeepseq-1.4.2.0-ghc8.0.1 -lHSarray-0.5.1.1-ghc8.0.1 -lHSghc-boot- th-8.0.1-ghc8.0.1 -lHSbase-4.9.0.0-ghc8.0.1 -lHSinteger- gmp-1.0.0.1-ghc8.0.1 -lHSghc-prim-0.5.0.0-ghc8.0.1 -liconv *** Deleting temp files: Deleting: /var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40913_0/ghc_8.rsp /var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40913_0/libghc_7.dylib *** Deleting temp dirs: Deleting: /var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40913_0 ghc: panic! (the 'impossible' happened) (GHC version 8.0.1 for x86_64-apple-darwin): Loading temp shared object failed: dlopen(/var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40913_0/libghc_7.dylib, 5): Symbol not found: _ShortText_ShortText_con_info Referenced from: /var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40913_0/libghc_7.dylib Expected in: flat namespace in /var/folders/8r/25v12lxd02zdm7lpky3xcnh80000gn/T/ghc40913_0/libghc_7.dylib }}} We can see here that the .o file for ShortText is not passed to gcc when building libghc_7.dylib. I suspect that libghc_7.dylib is used to run template haskell code, but this is only a guess. Looking at the .hi file for MetaHandler, we can see why gcc does not get the .o file for ShortText: ShortText is not included in the module dependencies: {{{ $ /Users/swehr/.stack/programs/x86_64-osx/ghc-8.0.1/bin/ghc --show-iface build/Main.hi Magic: Wanted 33214052, got 33214052 Version: Wanted [8, 0, 0, 1], got [8, 0, 0, 1] Way: Wanted [], got [d, y, n] interface Main 8001 interface hash: def8085a5dcb4980922089d5d21cf0f9 ABI hash: 1a46783c6a6c6b308503cf705db45d4d export-list hash: d2d75d71ebb6963f6e9c9a5b4050ef71 orphan hash: 693e9af84d3dfcc71e640e005bdc5e2e flag hash: 3468616c2ba87fe39e866ff30d651f01 sig of: Nothing used TH splices: True where exports: main string module dependencies: Hospital Types package dependencies: array-0.5.1.1 base-4.9.0.0 deepseq-1.4.2.0 ghc-boot-th-8.0.1 ghc-prim-0.5.0.0 integer- gmp-1.0.0.1 pretty-1.1.3.3 template-haskell-2.11.0.0 orphans: GHC.Base GHC.Float }}} I think the main problem is that src/Types.hs is not recompiled in the second run. I also experimented with compiling each file separately, using the -c flag. When I then compile src/Types.hs, ghc outputs "compilation NOT required" in the second run. If a compile src/Types.hs with -fforce- recomp in the 2nd run, everything works fine. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12983> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #12181: Multi-threaded code on ARM64 GHC runtime doesn't use all available cores
by GHC 27 Mar '17

27 Mar '17

#12181: Multi-threaded code on ARM64 GHC runtime doesn't use all available cores -------------------------------------+------------------------------------- Reporter: varosi | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Runtime | Version: 7.10.3 System | Keywords: | Operating System: Unknown/Multiple Architecture: arm | Type of failure: Runtime | performance bug Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- This is the machine: [http://www.cnx-software.com/2016/04/30/setup-guide-mini-review-of-bq- aquaris-m10-ubuntu-edition-tablet-from-a-developers-perspective/] Haskell ray-tracer that uses Control.Parallel.Strategies and parBuffer that is working well on x64 machine and using all the cores available use only 2 cores from 4 in total on that ARM machine. This machine usually work on two cores only and when it sees that they are used more - it enables two more to get total of four cores. If I give "+RTS -N4" it works just fine. So I think the problem is that the runtime doesn't check for all available cores, but only for enabled. In the first link you could see that "lscpu" returns 4 cores in total. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12181> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #9418: Warnings about "INLINE binder is (non-rule) loop breaker"
by GHC 27 Mar '17

