May 2015 - ghc-tickets

[GHC] #9600: Bad error message: Applicative is not a derivable class
by GHC 30 Jul '15

30 Jul '15

#9600: Bad error message: Applicative is not a derivable class -------------------------------------+------------------------------------- Reporter: nh2 | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.10.1 Component: Compiler | Version: 7.8.3 Keywords: | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Easy (less than 1 | Type of failure: hour) | None/Unknown Blocked By: | Test Case: Related Tickets: | Blocking: | Differential Revisions: -------------------------------------+------------------------------------- You have `GeneralizedNewtypeDeriving` on and use `deriving Applicative` with something that you cannot derive Applicative on, GHC correctly says `cannot eta-reduce the representation type enough`. If you don't have `GeneralizedNewtypeDeriving` on yet and deriving Applicative will work, GHC says something like "perhaps you want to enable GeneralizedNewtypeDeriving"? However, if you don't have `GeneralizedNewtypeDeriving` AND if GHC figures out that if you enable it, the next message will be `cannot eta-reduce the representation type enough`, then GHC will state the most unhelpful `Applicative is not a derivable class`, which suggests to me that '''in general''', Applicative is not derivable (which is wrong). rwbarton's suggestion: Suggest enabling `GeneralizedNewtypeDeriving` in any case and then let the user see why it cannot be derived. Some relevant #ghc chat: {{{ rwbarton, that's GND again rwbarton: I a ~ K a where K = ErrorT ImmError (ReaderT CliOptions (ReaderT Config IO)) is an instance of Applicative rwbarton: so it can coerce the instance for K to an instance for I nh2: so is it the error message that is the confusing part here? It suggests that "In general, Applicative is not derivable" nh2: I mean it as: Shouldn't it give an error message like "can't derive an Applicative instance in this case because ..."? rwbarton: well maybe it could add something like "Did you mean to use GeneralizedNewtypeDeriving?" rwbarton: oh I see what you mean, even if you turn that extension on it gives the same error message nh2: yes, because we already have such a message in some cases nh2: no, in this case I had really forgotten the extension. If I add it, I get the "cannot eta-reduce the representation type enough" error, but I haven't figured out why I get it in this case yet rwbarton: I wrote the condition above basically, "I a ~ K a where K = ... is an instance of Applicative" where ~ is the coercion/isomorphism introduced by the newtype rwbarton: if you can't write down K then you can't use GND because where would it get the instance? rwbarton: writing down K is eta reduction nh2: I'm wondering whether it doesn't suggest adding the extension because it already figured out I will not be able to eta-reduce? rwbarton: oh yeah, it seems so rwbarton: that seems a bit silly nh2: in this case, I think we should pimp the error message a bit nh2: (this is the thing I want to derive bwt: https://github.com /ganeti-github-testing/ganeti- test-1/blob/master/src/Ganeti/BasicTypes.hs#L147) nh2: what do you think would be the best error message to show in such a case? nh2: just suggest the GNTD suggestion regardless and let the user see afterwards that it cannot work? rwbarton: I would just have it always suggest GND ... yeah nh2: ok I file a ticket for that rwbarton: seems like more suggestions > fewer suggestions rwbarton: even if this is a weird second-order case where there are two errors }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9600> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #8767: Add rules involving `coerce` to the libraries
by GHC 28 Jul '15

28 Jul '15

#8767: Add rules involving `coerce` to the libraries ------------------------------------+------------------------------------- Reporter: nomeata | Owner: Type: task | Status: new Priority: normal | Milestone: Component: libraries/base | Version: 7.9 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: 8718 | Blocking: Related Tickets: #2110 | ------------------------------------+------------------------------------- With #2110 fixed, we can now add rules like {{{ {-# RULES "map/coerce" [0] map coerce = coerce #-} }}} to the standard libraries. But probably this should happen together or after #8718. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8767> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 27

[GHC] #8538: confusing specialization CORE warning, also can't mark type class instances INLINEABLE unless class defn is marked INLINEABLE
by GHC 27 Jul '15

