April 2015 - ghc-tickets

[GHC] #9249: Link to "latest" user's guide
by GHC 01 Nov '17

01 Nov '17

#9249: Link to "latest" user's guide ------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: bug | Status: new Priority: low | Milestone: Component: Documentation | 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: | ------------------------------------+------------------------------------- When I search (say, on Google) for a GHC-related bit of documentation, I variously get referred to old manuals. There doesn't seem to a good way for us, the GHC devs, to avoid this. But, perhaps we can include some information in the header of the pages to allow visitors to link to the latest manual, or perhaps to several versions. This header should also include the version number of the GHC that the page is documenting; currently, this info is only in the URL. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9249> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #9112: data families with representational matching
by GHC 17 Oct '17

17 Oct '17

#9112: data families with representational matching ------------------------------------+------------------------------------- Reporter: jwlato | Owner: Type: feature request | 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: | ------------------------------------+------------------------------------- Since ghc-7.8, the following is no longer possible: {{{ -- simplified example taken from the vector package class MVectorClass (v :: * -> * -> *) a where basicLength :: v s a -> Int data family MVector s a data instance MVector s Int -- implementation not important newtype Age = Age Int deriving (MVectorClass MVector) -- rejected }}} Following from discussion in #8177, to enable this ghc would need to support data families with representational matching, such that `MVector s Int` and `MVector s Age` are representationally equal. This has broken some code that previously worked, however as there are some workarounds I'm not sure how important it is. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9112> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 14

[GHC] #9274: GHC panic with UnliftedFFITypes+CApiFFI
by GHC 16 Oct '17

16 Oct '17

#9274: GHC panic with UnliftedFFITypes+CApiFFI -----------------------------------+--------------------------------------- Reporter: hvr | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.10.1 Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Compile-time crash Unknown/Multiple | Test Case: Difficulty: Unknown | Blocking: Blocked By: | Related Tickets: | -----------------------------------+--------------------------------------- The following code fails to compile with at least GHC 7.6, 7.8 and HEAD: {{{#!hs import GHC.Prim import Foreign.C type GmpLimb = CULong type GmpSize = CLong -- mp_limb_t mpn_add_1 (mp_limb_t *rp, const mp_limb_t *s1p, mp_size_t n, mp_limb_t s2limb) foreign import capi unsafe "gmp.h mpn_add_1" c_mpn_add_1 :: MutableByteArray# s -> ByteArray# -> GmpSize -> GmpLimb -> IO GmpLimb }}} with the following compile error message: {{{ ghc: panic! (the 'impossible' happened) (GHC version 7.9.20140704 for x86_64-unknown-linux): toCType MutableByteArray# s Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} Otoh, replacing `capi` with `ccall` makes the code compile just fine -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9274> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 7

[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] #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] #10229: setThreadAffinity assumes a certain CPU virtual core layout
by GHC 30 Sep '17

30 Sep '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] #9898: Couldn't match type `(Char, ())' with `()'
by GHC 27 Sep '17

27 Sep '17

#9898: Couldn't match type `(Char, ())' with `()' -------------------------------------+------------------------------------- Reporter: erisco | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler (Type | Version: 7.8.3 checker) | Operating System: Windows Keywords: | Type of failure: GHC Architecture: x86_64 (amd64) | rejects valid program Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Revisions: -------------------------------------+------------------------------------- 'test3' is expected to type check given that 'test1' and 'test2' type check. 'test4' is not expected to type check. {{{ ghc_bug2.hs:24:9: Couldn't match type `(Char, ())' with `()' Expected type: Filter (Equal Int) (Char, Filter (Equal Int) ()) Actual type: () In the expression: () :: Filter (Equal Int) (Char, Filter (Equal Int) ()) In an equation for `test2': test2 = () :: Filter (Equal Int) (Char, Filter (Equal Int) ()) ghc_bug2.hs:28:9: Couldn't match type `(Char, ())' with `()' Expected type: Filter (Equal Int) (Char, ()) Actual type: () In the expression: () In an equation for `test3': test3 = () }}} {{{#!hs {-# LANGUAGE TypeFamilies, UndecidableInstances, DataKinds #-} module Main where type family Filter f xs where Filter f (x, xs) = ApplyFilter (f x) (x, Filter f xs) Filter f () = () -- type family ApplyFilter p xs where ApplyFilter False (x, xs) = xs ApplyFilter p xs = xs -- type family Equal x y where Equal x x = True Equal x y = False -- -- Type checks test1 :: ApplyFilter ((Equal Int) Char) (Char, Filter (Equal Int) ()) test1 = () -- Couldn't match type `(Char, ())' with `()' test2 = () :: Filter (Equal Int) (Char, Filter (Equal Int) ()) -- Couldn't match type `(Char, ())' with `()' test3 :: Filter (Equal Int) (Char, ()) test3 = () -- Type checks, should not test4 :: Filter (Equal Int) (Char, ()) test4 = ('x', ()) }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9898> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #8919: Why is xhtml library installed but not exported to users?
by GHC 20 Sep '17

20 Sep '17

#8919: Why is xhtml library installed but not exported to users? ------------------------------------+------------------------------------- Reporter: simons | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.8.1 Component: Build System | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: Other Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- GHC 7.8.1-rc2 installs the xhtml library, but doesn't export it, i.e. it's completely invisible to users. Is there a particular reason why this setup is necessary? If GHC needs that library internally anywhere, would it be possible to link it statically and *not* install it? Or, if that's not possible, if the library needs to be installed, please make it visible to users. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8919> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 27

[GHC] #9281: Rewrite `integer-gmp` to use only non-allocating GMP functions
by GHC 19 Sep '17

19 Sep '17

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: new Priority: normal | Milestone: 7.10.1 Component: libraries (other) | Version: Keywords: integer-gmp | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | -------------------------------------+------------------------------------- After some promising results with a proof-of-concept implementation, I'm optimistic we can rewrite `integer-gmp` to use only non-allocating GMP lib functions without suffering from serious regressions. If successful, this would - allow to avoid the custom GMP allocator hack, and thus - avoid issues when linking against other C libraries using GMP, - simplify code, as we would perform all heap allocations in Haskell code (and never inside Cmm/C code as its done now), - and finally maybe even remove a few more superfluous temporary heap allocations. This rewrite can be done incrementally. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9281> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 32

[GHC] #9157: cmm common block not eliminated
by GHC 19 Sep '17

19 Sep '17

#9157: cmm common block not eliminated ------------------------------------+------------------------------------- Reporter: wojteknar | 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: Other Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- For the toTuple# function (in the attached file) GHC generates a cmm switch statement, with all the alternatives, unsurprisingly, the same. Yet, cmm common block elimination does not kick in. In this particular example, the whole case statement could be annihilated, because all alternatives lead to the same code. BTW, this is how they match in France. {{{ #!ocaml type alt = A of int | B of int | C of int | D of int let g x = match x with A v | B v | C v | D v -> v }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9157> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 12