March 2016 - ghc-tickets

[GHC] #8848: Warning: Rule too complicated to desugar
by GHC 01 Dec '16

01 Dec '16

#8848: Warning: Rule too complicated to desugar ------------------------------------+------------------------------------- Reporter: carter | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.1-rc2 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- I've a very very modest application of Specialize to fixed sized lists in some of my code which seems to trip up the specialization machinery. Is there any flags I can pass GHC to make sure it doesn't give up on these specialize calls? is the only work around to write my own monomorphic versions and add some hand written rewrite rules?! {{{ rc/Numerical/Types/Shape.hs:225:1: Warning: RULE left-hand side too complicated to desugar let { $dFunctor_a3XB :: Functor (Shape ('S 'Z)) [LclId, Str=DmdType] $dFunctor_a3XB = Numerical.Types.Shape.$fFunctorShape @ 'Z $dFunctor_a3Rn } in map2 @ a @ b @ c @ ('S ('S 'Z)) (Numerical.Types.Shape.$fApplicativeShape @ ('S 'Z) (Numerical.Types.Shape.$fFunctorShape @ ('S 'Z) $dFunctor_a3XB) (Numerical.Types.Shape.$fApplicativeShape @ 'Z $dFunctor_a3XB Numerical.Types.Shape.$fApplicativeShape0)) src/Numerical/Types/Shape.hs:226:1: Warning: RULE left-hand side too complicated to desugar let { $dFunctor_a3XG :: Functor (Shape ('S 'Z)) [LclId, Str=DmdType] $dFunctor_a3XG = Numerical.Types.Shape.$fFunctorShape @ 'Z $dFunctor_a3Rn } in let { $dFunctor_a3XF :: Functor (Shape ('S ('S 'Z))) [LclId, Str=DmdType] $dFunctor_a3XF = Numerical.Types.Shape.$fFunctorShape @ ('S 'Z) $dFunctor_a3XG } in map2 @ a @ b @ c @ ('S ('S ('S 'Z))) (Numerical.Types.Shape.$fApplicativeShape @ ('S ('S 'Z)) (Numerical.Types.Shape.$fFunctorShape @ ('S ('S 'Z)) $dFunctor_a3XF) (Numerical.Types.Shape.$fApplicativeShape @ ('S 'Z) $dFunctor_a3XF (Numerical.Types.Shape.$fApplicativeShape @ 'Z $dFunctor_a3XG Numerical.Types.Shape.$fApplicativeShape0))) }}} the associated code (smashed into a single module ) is {{{ {-# LANGUAGE DataKinds, GADTs, TypeFamilies, ScopedTypeVariables #-} {-# LANGUAGE DeriveDataTypeable #-} {-# LANGUAGE TypeOperators #-} {-# LANGUAGE BangPatterns #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE FlexibleContexts #-} {-# LANGUAGE FunctionalDependencies #-} {-# LANGUAGE UndecidableInstances #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE StandaloneDeriving #-} {-# LANGUAGE CPP #-} {-# LANGUAGE DeriveFunctor #-} {-# LANGUAGE TemplateHaskell #-} {-# LANGUAGE NoImplicitPrelude #-} module Numerical.Types.Shape where import GHC.Magic import Data.Data import Data.Typeable() import Data.Type.Equality import qualified Data.Monoid as M import qualified Data.Functor as Fun import qualified Data.Foldable as F import qualified Control.Applicative as A import Prelude hiding (foldl,foldr,init,scanl,scanr,scanl1,scanr1) data Nat = S !Nat | Z deriving (Eq,Show,Read,Typeable,Data) #if defined(__GLASGOW_HASKELL_) && (__GLASGOW_HASKELL__ >= 707) deriving instance Typeable 'Z deriving instance Typeable 'S #endif type family n1 + n2 where Z + n2 = n2 (S n1') + n2 = S (n1' + n2) -- singleton for Nat data SNat :: Nat -> * where SZero :: SNat Z SSucc :: SNat n -> SNat (S n) --gcoerce :: (a :~: b) -> ((a ~ b) => r) -> r --gcoerce Refl x = x --gcoerce = gcastWith -- inductive proof of right-identity of + plus_id_r :: SNat n -> ((n + Z) :~: n) plus_id_r SZero = Refl plus_id_r (SSucc n) = gcastWith (plus_id_r n) Refl -- inductive proof of simplification on the rhs of + plus_succ_r :: SNat n1 -> Proxy n2 -> ((n1 + (S n2)) :~: (S (n1 + n2))) plus_succ_r SZero _ = Refl plus_succ_r (SSucc n1) proxy_n2 = gcastWith (plus_succ_r n1 proxy_n2) Refl type N0 = Z type N1= S N0 type N2 = S N1 type N3 = S N2 type N4 = S N3 type N5 = S N4 type N6 = S N5 type N7 = S N6 type N8 = S N7 type N9 = S N8 type N10 = S N9 {- Need to sort out packed+unboxed vs generic approaches see ShapeAlternatives/ for -} infixr 3 :* {- the concern basically boils down to "will it specialize / inline well" -} newtype At a = At a deriving (Eq, Ord, Read, Show, Typeable, Functor) data Shape (rank :: Nat) a where Nil :: Shape Z a (:*) :: !(a) -> !(Shape r a ) -> Shape (S r) a --deriving (Show) #if defined(__GLASGOW_HASKELL_) && (__GLASGOW_HASKELL__ >= 707) deriving instance Typeable Shape #endif instance Eq (Shape Z a) where (==) _ _ = True instance (Eq a,Eq (Shape s a))=> Eq (Shape (S s) a ) where (==) (a:* as) (b:* bs) = (a == b) && (as == bs ) instance Show (Shape Z a) where show _ = "Nil" instance (Show a, Show (Shape s a))=> Show (Shape (S s) a) where show (a:* as) = show a ++ " :* " ++ show as -- at some point also try data model that -- has layout be dynamicly reified, but for now -- keep it phantom typed for sanity / forcing static dispatch. -- NB: may need to make it more general at some future point --data Strided r a lay = Strided { getStrides :: Shape r a } {-# INLINE reverseShape #-} reverseShape :: Shape n a -> Shape n a reverseShape Nil = Nil reverseShape list = go SZero Nil list where go :: SNat n1 -> Shape n1 a-> Shape n2 a -> Shape (n1 + n2) a go snat acc Nil = gcastWith (plus_id_r snat) acc go snat acc (h :* (t :: Shape n3 a)) = gcastWith (plus_succ_r snat (Proxy :: Proxy n3)) (go (SSucc snat) (h :* acc) t) instance Fun.Functor (Shape Z) where fmap = \ _ Nil -> Nil --{-# INLINE fmap #-} instance (Fun.Functor (Shape r)) => Fun.Functor (Shape (S r)) where fmap = \ f (a :* rest) -> f a :* Fun.fmap f rest --{-# INLINE fmap #-} instance A.Applicative (Shape Z) where pure = \ _ -> Nil --{-# INLINE pure #-} (<*>) = \ _ _ -> Nil --{-# INLINE (<*>) #-} instance A.Applicative (Shape r)=> A.Applicative (Shape (S r)) where pure = \ a -> a :* (A.pure a) --{-# INLINE pure #-} (<*>) = \ (f:* fs) (a :* as) -> f a :* (inline (A.<*>)) fs as --{-# INLINE (<*>) #-} instance F.Foldable (Shape Z) where foldMap = \ _ _ -> M.mempty --{-# fold #-} foldl = \ _ init _ -> init foldr = \ _ init _ -> init foldr' = \_ !init _ -> init foldl' = \_ !init _ -> init instance (F.Foldable (Shape r)) => F.Foldable (Shape (S r)) where foldMap = \f (a:* as) -> f a M.<> F.foldMap f as foldl' = \f !init (a :* as) -> let next = f init a in next `seq` F.foldl f next as foldr' = \f !init (a :* as ) -> f a $! F.foldr f init as foldl = \f init (a :* as) -> let next = f init a in F.foldl f next as foldr = \f init (a :* as ) -> f a $ F.foldr f init as -- map2 :: (A.Applicative (Shape r))=> (a->b ->c) -> (Shape r a) -> (Shape r b) -> (Shape r c ) map2 = \f l r -> A.pure f A.<*> l A.<*> r {-# SPECIALIZE map2 :: (a->b->c)-> (Shape Z a )-> Shape Z b -> Shape Z c #-} {-# SPECIALIZE map2 :: (a->b->c)-> (Shape (S Z) a )-> Shape (S Z) b -> Shape (S Z) c #-} {-# SPECIALIZE map2 :: (a->b->c)-> (Shape (S (S Z)) a )-> Shape (S (S Z)) b -> Shape (S (S Z)) c #-} {-# SPECIALIZE map2 :: (a->b->c)-> (Shape (S (S(S Z))) a )-> Shape (S (S (S Z))) b -> Shape (S (S(S Z))) c #-} -- {-# INLINABLE map2 #-} }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8848> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 17

