February 2017 - ghc-tickets

[GHC] #13280: Consider deriving more Foldable methods
by GHC 18 Jun '18

18 Jun '18

#13280: Consider deriving more Foldable methods -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: (none) Type: task | Status: new Priority: normal | Milestone: 8.2.1 Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Runtime Unknown/Multiple | performance bug Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- We currently derive `foldMap` and `foldr`, but some others may deserve attention as well. * The most critical spots are probably `length` and `null`. Deriving these instead of using the defaults can change the orders of growth! Given `data Tree a = Bin !(Tree a) a !(Tree a) | Tip`, we surely don't want `null` to traverse the whole spine of the tree when it's quite immediately obvious that `Bin` is never null. And if a constructor contains a type with a very efficient `length` or `null` implementation (e.g., one that stores its own size), we certainly want to use that. * `foldl` typically ends up with rather different code than `foldr` (for a recursive type) even after simplification. We need to check whether this has a performance impact, but my bet is that it will in cases where the optimizer can't understand what's going on well enough. * `foldl'` and `foldr'` are a bit tricky. Ideally, if we have something like `data T a = C (G a) a | ...` then we'd like to be able to use `G`'s `foldl'` definition to define `T`'s. Unfortunately, different types in the wild have different `foldl'` semantics (and indeed the semantics for `[]` changed by mistake fairly recently). So we have to decide to what extent the derived semantics should depend on the choices of included types. I think we should probably just depend, because that seems likely what people will expect and want. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13280> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #9106: GHC Panic related to functional dependencies - kindFunResult
by GHC 18 Jun '18

18 Jun '18

#9106: GHC Panic related to functional dependencies - kindFunResult ----------------------------+--------------------------------------- Reporter: yuriy | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.2 Keywords: | Operating System: Windows Architecture: x86 | Type of failure: Compile-time crash Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ----------------------------+--------------------------------------- Minimal example: {{{ {-# LANGUAGE MultiParamTypeClasses, DataKinds, FunctionalDependencies, TypeOperators, KindSignatures, PolyKinds, FlexibleInstances, FlexibleContexts, UndecidableInstances #-} import GHC.TypeLits data Proxy (a :: k) = Proxy class FunctorN n f a fa | n f a -> fa where fmapn :: Proxy n -> Proxy f -> (a -> a) -> fa -> fa instance FunctorN 0 f a a where fmapn _ _ a = a instance (Functor f, FunctorN (n - 1) f a fa) => FunctorN n f a (f fa) where fmapn _ pf f = fmap (fmapn (Proxy :: Proxy (n-1)) pf f) }}} Crashes with ghc and ghci: {{{ >ghc test [1 of 1] Compiling Main ( test.hs, test.o ) ghc.exe: panic! (the 'impossible' happened) (GHC version 7.8.20140130 for x86_64-unknown-mingw32): kindFunResult k{tv azb} [sk] >ghci test GHCi, version 7.8.20140130: http://www.haskell.org/ghc/ :? for help Loading package ghc-prim ... linking ... done. Loading package integer-gmp ... linking ... done. Loading package base ... linking ... done. [1 of 1] Compiling Main ( test.hs, interpreted ) ghc.exe: panic! (the 'impossible' happened) (GHC version 7.8.20140130 for x86_64-unknown-mingw32): kindFunResult k{tv aPm} [sk] }}} Removing the functional dependency makes the code compile. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9106> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #8619: Support anonymous string literals in C-- (OR) give better ASSERT failure messages in C--
by GHC 17 Jun '18

17 Jun '18

#8619: Support anonymous string literals in C-- (OR) give better ASSERT failure messages in C-- ------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: low | 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: | ------------------------------------+------------------------------------- So if I say "foobar" in C--, it turns into: {{{ section "rodata" { fresh_name : bits8[] "foobar\0"; } }}} This would let us fix the ASSERT macro in Cmm.h to also get the filename. Requires parser changes. Alternatively, we can redefine the ASSERT macro per C-- file to have a hardcoded filename with some preprocessor trickery. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8619> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #12970: Add default implementation for Bits.bitSize
by GHC 17 Jun '18

