February 2016 - ghc-tickets

[GHC] #9887: No message when GCHI reusing compiled code
by GHC 12 Mar '16

12 Mar '16

#9887: No message when GCHI reusing compiled code -------------------------------------+------------------------------------- Reporter: jsrice | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: GHCi | Version: 7.8.3 Keywords: | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Easy (less than 1 | Type of failure: Other hour) | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Revisions: -------------------------------------+------------------------------------- Running ":l Main.hs" at the GHCI prompt would load the compiled module. I am new to Haskell and this was a stumbling block that made it so that I could not interactively debug my code with GHCI. Deleting the object files was suggested to me on IRC and that allowed me to get the "*Main" prompt I was looking for. A message warning users "reusing compiled code" would likely save some future users some time. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9887> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #11395: The via-C code generation backend is incompatible with gcc 5.3.1 on m68k (ELF)
by GHC 12 Mar '16

12 Mar '16

#11395: The via-C code generation backend is incompatible with gcc 5.3.1 on m68k (ELF) -------------------------------------+------------------------------------- Reporter: mkarcher | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.2 Keywords: | Operating System: Linux Architecture: m68k | Type of failure: GHC doesn't work | at all Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- ghc is generating C code like in the following simplified example: demo1.c: {{{ #include <assert.h> extern void* int_returning_function(); typedef int (*int_f_ptr)(); int main(void) { int result; { int_f_ptr correctly_typed_pointer; correctly_typed_pointer = (int_f_ptr)int_returning_function; result = correctly_typed_pointer(); } assert(result == 42); return 0; } }}} demo2.c: {{{ int int_returning_function() { return 42; } }}} This code fails to work on m68k with gcc-5.3.1 if optimization is turned on. The real-life example from ghc is rts/Schedule.c:raiseExceptionHelper as int_returning_function and rts/Exception.cmm:stg_raisezh as main. The reason it fails is that on m68k, a function returning a pointer returns the result in %a0, while a function returning an integer returns it in %d0. If optimization is enabled, gcc elides the function pointer and calls the function as-if it were called directly. Without optimization, gcc expects the result to be in %d0, where it actually is. In my oppinion, gcc is allowed to do that, because according to N843 (C99 draft), 6.2.7 #2 "All declarations that refer to the same object or function shall have compatible type; otherwise, the behavior is undefined." is violated in this example. Compatibility of return types is defined in 6.7.5.3 #11 "For two function types to be compatible, both shall specify compatible return types. Moreover, [...]" which is clearly violated. If you just look at demo1.c and assume that 6.2.7 #2 is *not* violated, int_returning_function would in fact be a function returning a pointer. In that case, compiling demo1.c would violate 6.3.2.3 #8 "If a converted pointer is used to call a function whose type is not compatible with the pointed-to type, the behavior is undefined." and/or 6.5.2.2 "If the function is defined with a type that is not compatible with the type (of the expression) pointed to by the expression that denotes the called function, the behavior is undefined." You might want to discuss this issue with the gcc developers, but I don't see a standard claus justifying the approach of declaring all functions returning StgFunPtr and casting them to the right type. If you cross-compile the source tarball Debian claims to be 7.10.3-rc1 to m68k using gcc 5.3.1, ghc crashes on startup because stg_raisezh gets a wrong value as frame_type and tries to cancel an STM transaction that does not exist at all. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11395> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 30

[GHC] #11494: Add -Wcompat to -Wall
by GHC 11 Mar '16

11 Mar '16

#11494: Add -Wcompat to -Wall -------------------------------------+------------------------------------- Reporter: bgamari | Owner: Type: task | Status: new Priority: normal | Milestone: 8.0.1 Component: Compiler | Version: 8.0.1-rc1 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: -------------------------------------+------------------------------------- As discussed in the call today we will add the `-Wcompat` warnings to `-Wall` pending discussion with the CLC. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11494> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #11624: Cannot declare hs-boot declaration if there is already a value in scope
by GHC 11 Mar '16

