January 2016 - ghc-tickets

[GHC] #9253: "internal error: evacuate(static): strange closure type 0" when running in QEMU
by GHC 24 Apr '16

24 Apr '16

#9253: "internal error: evacuate(static): strange closure type 0" when running in QEMU ------------------------------------+------------------------------------- Reporter: ocharles | Owner: simonmar Type: bug | Status: new Priority: normal | Milestone: Component: Runtime System | Version: 7.8.2 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: Runtime crash Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- We're trying to use NixOS to automatically run our acceptance tests in a VM on commit, in order to guarantee system stability. However, we're periodically seeing our web server component, compiled with GHC 7.8.2 is failing to start up. When it fails, the error message is always: {{{ internal error: evacuate(static): strange closure type 0 (GHC version 7.8.2 for x86_64_unknown_linux) Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} Happy to do more debugging on this, but I'll need to be pointed in the right direction. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9253> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #10725: Figure out how to support type synonym implementions of abstract data
by GHC 21 Apr '16

21 Apr '16

#10725: Figure out how to support type synonym implementions of abstract data -------------------------------------+------------------------------------- Reporter: ezyang | Owner: ezyang Type: task | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.2 (Type checker) | Keywords: backpack | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- A classic example we have used when describing Backpack is that you might want to abstract over is different types of strings: String, ByteString, Text, etc. Unfortunately, with every implementation of Backpack we have had so far, you can't actually do this! For example, say you use a type synonym to get an existing data type to the "right name"; you get this error: {{{ A.hs-boot:2:6: Type constructor `T' has conflicting definitions in the module and its hs-boot file Main module: type T = String Boot file: abstract(False) T No C type associated RecFlag NonRecursive FamilyInstance: none }}} How can we make this work? Could it be as simple as just relaxing the restriction here? This seems to work OK with shaping, so long as we refer to the type synonym (and not the original type in question!) If some type checker experts could weight in that would be quite useful. ---- Probably the biggest danger is interaction with machinery that works with type families. For example, this is accepted: {{{ -- A.hs-boot module A where data T -- B.hs {-# LANGUAGE TypeFamilies #-} module B where import {-# SOURCE #-} qualified A data T = T type family F a :: * type instance F A.T = Bool type instance F T = Int }}} This doesn't lead to unsoundness today, because A.T must always be defined in A.hs (and thus have a unique nominal identity). But it's dangerous for this to be accepted today with signatures, since T could refer to B.T. (The moral of the story is that after shaping, you have to recheck all pairs of families; you can't assume that because they were typechecked earlier that they're OK.) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10725> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #8524: GHC is inconsistent with the Haskell Report on which Unicode characters are allowed in string and character literals
by GHC 18 Apr '16

18 Apr '16

#8524: GHC is inconsistent with the Haskell Report on which Unicode characters are allowed in string and character literals ----------------------------+---------------------------------------------- Reporter: oerjan | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: GHC rejects valid program Unknown/Multiple | Test Case: Difficulty: Unknown | Blocking: Blocked By: | Related Tickets: | ----------------------------+---------------------------------------------- GHC is inconsistent with the Haskell Report on which Unicode characters are allowed in string and character literals. (And I don't like either option, why leave out any characters in strings unnecessarily?) Examples from ghci 7.6.3 (also tested in lambdabot on irc): {{{ Prelude> "" -- Unicode char \8203, Format class. <interactive>:10:2: lexical error in string/character literal at character '\8203' Prelude> " " -- Unicode char \8202, Space class. "\8202" Prelude> "t\ \est" -- Unicode char \8202 in a string gap. <interactive>:14:4: lexical error in string/character literal at character '\8202' }}} My reading of http://www.haskell.org/onlinereport/haskell2010/haskellch2.html (section 2.2 and 2.6): * The report BNF token "graphic", which can be used in literals, includes indirectly many Unicode classes, but uniWhite is not one of them. Thus the only Unicode whitespace allowed to represent itself in literals is ASCII space. * Unicode formatting characters are not mentioned in the BNF that I can see, so are not allowed in literals. * String gaps are made out of the report BNF token whitespace, which ''does'' include uniWhite. Who wants what: || ||= GHC =||= Report =||= Me =|| || Format in string || No || No || Yes || || Space/uniWhite in string || Yes || No || Yes || || Space/uniWhite in string gap || No || Yes || Dunno || In short, GHC's behavior is buggy and/or annoying in two opposite ways: * It leaves out some Unicode characters as allowable in strings and character literals, presumably because the report says so. * It allows some characters the report says it ''shouldn't'', and refuses some characters the report says it ''should''. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8524> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #11180: A program writing to a read-only stdout should not succeed
by GHC 18 Apr '16

