December 2016 - ghc-tickets

[GHC] #11959: Importing doubly exported pattern synonym and associated pattern synonym panics
by GHC 04 Oct '17

04 Oct '17

#11959: Importing doubly exported pattern synonym and associated pattern synonym panics -------------------------------------+------------------------------------- Reporter: darchon | Owner: Type: bug | Status: new Priority: normal | Milestone: 8.0.1 Component: Compiler | Version: 8.0.1-rc3 Keywords: | Operating System: Unknown/Multiple PatternSynonyms | Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Given {{{#!hs {-# LANGUAGE PatternSynonyms, ViewPatterns #-} module Pat (Vec2(Nil,(:>)), pattern (:>)) where newtype Vec2 a = Vec2 {unvec2 :: [a]} pattern Nil :: Vec2 a pattern Nil = Vec2 [] pattern (:>) x xs <- ((\ys -> (head $ unvec2 ys,Vec2 . tail $ unvec2 ys)) -> (x,xs)) where (:>) x xs = Vec2 (x:unvec2 xs) }}} and {{{#!hs {-# LANGUAGE PatternSynonyms #-} module Main where import Pat (Vec2(..),pattern (:>)) }}} I get: {{{ $ ghci Main.hs GHCi, version 8.0.0.20160411: http://www.haskell.org/ghc/ :? for help [1 of 2] Compiling Pat ( Pat.hs, interpreted ) Pat.hs:2:29: warning: [-Wduplicate-exports] ‘:>’ is exported by ‘(:>)’ and ‘Vec2(Nil, type (:>))’ [2 of 2] Compiling Main ( Main.hs, interpreted ) ghc: panic! (the 'impossible' happened) (GHC version 8.0.0.20160411 for x86_64-unknown-linux): filterImports/combine (:>, :>, Nothing) (:>, Vec2{Vec2, :>, Nil}, Nothing) Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} Now, the warning about the duplicate export is of course correct. But the panic shouldn't happen when I try to import both the associated pattern synonym and the normal pattern synonym. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11959> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 17

[GHC] #12747: INLINE vs NOINLINE vs <nothing> give three different results; two would be better
by GHC 04 Oct '17

04 Oct '17

#12747: INLINE vs NOINLINE vs <nothing> give three different results; two would be better -------------------------------------+------------------------------------- Reporter: | Owner: MikolajKonarski | Type: feature | Status: new request | Priority: normal | Milestone: 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: -------------------------------------+------------------------------------- INLINE vs NOINLE vs <nothing> for this function https://github.com/LambdaHack/LambdaHack/blob/master/Game/LambdaHack/Client… produces distinct figures in all three cases. See https://ghc.haskell.org/trac/ghc/ticket/12603?replyto=8#comment:7 and some later comments, but I'd like to move here the discussion of the specific feature request that <nothing> should at any given point be equivalent to either INLINE or NOINLINE. Three different results are very unwieldy while optimizing code, because one cannot tweak the INLINE pragma for functions in isolation, via fixing INLINE or NOINLINE for all other functions and varying only one, because the optimal setting for some functions may be <nothing>, and it naturally varies (as it should; that is fine; but in each particular codebase/compilation flags state it should be equal to either of INLINE or NOINLINE). It seems almost as if GHC considers a function for inlining in some stages, optimizes accordingly, but changes its mind in other places, not being bound by either INLINE nor NOINLINE, and so the resulting stack of optimizations is different than in either case. IMHO that's too complex behaviour for the programmer to follow, it should irreversibly decide whether to inline at exactly the same place it first takes into account the NO/INLINE pragmas, if present. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12747> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #9725: Constraint deduction failure
by GHC 03 Oct '17

