April 2018 - ghc-tickets

[GHC] #14502: Build Alpine Linux binary distributions
by GHC 19 Jun '18

19 Jun '18

#14502: Build Alpine Linux binary distributions -------------------------------------+------------------------------------- Reporter: bgamari | Owner: bgamari Type: task | Status: new Priority: normal | Milestone: 8.4.1 Component: Continuous | Version: 8.2.1 Integration | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: 14057 Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- At least three people have come to me asking for Alpine Linux bindists. Given that Alpine is a fair bit different from most other distributions and can be useful in containers, this seems fairly reasonable to me. Terrorjack has some useful-looking automation at https://github.com/TerrorJack/meikyu. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14502> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #15043: Expand type synonym
by GHC 18 Jun '18

18 Jun '18

#15043: Expand type synonym -------------------------------------+------------------------------------- Reporter: domenkozar | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Compiler | Version: 8.2.2 (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: -------------------------------------+------------------------------------- Using GHC 8.2.2 and `stack build --ghc-options=-fprint-expanded-synonyms` I have the following type synonym for servant: {{{#!hs type Get302 (cts :: [*]) (hs :: [*]) = Verb 'GET 302 cts (Headers (Header "Location" String ': hs) NoContent) }}} and get the following error message when `String` is mismatched for `Text`: {{{ • Couldn't match type ‘Text’ with ‘[Char]’ Expected type: AsServerT App :- ("login" :> ("callback" :> (QueryParam "code" Text :> (QueryParam "state" Text :> MyApp.Types.Servant.Get302 '[PlainText] '[Header "Set-Cookie" SetCookie, Header "Set-Cookie" SetCookie])))) Actual type: Maybe Code -> Maybe Text -> App (Headers '[Header "Location" Text, Header "Set- Cookie" SetCookie, Header "Set-Cookie" SetCookie] NoContent) }}} The error is confusing as type synonym is not expanded and offending `Header` is missing from the output. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15043> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #14283: Remove the special case for tagToEnum# in the code generator?
by GHC 18 Jun '18

18 Jun '18

#14283: Remove the special case for tagToEnum# in the code generator? -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: (none) Type: task | Status: new Priority: normal | Milestone: 8.4.1 Component: Compiler | Version: 8.2.1 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: -------------------------------------+------------------------------------- The only remaining primop (aside from `unsafeCoerce#` and `seq`) that produces an enumeration type is `tagToEnum#`. There is some magic in the code generator to make garbage collection relating to this work as it has historically. We want to get rid of the special case. Simply removing it altogether actually does seem to work, but the code we generate likely isn't quite the same. The challenge here is that (despite what I thought earlier) we definitely ''can'' end up in code generation, under certain circumstances, with {{{#!hs case tagToEnum# @t x of ... }}} How does this happen? After all, in `caseRules` we carefully remove every case on `tagToEnum#` and `dataToTag#` applications! The trouble comes when CorePrep puts everything in A-normal form. Strict function applications are transformed into `case` forms, so {{{#!hs f (tagToEnum# x) }}} will become {{{#!hs case tagToEnum# x of y DEFAULT -> f y }}} Suddenly we have that ugly case! Hmph. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14283> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #13944: Introduce synchronized FFI
by GHC 18 Jun '18

18 Jun '18

#13944: Introduce synchronized FFI -------------------------------------+------------------------------------- Reporter: winter | Owner: (none) Type: feature | Status: new request | 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: -------------------------------------+------------------------------------- During hacking on ghc I/O libraries, i find an [https://github.com/winterland1989/stdio/tree/master/bench/diskIO interesting situation], to summarize: There're some FFI calls that vary a lot in time duration, this makes unsafe FFI bad because it make GC hard to synchronize all capacities. OTOH safe FFI 's overhead is too large to justify its usage. After some though i think we should Introduce a new FFI type, let call it `synchronized` for now, it should: 1. Have the same requirement as safe FFI of what kind of argument we can pass, e.g. pinned ByteArray, pass by value primitive type, etc. 2. It works like unsafe FFI call by directly jump to c code using current capacity's stack. 3. It should mark current capacity GC ready, and doesn't participate next GC. The point is that if a FFI call doesn't need GC guarantee, then it shouldn't stop GC running. The next GC will not be running with all threads, but that's not too terrible since it doesn't scale too much anyway. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13944> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 15

