December 2017 - ghc-tickets

[GHC] #14105: ApplicativeDo causes GHC panic on irrefutable list pattern match
by GHC 06 Mar '18

06 Mar '18

#14105: ApplicativeDo causes GHC panic on irrefutable list pattern match -------------------------------------+------------------------------------- Reporter: BoteboTsebo | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.2.1 Keywords: ApplicativeDo | Operating System: Unknown/Multiple Architecture: x86_64 | Type of failure: None/Unknown (amd64) | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- So this just happened: {{{#!hs {-# language ApplicativeDo #-} module Main where main :: IO () main = do [_] <- pure [] pure () }}} Compiling (under Stack) with `stack exec -- ghc --make src/Main.hs` yields {{{ [1 of 1] Compiling Main ( src/Main.hs, src/Main.o ) ghc: panic! (the 'impossible' happened) (GHC version 8.2.1 for x86_64-unknown-linux): isStrictPattern Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} This does not happen on GHC-8.0.2. It does not happen when `ApplicativeDo` is not activated. It also does not happen when the pattern match line is changed to {{{#!hs [] <- pure [] }}} For completeness, the compiler used by `stack` is that which comes with `nightly-2017-08-10`. I am on Linux x64 (Ubuntu). -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14105> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #12158: ghc: panic! (the 'impossible' happened) translateConPatVec: lookup
by GHC 05 Mar '18

05 Mar '18

#12158: ghc: panic! (the 'impossible' happened) translateConPatVec: lookup -------------------------------------+------------------------------------- Reporter: wozgonon | Owner: Type: bug | Status: new Priority: normal | Milestone: 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: -------------------------------------+------------------------------------- *Main> price Stock{name=name,ric=ric,price=price} = price ghc: panic! (the 'impossible' happened) (GHC version 8.0.1 for x86_64-unknown-mingw32): translateConPatVec: lookup Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12158> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 18

[GHC] #13193: Integer (gmp) performance regression?
by GHC 05 Mar '18

05 Mar '18

#13193: Integer (gmp) performance regression? -------------------------------------+------------------------------------- Reporter: j.waldmann | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Linux Architecture: x86_64 | Type of failure: Runtime (amd64) | performance bug Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- For a simple program that uses Integer, but actual data would fit in an Int, ghc-8.0.2 seems to have more runtime overhead than ghc-6.12.3 had: http://www.imn.htwk-leipzig.de/~waldmann/etc/mob/ I am not really sure what to make of it - an answer could be: "if you know that your Integer fits in an Int, then just use Int". -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13193> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 13

[GHC] #8772: ghci should safe history more often
by GHC 05 Mar '18

05 Mar '18

#8772: ghci should safe history more often ----------------------------------------------+---------------------------- Reporter: nomeata | Owner: Type: feature request | Status: new Priority: normal | Milestone: Component: GHCi | Version: 7.6.3 Keywords: | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Moderate (less than a day) | Type of failure: Blocked By: | None/Unknown Related Tickets: | Test Case: | Blocking: ----------------------------------------------+---------------------------- It is very annoying if `ghci` crashes (due to using a foreign library wrongly, for example), and then the history is lost. It should be simple to save each line to history before it is being executed. Maybe a good task for new contributors? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8772> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #12935: Object code produced by GHC is non-deterministic
by GHC 05 Mar '18

05 Mar '18

#12935: Object code produced by GHC is non-deterministic -------------------------------------+------------------------------------- Reporter: bgamari | Owner: Type: bug | Status: new Priority: normal | Milestone: 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: -------------------------------------+------------------------------------- This is a continuation of #4012, which made interface files deterministic. Now someone needs to make object code deterministic. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12935> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 22

[GHC] #11984: Pattern match incompleteness / inaccessibility discrepancy
by GHC 02 Mar '18

02 Mar '18

#11984: Pattern match incompleteness / inaccessibility discrepancy -------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.1 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: -------------------------------------+------------------------------------- Consider this module: {{{ {-# LANGUAGE PolyKinds, TypeOperators, DataKinds, TypeFamilies, GADTs #-} module Bug where data family Sing (a :: k) data Schema = Sch [Bool] data instance Sing (x :: Schema) where SSch :: Sing x -> Sing ('Sch x) data instance Sing (x :: [k]) where SNil :: Sing '[] SCons :: Sing a -> Sing b -> Sing (a ': b) data G a where GCons :: G ('Sch (a ': b)) eval :: G s -> Sing s -> () eval GCons s = case s of -- SSch SNil -> undefined SSch (SCons _ _) -> undefined }}} Upon seeing this, GHC says {{{ Bug.hs:21:9: warning: [-Wincomplete-patterns] Pattern match(es) are non-exhaustive In a case alternative: Patterns not matched: (SSch SNil) }}} So I uncomment the second-to-last line, inducing GHC to say {{{ Bug.hs:22:16: error: • Couldn't match type ‘a : b’ with ‘'[]’ Inaccessible code in a pattern with constructor: SNil :: forall k. Sing '[], in a case alternative • In the pattern: SNil In the pattern: SSch SNil In a case alternative: SSch SNil -> undefined }}} Thankfully, this pattern is much rarer than it once was, but it's a bit sad that it's still possible. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11984> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 14

