July 2017 - ghc-tickets

[GHC] #13993: Certain inter-module specializations run out of simplifier ticks
by GHC 20 Jul '17

20 Jul '17

#13993: Certain inter-module specializations run out of simplifier ticks -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: (none) Type: bug | Status: new Priority: high | Milestone: 8.4.1 Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Compile-time Unknown/Multiple | performance bug Test Case: | Blocked By: Blocking: | Related Tickets: #9630 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- A modification of ezyang's test case in #9630 yields the below. Triggering specialization in a separate module can run the simplifier out of ticks. Notably, moving the definition of `T` into `GenSpec` resolves the problem. Unlike #9630 proper, this seems to cause trouble going back as far as GHC 7.4. {{{#!hs module GenSpec where import Gen import GHC.Generics -- Trigger specialization tput :: T -> Put tput = gput . from }}} {{{#!hs {-# LANGUAGE DeriveGeneric #-} {-# LANGUAGE TypeOperators #-} {-# OPTIONS_GHC -fno-warn-missing-methods #-} module Gen where import GHC.Generics import Control.Monad import Control.Applicative import Data.Monoid data PairS a = PairS a !(() -> ()) newtype PutM a = Put { unPut :: PairS a } -- Use of this writer monad seems to be important; IO speeds it up type Put = PutM () --type Put = IO () -- binary has INLINE pragmas on most of the instances but you can still -- trigger bad behavior without them. instance Functor PutM where fmap f m = Put $ let PairS a w = unPut m in PairS (f a) w -- Just to appease AMP instance Applicative PutM where pure = return (<*>) = ap instance Monad PutM where return a = Put $ PairS a id m >>= k = Put $ let PairS a w = unPut m PairS b w' = unPut (k a) in PairS b (w . w') class GBinary f where gput :: f t -> Put -- Forcing the dictionary to have two elements hurts -- the optimizer a lot. not_used :: f t instance GBinary a => GBinary (M1 i c a) where gput = gput . unM1 instance Binary a => GBinary (K1 i a) where gput = put . unK1 instance (GBinary a, GBinary b) => GBinary (a :*: b) where gput (x :*: y) = gput x >> gput y class Binary t where put :: t -> Put instance Binary () where put () = return () data T = T () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () () deriving Generic }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13993> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

[GHC] #13805: GHC 8.0.2 fails to build on macOS 10.13/Xcode 9 - preprocessor error in ghc-pkg
by GHC 20 Jul '17

20 Jul '17

#13805: GHC 8.0.2 fails to build on macOS 10.13/Xcode 9 - preprocessor error in ghc-pkg -------------------------------------+------------------------------------- Reporter: mistydemeo | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: ghc-pkg | Version: 8.0.2 Keywords: | Operating System: MacOS X Architecture: x86_64 | Type of failure: Building GHC (amd64) | failed Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Building GHC 8.0.2 on macOS 10.13 fails with the following error when building {{{ utils/ghc-pkg/Main.hs:1269:40: error: error: editor placeholder in source file then termText (location db) <#> termText "\n (no packages)\n" ^ }}} I'm not very familiar with Haskell, but it looks to me like the C preprocessor is mistaking `<#>` for an invalid cpp directive instead of Haskell syntax. This is using the Xcode 9 beta (and its associated CLT), which ships "Apple LLVM version 9.0.0 (clang-900.0.22.8)". The same version should be available in the Xcode 9 beta for 10.12, but I haven't tested. The full build logs are available here: https://gist.github.com/anonymous/dc5f0c9d087f5d299f71393805c5d611 -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13805> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #13934: FP_CONVERT_CPU fails to convert CPU when invoked from bindist
by GHC 20 Jul '17

20 Jul '17

#13934: FP_CONVERT_CPU fails to convert CPU when invoked from bindist -------------------------------------+------------------------------------- Reporter: bgamari | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.2.1 Component: Build System | 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: -------------------------------------+------------------------------------- Trying to pass `--target=arm-linux-gnueabihf` to a ARM 8.0.2 bindist's `configure` on my AArch64 box fails with, {{{ $ ./configure --prefix=/mnt/ext/arm/roots/ghc/8.0.1-armv7 --target=arm- linux-gnueabihf checking for path to top of build tree... /mnt/ext/arm/ghc-8.0.2 Build platform inferred as: arm-unknown-linux Host platform inferred as: arm-unknown-linux Unknown CPU }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13934> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #13987: T13701 fails sporadically
by GHC 20 Jul '17

