May 2014 - ghc-tickets

[GHC] #9006: GHC accepts import of private data constructor if it has the same name as the type
by GHC 27 Oct '14

27 Oct '14

#9006: GHC accepts import of private data constructor if it has the same name as the type --------------------------+------------------------------------------------ Reporter: | Owner: Lemming | Status: new Type: bug | Milestone: Priority: normal | Version: 7.8.2 Component: | Operating System: Unknown/Multiple Compiler | Type of failure: GHC accepts invalid program Keywords: | Test Case: Architecture: | Blocking: Unknown/Multiple | Difficulty: | Unknown | Blocked By: | Related Tickets: | --------------------------+------------------------------------------------ {{{ $ cat A.hs B.hs module A (T) where data T = T module B where import A (T(T)) }}} `ghci-7.8.2` accepts module B, whereas `ghci-7.6.3` and `ghci-7.4.2` say: {{{ B.hs:3:11: Module `A' does not export `T(T)' }}} For this bug to happen it is important, that the data constructor has the same name as the type. I.e. if you rename the data constructor to, say, `Cons`, then `ghci-7.8.2` will emit the same error as earlier versions. The bug is not dramatic, since if you try to actually use the imported data constructor then ghc will say {{{ B.hs:20:5: Not in scope: data constructor ‘T’ }}} However this error message is very confusing, because it suggests that `T` is not in scope although you imported it correctly. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9006> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8345: A more efficient atomicModifyIORef'
by GHC 27 Oct '14

27 Oct '14

#8345: A more efficient atomicModifyIORef' ------------------------------------+------------------------------------- Reporter: parcs | Owner: Type: task | Status: new Priority: normal | Milestone: Component: libraries/base | 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: | ------------------------------------+------------------------------------- `atomicModifyIORef'` is currently defined as: {{{ #!haskell atomicModifyIORef' ref f = do b <- atomicModifyIORef ref (\x -> let (a, b) = f x in (a, a `seq` b)) b `seq` return b }}} This doesn't seem to be the best we can do, though. The following implementation exhibits a speedup of 1.7x (vanilla RTS) / 1.4x (threaded RTS) over the current implementation: {{{ #!haskell atomicModifyIORef' ref f = do b <- atomicModifyIORef ref $ \a -> case f a of v@(a',_) -> a' `seq` v b `seq` return b }}} The differences between this version and the current version are: 1. the lambda is now strict in `f a`, and 2. a new result tuple is no longer being allocated. Is there a good reason why `atomicModifyIORef'` is not already defined this way? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8345> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 18

[GHC] #9031: ghc panic: RegAlloc.Liveness.computeLivenss
by GHC 27 Oct '14

27 Oct '14

#9031: ghc panic: RegAlloc.Liveness.computeLivenss -----------------------------------+--------------------------------------- Reporter: jwlato | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.2 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Compile-time crash Unknown/Multiple | Test Case: Difficulty: Unknown | Blocking: Blocked By: | Related Tickets: | -----------------------------------+--------------------------------------- I have some code where ghc consistently panics with {{{ RegAlloc.Liveness.computeLivenss SCCs aren't in reverse dependent order bad blockId c1kek }}} The only reference I can find to that block in the dumped output is here: {{{ REC c1kej: movq $block{v c1keo}_info,(%rbp) movsd 7(%rbx),%vSSE_s1ivM movsd 15(%rbx),%vSSE_s1ivN movq 96(%rbp),%rbx movsd %vSSE_s1ivN,88(%rbp) movsd %vSSE_s1ivM,96(%rbp) testq $7,%rbx jne _c1keo jmp _c1kep c1kpo: movq $block{v c1kej}_info,(%rbp) movq %vI_s1it5,%rbx movq $0,120(%rbp) testq $7,%rbx jne _c1kej jmp _c1kek, }}} There doesn't appear to be any reference to {{{c1kpo}}} either. Unfortunately I don't yet have a test case I can provide; I'll keep working on that. Building with -dcore-lint and -dcmm-lint don't show any extra errors. This is with ghc-7.8.2 using NCG. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9031> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #9038: Foreign calls don't make their arguments look strict
by GHC 21 Oct '14

