May 2015 - ghc-tickets

[GHC] #10407: git head no longer builds with ghc < 7.10
by GHC 12 May '15

12 May '15

#10407: git head no longer builds with ghc < 7.10 -------------------------------------+------------------------------------- Reporter: erikd | Owner: erikd Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.11 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Building GHC Unknown/Multiple | failed Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- Building git head after commit 4fffbc34c0 with ghc-7.8 or ghc-7.6 I get: {{{ compiler/typecheck/TcInteract.hs:2034:13: Not in scope: ‘<$>’ Perhaps you meant one of these: ‘TcM.<$>’ (imported from TcRnMonad), ‘<+>’ (imported from Outputable), ‘TcM.<*>’ (imported from TcRnMonad) compiler/typecheck/TcInteract.hs:2034:29: Not in scope: ‘<$>’ Perhaps you meant one of these: ‘TcM.<$>’ (imported from TcRnMonad), ‘<+>’ (imported from Outputable), ‘TcM.<*>’ (imported from TcRnMonad) }}} The `<$>` operator is only available by default in ghc 7.10 and later. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10407> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

Re: [GHC] #2496: Invalid Eq/Ord instances in Data.Version
by GHC 12 May '15

12 May '15

#2496: Invalid Eq/Ord instances in Data.Version -------------------------------------+------------------------------------- Reporter: guest | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.12.1 Component: Core Libraries | Version: 6.8.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: None/Unknown | Unknown/Multiple Blocked By: | Test Case: Related Tickets: | Blocking: | Differential Revisions: Phab:D395 -------------------------------------+------------------------------------- Comment (by zilinc): At least from my own perspective, I think removing versionTags to retain some properties is rational (well to be honest in this particular case I don't even care). And, exactly as Edward said, I just need a replacement. So I simply did as little change as in https://github.com/haskell/cabal/pull/2584 to accomplish that. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/2496#comment:23> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #10307: Api Annotations: RdrHsSyn.mkAtDefault causes annotations to be disconnected.
by GHC 11 May '15

11 May '15

#10307: Api Annotations: RdrHsSyn.mkAtDefault causes annotations to be disconnected. -------------------------------------+------------------------------------- Reporter: alanz | Owner: alanz Type: bug | Status: new Priority: normal | Milestone: 7.10.2 Component: Compiler | Version: 7.10.1 Keywords: | Operating System: Unknown/Multiple ApiAnnotations | Type of failure: None/Unknown Architecture: | Blocked By: Unknown/Multiple | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- The code for `mkAtDefault` is as follows. {{{#!hs mkATDefault (L loc (TyFamInstDecl { tfid_eqn = L _ e })) | TyFamEqn { tfe_tycon = tc, tfe_pats = pats, tfe_rhs = rhs } <- e = do { tvs <- checkTyVars (ptext (sLit "default")) equalsDots tc (hswb_cts pats) ; return (L loc (TyFamEqn { tfe_tycon = tc , tfe_pats = tvs , tfe_rhs = rhs })) } }}} An associated type in a class of the form {{{#!hs type FoldableConstraint t x = () }}} has an `AnnEqual` attached to the location in `tfid_eqn`. Since the location is discarded, this annotation is then disconnected from the AST. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10307> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #10309: ApiAnnotations : mkGadtDecl discards annotations for HsFunTy
by GHC 11 May '15

11 May '15

#10309: ApiAnnotations : mkGadtDecl discards annotations for HsFunTy -------------------------------------+------------------------------------- Reporter: alanz | Owner: alanz Type: bug | Status: new Priority: normal | Milestone: 7.10.2 Component: Compiler | Version: 7.10.1 (Parser) | Operating System: Unknown/Multiple Keywords: | Type of failure: None/Unknown ApiAnnotations | Blocked By: Architecture: | Related Tickets: Unknown/Multiple | Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- When `mkGadtDecl` is presented wih a `HsFunTy` it discards the SrcSpan, thus disconnecting any annotations on the `HsFunTy`. {{{#!hs mkGadtDecl names (L ls (HsForAllTy imp Nothing qvars cxt tau)) = return $ mk_gadt_con names where (details, res_ty) -- See Note [Sorting out the result type] = case tau of L _ (HsFunTy (L l (HsRecTy flds)) res_ty) -> (RecCon (L l flds), res_ty) _other -> (PrefixCon [], tau) ... }}} This can be triggered by the following {{{#!hs {-# LANGUAGE GADTs #-} module GADTRecords2 (H1(..)) where -- | h1 data H1 a b where C3 :: (Num a) => { field :: a -- ^ hello docs } -> H1 Int Int }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10309> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #10287: ApiAnnotations : BooleanFormula construction discards original
by GHC 11 May '15