20 Jul '17

#13987: T13701 fails sporadically -------------------------------------+------------------------------------- Reporter: duog | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Test Suite | Version: 8.3 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Other Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- T13701 fails sporadically, comments in Phab:D3586 indicate that it happens while the system is under load. I reproduced this issue by duplicating T13701 7 times, making tests T13701[A-H], then running {{{ make test TEST="T13701A T13701B T13701C T13701D T13701E T13701F T13701G T13701H THREADS=8" }}} on my box with 4 cpus. Comparing profiling reports of the test under and not under load, I found the additional allocations to be in SysTools.builderMainLoop.loop. This function busy-waits from when 2 EOFs are received until getProcessExitCode returns a Just. That this is the cause was verified by changing {{{ | otherwise -> loop chan hProcess t p exitcode }}} to {{{ | otherwise -> threadDelay 10000 >> loop chan hProcess t p exitcode }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13987> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #13983: Type synonym referred to as datatype in error message
by GHC 19 Jul '17

19 Jul '17

#13983: Type synonym referred to as datatype in error message -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 (Type checker) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Poor/confusing Unknown/Multiple | error message Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- To reproduce the issue, compile this file with GHC 8.0.1, 8.0.2, 8.2.1, or HEAD: {{{#!hs {-# LANGUAGE PolyKinds #-} {-# LANGUAGE RankNTypes #-} module Bug where import Data.Proxy type Wat = forall (a :: k). Proxy a }}} {{{ GHCi, version 8.3.20170706: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/rgscott/.ghci [1 of 1] Compiling Bug ( Bug.hs, interpreted ) Bug.hs:7:1: error: • Kind variable ‘k’ is implicitly bound in datatype ‘Wat’, but does not appear as the kind of any of its type variables. Perhaps you meant to bind it (with TypeInType) explicitly somewhere? Type variables with inferred kinds: (k :: *) • In the type declaration for ‘Wat’ | 7 | type Wat = forall (a :: k). Proxy a | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ }}} The error message mistakenly refers to `Wat` as a datatype. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13983> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #13680: Can't use TypeApplications with empty list expression
by GHC 19 Jul '17

19 Jul '17

#13680: Can't use TypeApplications with empty list expression -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 (Type checker) | Keywords: | Operating System: Unknown/Multiple TypeApplications | Architecture: | Type of failure: GHC rejects Unknown/Multiple | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Why can't I do this? {{{#!hs {-# LANGUAGE TypeApplications #-} module Bug where foo :: [Int] foo = [] @Int }}} Compiling this with GHC 8.0.1, 8.0.2, 8.2.1, or HEAD gives me: {{{ $ /opt/ghc/head/bin/ghci Bug2.hs GHCi, version 8.3.20170509: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/rgscott/.ghci [1 of 1] Compiling Bug ( Bug2.hs, interpreted ) Bug2.hs:5:7: error: • Cannot apply expression of type ‘[a0]’ to a visible type argument ‘Int’ • In the expression: [] @Int In an equation for ‘foo’: foo = [] @Int | 5 | foo = [] @Int | ^^^^^^^ }}} This seems really strange, since I can use `TypeApplications` with expressions like `Nothing @Int` without issues. According to GHCi: {{{ $ /opt/ghc/head/bin/ghci GHCi, version 8.3.20170509: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/rgscott/.ghci λ> :set -XTypeApplications -fprint-explicit-foralls λ> :type +v [] [] :: forall {a}. [a] λ> :type +v Nothing Nothing :: forall a. Maybe a }}} The type variable for `[]` isn't visible! But I don't see any reason why it shouldn't be, especially since, conceptually, the data type declaration for lists is: {{{#!hs data [] a = [] | a : [a] }}} I suspect that `[]`'s tyvar binder visibility is not wired into GHC correctly. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13680> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 15

