April 2018 - ghc-tickets

Re: [GHC] #4020: Please consider adding support for local type synonyms
by GHC 09 Apr '18

09 Apr '18

#4020: Please consider adding support for local type synonyms -------------------------------------+------------------------------------- Reporter: nr | Owner: (none) Type: feature request | Status: new Priority: low | Milestone: Component: Compiler (Type | Version: 6.12.2 checker) | Resolution: | Keywords: type synonym Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by Tritlo): * cc: Tritlo (added) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/4020#comment:21> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #10424: Build path leaks into ABI hashes
by GHC 09 Apr '18

09 Apr '18

#10424: Build path leaks into ABI hashes -------------------------------------+------------------------------------- Reporter: nomeata | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.4 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Blocked By: Test Case: | Related Tickets: Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- The build path of GHC leaks into the ABI hashes, i.e. if you build GHC in /tmp/fooADDF and then in /tmp/fooSDFS, you get different ABI hashes. This is a serious problem for distributions, as it means that we cannot rebuild the GHC package without rebuilding all Haskell packages as well. The (or at least one) problem is include paths. We have {{{ include-dirs: /home/jojo/build/haskell/ghc/rts/dist/build /home/jojo/build/haskell/ghc/includes /home/jojo/build/haskell/ghc/includes/dist- derivedconstants/header }}} in `inplace/lib/package.conf.d/builtin_rts.conf`, so when running the preprocessor, GHC passes this full path to the preprocessor: {{{ /usr/bin/gcc -E -undef -traditional -DOPTIMISE_INTEGER_GCD_LCM -include libraries/base/dist-install/build/autogen/cabal_macros.h -I libraries/base /dist-install/build -I libraries/base/dist-install/build -I libraries/base /dist-install/build/autogen -I libraries/base/include -I /home/jojo/build/haskell/ghc/libraries/integer-gmp/include -I /home/jojo/build/haskell/ghc/rts/dist/build -I /home/jojo/build/haskell/ghc/includes -I /home/jojo/build/haskell/ghc/includes/dist-derivedconstants/header '-D__GLASGOW_HASKELL__=711' -include /home/jojo/build/haskell/ghc/includes/ghcversion.h '-Dlinux_BUILD_OS=1' '-Dx86_64_BUILD_ARCH=1' '-Dlinux_HOST_OS=1' '-Dx86_64_HOST_ARCH=1' '-D__GLASGOW_HASKELL_TH__=YES' '-D__SSE__=1' '-D__SSE2__=1' -x assembler- with-cpp libraries/base/System/Info.hs -o /tmp/ghc14804_0/ghc14804_1.hscpp }}} It then reads the included files from the `hscpp` file : {{{ $ grep home /tmp/ghc14830_0/ghc14830_1.hscpp # 1 "/home/jojo/build/haskell/ghc/includes/ghcversion.h" 1 # 1 "/home/jojo/build/haskell/ghc/includes/ghcplatform.h" 1 }}} and puts it into the interface, where it becomes part of the hash: {{{ $ ./inplace/bin/ghc-stage1 --show-iface ./libraries/base/dist- install/build/System/Info.hi |grep Dep addDependentFile "/home/jojo/build/haskell/ghc/includes/ghcplatform.h" addDependentFile "/home/jojo/build/haskell/ghc/includes/ghcversion.h" addDependentFile "libraries/base/dist- install/build/autogen/cabal_macros.h" addDependentFile "/usr/include/stdc-predef.h" }}} Clearly, the end result is undesirable (changing hashes due to path names) and actually useless, as the build path will not be there later anyways. I’m not sure what the best way to fix this is. Here are a few ideas: * Do not include the path of `addDependentFile` entries in the hash, but only the file hash. Should solve the ABI hash. * Use relative paths in `inplace/lib/package.conf.d/builtin_rts.conf`. Not sure what breaks. This bug appeared in 7.8.4 and is present in both 7.10 and current HEAD. Any other ideas? (Debian bugreport at http://bugs.debian.org/785282) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10424> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #15015: Parsing of exp_doc loses section markers
by GHC 09 Apr '18

