February 2019 - ghc-tickets

[GHC] #16360: GHC fails when GHC_PACKAGE_PATH contains trailing slash
by GHC 23 Feb '19

23 Feb '19

#16360: GHC fails when GHC_PACKAGE_PATH contains trailing slash -------------------------------------+------------------------------------- Reporter: nh2 | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: GHC doesn't work Unknown/Multiple | at all Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- When the directory given in `GHC_PACKAGE_PATH` contains a trailing slash, GHC stops working completely. For example, this works: {{{ env GHC_PACKAGE_PATH=/raid/src/ghc/ghc-atomic- writes/_build/stage1/lib/package.conf.d /raid/src/ghc/ghc-atomic- writes/_build/stage1/bin/ghc "--make" "-outputdir" "example-programs- build/" "example-programs/Hello.hs" "-o" "example-programs-build/haskell- hello" }}} but adding a trailing `/` like `GHC_PACKAGE_PATH=/path/to/package.conf.d/` fails: {{{ env GHC_PACKAGE_PATH=/raid/src/ghc/ghc-atomic- writes/_build/stage1/lib/package.conf.d/ /raid/src/ghc/ghc-atomic- writes/_build/stage1/bin/ghc "--make" "-outputdir" "example-programs- build/" "example-programs/Hello.hs" "-o" "example-programs-build/haskell- hello" }}} {{{ [1 of 1] Compiling Main ( example-programs/Hello.hs, example- programs-build/Main.o ) [Prelude changed] example-programs/Hello.hs:1:1: error: Could not find module ‘Prelude’ There are files missing in the ‘base-4.12.0.0’ package, try running 'ghc-pkg check'. Use -v (or `:set -v` in ghci) to see a list of the files searched for. | 1 | main :: IO () | ^ }}} This is surprising; most programs accept trailing slashes where directories are expected, and shells ofthen TAB-complete directories this way. It wastes developer time when they don't notice. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/16360> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #15029: haddock parsing fails with valid input
by GHC 23 Feb '19

23 Feb '19

#15029: haddock parsing fails with valid input -------------------------------------+------------------------------------- Reporter: lerkok | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Compiler | Version: 8.2.2 (Parser) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Other Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Reported here: https://github.com/haskell/haddock/issues/798 This is really a GHC parsing issue, even though it only shows up in the haddock "route." -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15029> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

[GHC] #16358: Confusing ghci error message
by GHC 23 Feb '19

23 Feb '19

#16358: Confusing ghci error message -------------------------------------+------------------------------------- Reporter: lerkok | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.6.3 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: -------------------------------------+------------------------------------- {{{ $ ghci GHCi, version 8.6.3: http://www.haskell.org/ghc/ :? for help Prelude> import qualified Data.Set as S Prelude S> S.foo <interactive>:2:1: error: Not in scope: ‘S.foo’ No module named ‘S’ is imported. }}} I'm curious why `ghci` says `No module named 'S' is imported`. I found it very confusing. I'd have expected it to say one of three things: 1. `Module Data.Set referenced via S does not export foo` 2. `Constructor S not in scope` 3. `Variable foo not in scope` The first one would be the most informative, but I can see that the second and the third messages might be quite reasonable as well due to the interpretation of dot as composition. But the message `No module S is imported` is quite confusing given I just issued `import qualified Data.Set as S`. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/16358> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #16353: ./validate --hadrian doesn't quite work on Windows
by GHC 22 Feb '19