27 Jul '15

#8538: confusing specialization CORE warning, also can't mark type class instances INLINEABLE unless class defn is marked INLINEABLE ------------------------------------+------------------------------------- Reporter: carter | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- As part of writing an array library for numerical purposes, I have a static sized strict list type for representing the Indexing tuple. One implementation strategy makes it a tad generic, and then I'm writing a bunch of manipulation utilities using Functor, Applicative and Foldable. I wrote it with all the functions marked INLINABLE so i could ensure that I could have the index arith calculations SPECIALIZE in the common cases. Unfortunately, it seems like type class instances can only marked INLINABLE if the type class definition is itself marked INLINE or INLINABLE. Considering i'm writing a numerical array library, I want to make sure that address arithmetic. This could could be argued to be by design (and thus a wont fix). That aside, I also get the following Warning for my "map2" function, which i found surprising. (and it perhaps is worth some investigation / cleanup ) {{{ src/Numerical/Types/Shape.hs:165:1: Warning: RULE left-hand side too complicated to desugar let { $dFunctor :: Functor (Shape ('S N0)) [LclId] $dFunctor = Numerical.Types.Shape.$fFunctorShape @ 'Z $dFunctor } in map2 @ a @ b @ c @ N2 (Numerical.Types.Shape.$fApplicativeShape @ ('S N0) (Numerical.Types.Shape.$fFunctorShape @ ('S N0) $dFunctor) (Numerical.Types.Shape.$fApplicativeShape @ 'Z $dFunctor Numerical.Types.Shape.$fApplicativeShape0)) }}} {{{ {-# LANGUAGE PolyKinds #-} {-# LANGUAGE BangPatterns #-} {-# LANGUAGE DataKinds #-} {-# LANGUAGE TypeFamilies #-} {-# LANGUAGE GADTs #-} {-# LANGUAGE DeriveDataTypeable #-} {-# LANGUAGE TypeOperators #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE FlexibleContexts #-} {-# LANGUAGE FunctionalDependencies #-} {-# LANGUAGE UndecidableInstances #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE StandaloneDeriving #-} {-# LANGUAGE CPP #-} {-# LANGUAGE TemplateHaskell #-} {-# LANGUAGE NoImplicitPrelude #-} module Numerical.Types.Shape where import GHC.Magic import Data.Monoid import Data.Functor import Data.Foldable import Control.Applicative -- import Numerical.Types.Nat import Prelude (seq, ($!),($),Show(..),Eq(),Int) data Nat = S !Nat | Z type N0 = Z type N1= S N0 type N2 = S N1 type N3 = S N2 type N4 = S N3 {- not doing the HLIST style shape because I don't want to have any pattern matchings going on. Also would play hell with locality quality in the address translation hackery, because there'd be an extra load to get those ints! -} infixr 3 :* {- the concern basically boils down to "will it specialize / inline well" -} data Shape (rank :: Nat) a where Nil :: Shape Z a (:*) :: !(a) -> !