27 Mar '17

#9418: Warnings about "INLINE binder is (non-rule) loop breaker" -------------------------------------+------------------------------------- Reporter: simonpj | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.2 Keywords: | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Unknown | Type of failure: Blocked By: | None/Unknown Related Tickets: | Test Case: | Blocking: | Differential Revisions: -------------------------------------+------------------------------------- Gabor rightly complains: I see literally ''thousands'' of these warnings (in yesterday's and) today's bootstraps: {{{ HC [stage 1] libraries/base/dist-install/build/GHC/Base.o *** Core Lint warnings : in result of Desugar (after optimization) *** {-# LINE 261 "libraries/base/GHC/Base.lhs #-}: Warning: [RHS of $c>>_arr :: forall r_agf a_adQ b_adR. (r_agf -> a_adQ) -> (r_agf -> b_adR) -> r_agf -> b_adR] INLINE binder is (non-rule) loop breaker: $c>>_arr {-# LINE 632 "libraries/base/GHC/Base.lhs #-}: Warning: [RHS of $c>>_apH :: forall a_adQ b_adR. GHC.Types.IO a_adQ -> GHC.Types.IO b_adR -> GHC.Types.IO b_adR] INLINE binder is (non-rule) loop breaker: $c>>_apH }}} }}} This is clearly unsatisfactory, because it is unsettling, and perhaps conceals more important warnings. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9418> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #12137: Warning about “INLINE binder is (non-rule) loop breaker” with `-dcore-lint`
by GHC 27 Mar '17

27 Mar '17

#12137: Warning about “INLINE binder is (non-rule) loop breaker” with `-dcore-lint` -------------------------------------+------------------------------------- Reporter: Iceland_jack | Owner: Type: bug | Status: new Priority: normal | Milestone: 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: -------------------------------------+------------------------------------- With this code using `lens` {{{#!hs {-# Language TemplateHaskell #-} import Control.Lens data Config = Config { _companyName :: String } data AppState = AppState { _asConfig :: Config } makeClassy ''Config instance HasConfig AppState where config = undefined }}} gives {{{ $ ghci -ignore-dot-ghci -dcore-lint /tmp/tvQq.hs GHCi, version 8.0.1: http://www.haskell.org/ghc/ :? for help [1 of 1] Compiling Main ( /tmp/tvQq.hs, interpreted ) *** Core Lint warnings : in result of Simplifier *** /tmp/tvQq.hs:10:10: warning: [RHS of $ccompanyName_a7EP :: Lens' AppState String] INLINE binder is (non-rule) loop breaker: $ccompanyName_a7EP *** Core Lint warnings : in result of Simplifier *** /tmp/tvQq.hs:10:10: warning: [RHS of $ccompanyName_a7EP :: Lens' AppState String] INLINE binder is (non-rule) loop breaker: $ccompanyName_a7EP Ok, modules loaded: Main. *Main> }}} ---- `makeClass` actually dumps {{{ <interactive>:64:54-72: Splicing declarations makeClassy ''Config ======> class HasConfig c_axo7 where config :: Lens' c_axo7 Config companyName :: Lens' c_axo7 String {-# INLINE companyName #-} companyName = (.) config companyName instance HasConfig Config where {-# INLINE companyName #-} config = id companyName = iso (\ (Config x_axo8) -> x_axo8) Config }}} Using that to create a small example ---- {{{#!hs data Config = Config { name :: String } class HasConfig a where config :: a -> Config companyName :: a -> String {-# INLINE companyName #-} companyName a = name (config a) instance HasConfig Config where {-# INLINE companyName #-} config = id companyName = name . config }}} which works perfectly fine without `-dcore-lint` but fails with {{{ $ ghci -ignore-dot-ghci -dcore-lint /tmp/tvQq.hs GHCi, version 8.0.1: http://www.haskell.org/ghc/ :? for help [1 of 1] Compiling Main ( /tmp/tvQq.hs, interpreted ) *** Core Lint warnings : in result of Simplifier *** /tmp/tvQq.hs:11:3: warning: [RHS of $ccompanyName_a18Q :: Config -> String] INLINE binder is (non-rule) loop breaker: $ccompanyName_a18Q *** Core Lint warnings : in result of Simplifier *** /tmp/tvQq.hs:11:3: warning: [RHS of $ccompanyName_a18Q :: Config -> String] INLINE binder is (non-rule) loop breaker: $ccompanyName_a18Q Ok, modules loaded: Main. }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12137> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #13313: cabal08 broken
by GHC 27 Mar '17