11 Mar '16

#11624: Cannot declare hs-boot declaration if there is already a value in scope -------------------------------------+------------------------------------- Reporter: ezyang | Owner: ezyang Type: bug | Status: new Priority: normal | Milestone: 8.0.1 Component: Compiler | Version: 8.0.1-rc2 (Type checker) | 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: -------------------------------------+------------------------------------- This code no longer works: {{{ -- A.hs-boot module A where concat :: Int -> Int -- overlaps with Prelude's concat -- B.hs module B where import Prelude () import {-# SOURCE #-} A x = concat 3 -- A.hs module A where import B concat n = n + 2 }}} Building `ghc --make A`, this crashes with: {{{ [1 of 3] Compiling A[boot] ( A.hs-boot, A.o-boot ) [2 of 3] Compiling B ( B.hs, B.o ) B.hs:4:5: error: • Can't find interface-file declaration for variable concat 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 • In the expression: concat 3 In an equation for ‘x’: x = concat 3 }}} The trouble is that renaming has botched up the top-level declaration for boot: {{{ ezyang@sabre:~$ ghc-8.0 -c A.hs-boot -ddump-rn -dppr-debug A.hs-boot:1:1: ==================== Renamer ==================== {A.hs-boot:2:1-20} base-4.9.0.0:Data.Foldable.concat{v r2S} :: {A.hs-boot:2:11-20} ghc-prim-0.5.0.0:GHC.Types.Int{(w) tc 3J} -> ghc-prim-0.5.0.0:GHC.Types.Int{(w) tc 3J} }}} Is a regression from 7.10. The dodginess is proximal to renameSig but I haven't quite narrowed it down yet. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11624> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #11608: Possible type-checker regression in GHC 8.0 when compiling `microlens`
by GHC 11 Mar '16

11 Mar '16

#11608: Possible type-checker regression in GHC 8.0 when compiling `microlens` -------------------------------------+------------------------------------- Reporter: hvr | Owner: Type: bug | Status: new Priority: highest | Milestone: 8.0.1 Component: Compiler | Version: (Type checker) | 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 just updated the GHC 8.0 snapshot I was using from 8.0.0.20160214-g977fb8 to 8.0.0.20160218-g23baff7, and suddenly hackage:microlens-0.4.2.0 fails to build with {{{ ... Preprocessing library microlens-0.4.2.0... [1 of 4] Compiling Lens.Micro.Type ( src/Lens/Micro/Type.hs, dist/build/Lens/Micro/Type.o ) [2 of 4] Compiling Lens.Micro.Internal ( src/Lens/Micro/Internal.hs, dist/build/Lens/Micro/Internal.o ) src/Lens/Micro/Internal.hs:184:3: error: • Couldn't match type ‘s’ with ‘g0 a’ ‘s’ is untouchable inside the constraints: (Traversable g, s ~ g a, t ~ g b, Applicative f) bound by the type signature for: each :: (Traversable g, s ~ g a, t ~ g b, Applicative f) => (a -> f b) -> s -> f t at src/Lens/Micro/Internal.hs:184:3-27 • In the ambiguity check for ‘each’ To defer the ambiguity check to use sites, enable AllowAmbiguousTypes When checking the class method: each :: forall s t a b. Each s t a b => Traversal s t a b In the class declaration for ‘Each’ }}} ...is this a regression or a feature? :-) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11608> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 19

[GHC] #10691: certain operations in new integer-gmp are too lazy
by GHC 11 Mar '16

11 Mar '16

#10691: certain operations in new integer-gmp are too lazy -------------------------------------+------------------------------------- Reporter: rwbarton | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: libraries | Version: 7.10.1 (other) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Incorrect result Unknown/Multiple | at runtime Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- This came up in #ghc the other day. {{{ rwbarton@morphism:~$ ghci-7.10.1 GHCi, version 7.10.1: http://www.haskell.org/ghc/ :? for help Prelude> undefined * 0 0 }}} For consistency not just with Int, but also other integer-* implementations, this should be undefined. Also affected is for example `andInteger`. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10691> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #11354: Install script installs docs without -version suffix
by GHC 11 Mar '16