(Shape r a ) -> Shape (S r) a --deriving instance (Show (Shape rank a)) -- deriving instance Eq (Shape rank a) -- #if defined( __GLASGOW_HASKELL__ ) && ( __GLASGOW_HASKELL__ >= 707) --deriving instance Typeable (Shape rank a) -- #endif instance Functor (Shape Z) where fmap = \ f Nil -> Nil {-# INLINABLE fmap #-} {-# SPECIALIZE fmap :: (a ->b )-> (Shape Z a)-> (Shape Z b) #-} instance (Functor (Shape r)) => Functor (Shape (S r)) where fmap = \ f (a :* rest) -> f a :* fmap f rest {-# INLINABLE fmap #-} {-# SPECIALIZE fmap :: (a ->b )-> (Shape N1 a)-> (Shape N1 b) #-} {-# SPECIALIZE fmap :: (a ->b )-> (Shape N2 a)-> (Shape N2 b) #-} instance Applicative (Shape Z) where pure = \ _ -> Nil {-# INLINE pure #-} (<*>) = \ _ _ -> Nil {-# INLINE (<*>) #-} {-# SPECIALIZE pure :: a -> Shape Z a #-} {-# SPECIALIZE (<*>) :: Shape Z (a -> b) -> Shape Z a -> Shape Z b #-} instance Applicative (Shape r)=> Applicative (Shape (S r)) where pure = \ a -> a :* (pure a) {-# INLINE pure #-} (<*>) = \ (f:* fs) (a :* as) -> f a :* (inline (<*>)) fs as {-# INLINE (<*>) #-} {-# SPECIALIZE pure :: a -> (Shape Z a) #-} {-# SPECIALIZE pure :: a -> (Shape N1 a) #-} {-# SPECIALIZE pure :: a -> (Shape N2 a )#-} instance Foldable (Shape Z) where foldMap = \ _ _ -> mempty {-# INLINE foldMap #-} foldl = \ _ init _ -> init {-# INLINE foldl #-} foldr = \ _ init _ -> init {-# INLINE foldr #-} foldr' = \_ !init _ -> init {-# INLINE foldr' #-} foldl' = \_ !init _ -> init {-# INLINE foldl' #-} -- {-# SPECIALIZE foldMap :: Monoid m => (a -> m)-> Shape Z a -> m #-} -- {-# SPECIALIZE foldl :: (a ->b -> a) -> a -> Shape Z b -> a #-} -- {-# SPECIALIZE foldr :: (b ->a -> a) -> a -> Shape Z b -> a #-} -- {-# SPECIALIZE foldl' :: (a ->b -> a) -> a -> Shape Z b -> a #-} -- {-# SPECIALIZE foldr' :: (b ->a -> a) -> a -> Shape Z b -> a #-} instance (Foldable (Shape r)) => Foldable (Shape (S r)) where foldMap = \f (a:* as) -> f a <> foldMap f as {-# INLINE foldMap #-} foldl' = \f !init (a :* as) -> let next = f init a in next `seq` (inline foldl) f next as {-# INLINE foldl' #-} foldr' = \f !init (a :* as ) -> f a $! (inline foldr) f init as {-# INLINABLE foldr' #-} foldl = \f init (a :* as) -> let next = f init a in (inline foldl) f next as {-# INLINE foldl #-} foldr = \f init (a :* as ) -> f a $ (inline foldr) f init as {-# INLINE foldr #-} {-# SPECIALIZE foldMap::Monoid m => (a -> m)-> Shape N1 a -> m #-} {-# SPECIALIZE foldl :: (a ->b -> a) ->a -> Shape N1 b -> a #-} {-# SPECIALIZE foldr :: (b ->a -> a) -> a -> Shape N1 b -> a #-} {-# SPECIALIZE foldl' :: (a ->b -> a) -> a -> Shape N1 b -> a #-} {-# SPECIALIZE foldr' :: (b ->a -> a) -> a -> Shape N1 b -> a #-} {-# SPECIALIZE foldMap :: Monoid m => (a -> m)-> Shape N2 a -> m #-} {-# SPECIALIZE foldl :: (a ->b-> a) ->a -> Shape N2 b -> a #-} {-# SPECIALIZE foldr :: (b ->a -> a) -> a -> Shape N2 b -> a #-} {-# SPECIALIZE foldl' :: (a ->b -> a) -> a -> Shape N2 b -> a #-} {-# SPECIALIZE foldr' :: (b ->a -> a) -> a -> Shape N2 b -> a #-} -- map2 :: (Applicative (Shape r))=> (a->b ->c) -> (Shape r a) -> (Shape r b) -> (Shape r c ) map2 = \f l r -> pure f <*> l <*> r {-# INLINABLE map2 #-} {-# SPECIALIZE map2 :: (a -> b -> c ) -> (Shape Z a) -> (Shape Z b) -> (Shape Z c )#-} {-# SPECIALIZE map2 :: (a -> b -> c ) -> (Shape N1 a) -> (Shape N1 b) -> (Shape N1 c )#-} {-# SPECIALIZE map2 :: (a -> b -> c ) -> (Shape N2 a) -> (Shape N2 b) -> (Shape N2 c )#-} }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8538> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #9982: cross building integer-gmp is running target program on build host
by GHC 26 Jul '15