[GHC] #9184: Allow the use of Python 3 when building GHC
by GHC 01 Dec '16

01 Dec '16

#9184: Allow the use of Python 3 when building GHC -------------------------------------------+------------------------------- Reporter: aspidites | Owner: Type: feature request | Status: new Priority: normal | Milestone: Component: Build System | Version: 7.8.2 Keywords: python | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Easy (less than 1 hour) | Type of failure: Blocked By: | None/Unknown Related Tickets: | Test Case: | Blocking: -------------------------------------------+------------------------------- According to the [https://ghc.haskell.org/trac/ghc/wiki/Building/Preparation/Windows wiki], it is preferred that a user have Python 2.7 installed before attempting to build GHC. As I'm on Arch Linux the default Python version is 3.4. As such, rather than simply changing the shebang line to explicitly ask for version 2, I made it so that the scripts all use syntax that works properly for versions 2.7 to 3.4. I don't think I've used anything that disallows 2.6, but don't have that version available to check against. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9184> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 24

[GHC] #11378: Use the compiler that built ghc for dynamic code loading, for cross-compiling
by GHC 01 Dec '16

01 Dec '16

#11378: Use the compiler that built ghc for dynamic code loading, for cross- compiling -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: feature | Status: new request | Priority: lowest | Milestone: ⊥ Component: Template | Version: 7.11 Haskell | 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: -------------------------------------+------------------------------------- In a non cross-compiled stage 2 compiler, this coincides with what we do today, because the stage 1 implementation of GHC is the same as the stage 2 implementation. But in a cross-compiler, these are NOT the same: the cross-compiler's ghc library handles the package database in the target platform, whereas the compiler that built stage 2 (stage 1)'s ghc library handles the package database for the compiler's platform. So, the correct architecture for supporting compiler plugins is that GHC should use the package-database/interface infrastructure from the compiler which built it (stage 1) to handle loading. A few things to note: 1. There is not a backwards-compatibility problem, because we can assume that stage 1 and stage 2 are built from the same codebase. In particular, this does NOT mean that TH/plugins start working on a stage 1 compiler. You have to have bootstrapped from the same codebase. 2. Plugins (and even Template Haskell) have to be a bit more careful to take advantage of this. We cannot just assume that the package database for plugins is the same as the package database for compiling (revisiting #11244). Template Haskell is even trickier: we should NOT be allowed to use regular imports in splices; they have to go through `TH_ONLY` imports, c.f. #11377 3. Can we actually link against the old version of `ghc`? Yes we can! All we need is (1) for it to have been built with a different IPID, and (2) to use module renaming to rename the relevant modules to a different module name, so that we can import them without a conflict. Stepping back, there have been two approaches for dealing with the problem of Template Haskell/compiler plugins not reliably working: 1. Run the code in an external process run in the target code, or 2. Fix GHC to load code which is appropriately built for the bootstrap platform. While (1) offers a more uniform environment for developers, (2) is absolutely necessary in some situations when you need to load code in the same address space as GHC, e.g. as mentioned in https://mail.haskell.org/pipermail/ghc-devs/2015-November/010478.html Thus, it seems worth pursuing both. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11378> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #11377: Template Haskell only imports
by GHC 01 Dec '16

