October 2015 - ghc-tickets

[GHC] #10845: Incorrect behavior when let binding implicit CallStack object
by GHC 14 Dec '15

14 Dec '15

#10845: Incorrect behavior when let binding implicit CallStack object -------------------------------------+------------------------------------- Reporter: | Owner: nitromaster101 | Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.11 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- I'm trying out ghc HEAD at 062feee4e7408ad5b9d882e5fed2c700e337db72. {{{#!hs {-# LANGUAGE ImplicitParams #-} import GHC.Types f :: (?loc :: CallStack) => Int f = let y = length $ getCallStack ?loc in if y < 2 then 3 else 4 f2 :: (?loc :: CallStack) => Int f2 = if (length $ getCallStack ?loc) < 2 then 3 else 4 f3 :: (?loc :: [Int]) => Int f3 = let y = length ?loc in if y < 2 then 3 else 4 }}} produces: {{{ [1 of 1] Compiling Main ( Bug.hs, interpreted ) Bug.hs:5:6: warning: Redundant constraint: ?loc::CallStack In the type signature for: f :: (?loc::CallStack) => Int }}} f3 doesn't produce an error which makes me think it's related to CallStack handling. The warning seems accurate: the returned callstack only contains the entry at f. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10845> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 19

[GHC] #10819: Can't generate data decls deriving multiparam type classes with TH
by GHC 14 Dec '15

14 Dec '15

#10819: Can't generate data decls deriving multiparam type classes with TH -------------------------------------+------------------------------------- Reporter: spinda | Owner: Type: feature | Status: new request | Priority: normal | Milestone: Component: Template | Version: 7.10.2 Haskell | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- This applies to both data declarations and (more commonly) newtypes. Example: {{{ {-# LANGUAGE GeneralizedNewtypeDeriving #-} newtype T a = X { unT :: State Int a } deriving (MonadState Int) }}} The newtype declaration is impossible to express in Template Haskell's AST, as the field for the derived classes is {{{[Name]}}} when it should really be {{{Cxt}}}. A workaround is to generate a {{{StandaloneDerivD}}} for each derived multiparam class, but this requires that the user enable the {{{StandaloneDeriving}}} extension which can be unintuitive, especially as there's no way (as far as I can tell) to detect whether this extension is enabled and report a proper error message if it isn't. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10819> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #10741: add flag to dump module and package dependencies
by GHC 13 Dec '15

13 Dec '15

#10741: add flag to dump module and package dependencies -------------------------------------+------------------------------------- Reporter: elaforge | Owner: Type: feature | Status: new request | Priority: normal | Milestone: Component: Compiler | Version: Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- You can already dump module and package dependencies with -M -include-pkg- deps, but it wants to modify a Makefile, so it's not so convenient for things that are not make. For other build systems, it would be convenient to get this on stdout in an easily parseable form. Say X/A.hs imports Y/B.hs, which imports Z/C.hs, and they all use package 'base' and 'containers': {{{ % ghc -Mstdout X/A.hs X/A.h Y/B.hs Y/B.hs Z/C.hs base containers % }}} I don't know about what the output format should be, but assuming no newlines in filenames, and package names never have spaces, it seems simplest to do something like: [src ++ "\n" ++ dest ++ "\n" | (src, dest) <- moduleDeps] ++ "\n" ++ unwords packageDeps It could use haskell syntax like --info, but then you'd need a full haskell parser just to read it, which is not a problem for shake, but probably inconvenient for other build systems. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10741> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

[GHC] #10732: Legal Bang Patterns cannot parse
by GHC 13 Dec '15

13 Dec '15

#10732: Legal Bang Patterns cannot parse -------------------------------------+------------------------------------- Reporter: rem | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.3 (Parser) | Keywords: Bang | Operating System: Unknown/Multiple Pattern Parser | Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- According to the syntax provided by GHC user guide pat ::= !pat Both !!pat and pat : !pat should be accepted, but GHC gives parsing error. For detail see http://stackoverflow.com/questions/31790327/where-can- bangpatterns-appear -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10732> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #10902: Refactor type families in Template Haskell
by GHC 12 Dec '15

12 Dec '15

#10902: Refactor type families in Template Haskell -------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: task | Status: new Priority: normal | Milestone: 8.0.1 Component: Template | Version: 7.11 Haskell | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- I see that we now have `OpenTypeFamilyD` and `ClosedTypeFamilyD` in Template Haskell. These constructors have quite a lot in common, and I think it's worth having a `TypeFamilyHead` datatype that contains the bits in common. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10902> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #10785: ghc-pkg appends colon to db path when piped/redirected
by GHC 12 Dec '15

12 Dec '15

#10785: ghc-pkg appends colon to db path when piped/redirected -------------------------------------+------------------------------------- Reporter: jgertm | Owner: jgertm Type: bug | Status: new Priority: normal | Milestone: Component: ghc-pkg | Version: 7.10.2 Keywords: colon pipe | Operating System: Unknown/Multiple redirect ghc-pkg list | Architecture: | Type of failure: Other Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- Quite simple. Compare the first line of output from {{{#!sh ghc-pkg list ghc-pkg list | head }}} In the second command's output, a colon gets appended at the end of the first line, compared to the first command. This is generally undesired behaviour, since the intent of redirection is to further process the output of the original command. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10785> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8566: Panic with kindFunResult
by GHC 11 Dec '15

