August 2016 - ghc-tickets

Re: [GHC] #5834: Allow both INLINE and INLINABLE for the same function
by GHC 27 Aug '16

27 Aug '16

#5834: Allow both INLINE and INLINABLE for the same function -------------------------------------+------------------------------------- Reporter: rl | Owner: Type: feature request | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.5 Resolution: | Keywords: inlining Operating System: Unknown/Multiple | Architecture: Type of failure: Runtime | Unknown/Multiple performance bug | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by thomie): * keywords: => inlining -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5834#comment:11> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #5927: A type-level "implies" constraint on Constraints
by GHC 27 Aug '16

27 Aug '16

#5927: A type-level "implies" constraint on Constraints -------------------------------------+------------------------------------- Reporter: illissius | Owner: Type: feature request | Status: new Priority: normal | Milestone: Component: Compiler (Type | Version: 7.4.1 checker) | Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by Iceland_jack): http://yav.github.io/publications/improving-smt-types.pdf > == 3.1 Implication Constraints == > During type inference, GHC uses ''implication constraints'', which do not appear in Haskell source code directly. An implication constraint is, roughly, of the form `G ⇒ W`, where `W` is a collection of constraints that need to be discharged, and `G` are assumptions that may be used while discarding `W`. In the GHC source code, the constraints in `G` are referred to as ''given constraints'', while the ones in `W` are known as wanted constraints. The intuition behind an implication constraint is that `W` contains the constraints that were collected while checking a program fragment, while `G` contains local assumptions that are available only in this particular piece of code. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5927#comment:21> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #12542: Unexpected failure.. (bug?)
by GHC 26 Aug '16

26 Aug '16

#12542: Unexpected failure.. (bug?) -------------------------------------+------------------------------------- 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: GHC rejects Unknown/Multiple | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- From [http://www.cse.chalmers.se/~emax/documents/axelsson2012generic- slides.pdf A Generic Abstract Syntax Model for Embedded Languages]([http://www.cse.chalmers.se/~emax/documents/axelsson2012generic.pdf paper]) {{{#!hs infixr :-> infixl 1 :$ data Full a data a :-> b data AST dom sig where Sym :: dom sig -> AST dom sig (:$) :: AST dom (a :-> sig) -> AST dom (Full a) -> AST dom sig class Binding dom where viewVar :: dom a -> Maybe Integer viewBnd :: dom (a :-> b) -> Maybe Integer freeVars :: Binding dom => AST dom a -> [Integer] freeVars = \case Sym (a -> Just v) -> [v] Sym (b -> Just v) :$ body -> undefined where (a, _) = (viewVar, undefined) (_, b) = (undefined, viewBnd) }}} and so does {{{#!hs where a = viewVar b = viewBnd }}} but {{{#!hs where (a, b) = (viewVar, viewBnd) }}} yields {{{ tSvp.hs:20:5-31: error: … • Could not deduce (Binding dom0) from the context: Binding dom bound by the type signature for: freeVars :: Binding dom => AST dom a -> [Integer] at /tmp/tSvp.hs:14:1-49 or from: Binding dom2 bound by the inferred type for ‘a’: Binding dom2 => dom2 a1 -> Maybe Integer at /tmp/tSvp.hs:20:5-31 The type variable ‘dom0’ is ambiguous • When checking that the inferred type a :: forall k k1 (dom :: * -> *) a (dom1 :: * -> *) (a1 :: k) (b :: k1). (Binding dom1, Binding dom) => dom a -> Maybe Integer is as general as its inferred signature a :: forall (dom :: * -> *) a. Binding dom => dom a -> Maybe Integer In an equation for ‘freeVars’: freeVars = \case { Sym (a -> Just v) -> [v] Sym (b -> Just v) :$ body -> undefined } where (a, b) = (viewVar, viewBnd) tSvp.hs:20:5-31: error: … • Could not deduce (Binding dom1) from the context: Binding dom bound by the type signature for: freeVars :: Binding dom => AST dom a -> [Integer] at /tmp/tSvp.hs:14:1-49 or from: Binding dom2 bound by the inferred type for ‘b’: Binding dom2 => dom2 (a1 :-> b) -> Maybe Integer at /tmp/tSvp.hs:20:5-31 The type variable ‘dom1’ is ambiguous • When checking that the inferred type b :: forall k k1 (dom :: * -> *) a (dom1 :: * -> *) (a1 :: k) (b :: k1). (Binding dom1, Binding dom) => dom1 (a1 :-> b) -> Maybe Integer is as general as its inferred signature b :: forall k k1 (dom :: * -> *) (a :: k) (b :: k1). Binding dom => dom (a :-> b) -> Maybe Integer In an equation for ‘freeVars’: freeVars = \case { Sym (a -> Just v) -> [v] Sym (b -> Just v) :$ body -> undefined } where (a, b) = (viewVar, viewBnd) Compilation failed. }}} Is this intentional? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12542> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #12544: Improved warning for redundant-constraints
by GHC 26 Aug '16