21 Oct '14

#9038: Foreign calls don't make their arguments look strict ------------------------------------+------------------------------------- Reporter: tibbe | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.2 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- Here's a function that takes a `Double` which is always passed to a `foreign import` expecting a `Double#`. Since the value is always unpacked, I'd expect `add3` to be strict in the `Double` argument and thereby avoid some allocation. {{{ add3 :: Distribution -> Double -> Int64 -> State# RealWorld -> (# State# RealWorld, () #) add3 = \ (distrib_a1uw :: Distribution) (val_a1ux :: Double) (n_a1uy :: Int64) (eta_X22 :: State# RealWorld) -> case myThreadId# eta_X22 of _ { (# ipv_i229, ipv1_i22a #) -> case threadStatus# ipv1_i22a ipv_i229 of _ { (# ipv2_i22e, _, ipv4_i22g, _ #) -> case modInt# ipv4_i22g 8 of ww1_i1NM { __DEFAULT -> case distrib_a1uw `cast` (<NTCo:Distribution> :: Distribution ~# Array Stripe) of _ { Array ds2_i22q -> case indexArray# ds2_i22q ww1_i1NM of _ { (# ipv6_i1Nz #) -> case ipv6_i1Nz of _ { Stripe ipv7_s22v ipv8_s22w ipv9_s22x ipv10_s22y -> case {__pkg_ccall main hs_lock Addr# -> State# RealWorld -> (# State# RealWorld #)}_i20J ipv9_s22x ipv2_i22e of _ { (# ds4_i20P #) -> case touch# ipv10_s22y ds4_i20P of s'_i20R { __DEFAULT -> case val_a1ux of _ { D# ds6_d1D8 -> case n_a1uy of _ { I64# ds8_d1Da -> case {__pkg_ccall main hs_distrib_add_n Addr# -> Double# -> Int# -> State# RealWorld -> (# State# RealWorld #)}_d1Dd ipv7_s22v ds6_d1D8 ds8_d1Da s'_i20R of _ { (# ds9_d1Db #) -> case {__pkg_ccall main hs_unlock Addr# -> State# RealWorld -> (# State# RealWorld #)}_i20m ipv9_s22x ds9_d1Db of _ { (# ds10_i20s #) -> case touch# ipv10_s22y ds10_i20s of s'1_i20u { __DEFAULT -> case touch# ipv8_s22w s'1_i20u of s'2_i1YT { __DEFAULT -> (# s'2_i1YT, () #) } } } } } } } } } } } } } } }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9038> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #8752: System.FilePath.Windows.combine does not really check isAbsolute
by GHC 19 Oct '14

19 Oct '14

#8752: System.FilePath.Windows.combine does not really check isAbsolute -------------------------------------+--------------------------------- Reporter: joeyhess | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: libraries (other) | Version: 7.6.3 Keywords: filepath | Operating System: Windows Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | -------------------------------------+--------------------------------- "Combine two paths, if the second path isAbsolute, then it returns the second.¨ But, the implementation of combine checks if the first character of a path is a path separator, which on Windows is not the same as checking if isAbsolute. This can have counterintuitive results. For example: {{{ import System.FilePath.Windows prop_windows_is_sane :: Bool prop_windows_is_sane = isAbsolute p || ("C:\\STUFF" </> p /= p) where p = "\\foo\\bar" }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8752> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8683: add "cpp program" and "cpp program options" to settings file
by GHC 14 Oct '14

14 Oct '14

#8683: add "cpp program" and "cpp program options" to settings file ------------------------------------+------------------------------------- Reporter: carter | Owner: Type: feature request | Status: new Priority: normal | Milestone: 7.8.1 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: | ------------------------------------+------------------------------------- so theres a pretty clear consensus that we need to move away from using our c compiler for CPP, so should we add that info to the settings file? I propose adding the following fields to the settings file: "use c compiler cpp", which can have either "YES" or "NO". If "YES", the settings file indicated C compiler is used unless the c compiler selection is overrided by ghc at runtime via -pgmc or -pgmp(p?) if "NO", then the "cpp program" and "cpp program options" fields would be used, unless overridden by -pgmp there'd be some associated patched in the compiler/main/DriverPipeline.hs too i think, but relatively minimal. this would make it much easier to distributed a GHC with its own CPP tool that works -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8683> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 69

