May 2014 - ghc-tickets

[GHC] #9082: Unexpected behavior involving closed type families and repeat occurrences of variables
by GHC 07 May '14

07 May '14

#9082: Unexpected behavior involving closed type families and repeat occurrences of variables -------------------------------------+------------------------------------- Reporter: haasn | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler (Type | Version: 7.8.2 checker) | Operating System: Unknown/Multiple Keywords: | Type of failure: GHC rejects Architecture: Unknown/Multiple | valid program Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | -------------------------------------+------------------------------------- I'd expect both of these to work fine, but GHC always picks the instance that is mentioned first: {{{ {-# LANGUAGE TypeFamilies, DataKinds #-} type family Bad a b where Bad (f a) a = False Bad a a = True bad :: ( Bad (f a) a ~ False , Bad a a ~ True ) => () bad = () -- Swapping the lines around does not change the behavior much: type family Bad' a b where Bad' a a = True Bad' (f a) a = False bad' :: ( Bad' (f a) a ~ False , Bad' a a ~ True ) => () bad' = () {- ctf.hs:7:8: Could not deduce (Bad a0 a0 ~ 'True) from the context (Bad (f a) a ~ 'False, Bad a a ~ 'True) bound by the type signature for bad :: (Bad (f a) a ~ 'False, Bad a a ~ 'True) => () at ctf.hs:(7,8)-(9,14) The type variable ‘a0’ is ambiguous In the ambiguity check for: forall (f :: * -> *) a. (Bad (f a) a ~ 'False, Bad a a ~ 'True) => () To defer the ambiguity check to use sites, enable AllowAmbiguousTypes In the type signature for ‘bad’: bad :: (Bad (f a) a ~ False, Bad a a ~ True) => () ctf.hs:18:9: Could not deduce (Bad' (f0 a0) a0 ~ 'False) from the context (Bad' (f a) a ~ 'False, Bad' a a ~ 'True) bound by the type signature for bad' :: (Bad' (f a) a ~ 'False, Bad' a a ~ 'True) => () at ctf.hs:(18,9)-(20,14) The type variables ‘f0’, ‘a0’ are ambiguous In the ambiguity check for: forall (f :: * -> *) a. (Bad' (f a) a ~ 'False, Bad' a a ~ 'True) => () To defer the ambiguity check to use sites, enable AllowAmbiguousTypes In the type signature for ‘bad'’: bad' :: (Bad' (f a) a ~ False, Bad' a a ~ True) => () -} }}} Trying to turn this into an open type family instead somewhat hints at the issue: {{{ type family Bad'' a b :: Bool type instance Bad'' a a = True type instance Bad'' (f a) a = False {- ctf.hs:58:15: Conflicting family instance declarations: Bad'' a a -- Defined at ctf.hs:58:15 Bad'' (f a) a -- Defined at ctf.hs:59:15 -} }}} It seems like GHC thinks these two family definitions conflict, yet to me they seem distinct - as a type that would match both would imply f a ~ a and hence be infinite. Can anybody offer insight into this issue? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9082> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #9083: symbol constructors declared prefix missing a leading colon are accepted
by GHC 07 May '14

07 May '14

#9083: symbol constructors declared prefix missing a leading colon are accepted -------------------------------------+------------------------------------- Reporter: aavogt | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler (Parser) | Version: 7.8.2 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | -------------------------------------+------------------------------------- {{{ data M = (++) Int data N = (⌐) Int data Ok = (:++) Int -- correctly accepted }}} I believe ghc should reject M and N. Instead `M` and `N` are accepted, but there seems to be no way to use that constructor (besides say a Data instance) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9083> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

Re: [GHC] #7730: :info and polykinds
by GHC 06 May '14

06 May '14

#7730: :info and polykinds --------------------------------------------+------------------------------ Reporter: monoidal | Owner: archblob Type: bug | Status: new Priority: normal | Milestone: 7.10.1 Component: Compiler (Type checker) | Version: 7.6.2 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: None/Unknown | Unknown/Multiple Test Case: | Difficulty: Unknown Blocking: | Blocked By: | Related Tickets: #8776 --------------------------------------------+------------------------------ Comment (by archblob): I have implemented what we have discussed in comment 28 as {{{ data IfaceType = .... | IfaceTyConApp IfaceTyCon [IfaceTcArg] | ... data IfaceTcArg = ITC_Type IfaceType | ITC_Kind IfaceKind }}} because as I understood it all we want are tags, so I have found this easier to work with than ad hoc lists, please correct me if I am wrong. I also saw that you have changed the way we print foralls, so I will implement that approach in this case too. We will also have to add extra information to one of {{{IfaceClosedSynFamilyTyCon}}} or {{{IfaceSyn}}} constructors in order to print closed type families the way we do now, but this should not be a problem. I have run the tests and the implementation works except when printing qualified names, for example: {{{ data Int = GHC.Types.I# GHC.Prim.Int# }}} prints like {{{data Int = I# GHC.Prim.Int#}}} this is T8469, and we have more of the same for ghci051 with {{{ -data Ghci1.T = A | ... -- Defined at <interactive>:3:10 -data Ghci4.T = B | ... -- Defined at <interactive>:6:12 +data T = A | ... -- Defined at <interactive>:3:10 +data T = B | ... -- Defined at <interactive>:6:12 }}} I will look into this after I sorted out all the other stuff, but in the mean time any suggestions would be nice. I wrote this as a progress update, and if everything goes well i'll have the patch ready in a couple of days. Andrei -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/7730#comment:35> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #6018: Injective type families
by GHC 05 May '14