18 Apr '16

#11180: A program writing to a read-only stdout should not succeed -------------------------------------+------------------------------------- Reporter: thomie | Owner: Type: bug | Status: new Priority: normal | 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): | Wiki Page: -------------------------------------+------------------------------------- In the shell, redirection of file descriptors can be read-only (<), write- only (>), or read-write (<>). A program trying to write to stdout, while stdout is redirected read-only, currently exits with exit code 0. {{{ $ cat hello.hs main = print "hello world" $ ghc hello.hs ... $ ./hello 1< /dev/null; echo Exit status: $? Exit status: 0 }}} This should however show some error message, and return some non-zero error message, because the write never succeeds. Note that this example is different from `./hello 1> /dev/null`, which rightfully does succeed. As suggested by this [http://roscidus.com/blog/blog/2013/06/09/choosing-a -python-replacement-for-0install/ blog post]: > "If you’re not sure why this is important, imagine the command is `dump-database > backup` and the filesystem is full." Supposedly OCaml gets this right. For reference, from my bash man page: > *Redirecting Input* > > Redirection of input causes the file whose name results from the expansion of word to be opened for reading on file descriptor n, or the standard input (file descriptor 0) if n is not specified. > > The general format for redirecting input is: > > [n]<word -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11180> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #10914: Bad symbol resolution on Darwin when using DYLD_FORCE_FLAT_NAMESPACE=1
by GHC 16 Apr '16

16 Apr '16

#10914: Bad symbol resolution on Darwin when using DYLD_FORCE_FLAT_NAMESPACE=1 -----------------------------------------+------------------------------- Reporter: jacereda | Owner: Type: bug | Status: new Priority: high | Milestone: Component: GHCi | Version: 7.10.2 Keywords: | Operating System: MacOS X Architecture: x86_64 (amd64) | Type of failure: GHCi crash Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -----------------------------------------+------------------------------- Starting ghci as: {{{ env DYLD_FORCE_FLAT_NAMESPACE=1 ghci }}} can cause symbol resolution to fail. I tried with Data.Hashable as follows: {{{ import Data.Hashable main = putStrLn $ show $ hashWithSalt 0 "foo" }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10914> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 13

[GHC] #11318: Data.Text.length allocates one closure per character
by GHC 15 Apr '16

15 Apr '16

#11318: Data.Text.length allocates one closure per character -------------------------------------+------------------------------------- Reporter: bgamari | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Runtime Unknown/Multiple | performance bug Test Case: | Blocked By: Blocking: | Related Tickets: #11284 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- In #11284 I noticed that `map Data.Text.length` applied to a list of `Text` values would result in code which would allocate one closure per character due to lack of lambda lifting. It turns out it's even worse than this. Consider this example (inspired by #11284), {{{#!hs module Hi2 (hello) where import qualified Data.Text as T hello :: Int hello = T.length hi hi :: T.Text hi = T.pack "hello" }}} When compiled with GHC 7.10.3 with `-O`, the following simplified Core is produced, {{{#!hs hello :: Int hello = case unpackCString# "hello"# of _ { Text dt_a33D dt1_a33E dt2_a33F -> let { a_a33C :: Int# a_a33C = +# dt1_a33E dt2_a33F } in letrec { $wloop_length_s3bt :: Int# -> Int# -> Int# $wloop_length_s3bt = \ (ww_s3bk :: Int#) (ww_s3bo :: Int#) -> case tagToEnum# @ Bool (>=# ww_s3bo a_a33C) of _ { False -> case indexWord16Array# dt_a33D ww_s3bo of r#_a35n { __DEFAULT -> case tagToEnum# @ Bool (geWord# r#_a35n (__word 55296)) of _ { False -> $wloop_length_s3bt (+# ww_s3bk 1) (+# ww_s3bo 1); True -> case tagToEnum# @ Bool (leWord# r#_a35n (__word 56319)) of _ { False -> $wloop_length_s3bt (+# ww_s3bk 1) (+# ww_s3bo 1); True -> $wloop_length_s3bt (+# ww_s3bk 1) (+# ww_s3bo 2) } } }; True -> ww_s3bk }; } in case $wloop_length_s3bt 0 dt1_a33E of ww_s3bs { __DEFAULT -> I# ww_s3bs } } }}} Even this simple example produces a `length` function which allocations -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11318> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #10318: Cycles in class declaration (via superclasses) sometimes make sense.
by GHC 13 Apr '16