11 Mar '16

#11354: Install script installs docs without -version suffix -------------------------------------+------------------------------------- Reporter: osa1 | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Build System | Version: 8.1 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: -------------------------------------+------------------------------------- This is potentially bad when multiple GHC versions are installed - last installed one overwrites all the docs. It seems like docs are installed under {{{$PACKAGE_TARNAME}}}, and tarname is initialized as {{{"ghc"}}} in the {{{ac}}} file. I guess one solution would be to add version suffix to tarname (it seems like tarname is only used when installing docs) or alternatively use {{{${PACKAGE_TARNAME}-${PACKAGE_VERSION}}}} to install docs. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11354> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #11475: Lint should check for inexhaustive alternatives
by GHC 10 Mar '16

10 Mar '16

#11475: Lint should check for inexhaustive alternatives -------------------------------------+------------------------------------- Reporter: simonpj | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 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: -------------------------------------+------------------------------------- GHC's simplifier prunes inaccessible alternatives from case expressions. E.g. {{{ case x of A y -> e1 _ -> ....(case x of A y -> r1 B -> r2 C -> r3) ... }}} It'll prune the `A` alternative from the inner case to get {{{ case x of A y -> e1 _ -> ....(case x of B -> r2 C -> r3) ... }}} BUT, there is no independent check from Lint that this pruning does the Right Thing. Trac #11172 turned out to be a bug in the pruning, which led to something like {{{ case x of Left y -> e1 }}} but in fact `x` ended up being bound to `(Right v)` for some value `v`. But because there is only a single alternative left, GHC does not test the tag, but instead goes ahead and accesses it the `Right` value as if it was a `Left` value. And because of pointer-tagging, it'll assume the wrong tag on the pointer, and very soon you are in seg-fault land. So this ticket is to suggest: make Core Lint do it's best to check that all cases are in fact exhaustive. There are two ways in which we prune cases: * Enclosing pattern matches (as above) * GADT pattern matches may be exhaustive even when they don't mention all constructors. For the latter, it's impossible for Lint to reproduce all the reasoning of the type checker, but I bet that in 99.5% of the cases the existing function `dataConCantMatch` will do the job. So if Lit * Maintains an environment saying what each variable can't match (as the simplifier does) * Does a simple-minded `dataConCantMatch` test based on the typ of the scrutinee I think we'd be a long way forward. It's possible that a valid program could trigger a Lint report, but let's see. Any volunteers? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11475> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #11661: Missing MonadFail instance for Q monad from TemplateHaskell
by GHC 08 Mar '16

08 Mar '16

#11661: Missing MonadFail instance for Q monad from TemplateHaskell -------------------------------------+------------------------------------- Reporter: PeterTrsko | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Template | Version: 8.0.1-rc2 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: -------------------------------------+------------------------------------- Since GHC 8.0.1 introduces MonadFailDesugaring one would expect Q monad to have an instance for MonadFail. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11661> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #11643: Core lint error in simplifier when compiling Rules1 with -O -prof -dcore-lint
by GHC 07 Mar '16

07 Mar '16

#11643: Core lint error in simplifier when compiling Rules1 with -O -prof -dcore- lint -------------------------------------+------------------------------------- Reporter: thomie | Owner: Type: bug | Status: new Priority: highest | Milestone: 8.0.1 Component: Compiler | Version: 8.0.1-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: -------------------------------------+------------------------------------- Running `make TEST=Rules1 WAY=profasm` results in a core lint error (see attachment) with ghc-8.0.1-rc2. This is the code: {{{#!hs {-# LANGUAGE TypeFamilies #-} module Rules1 where class C a where data T a instance (C a, C b) => C (a,b) where data T (a,b) = TPair (T a) (T b) mapT :: (C a, C b) => (a -> b) -> T a -> T b {-# NOINLINE mapT #-} -- Otherwwise we get a warning from the rule mapT = undefined zipT :: (C a, C b) => T a -> T b -> T (a,b) {-# NOINLINE [1] zipT #-} zipT = undefined {-# RULES "zipT/mapT" forall f x y. zipT (mapT f x) y = mapT (\(x,y) -> (f x, y)) (zipT x y) #-} }}} {{{ $ ghc-8.0.1 -O -prof Rules1.hs -dcore-lint }}} This is a regression from 7.10.3, and core lint errors are bad, so setting priority to highest. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11643> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8