June 2017 - ghc-tickets

[GHC] #12442: Pure unifier usually doesn't need to unify kinds
by GHC 30 Mar '18

30 Mar '18

#12442: Pure unifier usually doesn't need to unify kinds -------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Keywords: TypeInType | 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 pure unifier (in `types/Unify.hs`) is used to match up instances with actual types. Since GHC 8, it matches up the kinds with the types in a separate pass. But this is often wasteful, and sometimes downright wrong. It's wasteful because most invocations of the unifier on a list of types pass in well-kinded arguments to some type constructor. Because the kinds of type constructors are closed, if we process the list left-to-right, we will always unify the kinds of later arguments before we get to them. So we shouldn't take another pass on the kinds. It's wrong because it's conceivable for the kind to include a type family application, and using a type family application as a template in the pure unifier is very silly, indeed. I cam across this while trying to translate Idris's algebraic effects library to Haskell. My reduced test case is attached. Patch on the way. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12442> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 12

[GHC] #12176: Failure of bidirectional type inference at the kind level
by GHC 30 Mar '18

30 Mar '18

#12176: Failure of bidirectional type inference at the kind level -------------------------------------+------------------------------------- Reporter: goldfire | Owner: goldfire Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Keywords: TypeInType | 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 feeling particularly sadistic towards GHC at the moment, and so I fed it this: {{{ import Data.Kind data Proxy :: forall k. k -> Type where MkProxy :: forall k (a :: k). Proxy a data X where MkX :: forall (k :: Type) (a :: k). Proxy a -> X type Expr = (MkX :: forall (a :: Bool). Proxy a -> X) type family Foo (x :: forall (a :: k). Proxy a -> X) where Foo (MkX :: forall (a :: k). Proxy a -> X) = (MkProxy :: Proxy k) }}} Incredibly, GHC accepts it! And it even means what I wish it to, with GHCi telling me the following (with `-fprint-explicit-kinds`): {{{ λ> :i Foo type family Foo k (x :: forall (a :: k). Proxy k a -> X) :: Proxy * k where [k] Foo k ('MkX k) = 'MkProxy * k }}} That is, I wished to extract the kind parameter to `MkK`, matching against a partially-applied `MkX`, and GHC understood that intention. However, sadly, writing `Foo Expr` produces {{{ • Expected kind ‘forall (a :: k0). Proxy k0 a -> X’, but ‘Expr a0’ has kind ‘Proxy Bool a0 -> X’ • In the first argument of ‘Foo’, namely ‘Expr’ In the type ‘Foo Expr’ In the type family declaration for ‘XXX’ }}} For some reason, `Expr` is getting instantiated when it shouldn't be. Perhaps there's a simpler test case demonstrating the bug, but I feel so gleeful that this chicanery is accepted at all that I wanted to post. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12176> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #9406: unexpected failure for T7837(profasm)
by GHC 30 Mar '18

30 Mar '18

#9406: unexpected failure for T7837(profasm) -------------------------------------+------------------------------------- Reporter: jrp | Owner: Type: bug | Status: new Priority: low | Milestone: Component: Test Suite | Version: 7.8.3 Keywords: T7837 | Operating System: MacOS X Architecture: x86_64 (amd64) | Type of failure: Incorrect Difficulty: Unknown | result at runtime Blocked By: | Test Case: T7837 Related Tickets: | Blocking: | Differential Revisions: -------------------------------------+------------------------------------- This may be a test suite configuration error as the normal case works. When run with an llvm-perf build of the HEAD, I get: {{{ =====> T7837(normal) 1395 of 4068 [0, 0, 0] cd ./indexed-types/should_compile && '/Users/jrp/Projects/haskell/ghc/inplace/bin/ghc-stage2' -fforce-recomp -dcore-lint -dcmm-lint -dno-debug-output -no-user-package-db -rtsopts -fno-ghci-history -c T7837.hs -O -ddump-rule-firings >T7837.comp.stderr 2>&1 =====> T7837(profasm) 1395 of 4068 [0, 0, 0] cd ./indexed-types/should_compile && '/Users/jrp/Projects/haskell/ghc/inplace/bin/ghc-stage2' -fforce-recomp -dcore-lint -dcmm-lint -dno-debug-output -no-user-package-db -rtsopts -fno-ghci-history -c T7837.hs -O -prof -static -auto-all -O -ddump-rule- firings >T7837.comp.stderr 2>&1 Actual stderr output differs from expected: --- ./indexed-types/should_compile/T7837.stderr 2014-08-03 11:36:53.000000000 +0100 +++ ./indexed-types/should_compile/T7837.comp.stderr 2014-08-04 23:16:34.000000000 +0100 @@ -1,3 +1,4 @@ Rule fired: Class op abs Rule fired: Class op signum Rule fired: normalize/Double +Rule fired: Class op / *** unexpected failure for T7837(profasm) Unexpected results from: TEST="T7837" OVERALL SUMMARY for test run started at Mon Aug 4 23:16:32 2014 BST 0:00:02 spent to go through 4068 total tests, which gave rise to 19536 test cases, of which 19534 were skipped 0 had missing libraries 1 expected passes 0 expected failures 0 caused framework failures 0 unexpected passes 1 unexpected failures Unexpected failures: indexed-types/should_compile T7837 [stderr mismatch] (profasm) }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9406> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #13062: `opt' failed in phase `LLVM Optimiser'. (Exit code: -11)
by GHC 29 Mar '18