01 Dec '16

#11377: Template Haskell only imports -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: feature | Status: new request | Priority: low | Milestone: Component: Template | Version: 7.11 Haskell | 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: -------------------------------------+------------------------------------- This is the Template Haskell equivalent of #11244. It should be possible to use the plugin package database to make a Template Haskell only import; I suggest the syntax `import {-# TH_ONLY #-} Foo`. Such imports are only allowed to be used in splices, and cannot be quoted (they are not allowed to reside in the final code generated by the splice.) The reason a user may want to use such an import is that it means that they can declare that a library is necessary when compiling splices, but not necessary when actually linking against it. There are extra benefits. If a user is compiling the package with the splices with profiling, it ordinarily requires all imports to have been compiled with profiling. This would NOT be necessary for a `TH_ONLY` import, since we're only ever going to run the imported code in the (non- profiled) compiler, and not in the (profiled) executable.) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11377> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

[GHC] #11293: Compiler plugins don't work with profiling
by GHC 01 Dec '16

01 Dec '16

#11293: Compiler plugins don't work with profiling -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Profiling | Version: 7.11 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: -------------------------------------+------------------------------------- Suppose I have a normally built compiler plugin, and I'd like to use it while building some profiled code. This doesn't work: (non-profiled) GHC searches for the profiled version of the plugin. This is wrong wrong wrong: plugin searches should be with respect to how GHC was compiled. (Test case: take the `plugins01` test and add `-prof` to it) Now that we have `-plugin-package`, it should be a simple matter to ensure that plugin lookups follow a different codepath, though I'm not sure how to figure out what GHC is built with; this should just be however TH does it. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11293> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #10510: Testsuite driver does not run tests in parallel on Windows
by GHC 30 Nov '16

30 Nov '16

#10510: Testsuite driver does not run tests in parallel on Windows -------------------------------------+------------------------------------- Reporter: thomie | Owner: Type: bug | Status: new Priority: high | Milestone: Component: Test Suite | Version: 7.10.1 Keywords: | Operating System: Windows Architecture: | Type of failure: None/Unknown Unknown/Multiple | Blocked By: Test Case: | Related Tickets: Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- `testsuite/driver/runtests.py` contains the following 3 lines: {{{ if windows: print("Warning: Ignoring request to use threads as running on Windows") config.use_threads = 0 }}} They were introduced in commit e9b3d6cc65c0b69bd4a50bfa7bc7e1d3536446a8: {{{ Author: Ian Lynagh <> Date: Wed Oct 28 17:54:21 2009 +0000 Don't use threads on Windows It seems to cause some sort of deadlock }}} I tried removing them (Windows 64 bit, using msys2), but then running validate doesn't get any further than: {{{ ... =====> Overlap3(normal) 1605 of 4525 [2, 7, 0] =====> Overlap4(normal) 1606 of 4525 [2, 7, 0] =====> Overlap5(normal) 1607 of 4525 [2, 7, 0] =====> Overlap6(normal) 1608 of 4525 [2, 7, 0] =====> Overlap7(normal) 1609 of 4525 [2, 7, 0] =====> Overlap9(normal) 1610 of 4525 [2, 7, 0] }}} It would be great if we could run the testsuite in parallel again, since running validate takes forever currently. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10510> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #8974: 64 bit windows executable built with ghc-7.9.20140405+LLVM segfaults
by GHC 29 Nov '16