[GHC] #8980: ghc-7.8.1 -O2 eats excessive amounts of RAM, highlighting-kate and pandoc-citeproc
by GHC 07 Oct '14

07 Oct '14

#8980: ghc-7.8.1 -O2 eats excessive amounts of RAM, highlighting-kate and pandoc- citeproc ------------------------------------+------------------------------------- Reporter: slyfox | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.1 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- It's mostly a http://www.haskell.org/pipermail/ghc- devs/2014-March/004363.html rechecked on ghc-7.8.1: I attempted to leave original packages as-is. Steps to reproduce: {{{ $ cd /tmp/ $ wget -c http://code.haskell.org/~slyfox/selfcontained-eater- ghc-7.8-rc2.tar.gz $ tar -xzf selfcontained-eater-ghc-7.8-rc2.tar.gz $ cd selfcontained-eater-ghc-7.8-rc2 $ ./mk.sh ... [279 of 452] Compiling Text.Highlighting.Kate.Syntax.Asp ( highlighting- kate-0.5.6.1/Text/Highlighting/Kate/Syntax/Asp.hs, highlighting- kate-0.5.6.1/Text/Highlighting/Kate/Syntax/Asp.o ) stack overflow: use +RTS -K<size> to increase it }}} I did prepare this testcase for ghc-7.6.3 where it failed only on last 452 module. '''./mk.sh''' is a dumb script calling '''ghc -O2''' + cabal-generated bits. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8980> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8852: 7.8.1 uses a lot of memory when compiling attoparsec programs using <|>
by GHC 07 Oct '14

07 Oct '14

#8852: 7.8.1 uses a lot of memory when compiling attoparsec programs using <|> -------------------------+------------------------------------------------- Reporter: | Owner: joelteon | Status: new Type: bug | Milestone: Priority: | Version: 7.8.1-rc2 normal | Operating System: Unknown/Multiple Component: | Type of failure: Compile-time performance bug Compiler | Test Case: Keywords: | Blocking: Architecture: | Unknown/Multiple | Difficulty: | Unknown | Blocked By: | Related Tickets: | -------------------------+------------------------------------------------- To reproduce, install a pre-`0.11.2.1` version of attoparsec. This bug was worked around in 0.11.2.1 by removing the `INLINE` on `plus` in attoparsec. With this test program: {{{#!haskell {-# LANGUAGE OverloadedStrings #-} import Control.Applicative import Data.Attoparsec.Text import Data.Text (Text) parser :: Parser Text parser = string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" <|> string "a" main :: IO () main = parseTest parser "a" }}} and using GHC 7.8.1.rc2: Compiling using `-O2`, GHC tops out at ~1GB of RAM and takes 25s. Using `-O0`, GHC takes 0.47s and uses <150MB of RAM. Compare this with GHC 7.6.3: Compiling using `-O2`, GHC uses <150MB and takes 3.7s. Memory usage is similar with `-O0` although compile time goes down to 0.36s. An extreme version of this bug can be found in the `thyme` package here: https://github.com/liyang/thyme/blob/master/src/Data/Thyme/Format.hs#L589-L…. Compiling that module with an unfixed attoparsec makes GHC use all available memory and stall out, forcing kill -9. Replacing the function body with `undefined` makes the package compile as expected. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8852> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 10