29 Mar '18

#13062: `opt' failed in phase `LLVM Optimiser'. (Exit code: -11) -------------------------------------+------------------------------------- Reporter: 2bdkid | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 (LLVM) | Keywords: LLVM | Operating System: Linux Optimiser -11 | Architecture: arm | Type of failure: None/Unknown Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- While building [https://hackage.haskell.org/package/unicode-transforms] it fails with: {{{ cabal: Entering directory '/tmp/cabal-tmp-7667/unicode-transforms-0.2.0' Configuring unicode-transforms-0.2.0... Building unicode-transforms-0.2.0... Preprocessing library unicode-transforms-0.2.0... [ 1 of 12] Compiling Data.Unicode.Properties.DecompositionsK ( Data/Unicode/Properties/DecompositionsK.hs, dist/build/Data/Unicode/Properties/DecompositionsK.o ) [ 2 of 12] Compiling Data.Unicode.Properties.Decompositions ( Data/Unicode/Properties/Decompositions.hs, dist/build/Data/Unicode/Properties/Decompositions.o ) [ 3 of 12] Compiling Data.Unicode.Properties.DecomposableK ( Data/Unicode/Properties/DecomposableK.hs, dist/build/Data/Unicode/Properties/DecomposableK.o ) `opt' failed in phase `LLVM Optimiser'. (Exit code: -11) cabal: Leaving directory '/tmp/cabal-tmp-7667/unicode-transforms-0.2.0' }}} Other details: {{{ $ ghc --version The Glorious Glasgow Haskell Compilation System, version 8.0.1 $ opt --version LLVM (http://llvm.org/) LLVM version 3.7.1 Optimized build. Default target: armv7l-unknown-linux-gnueabihf Host CPU: (unknown) }}} I already submitted an issue to the package's bug tracker, but it is believed to be a GHC bug: [https://github.com/harendra-kumar/unicode- transforms/issues/10] -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13062> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #12915: cmmImplementSwitchPlans creates duplicate blocks
by GHC 27 Mar '18

27 Mar '18

#12915: cmmImplementSwitchPlans creates duplicate blocks -------------------------------------+------------------------------------- Reporter: alexbiehl | 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: -------------------------------------+------------------------------------- Given this c-- code *after* common block elimination and *before* cmmImplementSwitchPlans: {{{#!hs ... cfzc: _sfh3::P64 = R1; _cfB7::I64 = %MO_UU_Conv_W32_W64(I32[I64[_sfh3::P64 - 1] - 4]); switch [0 .. 20] _cfB7::I64 { case 2 : goto cfAR; case 4 : goto cfAV; case 6 : goto cfAZ; default: goto cfAC; } cfAZ: _sfh6::I64 = I64[_sfh3::P64 + 7]; _sfgp::I64 = 1; goto sfgo; cfAV: _sfh5::I64 = I64[_sfh3::P64 + 7]; _sfgp::I64 = 1; goto sfgo; cfAR: _sfh4::I64 = I64[_sfh3::P64 + 7]; _sfgp::I64 = 1; goto sfgo; sfgo: ... }}} Everything is fine, the blocks cfAZ, cfAV and cfAR are all a little different. But after cmmImplementSwitchPlans the code becomes: {{{#!hs ... cfzc: _sfgl::I64 = I64[Sp + 24]; _sfgn::P64 = P64[Sp + 16]; _cfB7::I64 = %MO_UU_Conv_W32_W64(I32[I64[R1 - 1] - 4]); if (_cfB7::I64 < 5) goto ufBe; else goto ufBg; ufBe: if (_cfB7::I64 < 4) goto ufBf; else goto cfAV; ufBf: if (_cfB7::I64 != 2) goto cfAC; else goto cfAR; cfAR: _sfgp::I64 = 1; goto sfgo; cfAV: _sfgp::I64 = 1; goto sfgo; ufBg: if (_cfB7::I64 != 6) goto cfAC; else goto cfAZ; cfAZ: _sfgp::I64 = 1; goto sfgo; sfgo: ... }}} We can now see that cfAZ, cfAV and cfAR are code duplicates! Now, https://github.com/ghc/ghc/blob/master/compiler/cmm/CmmPipeline.hs#L72 states {{{ ... ----------- Eliminate common blocks ------------------------------------- g <- {-# SCC "elimCommonBlocks" #-} condPass Opt_CmmElimCommonBlocks elimCommonBlocks g Opt_D_dump_cmm_cbe "Post common block elimination" -- Any work storing block Labels must be performed _after_ -- elimCommonBlocks g <- {-# SCC "createSwitchPlans" #-} runUniqSM $ cmmImplementSwitchPlans dflags g dump Opt_D_dump_cmm_switch "Post switch plan" g ... }}} For some reason I don't understand cmmImplementSwitchPlans is run *after* elimCommonBlocks. Although it seems elimCommonBlocks can deal with the constructs introduced in cmmImplementSwitchPlans. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12915> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #13643: Core lint error with TypeInType and TypeFamilyDependencies
by GHC 27 Mar '18