17 Jun '18

#12970: Add default implementation for Bits.bitSize -------------------------------------+------------------------------------- Reporter: txnull | Owner: Type: feature | Status: new request | Priority: normal | Milestone: Component: | Version: libraries/base | 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 think it would be a good idea to provide a default implementation for bitSize, example: {{{#!hs bitSize b = fromMaybe (error "bitSize is undefined") (bitSizeMaybe b) }}} The advantage is that from now on new instances of Bits no longer need to define bitSize. And once bitSize has been removed there will be no errors about it not being a method of class Bits. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12970> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 23

[GHC] #12951: Remove __USE_MINGW_ANSI_STDIO define from PosixSource.h
by GHC 16 Jun '18

16 Jun '18

#12951: Remove __USE_MINGW_ANSI_STDIO define from PosixSource.h -------------------------------------+------------------------------------- Reporter: bgamari | Owner: Type: task | Status: new Priority: normal | Milestone: 8.4.1 Component: Compiler | 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: -------------------------------------+------------------------------------- This was a stop-gap measure to allow users to build GHC 8.2 with 7.10.1 and 7.10.2. It can be removed for 8.4. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12951> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #13008: Cleanup backwards compatibility ifdefs due to stage1 external interpreter work
by GHC 16 Jun '18

16 Jun '18

#13008: Cleanup backwards compatibility ifdefs due to stage1 external interpreter work -------------------------------------+------------------------------------- Reporter: shlevy | Owner: Type: task | Status: new Priority: normal | Milestone: 8.4.1 Component: Compiler | 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: -------------------------------------+------------------------------------- https://phabricator.haskell.org/D2826 and https://phabricator.haskell.org/D2884 make some code paths that were incompatible with GHC 7.10.x newly built by the stage0 compiler, requiring some CPP hackery to keep our compatibility guarantees. When we can drop that dependency, we should be sure to remove the old code paths. Specifically, the checks for MIN_VERSION_base(4,9,0) and MIN_VERSION_process(1,4,2) can be removed once we're able to assume ghc 8.0.x as stage0. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13008> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #12999: Investigate use of floating-point arithmetic in I/O Event-handler
by GHC 16 Jun '18

16 Jun '18

#12999: Investigate use of floating-point arithmetic in I/O Event-handler -------------------------------------+------------------------------------- Reporter: hvr | Owner: Type: task | Status: new Priority: normal | Milestone: 8.2.1 Component: | Version: 8.0.1 libraries/base | 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 was asked/made aware of some questionable use of floating-point arithmetic in `GHC.Event.TimerManager`; Specifically, I can't explain why `getMonotonicTime` first gets a timestamp as `Word64`, only to perform a not-necessarily cheap floating- point division, and returns a `Double`... {{{#!hs -- | Return monotonic time in seconds, since some unspecified starting point getMonotonicTime :: IO Double getMonotonicTime = do w <- getMonotonicNSec return (fromIntegral w / 1000000000) foreign import ccall unsafe "getMonotonicNSec" getMonotonicNSec :: IO Word64 }}} which then gets used e.g. by {{{#!hs updateTimeout :: TimerManager -> TimeoutKey -> Int -> IO () updateTimeout mgr (TK key) us = do now <- getMonotonicTime let expTime = fromIntegral us / 1000000.0 + now editTimeouts mgr (Q.adjust (const expTime) key) wakeManager mgr }}} Whereas, the priority queue (`GHC.Event.PSQ`) then uses `Double`s for its priorities... For correctness and performance reasons, I'd expect the use of fixed-width integers (i.e. `Word64` or `Int64`) for such codepaths... -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12999> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 7

[GHC] #11259: Use system runtime linker in GHCi on PowerPC 64 bit
by GHC 14 Jun '18