[GHC] #14098: Incorrect pattern match warning on nested GADTs
by GHC 02 Mar '18

02 Mar '18

#14098: Incorrect pattern match warning on nested GADTs -------------------------------------+------------------------------------- Reporter: jmgrosen | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.2.1 Keywords: | Operating System: Unknown/Multiple PatternMatchWarnings | Architecture: | Type of failure: Incorrect Unknown/Multiple | error/warning at compile-time Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- {{{#!hs {-# Language GADTs #-} data App f a where App :: f a -> App f (Maybe a) data Ty a where TBool :: Ty Bool TInt :: Ty Int data T f a where C :: T Ty (Maybe Bool) -- Warning f :: T f a -> App f a -> () f C (App TBool) = () -- No warning g :: T f a -> App f a -> () g C (App x) = case x of TBool -> () }}} When compiling, with `-Wincomplete-patterns`: {{{ Foo.hs:15:1: warning: [-Wincomplete-patterns] Pattern match(es) are non-exhaustive In an equation for ‘f’: Patterns not matched: C (App TInt) | 15 | f C (App TBool) = () | ^^^^^^^^^^^^^^^^^^^^ }}} I'm sorry for such a complicated example, but I wasn't able to reduce it any further than this. The gist of the problem is that this code gives a pattern matching non- exhaustiveness warning when matching a nested pattern, when pulling out a value then matching on it produces no warning (correctly). It also seems to have to do with using a type constructor (`Maybe`) within the constructor definition of `App`, as changing it to {{{#!hs data App f a where App :: f a -> App f a ... data T f a where C :: T Ty Bool }}} does not give a warning, even when `App` is modified further to force it to be a proper GADT. This might be a known limitation of the checker, but given that it works fine when the nesting is removed, I would think it would be more of a bug. Thanks to Iavor for helping me minimize the test case. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14098> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 10

[GHC] #14544: doCorePass has at least one missing case
by GHC 02 Mar '18

02 Mar '18

#14544: doCorePass has at least one missing case -------------------------------------+------------------------------------- Reporter: mpickering | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.2.1 Keywords: newcomer | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- There is a catch all at the end of `doCorePass` which hides the fact that `doCorePass` should be total. In particular `doCorePass` is not implemented for `CoreOccurAnal` and so either `CoreOccurAnal` should be removed from `CoreToDo` or implemented properly in `doCorePass`. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14544> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #12706: Collecting type info is slow
by GHC 02 Mar '18

02 Mar '18

#12706: Collecting type info is slow -------------------------------------+------------------------------------- Reporter: vshabanov | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: GHCi | 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: -------------------------------------+------------------------------------- With `:set +c` option in GHCi, I'm waiting for 10-15 seconds for {{{ Collecting type info for 34 module(s) ... }}} after both first load and subsequent reloads, even if only top-level module was modified. And GHCi uses about 1.5-2x more memory than it uses without `:set +c`. My app has about 30k lines of code. Is it expected behavior and some performance improvements are planned or it's a bug? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12706> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #12790: GHC 8.0.1 uses copious amounts of RAM and time when trying to compile lambdabot-haskell-plugins
by GHC 02 Mar '18

02 Mar '18

#12790: GHC 8.0.1 uses copious amounts of RAM and time when trying to compile lambdabot-haskell-plugins -------------------------------------+------------------------------------- Reporter: clint | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Compile-time Unknown/Multiple | performance bug Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- When building lambdabot-haskell-plugins, Lambdabot.Plugin.Haskell.Pl.Parser compiles without incident on the non- profiling run, but cripples the build machine on the profiling run. An excerpt of -v3 output is below to show the number of terms skyrocketing. *** Common sub-expression [Lambdabot.Plugin.Haskell.Pl.Parser]: Result size of Common sub-expression = {terms: 39,049, types: 30,851, coercions: 990} !!! Common sub-expression [Lambdabot.Plugin.Haskell.Pl.Parser]: finished in 72.00 milliseconds, allocated 44.536 megabytes *** Float inwards [Lambdabot.Plugin.Haskell.Pl.Parser]: Result size of Float inwards = {terms: 39,049, types: 30,851, coercions: 990} !!! Float inwards [Lambdabot.Plugin.Haskell.Pl.Parser]: finished in 68.00 milliseconds, allocated 60.337 megabytes *** Simplifier [Lambdabot.Plugin.Haskell.Pl.Parser]: Result size of Simplifier iteration=1 = {terms: 977,045, types: 299,334, coercions: 1,058} Result size of Simplifier iteration=2 = {terms: 926,036, types: 263,306, coercions: 932} Result size of Simplifier = {terms: 922,724, types: 262,202, coercions: 932} !!! Simplifier [Lambdabot.Plugin.Haskell.Pl.Parser]: finished in 15480.00 milliseconds, allocated 9384.823 megabytes *** CoreTidy [Lambdabot.Plugin.Haskell.Pl.Parser]: Result size of Tidy Core = {terms: 922,623, types: 261,789, coercions: 865} !!! CoreTidy [Lambdabot.Plugin.Haskell.Pl.Parser]: finished in 2060.00 milliseconds, allocated 613.585 megabytes writeBinIface: 95 Names writeBinIface: 259 dict entries *** CorePrep [Lambdabot.Plugin.Haskell.Pl.Parser]: Result size of CorePrep = {terms: 1,128,145, types: 347,063, coercions: 865} !!! CorePrep [Lambdabot.Plugin.Haskell.Pl.Parser]: finished in 3468.00 milliseconds, allocated 1718.453 megabytes *** Stg2Stg: *** CodeGen [Lambdabot.Plugin.Haskell.Pl.Parser]: -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12790> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 20