April 2017 - ghc-tickets

[GHC] #13479: Core Lint issues during slowtest
by GHC 09 Apr '17

09 Apr '17

#13479: Core Lint issues during slowtest -------------------------------------+------------------------------------- Reporter: bgamari | Owner: (none) Type: bug | Status: new Priority: high | Milestone: 8.2.1 Component: Compiler | Version: 8.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Compile-time Unknown/Multiple | crash or panic Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- I'm am seeing Core Lint failurse from the `T5626` test, {{{ *** Core Lint errors : in result of Simplifier *** <no location info>: warning: [RHS of go_s54c :: [Int] -> (Int, Int, Int, Int)] idArity 2 exceeds typeArity 1: go_s54c }}} The relevant bits of the program are, {{{#!hs Rec { go_s54c [Occ=LoopBreaker] :: [Int] -> (Int, Int, Int, Int) [LclId, Arity=2, CallArity=2, Str=<B,U>b, Unf=Unf{Src=<vanilla>, TopLvl=True, Value=True, ConLike=True, WorkFree=True, Expandable=True, Guidance=NEVER}] go_s54c = \ (ds_X35o :: [Int]) -> case ds_X35o of { [] -> case lvl_s4Vm of wild_00 { }; : y_a34y [Dmd=<B,A>] ys_a34z [Dmd=<B,U>] -> go_s54c ys_a34z } end Rec } lvl_s54d :: (Int, Int, Int, Int) [LclId, Arity=1, Str=x, Unf=Unf{Src=<vanilla>, TopLvl=True, Value=True, ConLike=True, WorkFree=True, Expandable=True, Guidance=NEVER}] lvl_s54d = go_s54c ([] @ Int) }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13479> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 18

[GHC] #12793: Performs unaligned access on SPARC64
by GHC 09 Apr '17

09 Apr '17

#12793: Performs unaligned access on SPARC64 --------------------------------+---------------------------------------- Reporter: jrtc27 | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Linux Architecture: sparc | Type of failure: Building GHC failed Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: --------------------------------+---------------------------------------- (Filed under Architecture: sparc, since there's no sparc64 in the dropdown...) Trying to build GHC on sparc64 currently crashes with a SIGBUS (unaligned access). This is because sparc64 uses the C backend (only 32-bit sparc has native code gen), but is unaware that it can't perform unaligned accesses. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12793> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

Re: [GHC] #5553: sendWakeup error in simple test program with MVars and killThread
by GHC 09 Apr '17

09 Apr '17

#5553: sendWakeup error in simple test program with MVars and killThread -------------------------------------+------------------------------------- Reporter: bit | Owner: (none) Type: bug | Status: new Priority: high | Milestone: 7.4.2 Component: Runtime System | Version: 8.0.1 Resolution: | Keywords: Operating System: Linux | Architecture: x86 Type of failure: Incorrect result | Test Case: at runtime | Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by rrnewton): I'm getting this error too, in a different program (on some runs only), with GHC 8.0.2. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5553#comment:11> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #12396: Panic when specializing in another module
by GHC 08 Apr '17