03 Oct '17

#9725: Constraint deduction failure -------------------------------------+------------------------------------- Reporter: heisenbug | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler (Type | Version: 7.8.3 checker) | Operating System: Keywords: | Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: GHC Difficulty: Unknown | rejects valid program Blocked By: | Test Case: Related Tickets: | Blocking: | Differential Revisions: -------------------------------------+------------------------------------- I get a strange error message like this {{{ [1 of 1] Compiling Main ( 08-10.hs, interpreted ) 08-10.hs:254:33: Could not deduce (KnownLoc (ent par1 a)) arising from a use of ‘Hidden'’ from the context (List ConfFamily ts) bound by the type signature for locAware :: List ConfFamily ts => (forall (i :: k). KnownLoc (ent par i) => ConfFamily (Facility (ent par i)) ts) -> Con ConfFamily (On ent par) at 08-10.hs:251:13-160 or from (ent par a ~ 'Monitor par1 mp, KnownLoc ('Monitor par1 mp)) bound by a pattern with constructor MonitorF :: forall (par :: Entity) (mp :: MonitorPoint). KnownLoc ('Monitor par mp) => String -> String -> Observer -> Bool -> Facility ('Monitor par mp), in an equation for ‘h’ at 08-10.hs:254:19-28 In the expression: Hidden' (d f) In an equation for ‘h’: h (Hide f@(MonitorF {})) = Hidden' (d f) In an equation for ‘locAware’: locAware descr = Abstr h where h :: (ent `On` par) -> ConfFamily (ent `On` par) ts h (Hide f@(MonitorF {})) = Hidden' (d f) d :: KnownLoc (ent par i) => Facility (ent par i) -> ConfFamily (Facility (ent par i)) ts d _ = descr Failed, modules loaded: none. }}} I have many language extensions active, such as, `DataKinds`, `ScopedTypeVariables`, etc. If somebody says this indeed looks like a bug, then I'll try to post a minimal reproduction case. ** My reasoning goes like this ** Here are the relevant facts that GHC figured out from a GADT pattern match: `(ent par a ~ 'Monitor par1 mp, KnownLoc ('Monitor par1 mp))` From this I discover these equalities: {{{ ent === 'Monitor (*) par === par1 (**) a === mp (***) }}} Substituring `(*)`^-1^ and `(***)`^-1^ into the given `KnownLoc ('Monitor par1 mp)` results in the desired `KnownLoc (ent par1 a)` Looks like a bug, right? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9725> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 21

[GHC] #12759: Latest Debian GCC breajs GHC
by GHC 02 Oct '17

02 Oct '17

#12759: Latest Debian GCC breajs GHC ----------------------------------------+-------------------------- Reporter: erikd | Owner: Type: bug | Status: new Priority: normal | Milestone: 8.0.2 Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Linux Architecture: Unknown/Multiple | Type of failure: Other Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: ----------------------------------------+-------------------------- Just got `gcc (Debian 6.2.0-9) 6.2.0 20161019` in a recent upgrade (Debian Testing) and now: {{{ > cat hello.hs main :: IO () main = putStrLn "Hello" > ghc hello.hs -o hello [1 of 1] Compiling Main ( hello.hs, hello.o ) Linking hello ... /usr/bin/ld: hello.o: relocation R_X86_64_32S against symbol `stg_bh_upd_frame_info' \ can not be used when making a shared object; recompile with -fPIC /usr/bin/ld: /usr/lib/ghc-8.0/lib/base-4.9.0.0/libHSbase-4.9.0.0.a(Base__166.o): \ relocation R_X86_64_32S against symbol `stg_bh_upd_frame_info' can not be used when \ making a shared object; recompile with -fPIC .... }}} The previous version of gcc in Debian Testing (gcc (Debian 6.2.0-6) 6.2.0 20161010) is fine. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12759> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 27

[GHC] #9198: large performance regression in type checker speed in 7.8
by GHC 02 Oct '17