11 May '15

#10287: ApiAnnotations : BooleanFormula construction discards original -------------------------------------+------------------------------------- Reporter: alanz | Owner: alanz Type: bug | Status: new Priority: normal | Milestone: 7.10.2 Component: Compiler | Version: 7.10.1 Keywords: | Operating System: Unknown/Multiple ApiAnnotations | Type of failure: None/Unknown Architecture: | Blocked By: Unknown/Multiple | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- The MINIMAL pragma is captured in the parser using a `BooleanFormula`. The constructors (`mkBool`,`mkAnd`,`mkOr`) are smart and try to minimise the boolean formula as it is constructed. This discards the original information, making round tripping impossible. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10287> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #10357: ApiAnnotations : pquals production adds AnnVbar in the wrong place
by GHC 11 May '15

11 May '15

#10357: ApiAnnotations : pquals production adds AnnVbar in the wrong place -------------------------------------+------------------------------------- Reporter: alanz | Owner: alanz Type: bug | Status: new Priority: normal | Milestone: 7.10.2 Component: Compiler | Version: 7.10.1 (Parser) | Operating System: Unknown/Multiple Keywords: | Type of failure: None/Unknown ApiAnnotations | Blocked By: Architecture: | Related Tickets: Unknown/Multiple | Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- The `Parser.y` production for `pquals` is {{{ pquals :: { Located [[LStmt RdrName (LHsExpr RdrName)]] } : squals '|' pquals {% addAnnotation (gl $ last $ unLoc $1) AnnVbar (gl $2) >> return (sLL $1 $> (reverse (unLoc $1) : unLoc $3)) } | squals { L (getLoc $1) [reverse (unLoc $1)] } }}} The `squals` are returned in reverse order, so the `AnnVbar` should be attached to the `head` of the list, not the `last`. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10357> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #10358: ApiAnnotations : PatBind gives wrong SrcSpan for the pattern.
by GHC 11 May '15

11 May '15

#10358: ApiAnnotations : PatBind gives wrong SrcSpan for the pattern. -------------------------------------+------------------------------------- Reporter: alanz | Owner: alanz Type: bug | Status: new Priority: normal | Milestone: 7.10.2 Component: Compiler | Version: 7.10.1 (Parser) | Operating System: Unknown/Multiple Keywords: | Type of failure: None/Unknown ApiAnnotations | Blocked By: Architecture: | Related Tickets: Unknown/Multiple | Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- The production for `decl_no_th` starts {{{ decl_no_th :: { Located (OrdList (LHsDecl RdrName)) } : sigdecl { $1 } | '!' aexp rhs {% do { let { e = sLL $1 $> (SectionR (sL1 $1 (HsVar bang_RDR)) $2) }; pat <- checkPattern empty e; ... }}} The `e` value should be just the pattern, excluding the `rhs`, but the span created includes the `rhs`. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10358> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #10405: Pattern synonym fails with (Exp Bool) but works with (t ~ Bool => Exp Bool)
by GHC 11 May '15

11 May '15

#10405: Pattern synonym fails with (Exp Bool) but works with (t ~ Bool => Exp Bool) -------------------------------------+------------------------------------- Reporter: | Owner: Iceland_jack | Status: new Type: bug | Milestone: Priority: normal | Version: 7.10.1-rc1 Component: Compiler | Operating System: Linux (Type checker) | Type of failure: GHC rejects Keywords: | valid program Architecture: x86 | Blocked By: Test Case: | Related Tickets: Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- The following compiles just fine: {{{#!hs data Exp a where B :: Bool -> Exp Bool pattern Tru = B True pr :: Exp a -> String pr Tru = "true" }}} but add a signature to `Tru`… {{{#!hs pattern Tru :: Exp Bool pattern Tru = B True }}} …and it fails to compile {{{#!hs % ghci -ignore-dot-ghci /tmp/Bug.hs GHCi, version 7.10.0.20150316: http://www.haskell.org/ghc/ :? for help [1 of 1] Compiling Main ( /tmp/Bug.hs, interpreted ) /tmp/Bug.hs:11:4: Couldn't match type ‘a’ with ‘Bool’ ‘a’ is a rigid type variable bound by the type signature for pr :: Exp a -> String at /tmp/Bug.hs:10:7 Expected type: Exp a Actual type: Exp Bool Relevant bindings include pr :: Exp a -> String (bound at /tmp/Bug.hs:11:1) In the pattern: Tru In an equation for ‘pr’: pr Tru = "true" Failed, modules loaded: none. }}} Checking the inferred type of `pattern Tru = B True` I get {{{#!hs ghci> :i Tru pattern Tru :: t ~ Bool => Exp t }}} which works as expected {{{#!hs pattern Tru :: t ~ Bool => Exp t pattern Tru = B True pr :: Exp a -> String pr Tru = "true" }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10405> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