26 Jul '15

#9982: cross building integer-gmp is running target program on build host -------------------------------------+------------------------------------- Reporter: newsham | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Build | Version: 7.8.4 System | Operating System: Linux Keywords: | Type of failure: Building GHC Architecture: | failed Unknown/Multiple | Blocked By: Test Case: | Related Tickets: Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- I am trying to build ghc-7.8.4 as a cross compiler. After stage1 finishes and it tries to build libraries it detects that it can link against an existing -gmp and gmp.h (which is what I intended) and starts building the libraries/integer-gmp package against it. During the build it builds the inplace/lib/bin/mkGmpDerivedConstants binary for the target (genode) and then runs it on the build host (ubuntu linux), which leads to failure. The error emitted during the build is: "cp" -p libraries/integer- gmp/mkGmpDerivedConstants/dist/build/tmp/mkGmpDerivedConstants inplace/lib/bin/mkGmpDerivedConstants "rm" -f inplace/lib/mkGmpDerivedConstants echo '#!/bin/bash' >> inplace/lib/mkGmpDerivedConstants echo 'executablename="/u1/home/newsham/src/genode/ghc-7.8.4/inplace/lib/bin/mkGmpDerivedConstants"' >> inplace/lib/mkGmpDerivedConstants echo 'datadir="/u1/home/newsham/src/genode/ghc-7.8.4/inplace/lib"' >> inplace/lib/mkGmpDerivedConstants echo 'bindir="/u1/home/newsham/src/genode/ghc-7.8.4/inplace/bin"' >> inplace/lib/mkGmpDerivedConstants echo 'topdir="/u1/home/newsham/src/genode/ghc-7.8.4/inplace/lib"' >> inplace/lib/mkGmpDerivedConstants echo 'pgmgcc="/u1/home/newsham/src/genode/ghc-7.8.4/wrapper-gcc"' >> inplace/lib/mkGmpDerivedConstants echo 'export LD_LIBRARY_PATH=":$LD_LIBRARY_PATH"' >> inplace/lib/mkGmpDerivedConstants echo 'exec "$executablename" ${1+"$@"}' >> inplace/lib/mkGmpDerivedConstants chmod +x inplace/lib/mkGmpDerivedConstants inplace/lib/mkGmpDerivedConstants > libraries/integer- gmp/mkGmpDerivedConstants/dist/GmpDerivedConstants.h inplace/lib/mkGmpDerivedConstants: line 8: /u1/home/newsham/src/genode/ghc-7.8.4/inplace/lib/bin/mkGmpDerivedConstants: No such file or directory make[1]: *** [libraries/integer- gmp/mkGmpDerivedConstants/dist/GmpDerivedConstants.h] Error 127 make[1]: *** Deleting file `libraries/integer- gmp/mkGmpDerivedConstants/dist/GmpDerivedConstants.h' make: *** [all] Error 2 my build environment is large and complex so I'm not providing all details here, but I will be happy to provide any supporting material needed, on request. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9982> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #9267: Lack of type information in GHC error messages when the liberage coverage condition is unsatisfied
by GHC 25 Jul '15

25 Jul '15

#9267: Lack of type information in GHC error messages when the liberage coverage condition is unsatisfied -------------------------+------------------------------------------------- Reporter: | Owner: danilo2 | Status: new Type: bug | Milestone: Priority: | Version: 7.8.2 normal | Operating System: Unknown/Multiple Component: | Type of failure: Incorrect warning at Compiler | compile-time Keywords: | Test Case: Architecture: | Blocking: Unknown/Multiple | Difficulty: | Unknown | Blocked By: | Related Tickets: | -------------------------+------------------------------------------------- Hello! The problem is not a bug related to proper code execution, but it affects the development in such way, we can treat it as a bug. Lets think about such simple code (it can be further simplified of course): {{{#!haskell {-# LANGUAGE FunctionalDependencies #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE UndecidableInstances #-} data UnsafeBase base err val = Value val | Error err | Other (base val) deriving (Show, Eq) class MagicMerge m1 m2 | m1 -> m2 where magicMerge :: m1 a -> m2 a instance MagicMerge (UnsafeBase (UnsafeBase base err) err) (UnsafeBase base err) where magicMerge ma = case ma of Value a -> Value a Error e -> Error e Other o -> o instance MagicMerge (UnsafeBase (UnsafeBase base err1) err2) (UnsafeBase dstBase err1) where magicMerge (ma :: UnsafeBase (UnsafeBase base err1) err2 a) = case ma of Value a -> Value a Error e -> Other $ magicMerge (Error e :: UnsafeBase base err2 a) Other o -> case o of Value a -> Other $ magicMerge (Value a :: UnsafeBase base err2 a) Error e -> Error e Other o' -> Other $ magicMerge (Other o' :: UnsafeBase base err2 a) main = print "help me world!" }}} When trying to compile it using GHC-7.8, we get following error message: {{{#!haskell Illegal instance declaration for ‘MagicMerge (UnsafeBase (UnsafeBase base err1) err2) (UnsafeBase dstBase err1)’ The liberal coverage condition fails in class ‘MagicMerge’ for functional dependency: ‘m1 -> m2’ Reason: lhs type ‘UnsafeBase (UnsafeBase base err1) err2’ does not determine rhs type ‘UnsafeBase dstBase err1’ In the instance declaration for ‘MagicMerge (UnsafeBase (UnsafeBase base err1) err2) (UnsafeBase dstBase err1)’ }}} Which is of course right, but it is completely unhelpfull! When we run the same program against GHC-7.6, we get another message: {{{#!haskell No instance for (MagicMerge (UnsafeBase base err2) dstBase) arising from a use of `magicMerge' Possible fix: add an instance declaration for (MagicMerge (UnsafeBase base err2) dstBase) In the second argument of `($)', namely `magicMerge (Error e :: UnsafeBase base err2 a)' In the expression: Other $ magicMerge (Error e :: UnsafeBase base err2 a) In a case alternative: Error e -> Other $ magicMerge (Error e :: UnsafeBase base err2 a) }}} And it contains a very helpfull information about inferred types. The information is very helpfull, especialy with more complex instances. What is VERY IMPORTANT, when we use the inferred type, the code starts working in both, GHC-7.8 and GHC-7.6. Right now I have to use both versions of GHC to be able to develop my instances faster. So If we add the inferred by GHC-7.6 premise: {{{#!haskell {-# LANGUAGE FunctionalDependencies #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE UndecidableInstances #-} data UnsafeBase base err val = Value val | Error err | Other (base val) deriving (Show, Eq) class MagicMerge m1 m2 | m1 -> m2 where magicMerge :: m1 a -> m2 a instance MagicMerge (UnsafeBase (UnsafeBase base err) err) (UnsafeBase base err) where magicMerge ma = case ma of Value a -> Value a Error e -> Error e Other o -> o instance (MagicMerge (UnsafeBase base err2) dstBase) => MagicMerge (UnsafeBase (UnsafeBase base err1) err2) (UnsafeBase dstBase err1) where magicMerge (ma :: UnsafeBase (UnsafeBase base err1) err2 a) = case ma of Value a -> Value a Error e -> Other $ magicMerge (Error e :: UnsafeBase base err2 a) Other o -> case o of Value a -> Other $ magicMerge (Value a :: UnsafeBase base err2 a) Error e -> Error e Other o' -> Other $ magicMerge (Other o' :: UnsafeBase base err2 a) main = print "help me world!" }}} The code compiles and works ok in both GHC-7.6 and GHC-7.8. In such cases, GHC-7.8 should inform about lacking premises. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9267> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #10378: min/max for Double/Float instances are incorrect
by GHC 25 Jul '15