11 Dec '15

#8566: Panic with kindFunResult ------------------------------------+------------------------------------- Reporter: dreixel | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- The following program: {{{ {-# LANGUAGE DataKinds #-} {-# LANGUAGE PolyKinds #-} {-# LANGUAGE GADTs #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE TypeOperators #-} {-# LANGUAGE UndecidableInstances #-} {-# LANGUAGE MultiParamTypeClasses #-} module Bug where data U (s :: *) = forall k. AA k [U s] data I (u :: U *) (r :: [*]) :: * where A :: I (AA t as) r -- fs unused, but needs to be present for the bug class C (u :: U *) (r :: [*]) (fs :: [*]) where c :: I u r -> I u r instance (C (AA (t (I a ps)) as) ps fs) => C (AA t (a ': as)) ps fs where c A = c undefined }}} crashes a fresh copy of GHC HEAD with the following: {{{ ghc-stage2: panic! (the 'impossible' happened) (GHC version 7.7.20131126 for i386-unknown-linux): kindFunResult k1{tv a24f} [ssk] }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8566> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 21

[GHC] #9017: Confusing error message with PolyKinds
by GHC 11 Dec '15

11 Dec '15

#9017: Confusing error message with PolyKinds ------------------------------------+------------------------------------- Reporter: edsko | 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: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- Consider the following example code: {{{ import Control.Arrow foo :: a b (m b) foo = arr return }}} if we typecheck this we rightly get two errors about missing instances (Arrow a, Monad m). However, if we enable `PolyKinds` we get the following error message: {{{ T.hs:4:7: Kind incompatibility when matching types: a0 :: * -> * -> * a :: k1 -> k -> * Expected type: a b (m b) Actual type: a0 b0 (m0 b0) Relevant bindings include foo :: a b (m b) (bound at T.hs:4:1) In the expression: arr return In an equation for ‘foo’: foo = arr return }}} I can ''sort of'' see where this is coming from if I think really hard :), so perhaps it's not a bug per se, but it's definitely confusing; and it's a pity that a type error (forgotten type qualifier) is reported as a kind error. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9017> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #9389: Full Test Suite Failures
by GHC 11 Dec '15

11 Dec '15

#9389: Full Test Suite Failures ------------------------------------+---------------------------------- Reporter: jrp | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Test Suite | Version: 7.8.3 Keywords: | Operating System: MacOS X Architecture: Unknown/Multiple | Type of failure: Other Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Revisions: ------------------------------------+---------------------------------- This is just to record progress on running the full test suite on OS X for the current HEAD using a BuildFlavour = perf build. I'll check through these (some of which are configuration issues because of the non-standard build or some required libraries are not being pulled in). {{{ OVERALL SUMMARY for test run started at Thu Jul 31 23:03:39 2014 BST 0:32:31 spent to go through 4036 total tests, which gave rise to 19262 test cases, of which 4218 were skipped 358 had missing libraries 14485 expected passes 135 expected failures 0 caused framework failures 0 unexpected passes 66 unexpected failures Unexpected failures: ../../libraries/base/tests enum01 [exit code non-0] (normal,hpc,optasm,profasm,threaded1,threaded2,dyn,profthreaded,optllvm) ../../libraries/base/tests enum01 [bad stdout or stderr] (ghci) ../../libraries/base/tests enum02 [exit code non-0] (normal,hpc,optasm,profasm,threaded1,threaded2,dyn,profthreaded,optllvm) ../../libraries/base/tests enum02 [bad stdout or stderr] (ghci) ../../libraries/base/tests enum03 [exit code non-0] (normal,hpc,optasm,profasm,threaded1,threaded2,dyn,profthreaded,optllvm) ../../libraries/base/tests enum03 [bad stdout or stderr] (ghci) ../../libraries/unix/tests signals004 [bad exit code] (threaded2) concurrent/should_run T367_letnoescape [bad exit code] (optllvm) concurrent/should_run conc012 [bad exit code] (ghci) driver/objc objc-hi [bad profile] (profthreaded) driver/objc objc-hi [bad heap profile] (profasm) driver/objc objcpp-hi [bad profile] (profthreaded) driver/objc objcpp-hi [bad heap profile] (profasm) ffi/should_run ffi002 [bad exit code] (threaded2) indexed-types/should_compile T7837 [stderr mismatch] (profasm) perf/compiler T4801 [stat too good] (normal) perf/haddock haddock.base [stat not good enough] (normal) perf/should_run T9203 [stat too good] (normal) profiling/should_run callstack001 [bad stdout] (prof) quasiquotation/qq007 qq007 [exit code non-0] (profasm) quasiquotation/qq008 qq008 [exit code non-0] (profasm) rts T5435_dyn_asm [bad stdout] (normal) rts T7919 [exit code non-0] (profasm,profthreaded) rts T9078 [exit code non-0] (profasm) rts derefnull [bad profile] (profasm,profthreaded) rts divbyzero [bad profile] (profasm,profthreaded) rts overflow1 [bad exit code] (hpc,optasm,profasm,threaded2,dyn,profthreaded,optllvm) rts overflow2 [bad profile] (profasm,profthreaded) rts overflow3 [bad profile] (profasm,profthreaded) simplCore/should_compile T5550 [exit code non-0] (profasm) simplCore/should_compile T7702 [exit code non-0] (profasm) }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9389> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 23

[GHC] #10937: "ghc -no-link --make A.hs -o foo" does something silly
by GHC 11 Dec '15

11 Dec '15

#10937: "ghc -no-link --make A.hs -o foo" does something silly -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: lowest | Milestone: Component: Compiler | Version: 7.10.2 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | -------------------------------------+------------------------------------- If I run "ghc -no-link --make A.hs -o foo" it puts the first object file it compiles into the `-o` location, and then fails because the object is not where is expected. {{{ [ezyang@hs01 ghc-fat-interface]$ ghc -no-link --make B.hs -o foo [1 of 2] Compiling A ( A.hs, A.o ) ./A.o: getFileStatus: does not exist (No such file or directory) }}} Without `-no-link` I get a more reasonable warning, and object files are not redirected. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10937> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1