08 Apr '17

#12396: Panic when specializing in another module -------------------------------------+------------------------------------- Reporter: crockeea | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Compile-time Unknown/Multiple | crash Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- In the attached example (which requires `criterion` and `deepseq`), when compiled with `ghc -O2 -funfolding-use-threshold1000 Main.hs`, I get the following error on `Main.hs`: {{{ ghc: panic! (the 'impossible' happened) (GHC version 8.0.1 for x86_64-unknown-linux): StgCmmEnv: variable not found $d(%,%)_aaTF local binds for: $trModule $sfool $trModule1 $trModule2 $wlvl_rgY8 lvl_rgY9 lvl1_rgYa lvl3_rgYc lvl4_rgYd lvl5_rgYe lvl6_rgYf lvl7_rgYg lvl8_rgYh lvl9_rgYi lvl10_rgYj lvl11_rgYk lvl12_rgYl lvl13_rgYm lvl14_rgYn lvl15_rgYo lvl16_rgYp lvl17_rgYq lvl18_rgYr lvl19_rgYs lvl20_rgYt lvl21_rgYu lvl22_rgYv lvl23_rgYw lvl24_rgYx lvl25_rgYy lvl26_rgYz lvl27_rgYA lvl28_rgYB lvl29_rgYC lvl30_rgYD lvl31_rgYE lvl32_rgYF lvl33_rgYG lvl34_rgYH lvl35_rgYI lvl36_rgYJ lvl37_rgYK lvl38_rgYL lvl39_rgYM lvl40_rgYN lvl41_rgYO lvl42_rgYP lvl43_rgYQ lvl44_rgYR lvl45_rgYS lvl46_rgYT lvl47_rgYU lvl48_rgYV lvl49_rgYW lvl50_rgYX lvl51_rgYY lvl52_rgYZ lvl53_rgZ0 lvl54_rgZ1 lvl55_rgZ2 lvl56_rgZ3 lvl57_rgZ4 lvl58_rgZ5 lvl59_rgZ6 lvl60_rgZ7 lvl61_rgZ8 lvl62_rgZ9 lvl63_rgZa lvl64_rgZb lvl65_rgZc lvl66_rgZd ww1_rgZe ww2_rgZf ww3_rgZg ww4_rgZh pinfo_rgZi pinfo1_rgZj pinfo2_rgZk pinfo3_rgZl lvl67_rgZm lvl68_rgZn lvl69_rgZo $wgo12_rgZp eta_sh1k wild_sh1l v_sh1m sat_sh1n }}} I initially found this bug when I added a `SPECIALIZE` pragma for `fool` in `Main.hs`. However, I need `funfolding-use-threshold1000` and the constraint synonym `UCRTElt` in `Foo.hs` to trigger the bug. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12396> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #13548: the 'impossible' happened during building PureScript for profiling
by GHC 08 Apr '17

08 Apr '17

#13548: the 'impossible' happened during building PureScript for profiling -------------------------------------+------------------------------------- Reporter: hdgarrood | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.2 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Compile-time Unknown/Multiple | crash or panic Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Repro steps: * `git clone https://github.com/purescript/purescript` * `cd purescript` * `git checkout v0.11.2` * `stack build --profile` The error was: {{{ [ 23 of 145] Compiling Language.PureScript.Constants ( src/Language/PureScript/Constants.hs, .stack- work/dist/x86_64-linux/Cabal-1.24.2.0/build/Language/PureScript/Constants.p_o ) ghc: panic! (the 'impossible' happened) (GHC version 8.0.2 for x86_64-unknown-linux): kindPrimRep.go rep_sfuE Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug -- While building package purescript-0.11.2 using: /home/harry/.stack/setup-exe-cache/x86_64-linux/Cabal- simple_mPHDZzAJ_1.24.2.0_ghc-8.0.2 --builddir=.stack- work/dist/x86_64-linux/Cabal-1.24.2.0 build lib:purescript exe:purs --ghc- options " -ddump-hi -ddump-to-file" Process exited with code: ExitFailure 1 }}} Sorry that this is pretty much the complete opposite of a minimal reproducing example, but I have no clue how to go about isolating the part that's actually causing the problem. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13548> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #13544: ghc panic
by GHC 07 Apr '17

07 Apr '17

#13544: ghc panic -------------------------------------+------------------------------------- Reporter: shauryab98 | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.2 Keywords: | Operating System: Windows Architecture: x86_64 | Type of failure: Compile-time (amd64) | crash or panic Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- I got this error: {{{ ghc: panic! (the 'impossible' happened) (GHC version 8.0.2 for x86_64-unknown-mingw32): initTc: unsolved constraints WC {wc_insol = [W] %_a1El :: t_a1Ek[tau:1] (CHoleCan: %)} }}} while compiling the following code {{{#!hs halveEvens :: [Integer] -> [Integer] halveEvens = map . filter (\c -> c%2 == 0) }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13544> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

[GHC] #13537: Adding underscore (m :: _) to pattern variable makes it fail
by GHC 07 Apr '17