26 Aug '16

#12544: Improved warning for redundant-constraints -------------------------------------+------------------------------------- Reporter: crockeea | Owner: Type: feature | Status: new request | 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: -------------------------------------+------------------------------------- Currently, {{{ foo :: (Num a, Integral a) => a -> a foo = id }}} produces the warning {{{ Main.hs:1:1: warning: [-Wredundant-constraints] • Redundant constraints: (Num a, Integral a) • In the type signature for: foo :: (Num a, Integral a) => a -> a }}} Since GHC can detect there is a redundancy, it would be nice it also told me which constraint was redundant. In #9939, I suggested that the warning include something to the effect of `(Num a) is implied by (Integral a)`. Lets make this as easy for the user as possible. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12544> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #12543: Warning for duplicate language extensions
by GHC 26 Aug '16

26 Aug '16

#12543: Warning for duplicate language extensions -------------------------------------+------------------------------------- Reporter: crockeea | Owner: Type: feature | Status: new request | 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 recently discovered that `-XRebindableSyntax` implies `-XNoImplicitPrelude`. Since I frequently listed both extensions at the top of my file with a `LANGUAGE` pragma, this knowledge allowed me to remove many superfluous `-XNoImplicitPrelude` pragmas. I wouldn't be surprised if there are other unused/implied pragmas floating around in my code. It would be nice to have a warning when two pragmas are explicitly listed, but one implies the other. The warning format could closely follow the warning for #9939. For example: `{-# LANGUAGE NoImplicitPrelude, RebindableSyntax #-}` would produce the warning {{{ Main.hs:1:1: warning: [-Wredundant-pragmas] • Redundant pragmas: RebindableSyntax => NoImplicitPrelude • In the language pragmas: {-# LANGUAGE NoImplicitPrelude, RebindableSyntax #-} }}} Although it might be harder (impossible?) to detect, it would also be nice if there was a similar warning for unused pragmas. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12543> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #10547: feature request: expanding type synonyms in error messages
by GHC 26 Aug '16

26 Aug '16

#10547: feature request: expanding type synonyms in error messages -------------------------------------+------------------------------------- Reporter: osa1 | Owner: osa1 Type: feature | Status: new request | Milestone: 7.12.1 Priority: low | Version: Component: Compiler | Operating System: Unknown/Multiple (Type checker) | Type of failure: None/Unknown Keywords: | Blocked By: Architecture: | Related Tickets: Unknown/Multiple | Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Type errors sometimes becoming confusing because of synonyms, for example, expected type and found type sometimes use different synonyms or sometimes only one of them use a synonym etc. We discussed this in the mailing list: https://mail.haskell.org/pipermail /ghc-devs/2015-June/009247.html And also moved it to the Wiki to collect examples: https://wiki.haskell.org/Expanding_type_synonyms_in_error_messages_proposal I'm currently working on a patch, I made some progress and I'll hopefully submit something for reviews in a couple of days. This ticket is for further discussions and keep track of the progress. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10547> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 23

[GHC] #12407: Template Haskell thinks an unboxed tuple name is illegal to splice in
by GHC 26 Aug '16

26 Aug '16

#12407: Template Haskell thinks an unboxed tuple name is illegal to splice in -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: Type: bug | Status: new Priority: normal | Milestone: 8.0.2 Component: Template | Version: 8.0.1 Haskell | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: GHC rejects Unknown/Multiple | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- {{{#!hs {-# LANGUAGE TemplateHaskell #-} {-# LANGUAGE UnboxedTuples #-} module Bug where import Language.Haskell.TH.Lib import Language.Haskell.TH.Syntax $(do let ubxTup = conT (unboxedTupleTypeName 2) `appT` conT ''Int `appT` conT ''Int x <- newName "x" y <- newName "y" [d| f :: $(ubxTup) -> $(ubxTup) f $(conP (unboxedTupleDataName 2) [varP x, varP y]) = $(conE (unboxedTupleDataName 2) `appE` varE x `appE` varE y) |]) }}} {{{ $ /opt/ghc/8.0.1/bin/ghc Bug.hs [1 of 1] Compiling Bug ( Bug.hs, Bug.o ) Bug.hs:8:3: error: Illegal type constructor or class name: ‘(#,#)’ When splicing a TH declaration: f_0 :: GHC.Tuple.(#,#) GHC.Types.Int GHC.Types.Int -> GHC.Tuple.(#,#) GHC.Types.Int GHC.Types.Int }}} Patch coming soon. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12407> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #12401: GHC panic! Template variable unbound in rewrite rule
by GHC 26 Aug '16

