August 2013 - ghc-tickets

Re: [GHC] #5144: Pattern synonyms
by GHC 22 Aug '13

22 Aug '13

#5144: Pattern synonyms -------------------------------------+------------------------------------ Reporter: simonpj | Owner: cactus Type: feature request | Status: new Priority: normal | Milestone: _|_ Component: Compiler | Version: Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by simonmar): We need to think about the import/export situation. One of the goals of pattern synonyms is to abstract constructors, so that a library can change a data type representation while allowing clients to continue to use pattern matching. Clearly we can't use the `T(P1,P2,..)` syntax for exporting patterns (what is `T`?), but perhaps we can use the `pattern` syntax for exporting constructors. I haven't thought about this very hard. When a pattern is imported, can it be used in an expression? Does the client know whether it can be used in an expression or not? What do the Haddock docs for pattern synonyms look like? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5144#comment:21> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #8154: Possible bug in open type familes: Conflicting (a->a) and (a->a->a) instances
by GHC 22 Aug '13

22 Aug '13

#8154: Possible bug in open type familes: Conflicting (a->a) and (a->a->a) instances ----------------------------+---------------------------------------------- Reporter: nh2 | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.8.1 Component: Template | Version: 7.7 Haskell | Operating System: Unknown/Multiple Keywords: type | Type of failure: GHC rejects valid program families | Test Case: Architecture: | Blocking: Unknown/Multiple | Difficulty: Unknown | Blocked By: | Related Tickets: | ----------------------------+---------------------------------------------- {{{ {-# LANGUAGE TypeFamilies #-} module Test where type family BoundsOf x type instance BoundsOf (a->a) = Int type instance BoundsOf (a->a->a) = (Int,Int) }}} This worked with GHC 7.6, but not with 7.8 HEAD (currently at 6cc7d3f). To check: {{{ wget https://gist.github.com/nh2/6302087/raw/8167e7a1c8613aa384c2e8ca2f4ea9ade87… ghci ghc-7.7-type-a-a-a-families.hs # 7.6, all fine ghci ghc-7.7-type-a-a-a-families.hs # 7.7, breaks }}} On #ghc, we don't really understand whether this is the right thing to happen or not. {{{ <rwbarton> see http://comments.gmane.org/gmane.comp.lang.haskell.glasgow.user/23734 <carter> c_wraith: im not sure why thats NOT working for open type familes too though <carter> a->a and a->a->a don't overlap... <rwbarton> ah it's in that thread. "Open (normal, old-fashioned) type families are essentially unchanged. In particular, coincident overlap and non-linear patterns *are* allowed. The overlap check between open type family instances now does a unification without an "occurs check" to mark (x, x) and ([y], y) as overlapping, as necessary for type soundness." }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8154> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #8077: Make write_barrier and friends show up with KEEP_INLINES
by GHC 21 Aug '13

21 Aug '13

#8077: Make write_barrier and friends show up with KEEP_INLINES -------------------------------------+------------------------------------- Reporter: rrnewton | Owner: Type: bug | Status: new Priority: low | Milestone: Component: Runtime System | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: Compile-time Difficulty: Easy (less than 1 | crash hour) | Test Case: Blocked By: | Blocking: Related Tickets: | -------------------------------------+------------------------------------- The three barrier functions (write_barrier, store_load_barrier, load_load_barrier) are currently #defined macros when the non-threaded RTS is built. This means they are missed by the Inlines.c trick and can cause link errors in libraries that depend on them. The fix for this is obvious and this ticket is just a place to put the patch. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8077> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

Re: [GHC] #3588: ghc -M should emit dependencies on CPP headers
by GHC 21 Aug '13

21 Aug '13

#3588: ghc -M should emit dependencies on CPP headers -------------------------------------+------------------------------------ Reporter: simonmar | Owner: Type: bug | Status: new Priority: low | Milestone: 7.6.2 Component: Compiler | Version: 6.10.4 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: -------------------------------------+------------------------------------ Changes (by nh2): * cc: mail@… (added) * failure: => None/Unknown -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/3588#comment:7> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #7794: GHCi "Prelude.undefined" exceptions on ARM; ByteCodeItbls.mkJumpToAddr unimplemented
by GHC 21 Aug '13