02 Oct '17

#9198: large performance regression in type checker speed in 7.8 -------------------------------------+------------------------------------- Reporter: carter | Owner: Type: bug | Status: new Priority: high | Milestone: Component: Compiler (Type | Version: 7.8.2 checker) | Operating System: Unknown/Multiple Keywords: | Type of failure: Compile-time Architecture: Unknown/Multiple | performance bug Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | -------------------------------------+------------------------------------- it was recently demonstrated on the haskell-cafe mailing list that the following code takes measurably longer to type check in GHC 7.8.2 than in GHC 7.6. http://www.haskell.org/pipermail/haskell-cafe/2014-June/114562.html the program example is as follows {{{ begin cont = cont (return ()) a m =cont (m >> putstrn "a") end m = m main = begin a a a a a a a a a a a a a a a a a end }}} takes about a second to type check in ghc 7.8, and is "instant" in 7.6 () each additional a doubles the time to type check the program {{{ main = begin a a a a a a a a a a a a a a a a a a a a a a a a end }}} takes longer than I have the patience to wait :) (so more than 5 seconds) the author of the email notes that a related program {{{ main = id id id id id id id id id id id id id id id id id id id id id id (return ()) }}} has the exponential complexity in both 7.8.2 and 7.6.2 This It could very well be that some of the type checker changes between 7.8 and 7.6 enlarged the space of programs that trip up exponential worst case behavior, but one way or another understanding *why* this happened -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9198> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 10

[GHC] #11771: ghc.exe: `panic'! (the 'impossible' happened); thread blocked indefinitely in an MVar operation
by GHC 02 Oct '17

02 Oct '17

#11771: ghc.exe: `panic'! (the 'impossible' happened); thread blocked indefinitely in an MVar operation ----------------------------------------+------------------------------- Reporter: YoYoYonnY | Owner: Type: bug | Status: new Priority: low | Milestone: Component: GHCi | Version: 7.10.3 Keywords: | Operating System: Windows Architecture: Unknown/Multiple | Type of failure: GHCi crash Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: ----------------------------------------+------------------------------- Here's a "fun" one: {{{#!hs *Main> main Interrupted. *Main> *Main> !cl ghc.exe: panic! (the 'impossible' happened) (GHC version 7.10.3 for x86_64-unknown-mingw32): thread blocked indefinitely in an MVar operation Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug > > ghci --version The Glorious Glasgow Haskell Compilation System, version 7.10.3 }}} Don't mind the {{{!cl}}}, GHC crashed as soon as the interupt was registered, I usually {{{:!cls}}} after I execute main. I will backup the code in case someone wants to see if it's a bug in my program, but I didn't use multithreading myself, and I don't think the vector, unordered-containers, bytestring and parsec packages use threading either (all heavily at work, though)... The interupt was nececairy because my code was generating a infinite (oops) collection of these bad boys, using all sorts of folds, maps, etc.: {{{#!hs newtype ParserData = ParserData { unpackParserData :: HashMap ByteString (ParserData -> Either ByteString (Parser ())) } }}} I am wondering if this could be a problem with Windows, because the system was running slow at the time of the interupt; The {{{Alt+Tab}}} screen didn't appear until after the interupt. I don't think I can manage to reproduce this bug, which is unfortunate aswell (low priority because of that :c). Either way, it crashed, that's all I can tell you for sure. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11771> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #11963: GHC introduces kind equality without TypeInType
by GHC 02 Oct '17

02 Oct '17

#11963: GHC introduces kind equality without TypeInType -------------------------------------+------------------------------------- Reporter: ezyang | Owner: goldfire Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1-rc2 (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: -------------------------------------+------------------------------------- The following program is accepted by GHC 8.0 rc2 {{{ {-# LANGUAGE GADTs, PolyKinds, RankNTypes #-} data Typ k t where Typ :: (forall (a :: k -> *). a t -> a t) -> Typ k t }}} This probably shouldn't be accepted, because this is a circuitous way of saying: {{{ {-# LANGUAGE TypeInType #-} data Typ k (t :: k) = Typ }}} which does not work without `TypeInType`. Or perhaps both should be accepted without `TypeInType`? Printing with explicit kinds makes it clear why the obvious check didn't fire: {{{ ezyang@sabre:~$ ghc-8.0 --interactive B.hs -fprint-explicit-kinds GHCi, version 8.0.0.20160204: http://www.haskell.org/ghc/ :? for help [1 of 1] Compiling Main ( B.hs, interpreted ) Ok, modules loaded: Main. *Main> :info Typ type role Typ nominal nominal nominal data Typ k k1 (t :: k) where Typ :: forall k (t :: k). (forall (a :: k -> *). a t -> a t) -> Typ k k t -- Defined at B.hs:2:1 }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11963> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #10229: setThreadAffinity assumes a certain CPU virtual core layout
by GHC 01 Oct '17