Re: [GHC] #2496: Invalid Eq/Ord instances in Data.Version
by GHC 11 May '15

11 May '15

#2496: Invalid Eq/Ord instances in Data.Version -------------------------------------+------------------------------------- Reporter: guest | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.12.1 Component: Core Libraries | Version: 6.8.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: None/Unknown | Unknown/Multiple Blocked By: | Test Case: Related Tickets: | Blocking: | Differential Revisions: Phab:D395 -------------------------------------+------------------------------------- Comment (by ekmett): One reason for not following semver exactly is that the PVP doesn't line up exactly with semver. Both of our first two digits are "major", and we allow as many subsequent digits as we want. We also don't have an observable prerelease tag for better or worse. As for version tags being ignored by the `Eq` instance. That would have been a viable alternative, but it would have meant that `Version` still didn't line up with what we use it for in ghc, cabal, and elsewhere. IIRC the main use of the tags currently is to talk about a version "*" in some internals in places. A "structural" `Eq` has the benefit that `x == y` implies `f x == f y`, and doesn't require the user to track the exception to the semantics they expect for equality, whereas under your suggestion we'd lose that. It'd just be 'some place to shove extra stuff' bolted onto a data type that could be fundamentally simpler. Then if users want to work with a tagged version they can build it by one of several means, for whatever notions of tags they want to allow, using the existing `Version` type as a primitive component in their own tagged version type with tags being relevant or not for equality as they choose. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/2496#comment:22> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #10384: "Can't splice the polymorphic local variable" check looks dead
by GHC 11 May '15

11 May '15

#10384: "Can't splice the polymorphic local variable" check looks dead -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: task | Status: new Priority: normal | Milestone: Component: Template | Version: 7.11 Haskell | Operating System: Unknown/Multiple Keywords: | Type of failure: None/Unknown Architecture: | Blocked By: Unknown/Multiple | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- While I was looking for TH checks which might want to be checked when `Quotes` are enabled ala #10382, I found the "Can't splice the polymorphic local variable" check in `checkCrossStageLifting` in `TcExpr`. I thought this was a bit odd, because there is another implementation of this very function in `RnSplice`. So I went ahead and removed this check from the compiler, and it validated fine. I then went and tried to see if I could tickle the problem. An old mailing list suggested the following test program to induce the error: {{{ module TH_polymorphic where import Language.Haskell.TH import Language.Haskell.TH.Syntax -- See https://mail.haskell.org/pipermail/template- haskell/2006-April/000552.html test2 () = runQ [| foldr f z xs |] where (f,z,xs) = undefined }}} but I get this error: {{{ TH_polymorphic.hs:8:17: error: Could not deduce (Lift t0) arising from a use of ‘lift’ from the context: Quasi m bound by the inferred type of test2 :: Quasi m => () -> m Exp at TH_polymorphic.hs:(8,1)-(9,30) The type variable ‘t0’ is ambiguous Note: there are several potential instances: instance (Lift a, Lift b) => Lift (Either a b) -- Defined in ‘Language.Haskell.TH.Syntax’ instance Lift a => Lift (Maybe a) -- Defined in ‘Language.Haskell.TH.Syntax’ instance Lift Int16 -- Defined in ‘Language.Haskell.TH.Syntax’ ...plus 24 others In the expression: lift xs In the first argument of ‘runQ’, namely ‘[| foldr f z xs |] pending(rn) [<xs, lift xs>, <z, lift z>, <f, lift f>]’ In the expression: runQ [| foldr f z xs |] pending(rn) [<xs, lift xs>, <z, lift z>, <f, lift f>] }}} which seems to be induced the check in `RnSplice`. So is it dead? If so, let's remove it! -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10384> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11