26 Aug '16

#12401: GHC panic! Template variable unbound in rewrite rule -------------------------------------+------------------------------------- Reporter: vagarenko | Owner: Type: bug | Status: new Priority: high | Milestone: Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Windows Architecture: x86_64 | Type of failure: Compile-time (amd64) | crash Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- When compiling following code with `-O`: {{{#!hs {-# LANGUAGE FlexibleContexts, TypeFamilies, TypeSynonymInstances #-} module Bug where import Data.Word import Foreign.Storable import qualified Data.Vector.Storable as VS data Image a = Image { imageWidth :: !Int , imageHeight :: !Int , imageData :: VS.Vector (PixelBaseComponent a) } class (Storable (PixelBaseComponent a)) => Pixel a where type PixelBaseComponent a :: * instance Pixel Pixel8 where type PixelBaseComponent Pixel8 = Word8 instance Pixel Pixel16 where type PixelBaseComponent Pixel16 = Word16 data DynamicImage = ImageY8 (Image Pixel8) | ImageY16 (Image Pixel16) type Pixel8 = Word8 type Pixel16 = Word16 -- | imageMirrorY :: DynamicImage -> DynamicImage imageMirrorY dynImg = case dynImg of ImageY8 img -> ImageY8 $ mirror img 1 ImageY16 img -> ImageY16 $ mirror img 1 where mirror img channels = img { imageData = VS.concat $ reverse $ map (\y -> VS.slice (y * rowLen) rowLen (imageData img)) [0 .. imageHeight img - 1] } where rowLen = channels * imageWidth img }}} I get error: {{{ > ghc Bug -O [1 of 1] Compiling Bug ( Bug.hs, Bug.o ) ghc: panic! (the 'impossible' happened) (GHC version 8.0.1 for x86_64-unknown-mingw32): Template variable unbound in rewrite rule Variable: cobox_s5Rv Rule "SPEC mirror @ Pixel8 @ Pixel8" Rule bndrs: [cobox_s5Rv, $dStorable_s5Rw] LHS args: [TYPE: Pixel8, TYPE: Pixel8, CO: <PixelBaseComponent Pixel8>_N, $dStorable_s5Rw] Actual args: [TYPE: Pixel8, TYPE: Pixel8, CO: <PixelBaseComponent Pixel8>_N, $fStorableWord8 `cast` ((Storable (Sym D:R:PixelBaseComponentWord8[0]))_R :: (Storable Word8 :: Constraint) ~R# (Storable (PixelBaseComponent Pixel8) :: Constraint)), img_a4IS, lvl_s7o4] Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12401> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #12125: Field accessors unnecessarily kept alive
by GHC 26 Aug '16

26 Aug '16

#12125: Field accessors unnecessarily kept alive -------------------------------------+------------------------------------- Reporter: osa1 | Owner: osa1 Type: feature | Status: new request | 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): Phab:D2270 | Wiki Page: -------------------------------------+------------------------------------- Suppose you have a record with some non-exported field accessors, like so: {{{#!haskell module Lib (Rec) where data Rec = Rec { f1 :: Int , f2 :: Int , f3 :: Int , f4 :: Int , f5 :: Int } }}} Code for fields `f1`, `f2` ... are redundantly generated, optimized etc. and they made it to the object file, causing wasted compile times and bloated object files. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12125> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #11442: Segfault when showing (undefined :: Type)
by GHC 26 Aug '16

26 Aug '16

#11442: Segfault when showing (undefined :: Type) -------------------------------------+------------------------------------- Reporter: Iceland_jack | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.1 Keywords: | Operating System: Linux Architecture: x86_64 | Type of failure: Compile-time (amd64) | crash Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- This [https://ghc.haskell.org/trac/ghc/ticket/11311?replyto=3#comment:3 comment] made me wonder about the relationship between `Void` and `Type`, `id :: Void -> Void` and `id :: Type -> Type`: https://ghc.haskell.org/trac/ghc/ticket/11311?replyto=3#comment:3 {{{#!hs {-# LANGUAGE TypeSynonymInstances, FlexibleInstances #- import Data.Kind (Type) instance Show Type where show _ = "..." main = print (undefined :: Type) }}} running gives: {{{ $ runghc --version runghc 8.1.20160113 $ runghc -ignore-dot-ghci Segfault.hs Segmentation fault (core dumped) }}} Verbose log attached. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11442> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3