01 Oct '17

#10229: setThreadAffinity assumes a certain CPU virtual core layout -------------------------------------+------------------------------------- Reporter: nh2 | Owner: simonmar Type: bug | Status: new Priority: normal | Milestone: Component: Runtime | Version: 7.10.1 System | Operating System: Unknown/Multiple Keywords: | Type of failure: Runtime Architecture: | performance bug Unknown/Multiple | Blocked By: Test Case: | Related Tickets: Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- The {{{RTS -qa}}} option that can set thread affinity was implemented in https://git.haskell.org/ghc.git/commitdiff/31caec794c3978d55d79f715f21fb729… {{{ // Schedules the thread to run on CPU n of m. m may be less than the // number of physical CPUs, in which case, the thread will be allowed // to run on CPU n, n+m, n+2m etc. void setThreadAffinity (nat n, nat m) }}} Today I discovered that on some machines, this option helps parallel performance (e.g. {{{+RTS -N4}}}) a lot, while on others it doesn't. Together with thomie on #ghc, I found out the reason: Lets assume I have 4 real cores with hyperthreading, so 8 virtual cores. The mapping of hyperthreading cores to physical cores is different across machines. On my one machine (Intel i5), the layout is 11223344, meaning that the first two vCPUs (hyperthreads) that the OS announces (visible e.g. in HTOP) map to the first physical core in the system, and so on. On my other machine (Intel Xeon), the layout is 12341234; here the 1st and the 5th vCPU map to the same physical core. This layout can be (on Linux) observed by running: {{{ cat /proc/cpuinfo|egrep "processor|physical id|core id" |sed 's/^processor/\nprocessor/g' }}} I do not know whether this layout is dictated by the processor, chosen by the OS, or even changing across reboots; what is clear is that the layout can vary across machines. Now, as explained by thomie: {{{ -qa will set your 4 capabilities to cores [(1,5), (2,6), (3,7), (4,8)], and then the os randomly chooses out of those tuples }}} This strategy is optimal for the 12341234 layout; for example, when running with -N4, it ensures that two threads are not scheduled onto vCPUs that are on the same physical core. The possible {{{+RTS -aq}}} choice {{{1__4_23_}}} is a great assignment in this case, as is {{{1234____}}} ({{{_}}} means the vCPU is not chosen). But for the 11223344, the choice {{{1234____}}} isn't good, because it uses only 2 of our 4 physical cores; our program now takes twice as long to run. ---- It seems likely to me that {{{setThreadAffinity}}} was written on a machine with 12341234 layout, and with the assumption that all machines have this layout. It would be great if we could change it to take the actual layout into account. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10229> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #10941: Large heap address space breaks valgrind
by GHC 29 Sep '17

29 Sep '17

#10941: Large heap address space breaks valgrind -------------------------------------+------------------------------------- Reporter: nomeata | Owner: Type: task | Status: new Priority: normal | Milestone: Component: Runtime | Version: 7.11 System | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: #9706 Differential Rev(s): | -------------------------------------+------------------------------------- It seems that programs compiled with a GHC `--enable-large-address-space` (see #9706) no longer run under valgrind, which is a pitty, as that is useful to get more reliable numbers out of nofib. It fails with `internal error: getMBlock: mmap: Invalid argument` This bug is a general “Can we do anything about this”. In particular, is there a possibility to make this a RTS option instead of a compile time option? Or maybe it is possible to talk to the valgrind authors if they can support such large allocations? If neither, can we maybe extend the error message to say “If you are using a tool like valgrind, you might want to recompile the compiler with `--disable-large-address-space.” -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10941> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #11228: Interaction between ORF and record pattern synonyms needs to be resolved.
by GHC 29 Sep '17

29 Sep '17

#11228: Interaction between ORF and record pattern synonyms needs to be resolved. -------------------------------------+------------------------------------- Reporter: mpickering | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 Keywords: | Operating System: Unknown/Multiple PatternSynonyms, orf | Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Currently the two extensions do not work well together. When defining a record pattern synonym all field names must be distinct even with the extension turned on. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11228> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 7