27 Mar '18

#13643: Core lint error with TypeInType and TypeFamilyDependencies -------------------------------------+------------------------------------- Reporter: Iceland_jack | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Unknown/Multiple InjectiveFamilies, TypeInType | Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: 12102 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- In the code {{{#!hs {-# Language TypeFamilyDependencies #-} {-# Language RankNTypes #-} {-# Language KindSignatures #-} {-# Language DataKinds #-} {-# Language TypeInType #-} {-# Language GADTs #-} import Data.Kind (Type) data Code = I type family Interp (a :: Code) = (res :: Type) | res -> a where Interp I = Bool data T :: forall a. Interp a -> Type where MkNat :: T False instance Show (T a) where show _ = "MkNat" main = do print MkNat }}} but add `{-# Options_GHC -dcore-lint #-}` and we get the attached log from running `runghc /tmp/tPb2.hs > /tmp/tPb2.log`. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13643> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #10536: Clear up how to turn off dynamic linking in build.mk
by GHC 26 Mar '18

26 Mar '18

#10536: Clear up how to turn off dynamic linking in build.mk -------------------------------------+------------------------------------- Reporter: thomie | Owner: Type: task | Status: new Priority: normal | Milestone: 7.12.1 Component: Build | Version: 7.10.1 System | Operating System: Unknown/Multiple Keywords: | Type of failure: None/Unknown Architecture: | Blocked By: Unknown/Multiple | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- See https://mail.haskell.org/pipermail/ghc-devs/2014-December/007725.html -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10536> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #11188: Confusing "parse error in pattern" for missing indentation.
by GHC 26 Mar '18

26 Mar '18

#11188: Confusing "parse error in pattern" for missing indentation. -------------------------------------+------------------------------------- Reporter: andreas.abel | Owner: Type: bug | Status: new Priority: high | Milestone: Component: Compiler | Version: 7.10.1 (Parser) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Other Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- The following problem exists at least in ghc 7.4.2 through 7.10.1: {{{#!hs main = do putStrLn "Hello, lots of crap" $ do a <- return 3 -- The following line is mis-indented, the error is incomprehensible c <- do return 5 }}} Error: {{{ <file>:2:3: Parse error in pattern: putStrLn Possibly caused by a missing 'do'? }}} The hint about "missing do" exists from ghc 7.8, but is wrong in this case. Problematic about this bug is that the wrong indentation can be arbitrary many lines from where ghc reports the (totally incomprehensible) error. My original problem is to find the syntax error here: {{{#!hs -- | Type check a function clause. checkClause :: Type -> A.SpineClause -> TCM Clause checkClause t c(a)(A.Clause (A.SpineLHS i x aps withPats) rhs0 wh) = do unless (null withPats) $ do typeError $ UnexpectedWithPatterns withPats traceCall (CheckClause t c) $ do aps <- expandPatternSynonyms aps checkLeftHandSide (CheckPatternShadowing c) (Just x) aps t $ \ (LHSResult delta ps trhs perm) -> do -- Note that we might now be in irrelevant context, -- in case checkLeftHandSide walked over an irrelevant projection pattern. -- As we will be type-checking the @rhs@ in @delta@, but the final -- body should have bindings in the order of the pattern variables, -- we need to apply the permutation to the checked rhs @v@. let mkBody v = foldr (\ x t -> Bind $ Abs x t) b xs where b = Body $ applySubst (renamingR perm) v xs = [ stringToArgName $ "h" ++ show n | n <- [0..permRange perm - 1] ] -- introduce trailing implicits for checking the where decls TelV htel t0 <- telViewUpTo' (-1) (not . visible) $ unArg trhs (body, with) <- do let n = size htel addCtxTel htel $ checkWhere (size delta + n) wh $ -- for the body, we remove the implicits again escapeContext n $ handleRHS aps (unArgs trhs) rhs0 escapeContext (size delta) $ checkWithFunction with reportSDoc "tc.lhs.top" 10 $ escapeContext (size delta) $ vcat [ text "Clause before translation:" , nest 2 $ vcat [ text "delta =" <+> prettyTCM delta , text "perm =" <+> text (show perm) , text "ps =" <+> text (show ps) , text "body =" <+> text (show body) , text "body =" <+> prettyTCM body ] ] return $ Clause { clauseRange = getRange i , clauseTel = killRange delta , clausePerm = perm , namedClausePats = ps , clauseBody = body , clauseType = Just trhs } }}} {{{ src/full/Agda/TypeChecking/Rules/Def.hs:328:5: Parse error in pattern: checkLeftHandSide Possibly caused by a missing 'do'? }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11188> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #13324: Allow PartialTypeSignatures in the instance context of a standalone deriving declaration
by GHC 24 Mar '18