09 Apr '18

#15015: Parsing of exp_doc loses section markers -------------------------------------+------------------------------------- Reporter: lerkok | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.4.3 Component: Compiler | Version: 8.2.2 (Parser) | 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 originally filed as a Haddock issue, but was traced back to a parsing issue with GHC itself. See here for details: https://github.com/haskell/haddock/issues/797 There's a work-around (by putting extra empty lines) but that's quite fragile and would be best to fix directly in GHC. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15015> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #15008: Type synonyms with hidden, determined type variables
by GHC 08 Apr '18

08 Apr '18

#15008: Type synonyms with hidden, determined type variables -------------------------------------+------------------------------------- Reporter: nomeata | Owner: (none) Type: feature | Status: new request | Priority: normal | Milestone: Research needed Component: Compiler | Version: 8.5 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 found myself wanting to write (simplified) {{{ type MyMonad m = (MonadReader r m, Show r) }}} but GHC does not accept this, it wants `r` as part of the type synonym. So why did I think this could work? Because `MonadReader` has a functional dependency: {{{ class Monad m => MonadReader r m | m -> r }}} so if `m` determines `r`, why do I have to write `MyMonad r m` and not just `MyMonad m`? Can I not “hide” the fact that, somewhere internal to the `MyMonad` synonym, there is an additional variable `r` around? Clearly supporting this would need a language proposal, but I wanted to open a ticket here first to hear if this would even be possible to implement. ([https://stackoverflow.com/questions/29686920/is-it-possible-to-have- forgetful-type-synonyms-in-haskell Others have wanted this before as well.]) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15008> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #15014: Exhaustivity check should suggest when COMPLETE could be helpful
by GHC 08 Apr '18

08 Apr '18

#15014: Exhaustivity check should suggest when COMPLETE could be helpful -------------------------------------+------------------------------------- Reporter: ezyang | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.4.3 Component: Compiler | Version: 8.5 (Type checker) | Keywords: | Operating System: Unknown/Multiple PatternSynonyms, | PatternMatchWarnings | Architecture: | Type of failure: Poor/confusing Unknown/Multiple | error message Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- {{{ MacBook-Pro-97:ghc ezyang$ cat A.hs {-# LANGUAGE PatternSynonyms #-} module A where pattern F :: a -> b -> (a, b) pattern F x y = (x, y) g :: (a, b) -> (a, b) g (F x y) = (x, y) MacBook-Pro-97:ghc ezyang$ inplace/bin/ghc-stage2 -c A.hs -Wall -fforce- recomp A.hs:8:1: warning: [-Wincomplete-patterns] Pattern match(es) are non-exhaustive In an equation for ‘g’: Patterns not matched: _ | 8 | g (F x y) = (x, y) | ^^^^^^^^^^^^^^^^^^ MacBook-Pro-97:ghc ezyang$ inplace/bin/ghc-stage2 --version The Glorious Glasgow Haskell Compilation System, version 8.5.20180304 }}} Any time the exhaustiveness checker throws up its hands and says that `_` is not matched, we ought to give a hint that this may have occurred due to pattern synonyms, and that the author of the pattern synonyms can help out by specifying a COMPLETE pragma. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15014> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #15013: Building nofib using llvm and HEAD fails
by GHC 08 Apr '18

08 Apr '18

#15013: Building nofib using llvm and HEAD fails -------------------------------------+------------------------------------- Reporter: AndreasK | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.4.2 Component: Compiler | Version: (LLVM) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Compile-time Unknown/Multiple | crash or panic Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- {{{ == make all ; in /home/Andi/ghc_head/nofib/shootout/pidigits ------------------------------------------------------------------------ HC = /home/Andi/ghc_head/inplace/bin/ghc-stage2 HC_OPTS = -O2 -Rghc-timing -H32m -hisuf hi -fllvm -rtsopts -O2 RUNTEST_OPTS = -ghc-timing ==nofib== pidigits: time to compile Main follows... /home/Andi/ghc_head/inplace/bin/ghc-stage2 -O2 -Rghc-timing -H32m -hisuf hi -fllvm -rtsopts -O2 -c Main.hs -o Main.o ghc-stage2.exe: panic! (the 'impossible' happened) (GHC version 8.5.20180401 for x86_64-unknown-mingw32): Each block should be reachable from only one ProcPoint Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug <<ghc: 506688344 bytes, 156 GCs, 7738569/16639872 avg/max bytes residency (5 samples), 37M in use, 0.000 INIT (0.000 elapsed), 0.344 MUT (0.333 elapsed), 0.062 GC (0.112 elapsed) :ghc>> make[1]: *** [../../mk/suffix.mk:23: Main.o] Error 1 Failed making all in pidigits: 1 make: *** [../mk/ghc-recurse.mk:69: all] Error 1 $ opt --version LLVM (http://llvm.org/) LLVM version 5.0.1 Optimized build. Default target: x86_64-w64-windows-gnu Host CPU: skylake }}} This is on commit 8b823f270e53627ddca1a993c05f1ab556742d96 -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15013> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #15011: Automate update of VersionHistory table
by GHC 07 Apr '18