[GHC] #13434: hs_try_putmvar003 is timing out / segfaulting
by GHC 19 Jul '17

19 Jul '17

#13434: hs_try_putmvar003 is timing out / segfaulting -------------------------------------+------------------------------------- Reporter: ezyang | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Runtime | Version: 8.1 System | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Runtime crash Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- This is basically reproduceable every time I validate. During validation, the program times out, but when I run it by hand, it segfaults. Running in debug sanity mode, I get: {{{ hstry: internal error: MVAR_CLEAN on mutable list (GHC version 8.3.20170316 for x86_64_unknown_linux) Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} Simon Marlow, I'm happy to investigate more if you can't repro. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13434> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #13915: GHC 8.2 regression: "Can't find interface-file declaration" for promoted data family instance
by GHC 19 Jul '17

19 Jul '17

#13915: GHC 8.2 regression: "Can't find interface-file declaration" for promoted data family instance -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: high | Milestone: 8.2.1 Component: Compiler | Version: 8.2.1-rc2 Keywords: TypeInType, | Operating System: Unknown/Multiple TypeFamilies | Architecture: | Type of failure: GHC rejects Unknown/Multiple | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Due to #12088, you can't define a data family instance and promote it in the same module. One could, up until GHC 8.2, work around this using (somewhat obscure) wisdom: define the data family instance in a separate module from where it's promoted. For example, `Bug` typechecks in GHC 8.0.1 and 8.0.2: {{{#!hs {-# LANGUAGE TypeFamilies #-} module Foo where data family T a data instance T Int = MkT }}} {{{#!hs {-# LANGUAGE TypeInType #-} module Bug where import Foo data Proxy (a :: k) data S = MkS (Proxy 'MkT) }}} However, this stopped typechecking in GHC 8.2: {{{ $ /opt/ghc/8.2.1/bin/ghci Bug.hs GHCi, version 8.2.0.20170623: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/rgscott/.ghci [1 of 2] Compiling Foo ( Foo.hs, interpreted ) [2 of 2] Compiling Bug ( Bug.hs, interpreted ) Bug.hs:1:1: error: Can't find interface-file declaration for variable Foo.$tc'MkT Probable cause: bug in .hi-boot file, or inconsistent .hi file Use -ddump-if-trace to get an idea of which file caused the error | 1 | {-# LANGUAGE TypeInType #-} | ^ }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13915> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #13948: GHC incorrectly suggests a data constructor is in scope in 8.2
by GHC 19 Jul '17

19 Jul '17

#13948: GHC incorrectly suggests a data constructor is in scope in 8.2 -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.2.1-rc2 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Poor/confusing Unknown/Multiple | error message Test Case: | Blocked By: Blocking: | Related Tickets: #13568 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- After commit 343cb32d0983f576d344a2d04a35c3fd6eecf2c5 (the fix for #13568), which debuted in GHC 8.2, GHC thinks that out-of-scope identifiers are actually in-scope! {{{#!hs module Bug where f :: T f = undefined }}} In GHC 8.2.1-rc3, the error message is: {{{ $ /opt/ghc/8.2.1/bin/ghc Bug.hs [1 of 1] Compiling Bug ( Bug.hs, Bug.o ) Bug.hs:3:6: error: Not in scope: type constructor or class ‘T’ A data constructor of that name is in scope; did you mean DataKinds? | 3 | f :: T | ^ }}} Compare this to GHC 8.0.2, which gives a more sensible error message: {{{ $ /opt/ghc/8.0.2/bin/ghc Bug.hs [1 of 1] Compiling Bug ( Bug.hs, Bug.o ) Bug.hs:3:6: error: Not in scope: type constructor or class ‘T’ }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13948> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #13899: Improve [-Wmissing-home-modules] error message
by GHC 19 Jul '17

19 Jul '17

#13899: Improve [-Wmissing-home-modules] error message -------------------------------------+------------------------------------- Reporter: alanz | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.2.1 Component: Compiler | Version: 8.2.1-rc2 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: -------------------------------------+------------------------------------- At the moment when this warning is triggered, GHC emits something like {{{#!hs <no location info>: warning: [-Wmissing-home-modules] Modules are not listed in command line: Language.Haskell.GHC.ExactPrint }}} It is not clear that this is actually reporting a problem in the cabal file which has incomplete "other-modules" for one or other target. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13899> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5