22 Feb '19

#16353: ./validate --hadrian doesn't quite work on Windows -------------------------------------+------------------------------------- Reporter: alpmestan | Owner: (none) Type: bug | Status: new Priority: high | Milestone: Component: Build System | Version: 8.7 (Hadrian) | Keywords: | Operating System: Windows Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- On my Windows machine, using an msys2 shell as advised by the Windows- specific pages of GHC's trac, I managed to get `./validate --hadrian` (from [https://gitlab.haskell.org/ghc/ghc/merge_requests/326 this MR]) to go pretty far. I did manage to make it produce a binary distribution after some serious wrangling, to have it installed somewhere nearby and it ends up choking on [https://gitlab.haskell.org/alp/ghc/blob/wip/alp/hadrian- validate/validate#L258 this command] which tries to configure an `xhtml` build using `Setup.hs` scripts. When I looked at the verbose output, I ended up seeing that Cabal calls `findProgramVersion` for the ghc that I pass using `--with-ghc=...`, with `--numeric-version` as the only argument. And this ends up returning `""` -- which is what's returned when exceptions are thrown apparently, judging from [https://github.com/haskell/cabal/blob/c2cf30cc50109f2ffa30fd7affe9c22e1adf9… findProgramVersion's implementation]. Now, where this gets "interesting" is that I've tried to run `./Setup configure --with-ghc=X` for different values of `X`... - (1) `./Setup configure --with- ghc=/path/to/ghc/tree/_validatebuild/stage1/bin/ghc` works - (2) `./Setup configure --with- ghc=/path/to/ghc/tree/_validatebuild/bindist/ghc-8.6.20190220/bin/ghc` works - (3) `./Setup configure --with-ghc=/path/to/ghc/tree/bindisttest/install dir/bin/ghc` **doesn't work** (3 spaces between `install` and `dir`) - (4) `./Setup configure --with-ghc=/path/to/ghc/tree/a/b/bin/ghc` **doesn't work** (1) is our stage 2 GHC executable, the real one, produced right at this path. (2) is a copy of (1) that we put under this `bindist/ghc-X.Y.Z/` dir, which we end up putting wholesale in an archive. (3) is where `./validate` (whether with `--hadrian` or not) installs binary distributions. The install process goes fine, I can even use the GHC in question and evaluate `.../bin/ghc -e 'Data.Text.IO.putStrLn (Data.Text.pack "hello")'` without any problem, or get the expected output with `.../bin/ghc --numeric-version`. But the `configure` step just won't have it. I initially thought this was due to the directory with a space in it, one way or another, so I decided to try and install the bindist manually in a place with no spaces in the directory names, (4). (4) doesn't work either though. The main difference between 1) 2) on one hand, and 3) 4) on the other, is that in the first two cases I'm passing the path to an actual stage 2 GHC executable, while in the case of 3) and 4) I'm pointing to wrapper scripts, with the actual executables being under `$topdir/lib/bin/`. I'm however not sure of what's going on, it seems to be like this should work, any help is appreciated. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/16353> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #16343: Changing an irrelevant constraint leads to inlining failure
by GHC 22 Feb '19