07 Apr '18

#15011: Automate update of VersionHistory table -------------------------------------+------------------------------------- Reporter: bgamari | Owner: bgamari Type: task | Status: new Priority: normal | Milestone: 8.6.1 Component: Continuous | Version: 8.2.2 Integration | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Other Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- wiki:Commentary/Libraries/VersionHistory should really be updated automatically from CI builds. This is currently generated manually with https://github.com/bgamari/ghc-utils/tree/master/library-versions. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15011> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #1965: Allow unconstrained existential contexts in newtypes
by GHC 07 Apr '18

07 Apr '18

#1965: Allow unconstrained existential contexts in newtypes -------------------------------------+------------------------------------- Reporter: guest | Owner: (none) Type: feature request | Status: new Priority: normal | Milestone: Component: Compiler | Version: 6.8.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by glaebhoerl): * cc: glaebhoerl (added) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/1965#comment:36> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #12891: Automate symbols inclusion in RtsSymbols.c from Rts.h
by GHC 07 Apr '18

07 Apr '18

#12891: Automate symbols inclusion in RtsSymbols.c from Rts.h -------------------------------------+------------------------------------- Reporter: Phyx- | Owner: Type: task | Status: new Priority: normal | Milestone: Component: Build System | Version: 8.0.1 Keywords: newcomer | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: #12846 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Public symbols that we export in the public headers (rooted from Rts.h) need to be added to the symbols table of the RTS. We currently do this by hand but the list can quickly grow out of sync. Investigate a way to automate either the inclusion or a check during build time. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12891> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 18

[GHC] #14238: `:kind` suppresses visible dependent quantifiers by default in GHCi 8.2.1
by GHC 07 Apr '18

07 Apr '18

#14238: `:kind` suppresses visible dependent quantifiers by default in GHCi 8.2.1 -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.2.1 (Type checker) | Keywords: TypeInType | Operating System: Unknown/Multiple Architecture: | Type of failure: Incorrect result Unknown/Multiple | at runtime Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Load this program into GHCi on 8.2.1 or later: {{{#!hs {-# LANGUAGE GADTs #-} {-# LANGUAGE TypeInType #-} import Data.Kind data Foo (k :: Type) :: k -> Type where MkFoo :: Foo (k1 -> k2) f -> Foo k1 a -> Foo k2 (f a) }}} And ask it what the kind of `Foo` is: {{{ GHCi, version 8.2.1: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/rgscott/.ghci [1 of 1] Compiling Main ( Bug.hs, interpreted ) Ok, 1 module loaded. λ> :k Foo Foo :: k -> * }}} This is just plain wrong: the actual kind of `Foo` should be `forall k -> k -> *`. Normally, one can omit `forall`s from a kind signature, but this is a special case where we're using a //visible// forall. In other words, `forall k -> k -> *` is not the same as `k -> *`, as the former takes two arguments, whereas the latter takes one. A workaround is to force GHCi to come to its senses by explicitly enabling `-fprint-explicit-foralls`: {{{ λ> :set -fprint-explicit-foralls λ> :k Foo Foo :: forall k -> k -> * }}} This is actually a regression since GHC 8.0.2, since GHCi did the right thing by default then: {{{ GHCi, version 8.0.2: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/rgscott/.ghci [1 of 1] Compiling Main ( Bug.hs, interpreted ) Ok, modules loaded: Main. λ> :k Foo Foo :: forall k -> k -> Type }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14238> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5