05 May '14

#6018: Injective type families -------------------------------+------------------------------------------- Reporter: lunaris | Owner: simonpj Type: feature | Status: new request | Milestone: 7.10.1 Priority: normal | Version: 7.4.1 Component: Compiler | Keywords: TypeFamilies, Injective Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #4259 None/Unknown | Test Case: | Blocking: | -------------------------------+------------------------------------------- Comment (by ocharles): It's a little hard to say whether or not injective type families would actually help with what I'm doing - it could well be that I don't just understand this type level magic enough! However, in a few cases where I've got stuck I've moved things to a GADT to witness the relation between types instead, which I believe does get me injectivity and from that points things seemed to "just work" - albeit requiring a bit more effort. I'll keep my eyes out, and try and provide examples here if I find anything concrete. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/6018#comment:24> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #6018: Injective type families
by GHC 05 May '14

05 May '14

#6018: Injective type families -------------------------------+------------------------------------------- Reporter: lunaris | Owner: simonpj Type: feature | Status: new request | Milestone: 7.10.1 Priority: normal | Version: 7.4.1 Component: Compiler | Keywords: TypeFamilies, Injective Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #4259 None/Unknown | Test Case: | Blocking: | -------------------------------+------------------------------------------- Comment (by simonpj): ollie, acowley: if you have a real-life use case for injective type families, do give the details here. Currently we don't have many compelling examples. Thanks Simon -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/6018#comment:23> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #9072: RankNTypes composition now requires ImpredicativeTypes
by GHC 05 May '14

05 May '14

#9072: RankNTypes composition now requires ImpredicativeTypes -------------------------------------+------------------------------------- Reporter: gelisam | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler (Type | Version: 7.8.2 checker) | Operating System: Unknown/Multiple Keywords: | Type of failure: GHC rejects Architecture: Unknown/Multiple | valid program Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | -------------------------------------+------------------------------------- {{{ {-# LANGUAGE RankNTypes #-} instantiateId :: (forall a. a -> a) -> Int -> Int instantiateId f = f passes :: [Int] -> [Int] passes = map (instantiateId id) -- Couldn't match type ‘a0 -> a0’ with ‘forall a. a -> a’ fails :: [Int] -> [Int] fails = (map . instantiateId) id }}} works with ghc 7.6, fails with ghc 7.8, works with ghc 7.8 + `ImpredicativeTypes`. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9072> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

Re: [GHC] #6018: Injective type families
by GHC 04 May '14

04 May '14

#6018: Injective type families -------------------------------+------------------------------------------- Reporter: lunaris | Owner: simonpj Type: feature | Status: new request | Milestone: 7.10.1 Priority: normal | Version: 7.4.1 Component: Compiler | Keywords: TypeFamilies, Injective Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #4259 None/Unknown | Test Case: | Blocking: | -------------------------------+------------------------------------------- Changes (by acowley): * cc: acowley@… (added) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/6018#comment:22> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #6018: Injective type families
by GHC 04 May '14

04 May '14

#6018: Injective type families -------------------------------+------------------------------------------- Reporter: lunaris | Owner: simonpj Type: feature | Status: new request | Milestone: 7.10.1 Priority: normal | Version: 7.4.1 Component: Compiler | Keywords: TypeFamilies, Injective Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #4259 None/Unknown | Test Case: | Blocking: | -------------------------------+------------------------------------------- Changes (by ocharles): * cc: ollie@… (added) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/6018#comment:21> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #9076: GHC.Exts docs don't contain all primops
by GHC 04 May '14

04 May '14

#9076: GHC.Exts docs don't contain all primops ------------------------------------+------------------------------------- Reporter: tibbe | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Documentation | Version: 7.8.2 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- I would expect the docs for `GHC.Exts`: https://www.haskell.org/ghc/docs/7.8.2/html/libraries/base-4.7.0.0/GHC- Exts.html to contain all the primops from `GHC.Prim`: http://hackage.haskell.org/package/ghc-prim-0.3.1.0/docs/GHC-Prim.html as `GHC.Exts` is the recommended module to import the primops from. However, `GHC.Exts` are missing loads of the primops. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9076> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #703: all binaries built by ghc have executable stacks
by GHC 04 May '14

04 May '14

#703: all binaries built by ghc have executable stacks ----------------------------+---------------------------------------------- Reporter: duncan | Owner: ezyang Type: merge | Status: new Priority: normal | Milestone: 6.6.1 Component: | Version: 7.6.3 Compiler (NCG) | Keywords: Resolution: | Architecture: Unknown/Multiple Operating System: Linux | Difficulty: Moderate (less than a day) Type of failure: | Blocked By: None/Unknown | Related Tickets: Test Case: N/A | Blocking: | ----------------------------+---------------------------------------------- Comment (by ezyang): Jens reported to me that 7.8 appears to fix the problem on 64-bit as well. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/703#comment:35> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0