22 Feb '19

#16343: Changing an irrelevant constraint leads to inlining failure -------------------------------------+------------------------------------- Reporter: mpickering | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.6.3 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 is an example Andres sent me but I can't work out precisely what is going on. {{{ {-# LANGUAGE AllowAmbiguousTypes #-} {-# LANGUAGE ConstraintKinds #-} {-# LANGUAGE DataKinds #-} {-# LANGUAGE FlexibleContexts #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE GADTs #-} {-# LANGUAGE MultiParamTypeClasses #-} {-# LANGUAGE PolyKinds #-} {-# LANGUAGE QuantifiedConstraints #-} {-# LANGUAGE RankNTypes #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE TemplateHaskell #-} {-# LANGUAGE TypeApplications #-} {-# LANGUAGE TypeOperators #-} module Tests where import Data.Kind import Data.Maybe (maybeToList) data Product (f :: a -> Type) (xs :: [a]) :: Type where Nil :: Product f '[] (:*) :: f x -> !(Product f xs) -> Product f (x : xs) infixr 5 :* class Top a instance Top a class (forall d . (forall x . c x => d x) => All d xs) => All (c :: a -> Constraint) (xs :: [a]) where ccataList :: f '[] -> (forall y ys . (c y, All c ys) => f ys -> f (y : ys)) -> f xs instance All c '[] where ccataList nil _cons = nil {-# INLINE ccataList #-} instance (c x, All c xs) => All c (x : xs) where ccataList nil cons = cons (ccataList @_ @c nil cons) {-# INLINE ccataList #-} newtype (f ~> g) x = Lambda { apply :: f x -> g x } cpureProduct :: forall c xs f . All c xs => (forall x . c x => f x) -> Product f xs cpureProduct p = ccataList @_ @c Nil (p :*) {-# INLINE cpureProduct #-} newtype Apply f g x = MkApply { unApply :: Product (f ~> g) x -> Product f x -> Product g x } capplyProduct :: forall c xs f g . All c xs => Product (f ~> g) xs -> Product f xs -> Product g xs capplyProduct = unApply $ ccataList @_ @c (MkApply (\ Nil Nil -> Nil)) (\ rec -> MkApply (\ (f :* fs) (x :* xs) -> apply f x :* unApply rec fs xs)) {-# INLINE capplyProduct #-} -- The addition of the constraint makes the test pass. mapProduct :: forall c xs f g . ({- All Top xs, -} All c xs) => (forall x . c x => f x -> g x) -> Product f xs -> Product g xs mapProduct p q = capplyProduct @Top (cpureProduct @c (Lambda p)) q {-# INLINE mapProduct #-} test :: Product [] [ (), (), () ] test = mapProduct @Top maybeToList (cpureProduct @Top Nothing) reference :: Product [] [ (), (), () ] reference = [] :* [] :* [] :* Nil }}} `test` should simplify to something close to `reference. There are a few ways to make it do so. 1. Uncomment the constraint in `mapProduct` 2. Change `Top` to `c` in `capplyProduct @Top`. From looking at `-ddump-inlinings`, it seems that things start to go wrong when `$cccataList` is inlined. In the good case, it looks like this. {{{ Considering inlining: $cccataList_a1Bi arg infos [ValueArg, ValueArg, ValueArg, ValueArg, NonTrivArg, NonTrivArg] interesting continuation RhsCtxt some_benefit True is exp: True is work-free: True guidance ALWAYS_IF(arity=4,unsat_ok=False,boring_ok=False) ANSWER = YES }}} In the bad case it looks like {{{ Considering inlining: $cccataList_a1Bv arg infos [ValueArg, ValueArg, ValueArg, ValueArg] interesting continuation RuleArgCtxt some_benefit True is exp: True is work-free: True guidance ALWAYS_IF(arity=4,unsat_ok=False,boring_ok=False) ANSWER = YES }}} Notice the continuation and arg infos are different for the two cases. The first case seems to be eta expanded? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/16343> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

[GHC] #16352: Hadrian: generation of PDF documents doesn't work
by GHC 22 Feb '19