07 Apr '17

#13537: Adding underscore (m :: _) to pattern variable makes it fail -------------------------------------+------------------------------------- Reporter: Iceland_jack | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 (Type checker) | 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: -------------------------------------+------------------------------------- This works fine {{{#!hs {-# Language RankNTypes, TypeOperators, KindSignatures, GADTs, ScopedTypeVariables, PolyKinds #-} import Data.Kind data CodTy :: (k -> Type) -> Type where CodTy :: { runCodTy :: (forall r. (f ty -> r) -> r) } -> CodTy f type f ~> g = forall xx. f xx -> g xx fmap' :: (f ~> f') -> (CodTy f -> CodTy f') fmap' f (CodTy m) = CodTy $ \ c -> m (c . f) }}} but adding {{{#!hs fmap' f (CodTy (m :: _)) = CodTy $ \ c -> m (c . f) }}} makes it fail (also fails without `PolyKinds`) {{{ $ ghci -ignore-dot-ghci tNT2.hs GHCi, version 8.0.1: http://www.haskell.org/ghc/ :? for help [1 of 1] Compiling Main ( tNT2.hs, interpreted ) tNT2.hs:11:22: error: • Found type wildcard ‘_’ standing for ‘(f ty -> r0) -> r0’ Where: ‘r0’ is an ambiguous type variable ‘f’ is a rigid type variable bound by the type signature for: fmap' :: forall k (f :: k -> *) (f' :: k -> *). f ~> f' -> CodTy f -> CodTy f' at tNT2.hs:10:10 ‘ty’ is a rigid type variable bound by a pattern with constructor: CodTy :: forall k (f :: k -> *) (ty :: k). (forall r. (f ty -> r) -> r) -> CodTy f, in an equation for ‘fmap'’ at tNT2.hs:11:10 ‘k’ is a rigid type variable bound by the type signature for: fmap' :: forall k (f :: k -> *) (f' :: k -> *). f ~> f' -> CodTy f -> CodTy f' at tNT2.hs:10:10 To use the inferred type, enable PartialTypeSignatures • In a pattern type signature: _ In the pattern: m :: _ In the pattern: CodTy (m :: _) • Relevant bindings include f :: f ~> f' (bound at tNT2.hs:11:7) fmap' :: f ~> f' -> CodTy f -> CodTy f' (bound at tNT2.hs:11:1) tNT2.hs:11:46: error: • Couldn't match type ‘r0’ with ‘r’ because type variable ‘r’ would escape its scope This (rigid, skolem) type variable is bound by a type expected by the context: (f' ty -> r) -> r at tNT2.hs:11:28-51 Expected type: f ty -> r0 Actual type: f ty -> r • In the first argument of ‘m’, namely ‘(c . f)’ In the expression: m (c . f) In the second argument of ‘($)’, namely ‘\ c -> m (c . f)’ • Relevant bindings include c :: f' ty -> r (bound at tNT2.hs:11:38) m :: (f ty -> r0) -> r0 (bound at tNT2.hs:11:17) Failed, modules loaded: none. Prelude> }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13537> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

Re: [GHC] #5218: Add unpackCStringLen# to create Strings from string literals
by GHC 07 Apr '17

07 Apr '17

#5218: Add unpackCStringLen# to create Strings from string literals -------------------------------------+------------------------------------- Reporter: tibbe | Owner: thoughtpolice Type: feature request | Status: patch Priority: normal | Milestone: Component: Compiler | Version: 7.0.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Runtime | Unknown/Multiple performance bug | Test Case: Blocked By: | Blocking: Related Tickets: #5877 #10064 | Differential Rev(s): Phab:D2443 Wiki Page: | -------------------------------------+------------------------------------- Comment (by winter): > This would essentially amount to making `ByteArray#` a sum internally. You would simply add a closure type which would carry a pointer to the payload instead of the VLA carried by `StgArrBytes` currently. I'm not an expert on this, so maybe this sounds ridiculous: If we add a new closure type for `ByteArray#`, could it be used in runtime also? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5218#comment:65> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #13538: Weird kind inference problems with closed type families
by GHC 07 Apr '17