[GHC] #8790: rts: unrust 'libbfd' debug symbols parser
by GHC 05 Oct '14

05 Oct '14

#8790: rts: unrust 'libbfd' debug symbols parser ------------------------------------+------------------------------------- Reporter: slyfox | 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: | ------------------------------------+------------------------------------- For quite a long while you might have seen such ./configure output: {{{ checking bfd.h usability... yes checking bfd.h presence... yes checking for bfd.h... yes ... checking for bfd_uncompress_section_contents in -lbfd... no }}} which means very interesting debug bit in rts (+RTS -Di) is disabled. It's because most of functions in bfd.h are actually macros, thus test fails: config.log: {{{ configure:9545: checking for bfd_uncompress_section_contents in -lbfd configure:9570: x86_64-pc-linux-gnu-gcc -o conftest -march=corei7-avx -mtune=corei7-avx -O2 -pipe -fdiagnostics-show-option -frecord-gcc- switches -ggdb3 -march=corei7-avx -march=corei7-avx -m tune=corei7-avx -mtune=corei7-avx -ggdb3 -fno-stack-protector -Wa,-- noexecstack -Wa,--noexecstack -fno-stack-protector -Wl,-O1 -Wl,--as- needed -Wl,--hash-style=gnu -Wl,--hash-size=31 -Wl,--r educe-memory-overheads conftest.c -lbfd -liberty >&5 /tmp/portage/dev-lang/ghc-7.6.3-r1/temp/ccCEvDZ7.o: In function `main': /tmp/portage/dev-lang/ghc-7.6.3-r1/work/ghc-7.6.3/conftest.c:99: undefined reference to `bfd_uncompress_section_contents' }}} We don't use that function anyways, thus i've picked '''bfd_init'''. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8790> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 7

[GHC] #8374: `tcIfaceGlobal (local): not found` while compiling
by GHC 29 Sep '14

29 Sep '14

#8374: `tcIfaceGlobal (local): not found` while compiling ----------------------------+---------------------------------------- Reporter: bgamari | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: arm | Type of failure: Building GHC failed Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ----------------------------+---------------------------------------- On commit fa3ffb43144eadc406031110b01ba3dc4f9bd94e on compiling on ARM, ``` "rm" -f libraries/integer-gmp/dist-install/build/libHSinteger- gmp-0.5.1.0.a.contents "/usr/bin/ld" -r -o libraries/integer-gmp/dist-install/build/HSinteger- gmp-0.5.1.0.o libraries/integer-gmp/dist-install/build/GHC/Integer.o libraries/integer-gmp/dist-install/build/GHC/Integer/GMP/Internals.o libraries/integer-gmp/dist-install/build/GHC/Integer/GMP/Prim.o libraries /integer-gmp/dist-install/build/GHC/Integer/Logarithms.o libraries /integer-gmp/dist-install/build/GHC/Integer/Logarithms/Internals.o libraries/integer-gmp/dist-install/build/GHC/Integer/Type.o libraries /integer-gmp/dist-install/build/cbits/gmp-wrappers.o libraries/integer- gmp/dist-install/build/cbits/cbits.o ghc-stage1: panic! (the 'impossible' happened) (GHC version 7.7.20130924 for arm-unknown-linux): tcIfaceGlobal (local): not found: base:GHC.Base.$fMonadIO{v r1} [(02K, Identifier ‛base:GHC.Base.fail{v 02K}’), (02L, Identifier ‛base:GHC.Base.>>={v 02L}’), (02M, Identifier ‛base:GHC.Base.>>{v 02M}’), (02N, Identifier ‛base:GHC.Base.fmap{v 02N}’), (02O, Identifier ‛base:GHC.Base.return{v 02O}’), (28, Class ‛base:GHC.Base.Monad{tc 28}’), (2a, Class ‛base:GHC.Base.Functor{tc 2a}’), (36r, Type constructor ‛base:GHC.Base.Opaque{tc 36r}’), (36u, Type constructor ‛base:GHC.Base.String{tc 36u}’), (rB, Data constructor ‛base:GHC.Base.O{d rB}’), (rD, Identifier ‛base:GHC.Base.<${v rD}’), (rhP, Identifier ‛base:GHC.Base.O{v rhP}’), (ri5, Data constructor ‛base:GHC.Base.D:Functor{d ri5}’), (rit, Identifier ‛base:GHC.Base.$dm<${v rit}’), (riv, Identifier ‛base:GHC.Base.D:Functor{v riv}’), (riX, Data constructor ‛base:GHC.Base.D:Monad{d riX}’), (rjw, Identifier ‛base:GHC.Base.$dm>>{v rjw}’), (rjx, Identifier ‛base:GHC.Base.$dmfail{v rjx}’), (rjz, Identifier ‛base:GHC.Base.D:Monad{v rjz}’)] ``` -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8374> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 36