29 Nov '16

#8974: 64 bit windows executable built with ghc-7.9.20140405+LLVM segfaults -----------------------------------+---------------------------------- Reporter: awson | Owner: Type: bug | Status: new Priority: high | Milestone: 7.8.2 Component: Compiler (LLVM) | Version: 7.9 Keywords: | Operating System: Windows Architecture: x86_64 (amd64) | Type of failure: Runtime crash Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | -----------------------------------+---------------------------------- {{{ -- test.hs import System.Mem (performMajorGC) main = performMajorGC >> putStrLn "Done" }}} built with `ghc -pgmlo opt -pgmlc llc -fllvm --make test.hs` segfaults both for LLVM 3.4 and 3.5svn (taken from http://sourceforge.net/projects/msys2/files/REPOS/MINGW/x86_64) 32-bit ghc-7.9.20140404+llvm produces good executable. Adding `ArchX86_64` `OSMinGW32` case to `moduleLayout` in `compiler/llvmGen/LlvmCodeGen/Ppr.hs` improves things slightly (some code segfaulting without it starts to work) but still does not cure the code above. Also I've found the Cmm produced for LLVM CG differs from that produced for NCG. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8974> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 52

[GHC] #11219: Implement fine-grained `-Werror=...` facility
by GHC 29 Nov '16

29 Nov '16

#11219: Implement fine-grained `-Werror=...` facility -------------------------------------+------------------------------------- Reporter: hvr | Owner: quchen Type: feature | Status: new request | Priority: normal | Milestone: 8.2.1 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: Design/Warnings -------------------------------------+------------------------------------- ...copied from Design/Warnings: Introduce variant of `-Werror` (c.f. GCC's `-Werror=*`) which allows to specify the individual warnings to be promoted to errors, e.g. - `-Wall -Werror=orphans` would only promote `-Worphans` warnings into errors - `-Wall -Werror -Wno-error=missing-methods` would promote all warnings //except// `-Wmissing-methods` into errors TODO: specify semantics a bit more exactly (the basic idea is to follow gcc/clang as much as sensible, and only deviate where it's necessary) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11219> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 10

[GHC] #11629: reify returns Dec that use ConT instead of PromotedT
by GHC 29 Nov '16

29 Nov '16

#11629: reify returns Dec that use ConT instead of PromotedT -------------------------------------+------------------------------------- Reporter: aavogt | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Template | Version: 7.10.2 Haskell | 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: -------------------------------------+------------------------------------- -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11629> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #9007: relocation R_X86_64_32 against `stg_CHARLIKE_closure' can not be used when making a shared object
by GHC 29 Nov '16

29 Nov '16

#9007: relocation R_X86_64_32 against `stg_CHARLIKE_closure' can not be used when making a shared object ----------------------------------+---------------------------------------- Reporter: nomeata | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.9 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Building GHC failed Unknown/Multiple | Test Case: Difficulty: Unknown | Blocking: Blocked By: | Related Tickets: | ----------------------------------+---------------------------------------- Building HEAD on Debian wheezy with the Debian packaging setup fails: {{{ configure: Building in-tree ghc-pwd /usr/bin/ld: utils/ghc-pwd/dist-boot/Main.o: relocation R_X86_64_32 against `stg_CHARLIKE_closure' can not be used when making a shared object; recompile with -fPIC utils/ghc-pwd/dist-boot/Main.o: could not read symbols: Bad value collect2: error: ld returned 1 exit status configure: error: Building ghc-pwd failed }}} http://deb.haskell.org/dailies/2014-04-16/ghc_7.9.20140416-0.daily_amd64.bu… This is possibly related to the Hardening flags added by Debian: {{{ ~ $ dpkg-buildflags --get LDFLAGS -Wl,-z,relro ~ $ dpkg-buildflags --get CFLAGS -g -O2 -fstack-protector --param=ssp-buffer-size=4 -Wformat -Werror =format-security }}} Other related bug might be #3668 and https://bugs.debian.org/712228 I’ll disable the hardening flags and see if it helps. The error first occurred 2014-03-27. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9007> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 24