14 Jun '18

#11259: Use system runtime linker in GHCi on PowerPC 64 bit -------------------------------------+------------------------------------- Reporter: trommler | Owner: trommler Type: bug | Status: new Priority: normal | Milestone: 8.2.1 Component: GHCi | Version: 7.11 Keywords: | Operating System: Linux Architecture: powerpc64 | Type of failure: GHCi crash Test Case: ghcilink004, | Blocked By: prog001, and 11 more | Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Teach GHCi to use the system runtime linker to load Haskell libraries on powerpc64. See the following test failures. ghcilink004: {{{ ghc-stage2: dir004/libfoo.a: unknown architecture (e_machine == 21) ghc-stage2: panic! (the 'impossible' happened) (GHC version 7.11.20151219 for powerpc64-unknown-linux): loadArchive "dir004/libfoo.a": failed CallStack (from ImplicitParams): error, called at libraries/ghci/GHCi/ObjLink.hs:91:21 in ghci-0.0:GHCi.ObjLink Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} prog001: {{{ ghc-iserv.bin: /home/peter/ghc/libraries/ghc-prim/dist-install/build /libHSghc-prim-0.5.0.0.a: unknown architecture (e_machine == 21) ghc-iserv.bin: loadArchive "/home/peter/ghc/libraries/ghc-prim/dist- install/build/libHSghc-prim-0.5.0.0.a": failed CallStack (from ImplicitParams): error, called at libraries/ghci/GHCi/ObjLink.hs:91:21 in ghci-0.0:GHCi.ObjLink ghc-stage2: ghc-iserv terminated (1) }}} Using the system runtime linker fixes 13 failing tests. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11259> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #12903: SIGINT is not handled in forked process
by GHC 14 Jun '18

14 Jun '18

#12903: SIGINT is not handled in forked process -------------------------------------+------------------------------------- Reporter: qnikst | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Runtime | Version: 8.0.1 System | 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: -------------------------------------+------------------------------------- If haskell program is forked in a toplevel hander that sends UserInterrupt in the main thread doesn't work, because Thread 1 doesn't exists in forked program, so that action is noop. Minimal example: {{{#!hs import Control.Concurrent import Control.Exception import System.Posix main = do pid <- forkProcess $ do handle (\UserInterrupt{} -> putStrLn "caught") $ threadDelay 2000000 signalProcess sigINT pid threadDelay 2000000 }}} Expected result: "caught" printed to stdout. Actual result: nothing is printed. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12903> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #9539: TQueue can lead to thread starvation
by GHC 12 Jun '18

12 Jun '18

#9539: TQueue can lead to thread starvation -------------------------------------+------------------------------------- Reporter: jwlato | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: libraries (other) | Version: 7.8.2 Keywords: stm | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Unknown | Type of failure: Blocked By: | None/Unknown Related Tickets: | Test Case: | Blocking: | Differential Revisions: -------------------------------------+------------------------------------- For background, please see [http://stackoverflow.com/questions/25536604 /huge-memory-consumption-for-simple-multithreaded- haskell/25560047#25560047 this SO question] When using TQueue, a sufficiently fast/persistent writer can result in the reader never getting scheduled, causing the queue to continually grow and lost concurrency. I believe the issue is caused by the definition of readTQueue: {{{ readTQueue :: TQueue a -> STM a readTQueue (TQueue read write) = do xs <- readTVar read case xs of (x:xs') -> do writeTVar read xs' return x [] -> do ys <- readTVar write case ys of [] -> retry _ -> case reverse ys of [] -> error "readTQueue" (z:zs) -> do writeTVar write [] writeTVar read zs return z }}} Note that the last `case` needs to traverse the write-end of the queue within the STM transaction. If the list is sufficiently large, a writer can commit a new transaction much more quickly, invalidating the read transaction. If threads continue to write to the queue, the reader will never get an opportunity to commit (and the list will grow, exacerbating the problem). This alternative definition seems to fix the problem, but I don't know if there are other drawbacks to using it. {{{ readTQueue' :: TQueue a -> STM a readTQueue' (TQueue read write) = do xs <- readTVar read case xs of (x:xs') -> do writeTVar read xs' return x [] -> do ys <- readTVar write case ys of [] -> retry _ -> do writeTVar write [] let (z:zs) = reverse ys writeTVar read zs return z }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9539> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 23