24 Mar '18

#13324: Allow PartialTypeSignatures in the instance context of a standalone deriving declaration -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: feature | Status: new request | 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: #10607 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Currently, if you try to use a wildcard anywhere in an instance context, it will fail immediately: {{{ $ ghci GHCi, version 8.0.2: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/rgscott/.ghci λ> instance _ => Show (Maybe a) <interactive>:1:10: error: Wildcard ‘_’ not allowed in an instance declaration for ‘Show’ }}} But there's one situation where we could lift this restriction: the context of a standalone, derived instance declaration. That is, something like this: {{{#!hs deriving instance _ => Show (Maybe a) }}} Why? Because GHC already has the machinery needed to infer what the context should be (see [http://git.haskell.org/ghc.git/blob/8a6b8c5fb96472d7c2e2cd1918950dd50f2fef7… this part] of `TcDerivInfer`), so if a user turned on `PartialTypeSignatures`, GHC could just fill in the wildcard with the inferred constraints. The implementation won't be //that// easy, however, since we'd also have to watch out for trickery such as: {{{#!hs deriving instance (C a, _) => C (T a b c) }}} I only mentioned putting wildcards in a derived instance //context//, because I think allowing the use of wildcards elsewhere in the instance head might be too difficult to deal with. I mean, how would you fill in this, for example? {{{#!hs instance (_ a) }}} This mini-feature has a very practical application: it would allow users to wield the flexibility of `StandaloneDeriving` without having to manually type in the instance context every time. That is, users could type in these instances: {{{#!hs deriving instance _ => Data (T a b c) deriving instance _ => Eq (T a b c) }}} Instead of their fully spelled-out, more laborious counterparts. This would be crucial for Template Haskell, as its ability to infer these contexts is quite limited (see #10607 for an example where this cropped up). Idle thought: could this be generalized to work for the instance context of //any// instance declaration (and not just derived ones)? From an outside perspective, it seems like typechecking other instances would require inferring constraints in a fashion quite similar to that of derived instances. But I am not at all familiar with that part of the typechecker, so I might be totally off here. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13324> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 12

[GHC] #13650: Implement KPush in types
by GHC 20 Mar '18

20 Mar '18

#13650: Implement KPush in types -------------------------------------+------------------------------------- Reporter: goldfire | Owner: (none) Type: task | Status: new Priority: high | Milestone: 8.4.1 Component: Compiler | Version: 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: -------------------------------------+------------------------------------- A recent commit contributed a [https://phabricator.haskell.org/diffusion/GHC/browse/master/compiler/types/… Note] that explains why we need the dreaded KPush rule to be implemented in `splitTyConApp`. Without KPush there, it's possible that we can have two types t1 and t2 such that {{{t1 `eqType` t2}}} and yet they respond differently to `splitTyConApp`: t1 = `(T |> co1) (a |> co2)` and t2 = `T a`. Both t1 and t2 are well-kinded and can have the same kind. But one is a `TyConApp` and one is an `AppTy`. (Actually, looking at this, perhaps the magic will be in `mkAppTy`, not `splitTyConApp`.) But I have to look closer. This ticket serves as a reminder to do so. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13650> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5