[GHC] #13709: Drop GCC Driver
by GHC 18 Jun '18

18 Jun '18

#13709: Drop GCC Driver ----------------------------------------+--------------------------------- Reporter: Phyx- | Owner: (none) Type: task | Status: new Priority: normal | Milestone: 8.4.1 Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Windows Architecture: Unknown/Multiple | Type of failure: None/Unknown Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: ----------------------------------------+--------------------------------- The original reason for the Windows GCC driver was to correct hard-coded paths in the GCC build from the `msys` project. The `msys2` builds we now exclusively use no longer have these issues. Which means 90% of the reason for the driver is gone. Given the recent issues with the driver. We can/should simplify things and drop it. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13709> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[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] #14978: GADTs don't seem to unpack properly
by GHC 18 Jun '18

18 Jun '18

#14978: GADTs don't seem to unpack properly -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Compiler | Version: 8.2.2 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: -------------------------------------+------------------------------------- Consider this simple code: {{{#!hs type family Foo a b data Goof a b where Goof :: !(Foo a b :~: Int) -> Goof a b }}} `(:~:)` has just one constructor (`Refl`) with no arguments, so I'd expect it to unpack to a void representation: {{{#!hs GoofConstr :: Foo a b ~N# Int => Goof a b WGoof :: Foo a b :~: Int -> Goof a b WGoof eq = case eq of Refl -> GoofConstr }}} But it doesn't seem to unpack at all. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14978> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #14904: Compiler panic (piResultTy)
by GHC 18 Jun '18

18 Jun '18

#14904: Compiler panic (piResultTy) -------------------------------------+------------------------------------- Reporter: kcsongor | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.2.2 (Type checker) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: 14873 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Type-checking the following type family {{{#!hs type family F (f :: forall a. g a) :: Type where F (f :: forall a. g a) = Int }}} panics with the message: {{{#!txt ghc: panic! (the 'impossible' happened) (GHC version 8.4.0.20180118 for x86_64-apple-darwin): piResultTy k_aVM[tau:1] a_aVF[sk:1] Call stack: CallStack (from HasCallStack): callStackDoc, called at compiler/utils/Outputable.hs:1150:37 in ghc:Outputable pprPanic, called at compiler/types/Type.hs:950:35 in ghc:Type }}} The panic happens with HEAD, 8.4 and 8.2. 8.0 rejects the program with an error message, but even it panics on the following version: {{{#!hs type family F f :: Type where F ((f :: forall a. g a) :: forall a. g a) = Int }}} #14873 seemed somewhat related, so I tried with the patch implemented in 3d252037234ce48f9bdada7d5c9b1d8eba470829, but that fails with the same panic too. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14904> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #14845: TypeInType, index by GADT witnessing constraint
by GHC 18 Jun '18

18 Jun '18

#14845: TypeInType, index by GADT witnessing constraint -------------------------------------+------------------------------------- Reporter: Iceland_jack | Owner: (none) Type: feature | Status: new request | Priority: normal | Milestone: Component: Compiler | Version: 8.5 Keywords: TypeInType | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Just wondering if it would ever make sense to allow or use constraints like this {{{#!hs {-# Language PolyKinds, DataKinds, KindSignatures, GADTs, TypeInType, ConstraintKinds #-} import Data.Kind data Struct :: (k -> Constraint) -> (k -> Type) where S :: cls a => Struct cls a data Foo a where F :: Foo (S::Struct Eq a) }}} {{{ $ ghci -ignore-dot-ghci /tmp/test.hs GHCi, version 8.5.20180105: http://www.haskell.org/ghc/ :? for help [1 of 1] Compiling Main ( /tmp/test.hs, interpreted ) /tmp/test.hs:9:13: error: • Illegal constraint in a type: cls0 a0 • In the first argument of ‘Foo’, namely ‘(S :: Struct Eq a)’ In the type ‘Foo (S :: Struct Eq a)’ In the definition of data constructor ‘F’ | 9 | F :: Foo (S::Struct Eq a) | ^ Failed, no modules loaded. Prelude> }}} Please close the ticket if it doesn't -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14845> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 22

[GHC] #14487: Can't Hide Field When DuplicateRecordFields Is Enabled
by GHC 18 Jun '18

18 Jun '18

#14487: Can't Hide Field When DuplicateRecordFields Is Enabled -------------------------------------+------------------------------------- Reporter: iansullivan88 | Owner: (none) Type: bug | Status: new Priority: lowest | Milestone: Component: Compiler | Version: 8.0.2 Keywords: | Operating System: Linux DuplicateRecordFields | Architecture: x86_64 | Type of failure: GHC rejects (amd64) | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- A.hs fails to compile when DuplicateRecordFields is enabled in B.hs. A.hs {{{#!hs module A where import B hiding (duplicateName) test = X duplicateName duplicateName = 5 }}} B.hs {{{#!hs {-# LANGUAGE DuplicateRecordFields #-} module B where data X = X { duplicateName :: Int } }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14487> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6