25 Jul '15

#10378: min/max for Double/Float instances are incorrect -------------------------------------+------------------------------------- Reporter: lerkok | Owner: Type: bug | Status: new Priority: high | Milestone: Component: Compiler | Version: 7.10.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Blocked By: Test Case: | Related Tickets: Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- This is similar to many other numeric issues around `Double`s and `Float`s. The IEEE754 requires `min` and `max` on floats to return the "other" number, if one of the arguments is `NaN`. The default definitions used in Haskell does not satisfy this property. Furthermore, the current definitions are not commutative, when given `NaNs` and `-0` arguments. The following cases demonstrate the issue with `max`, but `min` has the exact same problems. {{{#!hs Prelude> (0/0) `max` 5 NaN Prelude> 5 `max` (0/0) 5.0 Prelude> isNegativeZero ((-0) `max` 0) False Prelude> isNegativeZero (0 `max` (-0)) True }}} It wouldn't be hard to fix the definitions appropriately; here are reference implementations that are IEEE754 compliant: {{{#!hs maxH x y | isNaN x = y | isNaN y = x | x > y || (x == y && isNegativeZero y) = x | True = y minH x y | isNaN x = y | isNaN y = x | x < y || (x == y && isNegativeZero x) = x | True = y }}} Note that these reference implementations would be quite expensive if GHC were to use them directly. Luckily, IEEE754 compliant min/max operations are supported by all modern CPUs, so doing the "right" thing here is also the cheaper thing as well. (i.e., it'll even be faster than the current incorrect implementation.) On platforms that do not have hardware implementations, the above definitions can serve as the correct implementations, albeit they'll be slower. (Embedded devices, perhaps.) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10378> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #10313: ApiAnnotations : strings in warnings do not return SourceText
by GHC 25 Jul '15