21 Aug '13

#7794: GHCi "Prelude.undefined" exceptions on ARM; ByteCodeItbls.mkJumpToAddr unimplemented -------------------------------+--------------------------- Reporter: cjwatson | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.8.1 Component: GHCi | Version: 7.6.2 Resolution: | Keywords: Operating System: Linux | Architecture: arm Type of failure: GHCi crash | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: -------------------------------+--------------------------- Comment (by RoboTux): Inlining the call of the cacheflush system call (the call behind __clear_cache) should work but it means the cache would be flush every time mkJumpToAddr is called. I don't know enough the context of the code generation but if that function is called several time after 1 code generation then it's suboptimal. The patch then become: +#elif arm_TARGET_ARCH +type ItblCode = Word32 +mkJumpToAddr a + = [ 0xe92d0080 -- push {r7} + , 0xe3a0780f -- mov r7, #983040 ; 0xf0000 + , 0xe2877002 -- add r7, r7, #2 + , 0xe3a02000 -- mov r2, #0 + , 0xef000000 -- swi 0x0 + , 0xe8bd0080 -- pop {r7} + , 0xe51ff004 -- ldr pc, [pc, #-4] # pc reads as <current insn>+8 + , fromIntegral (ptrToInt a) ] -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/7794#comment:4> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #8145: can't build ghc on OS X 10.8 with XCODE 5 installed
by GHC 21 Aug '13

21 Aug '13

#8145: can't build ghc on OS X 10.8 with XCODE 5 installed ------------------------------------+------------------------------------- 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: | ------------------------------------+------------------------------------- see #8143 http://ghc.haskell.org/trac/ghc/ticket/8143 Oddly, I had this problem even when I was using gcc-4.8 to do the cpp when I had the Xcode 5 tools installed. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8145> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #8143: File name module name mismatch in Parser.hs in HEAD
by GHC 21 Aug '13

21 Aug '13

#8143: File name module name mismatch in Parser.hs in HEAD ------------------------------------+------------------------------------- Reporter: carter | 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: | ------------------------------------+------------------------------------- {{{ compiler/stage1/build/Parser.hs:1:1: File name does not match module name: Saw: `Main' Expected: `Parser' make[1]: *** [compiler/stage1/build/.depend-v.haskell] Error 1 make: *** [all] Error 2 }}} on current head c384bb1e30499bb4809dca60803a4066762ce5f4 and i tried a random recent commit 2e41f2fa91c833a4420ac273254e49468044bc4b -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8143> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #8138: Core lint fails with error "Role incompatibility: expected N, got R"
by GHC 20 Aug '13

20 Aug '13

#8138: Core lint fails with error "Role incompatibility: expected N, got R" ------------------------------------+------------------------------------- Reporter: gmainland | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: Other Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- I'm seeing a core lint error, "Role incompatibility: expected N, got R", on 333273792479695cf79a72cffa4c579d0c9deb21. The attached program will reproduce the error when compiled as: {{{ ghc -O2 -dcore-lint Role.hs }}} Note that `-O2` is required for core lint to complain. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8138> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

Re: [GHC] #5144: Pattern synonyms
by GHC 20 Aug '13

20 Aug '13

#5144: Pattern synonyms -------------------------------------+------------------------------------ Reporter: simonpj | Owner: cactus Type: feature request | Status: new Priority: normal | Milestone: _|_ Component: Compiler | Version: Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by cactus): Exports now work with the following syntax: {{{ module Foo (pattern P) where pattern P x y = (x, y, 42) }}} This exports `P` both as a pattern and as an expression (if it's bidirectional, like in the case here) The extra `pattern` keyword is needed because pattern names live in the constructor namespace, so it's perfectly valid to have {{{ data P = C pattern P = () }}} (due to entirely non-interesting reasons, the implementation of exporting is not yet pushed to https://github.com/gergoerdi/ghc/tree/pattern- synonyms) Remaining work to do: * Sort out recursive pattern definitions. Currently these are rejected implicitly by an internal GHC error, not exactly ideal... also, non- recursive usages of pattern synonyms mentioning each other don't work, but I see no reason why we wouldn't want to allow it: {{{ pattern P1 = P2 pattern P2 = () }}} * The typechecker for pattern synonym definitions is wrong: it returns monomorphic types (e.g. for `pattern Head x = x:_`, it infers type `t -> [t]` instead of `forall a. a -> [a]`. This is worked around by a horrible hack when typechecking pattern synonym usages in patterns. Richard Eisenberg and Simon Peyton Jones have already offered to help with this. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5144#comment:20> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #7958: 'Cannot continue after interface file error' during compilation
by GHC 20 Aug '13

20 Aug '13

#7958: 'Cannot continue after interface file error' during compilation ----------------------------------------------+---------------------------- Reporter: jstolarek | Owner: Type: bug | jstolarek Priority: normal | Status: new Component: Compiler | Milestone: Resolution: | Version: 7.7 Operating System: Unknown/Multiple | Keywords: Type of failure: GHC rejects valid program | Architecture: Test Case: | Unknown/Multiple Blocking: | Difficulty: Unknown | Blocked By: | Related Tickets: ----------------------------------------------+---------------------------- Changes (by jstolarek): * owner: => jstolarek -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/7958#comment:3> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0