07 Apr '17

#13538: Weird kind inference problems with closed type families -------------------------------------+------------------------------------- Reporter: achirkin | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.2 (Type checker) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Incorrect Unknown/Multiple | error/warning at compile-time Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Consider a problem of injective cons on a type-level list [https://ghc.haskell.org/trac/ghc/ticket/12114] I made a small workaround by adding an intermediate data type `List1 k`. Here is my code: {{{#!hs {-# LANGUAGE UndecidableInstances #-} {-# LANGUAGE TypeFamilies, TypeFamilyDependencies #-} {-# LANGUAGE KindSignatures, DataKinds, PolyKinds #-} {-# LANGUAGE TypeOperators #-} {-# LANGUAGE TypeApplications #-} module TFTest where import GHC.TypeLits import Data.Proxy -- | Synonym for a type-level snoc (injective!) type (ns :: [k]) +: (n :: k) = GetList1 (SinkFirst (n ': ns)) infixl 5 +: -- | A weird data type used to make `(+:)` operation injective. -- `List k [k]` must have at least two elements. data List1 k = L1Single k | L1Head k [k] -- | Sink first element of a list to the end of the list type family SinkFirst (xs :: [k]) = (ys :: List1 k) | ys -> xs where SinkFirst '[y] = 'L1Single y -- SinkFirst (y ': x ': xs :: [Nat]) -- = ('L1Head x (GetList1Nat (SinkFirst (y ': xs))) :: List1 Nat) SinkFirst (y ': x ': xs :: [k]) = ('L1Head x (GetList1 (SinkFirst (y ': xs))) :: List1 k) type family GetList1 (ts :: List1 k) = (rs :: [k]) | rs -> ts where GetList1 ('L1Single x) = '[x] GetList1 ('L1Head y (x ':xs)) = y ': x ': xs type family GetList1Nat (ts :: List1 Nat) = (rs :: [Nat]) | rs -> ts where GetList1Nat ('L1Single x) = '[x] GetList1Nat ('L1Head y (x ': xs)) = y ': x ': xs type family (++) (as :: [k]) (bs :: [k]) :: [k] where '[] ++ bs = bs (a ': as) ++ bs = a ': (as ++ bs) ff :: Proxy k -> Proxy (as +: k) -> Proxy (k ': bs) -> Proxy (as ++ bs) ff _ _ _ = Proxy yy :: Proxy '[3,7,2] yy = ff (Proxy @5) (Proxy @'[3,7,5]) (Proxy @'[5,2]) }}} This gives me the following error: {{{ ~/TFTest.hs: 47, 21 • Couldn't match kind ‘k’ with ‘Nat’ When matching the kind of ‘7 : xs’ Expected type: Proxy ((3 : (7 : xs)) +: 5) Actual type: Proxy '[3, 7, 5] • In the second argument of ‘ff’, namely ‘(Proxy @'[3, 7, 5])’ In the expression: ff (Proxy @5) (Proxy @'[3, 7, 5]) (Proxy @'[5, 2]) In an equation for ‘yy’: yy = ff (Proxy @5) (Proxy @'[3, 7, 5]) (Proxy @'[5, 2]) }}} However, if I uncomment lines `26-27`, the code works perfectly fine! This behaviour feels like a bug in type checker, though I am not sure. If it is not, please, explain me what happens here? :) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13538> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

Re: [GHC] #5218: Add unpackCStringLen# to create Strings from string literals
by GHC 07 Apr '17

07 Apr '17

#5218: Add unpackCStringLen# to create Strings from string literals -------------------------------------+------------------------------------- Reporter: tibbe | Owner: thoughtpolice Type: feature request | Status: patch Priority: normal | Milestone: Component: Compiler | Version: 7.0.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Runtime | Unknown/Multiple performance bug | Test Case: Blocked By: | Blocking: Related Tickets: #5877 #10064 | Differential Rev(s): Phab:D2443 Wiki Page: | -------------------------------------+------------------------------------- Comment (by bgamari): > But i fail to understand how can we provide `unsafeMkByteArray#` in an obvious way. This would essentially amount to making `ByteArray#` a sum internally. You would simply add a closure type which would carry a pointer to the payload instead of the VLA carried by `StgArrBytes` currently. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5218#comment:64> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0