27 Mar '17

#13313: cabal08 broken -------------------------------------+------------------------------------- Reporter: bgamari | Owner: (none) Type: bug | Status: new Priority: high | 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: -------------------------------------+------------------------------------- Upon updating the `Cabal` submodule from 19c278feb732f5797256ff28eb671440f5511c6c to e4c36b9dd51820f2380ce7a66f980c4e7b2e96fc the `cabal08` testcase broke. I've marked it as `expect_broken` for now. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13313> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #13219: CSE for join points
by GHC 22 Mar '17

22 Mar '17

#13219: CSE for join points -------------------------------------+------------------------------------- Reporter: lukemaurer | Owner: lukemaurer Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.1 Keywords: JoinPoints | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Join points are hazardous for CSE: {{{ join j = e in ... f e ... }}} The temptation is of course to rewrite to {{{ join j = e in ... f (jump j) ... }}} which doesn't typecheck because you can't have a jump in an argument. Similarly: {{{ join j x = e in let y = (... join j2 x = e in ...) in ... }}} can't be rewritten to {{{ join j x = e in let y = (... join j2 x = jump j x ...) in ... }}} because you can't jump out of a value binding. This doesn't seem to come up very often (took a long time for me to trip over it). The simple solution is simply to ignore CSE for join points. That's what we do at the moment. But this is saddening in some cases: {{{ join j x = e in join j2 x = e in ... }}} can perfectly well become {{{ join j x = e in join j2 x = jump j x in ... }}} We just need to track which join points are still currently valid. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13219> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #13343: Levity polymorphism-related GHC panic: expectJust zonkTcTyVarToVar
by GHC 21 Mar '17

21 Mar '17

#13343: Levity polymorphism-related GHC panic: expectJust zonkTcTyVarToVar -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 (Type checker) | Keywords: | Operating System: Unknown/Multiple LevityPolymorphism | Architecture: | Type of failure: Compile-time Unknown/Multiple | crash or panic Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Reproducible with GHC 8.0.2 and HEAD: {{{#!hs {-# LANGUAGE RankNTypes #-} {-# LANGUAGE TypeInType #-} module Bug where import GHC.Exts type Bad = forall (v1 :: RuntimeRep) (a1 :: TYPE v). a1 }}} {{{ [1 of 1] Compiling Bug ( Bug.hs, Bug.o ) ghc-stage2: panic! (the 'impossible' happened) (GHC version 8.1.20170223 for x86_64-unknown-linux): expectJust zonkTcTyVarToVar CallStack (from HasCallStack): error, called at compiler/utils/Maybes.hs:53:27 in ghc:Maybes expectJust, called at compiler/typecheck/TcType.hs:1576:21 in ghc:TcType }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13343> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #12709: GHC panic
by GHC 21 Mar '17