25 Jul '15

#10313: ApiAnnotations : strings in warnings do not return SourceText -------------------------------------+------------------------------------- Reporter: alanz | Owner: alanz Type: bug | Status: new Priority: normal | Milestone: 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 `strings` used in a `WARNING` pragma are captured via {{{ strings :: { Located ([AddAnn],[Located FastString]) } : STRING { sL1 $1 ([],[L (gl $1) (getSTRING $1)]) } .. }}} The `STRING` token has a method `getSTRINGs` that returns the original source text for a string. A warning of the form {{{#!hs {-# WARNING Logic , mkSolver , mkSimpleSolver , mkSolverForLogic , solverSetParams , solverPush , solverPop , solverReset , solverGetNumScopes , solverAssertCnstr , solverAssertAndTrack , solverCheck , solverCheckAndGetModel , solverGetReasonUnknown "New Z3 API support is still incomplete and fragile: \ \you may experience segmentation faults!" #-} }}} returns the concatenated warning string rather than the original source. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10313> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #8327: Cmm sinking does not eliminate dead code in loops
by GHC 25 Jul '15

25 Jul '15

#8327: Cmm sinking does not eliminate dead code in loops ------------------------------------+------------------------------------- Reporter: jstolarek | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- Cmm code shown in #8326 is suboptimal. For this Haskell code: {{{ isTrue# :: Int# -> Bool isTrue# 1# = True isTrue# _ = False f :: Int# -> Int f x | isTrue# (x ># 0#) = I# x | otherwise = -(I# x) }}} we get Cmm that contains dead code in a loop: {{{ cFg: // False branch Hp = Hp + 16; if (Hp > HpLim) goto cFy; else goto cFx; cFy: // not enough heap, call GC HpAlloc = 16; I64[Sp - 16] = cFf; R1 = _sEV::I64; I64[Sp - 8] = _sEU::I64; Sp = Sp - 16; call stg_gc_unbx_r1(R1) returns to cFf, args: 8, res: 8, upd: 8; cFf: // re-do the False branch _sEU::I64 = I64[Sp + 8]; Sp = Sp + 16; _sEV::I64 = R1; goto cFg; cFx: // RHS of False branch I64[Hp - 8] = GHC.Types.I#_con_info; I64[Hp] = -_sEU::I64; R1 = Hp - 7; call (P64[Sp])(R1) args: 8, res: 0, upd: 8; }}} Notice that `_sEV` is dead - we should optimize it away. This is a known deficiency in the Cmm sinking pass (i.e. it does not work for loops). I'm putting this on Trac so we have this documented. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8327> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #8604: Some stack/vmem limits (ulimit) combinations causing GHC to fail
by GHC 24 Jul '15

24 Jul '15

#8604: Some stack/vmem limits (ulimit) combinations causing GHC to fail ----------------------------------+---------------------------------- Reporter: clavin | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Other Architecture: x86_64 (amd64) | Type of failure: Runtime crash Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ----------------------------------+---------------------------------- I have encountered a strange occurrence with GHC that was causing several GHC job failures on an SGE cluster. It turned out that there were other SGE users who needed an absurdly large stack size limit (set via 'ulimit -s') in the several gigabytes range. The default stack size limit had to be raised for the entire cluster. Any job that was run on a machine where the virtual memory limit was less than or equal to 2X the stack size limit, GHC would crash with the following message: ghc: failed to create OS thread: Cannot allocate memory I am running on GHC 7.6.3 with a 64-bit RedHat Enterprise OS, version 5.5. To reproduce the error, I was able to create the following test case: [ ~]$ ulimit -v unlimited [ ~]$ ulimit -s 10240 [ ~]$ ghc --version The Glorious Glasgow Haskell Compilation System, version 7.6.3 [ ~]$ ulimit -v 200000 [ ~]$ ulimit -s 100000 [ ~]$ ghc --version ghc: failed to create OS thread: Cannot allocate memory Several other programs work find using these settings, but GHC consistently has problems. Is this a fundamental issue with how GHC operates or can this be addressed? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8604> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 7

[GHC] #9229: Compiler memory use regression
by GHC 24 Jul '15

24 Jul '15

#9229: Compiler memory use regression -------------------------+------------------------------------------------- Reporter: | Owner: augustss | Status: new Type: bug | Milestone: Priority: | Version: 7.8.2 normal | Operating System: Windows Component: | Type of failure: Compile-time performance bug Compiler | Test Case: Keywords: | Blocking: Architecture: | Unknown/Multiple | Difficulty: | Unknown | Blocked By: | Related Tickets: | -------------------------+------------------------------------------------- I have a module which compiles fine with ghc 7.6, but where ghc 7.8.2 runs out of memory (32-bit Windows). Sadly, I cannot supply the source code since it's proprietary, so I realize that this bug will be almost impossible to fix. It is possible to compile the module with -O0 and -O1, but not -O2. (This is a very performance critical module so I'd like -O2.) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9229> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5