22 Feb '19

#16352: Hadrian: generation of PDF documents doesn't work -------------------------------------+------------------------------------- Reporter: alpmestan | Owner: (none) Type: bug | Status: new Priority: high | Milestone: Component: Build System | Version: 8.7 (Hadrian) | Keywords: | Operating System: Linux Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- 1) With [https://github.com/alpmestan/ghc.nix ghc.nix] While trying to produce `_build/docs/pdfs/Haddock.pdf`: {{{ kpathsea: Running mktexmf pcrr8t ! I can't find file `pcrr8t'. <*> ...:=ljfour; mag:=1; nonstopmode; input pcrr8t Please type another input file name ! Emergency stop. <*> ...:=ljfour; mag:=1; nonstopmode; input pcrr8t Transcript written on mfput.log. grep: pcrr8t.log: No such file or directory }}} 2) With the CI docker images (e.g [https://gitlab.haskell.org/ghc/ghc/blob/master/.circleci/images/x86_64 -linux-deb8/Dockerfile this x86_64 debian8 one]) Agian, while trying to product the PDF version of the haddock manual: {{{ | Run Xelatex: Haddock.tex => /tmp/extra-dir-87721391039866 This is XeTeX, Version 3.14159265-2.6-0.99992 (TeX Live 2015/dev/Debian) (preloaded format=xelatex) restricted \write18 enabled. entering extended mode (./Haddock.tex LaTeX2e <2014/05/01> Babel <3.9l> and hyphenation patterns for 2 languages loaded. (./sphinxmanual.cls Document Class: sphinxmanual 2009/06/02 Document class (Sphinx manual) (/usr/share/texlive/texmf-dist/tex/latex/base/report.cls Document Class: report 2014/09/29 v1.4h Standard LaTeX document class (/usr/share/texlive/texmf-dist/tex/latex/base/size10.clo))) (/usr/share/texlive/texmf-dist/tex/latex/base/inputenc.sty Package inputenc Warning: inputenc package ignored with utf8 based engines. ) ! Undefined control sequence. <recently read> \DeclareUnicodeCharacter l.5 \DeclareUnicodeCharacter {00A0}{\nobreakspace} No pages of output. Transcript written on Haddock.log. }}} [https://tex.stackexchange.com/a/378643 This StackExchange post] that Matthew pointed me to gives a clue as to as why this happens. --- I am going to comment out the `need`s that make us produce these files in Hadrian when we ask for the docs, for now, leaving a comment pointing to this ticket. Then as we figure out solutions to these issues (especially 2), since that's GHC's CI system), we can put those PDFs back in and mark this ticket as solved. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/16352> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #16351: Extend constant folding operations to bit manipulations
by GHC 22 Feb '19

22 Feb '19

#16351: Extend constant folding operations to bit manipulations -------------------------------------+------------------------------------- Reporter: Fuuzetsu | Owner: (none) Type: feature | Status: new request | Priority: normal | Milestone: Component: Compiler | Version: 8.6.3 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: -------------------------------------+------------------------------------- In #9136 GHC has greatly improved how it folds expressions for Int/Word when the basic arithmetic operations {{{+, -, *}}} are involved. I think that somewhat analogously this extend should be extended to binary manipulation operations. For example, consider the below snippet: {{{#!hs {-# OPTIONS_GHC -O2 -ddump-simpl -ddump-to-file -ddump-asm #-} module G (bar, baz) where import Data.Bits ((.|.)) import Data.Word (Word16) bar :: Word16 -> Int bar w = -9223372036854775808 .|. (281474976710656 .|. fromIntegral w) baz :: Word16 -> Int baz w = -9223372036854775808 + (281474976710656 + fromIntegral w) }}} Let's peek into the Core: {{{#!hs bar = \ (w_a1hX :: Word16) -> case w_a1hX of { GHC.Word.W16# x#_a1Lu -> GHC.Types.I# (GHC.Prim.orI# -9223372036854775808# (GHC.Prim.orI# 281474976710656# (GHC.Prim.word2Int# x#_a1Lu))) } baz = \ (w_a1sr :: Word16) -> case w_a1sr of { GHC.Word.W16# x#_a1Lu -> GHC.Types.I# (GHC.Prim.+# -9223090561878065152# (GHC.Prim.word2Int# x#_a1Lu)) } }}} Due to #9136, {{{bar}}} gets nicely folded away into a single addition. {{{baz}}} however is still in its "naive" form. The above example was extracted from an actual program where there may be long such chains of {{{orI#}}} and similar operations. I think it's worth pointing out that if we peek in the ASM, {{{baz}}} seems to get folded away eventually but I don't trust it to always happen properly at that level and I don't particularly fancy trawling through the ASM output every time. https://stackoverflow.com/a/45909278 has some laws we could use. I'm sure we could pick in LLVM source &c. too if we want to. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/16351> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #14198: Inconsistent treatment of implicitly bound kind variables as free-floating
by GHC 21 Feb '19