21 Mar '17

#12709: GHC panic -------------------------------------+------------------------------------- Reporter: Iceland_jack | Owner: Type: bug | Status: new Priority: normal | Milestone: 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: -------------------------------------+------------------------------------- I get {{{ $ ./ghc-stage2 -ignore-dot-ghci --interactive /tmp/tEZm.hs GHCi, version 8.1.20160930: http://www.haskell.org/ghc/ :? for help [1 of 1] Compiling Main ( /tmp/tEZm.hs, interpreted ) ghc-stage2: panic! (the 'impossible' happened) (GHC version 8.1.20160930 for x86_64-unknown-linux): kindPrimRep.go rep_a13E Call stack: CallStack (from HasCallStack): prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1076:58 in ghc:Outputable callStackDoc, called at compiler/utils/Outputable.hs:1080:37 in ghc:Outputable pprPanic, called at compiler/simplStg/RepType.hs:369:9 in ghc:RepType kindPrimRep, called at compiler/simplStg/RepType.hs:337:18 in ghc:RepType typePrimRep, called at compiler/ghci/ByteCodeGen.hs:1541:15 in ghc:ByteCodeGen Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug > }}} running {{{#!hs {-# Language MagicHash, PolyKinds, ViewPatterns, TypeInType, RebindableSyntax, NoImplicitPrelude #-} import GHC.Types import Prelude hiding (Num (..)) import qualified Prelude as P import GHC.Prim data BoxUnbox = BUB Int Int# class Num (a :: TYPE rep) where (+) :: a -> a -> a fromInteger :: Integer -> a instance Num Int where (+) = (P.+) fromInteger = P.fromInteger instance Num Int# where (+) = (+#) fromInteger (fromInteger -> I# n) = n a :: BoxUnbox a = let u :: Num (a :: TYPE rep) => a u = 1 + 2 + 3 + 4 in BUB u u }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12709> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 7

[GHC] #13202: Levity polymorphism panic in GHCi
by GHC 21 Mar '17

21 Mar '17

#13202: Levity polymorphism panic in GHCi -------------------------------------+------------------------------------- Reporter: bgamari | Owner: Type: bug | Status: new Priority: high | 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: -------------------------------------+------------------------------------- With current `master`, {{{ $ ghci GHCi, version 8.1.20170126: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/ben/.ghci λ> import GHC.Exts λ> :set -XTypeApplications -XMagicHash -XTypeInType λ> data TypeRep (a :: k) = TypeRep λ> let typeRepKind = undefined :: TypeRep (a :: k) -> TypeRep k λ> let typeRep = undefined :: TypeRep (a :: k) λ> let x = typeRepKind (typeRep @(Maybe Int#)) ghc: panic! (the 'impossible' happened) (GHC version 8.1.20170126 for x86_64-unknown-linux): isUnliftedType forall (a :: TYPE q). a :: TYPE q Call stack: CallStack (from HasCallStack): prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1166:58 in ghc:Outputable callStackDoc, called at compiler/utils/Outputable.hs:1170:37 in ghc:Outputable pprPanic, called at compiler/types/Type.hs:1869:10 in ghc:Type Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13202> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #9701: GADTs not specialized properly
by GHC 21 Mar '17

21 Mar '17

#9701: GADTs not specialized properly -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.9 Keywords: | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Unknown | Type of failure: Runtime Blocked By: | performance bug Related Tickets: | Test Case: | Blocking: | Differential Revisions: -------------------------------------+------------------------------------- This has probably been raised before, but I can't find a ticket. If I write {{{#!hs data Silly a where Silly :: Ord a => a -> Silly a isItSilly :: a -> Silly a -> Bool isItSilly a (Silly x) = a < x isItSillyInt :: Int -> Silly Int -> Bool isItSillyInt = isItSilly }}} then I get {{{ isItSilly isItSilly = \ @ a_aBq eta_B2 eta1_B1 -> case eta1_B1 of _ { Silly $dOrd_aBs x_aAy -> < $dOrd_aBs eta_B2 x_aAy } isItSillyInt isItSillyInt = isItSilly }}} Although GHC knows that `eta_B2` is an `Int`, and `x_aAy` therefore must be one as well, it looks up the `<` method in the `eta1_B1` dictionary instead of just using `ltInt`. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9701> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 14