21 Feb '19

#14198: Inconsistent treatment of implicitly bound kind variables as free-floating -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.2.1 (Type checker) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: #7873 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- (Spun off from the discussion starting at https://phabricator.haskell.org/D3872#109927.) This program is accepted: {{{#!hs {-# LANGUAGE ExistentialQuantification #-} {-# LANGUAGE PolyKinds #-} {-# LANGUAGE RankNTypes #-} import Data.Proxy data Foo = MkFoo (forall a. Proxy a) }}} There's something interesting going on here, however. Because `PolyKinds` is enabled, the kind of `a` is generalized to `k`. But where does `k` get quantified? It turns out that it's implicitly quantified as an existential type variable to `MkFoo`: {{{ λ> :i Foo data Foo = forall k. MkFoo (forall (a :: k). Proxy a) }}} This was brought up some time ago in #7873, where the conclusion was to keep this behavior. But to make things stranger, it you write out the kind of `a` explicitly: {{{#!hs {-# LANGUAGE ExistentialQuantification #-} {-# LANGUAGE PolyKinds #-} {-# LANGUAGE RankNTypes #-} import Data.Proxy data Foo2 = MkFoo2 (forall (a :: k). Proxy a) }}} Then GHC will reject it: {{{ Bug.hs:7:1: error: Kind variable ‘k’ is implicitly bound in data type ‘Foo2’, but does not appear as the kind of any of its type variables. Perhaps you meant to bind it (with TypeInType) explicitly somewhere? | 7 | data Foo2 = MkFoo2 (forall (a :: k). Proxy a) | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ }}} So GHC's treatment is inconsistent. What should GHC do? We could: * Have both be an error. * Have both be accepted, and implicitly quantify `k` as an existential type variable * Have both be accepted, and implicitly quantify `k` in the `forall` itself. That is: {{{#!hs MkFoo :: (forall {k} (a :: k). Proxy k a) -> Foo }}} * Something else. When you try a similar trick with type synonyms: {{{#!hs {-# LANGUAGE ExistentialQuantification #-} {-# LANGUAGE PolyKinds #-} {-# LANGUAGE RankNTypes #-} import Data.Proxy type Foo3 = forall a. Proxy a }}} Instead of generalizing the kind of `a`, its kind will default to `Any`: {{{ λ> :i Foo3 type Foo3 = forall (a :: GHC.Types.Any). Proxy a }}} Would that be an appropriate trick for data types as well? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14198> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #15227: Add PrelRules for par#
by GHC 21 Feb '19

21 Feb '19

#15227: Add PrelRules for par# -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: (none) Type: bug | Status: new Priority: low | Milestone: 8.6.1 Component: Compiler | Version: 8.4.3 Keywords: Newcomer | 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: -------------------------------------+------------------------------------- `seq#` and `spark#` are elided when their arguments are known to be evaluated. `par#`, however, is not. Presumably this leads to sparks being created and then immediately going away. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15227> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #16140: Cannot create type synonym for quantified constraint without ImpredicativeTypes
by GHC 21 Feb '19

21 Feb '19

#16140: Cannot create type synonym for quantified constraint without ImpredicativeTypes -------------------------------------+------------------------------------- Reporter: Ashley | Owner: (none) Yakeley | Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: GHC rejects Unknown/Multiple | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- {{{#!hs {-# LANGUAGE KindSignatures, RankNTypes, ConstraintKinds, QuantifiedConstraints, FlexibleInstances, UndecidableInstances #-} module Bug where type F1 (f :: * -> *) = forall a. Eq (f a) class (Functor f, F1 f) => C f instance (Functor f, F1 f) => C f type F2 f = (Functor f, F1 f) }}} {{{ Bug.hs:7:1: error: • Illegal polymorphic type: F1 f GHC doesn't yet support impredicative polymorphism • In the type synonym declaration for ‘F2’ | 7 | type F2 f = (Functor f, F1 f) | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ }}} (GHC accepts the program with ImpredicativeTypes.) `(Functor f, F1 f)` is allowed as a superclass constraint, and as an instance constraint, but a type synonym cannot be made for it. Not sure if this really counts as a bug ("just switch on ImpredicativeTypes"), but I think it's worth discussing. I prefer to keep ImpredicativeTypes switched off, but if something can be a constraint, shouldn't I be able to create a type synonym of it? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/16140> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 20