13 Apr '16

#10318: Cycles in class declaration (via superclasses) sometimes make sense. -------------------------------------+------------------------------------- Reporter: ekmett | Owner: Type: feature | Status: new request | Milestone: Priority: normal | Version: 7.10.1 Component: Compiler | Operating System: Unknown/Multiple (Type checker) | Type of failure: GHC rejects Keywords: | valid program Architecture: | Blocked By: Unknown/Multiple | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- I'd like to be able to say the following, to describe the notion of an integral domain in Haskell: {{{ -- | Product of non-zero elements always non-zero. -- Every integral domain has a field of fractions. -- The field of fractions of any field is itself. class (Frac (Frac a) ~ Frac a, Fractional (Frac a), IntegralDomain (Frac a)) => IntegralDomain a where type Frac a :: * embed :: a -> Frac a instance IntegralDomain Integer where type Frac Integer = Rational embed = fromInteger instance IntegralDomain Rational where type Frac Rational = Rational embed = id }}} But GHC gets scared when it sees the cyclic reference that `IntegralDomain` instances depend on an IntegralDomain superclass, which really is cyclic in the (Frac a) case here, and that is kind of the point. =) Right now the best approximation of the correct answer that I have for this situation is to lie and claim the constraint is weaker: {{{ -- | Product of non-zero elements always non-zero class (Frac (Frac a) ~ Frac a, Fractional (Frac a)) => AlmostIntegralDomain a where type Frac a :: * embed :: a -> Frac a class (AlmostIntegralDomain a, AlmostIntegralDomain (Frac a)) => IntegralDomain a instance (AlmostIntegralDomain a, AlmostIntegralDomain (Frac a)) => IntegralDomain a instance AlmostIntegralDomain Integer where type Frac Integer = Rational embed = fromInteger instance AlmostIntegralDomain Rational where type Frac Rational = Rational embed = id }}} Now the user is stuck defining a different class than the one they consume. Alternately, with `ConstraintKinds`, I can encode: {{{ data Dict p where Dict :: p => Dict p class (Frac (Frac a) ~ Frac a, Fractional (Frac a)) => IntegralDomain a where type Frac a :: * embed :: a -> Frac a proofFracIsIntegral :: p a -> Dict (IntegralDomain (Frac a)) default proofFracIsIntegral :: IntegralDomain (Frac a) => p a -> Dict (IntegralDomain (Frac a)) proofFracIsIntegral _ = Dict }}} but now whenever I need to get from `IntegralDomain a` to `IntegralDomain (Frac a)` I need to explicitly open the `proofFracIsIntegral` with a rats' nest of `ScopedTypeVariables`. It would be really really nice if I could get GHC to deal with this for me as I currently have a few thousand lines of code hacking around this limitation. =/ -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10318> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 16

[GHC] #11213: Incorrect reported pattern synonym signature
by GHC 12 Apr '16

12 Apr '16

#11213: Incorrect reported pattern synonym signature -------------------------------------+------------------------------------- Reporter: mpickering | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 Keywords: | Operating System: Unknown/Multiple PatternSynonyms | Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- {{{ data T where MkT :: (Show b) => b -> T --pattern ExNumPat :: () => Show b => b -> T pattern ExNumPat x = MkT x }}} GHC reports that `ExNumPat` is missing a signature (correctly) but it reports the wrong type. The correct type is the one commented out. {{{ pstest.hs:12:1: warning: Top-level binding with no type signature: ExNumPat :: forall b. Show b => b -> T }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11213> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #10001: GHC crash trying to build a project within Nix-shell
by GHC 11 Apr '16

11 Apr '16

#10001: GHC crash trying to build a project within Nix-shell -------------------------------------+------------------------------------- Reporter: wolftune | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.4 Keywords: | Operating System: Linux Architecture: x86_64 | Type of failure: Compile-time (amd64) | crash Test Case: | Blocked By: Blocking: | Related Tickets: 9825 Differential Revisions: | -------------------------------------+------------------------------------- {{{ <no location info>: ghc: panic! (the 'impossible' happened) (GHC version 7.8.4 for x86_64-unknown-linux): Loading temp shared object failed: /run/user/1000/ghc19915_0/ghc19915_20.so: failed to map segment from shared object Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug [20 of 99] Compiling Settings.StaticFiles ( Settings/StaticFiles.hs, dist/build/Settings/StaticFiles.o ) <no location info>: ghc: panic! (the 'impossible' happened) (GHC version 7.8.4 for x86_64-unknown-linux): Loading temp shared object failed: /run/user/1000/ghc19915_0/ghc19915_21.so: failed to map segment from shared object }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10001> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 7

[GHC] #11463: Template Haskell applies too many arguments to kind synonym
by GHC 10 Apr '16

10 Apr '16

#11463: Template Haskell applies too many arguments to kind synonym -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Template | Version: 8.0.1-rc1 Haskell | Keywords: TypeInType, | Operating System: Unknown/Multiple TypeApplications | Architecture: | Type of failure: Incorrect result Unknown/Multiple | at runtime Test Case: | Blocked By: Blocking: | Related Tickets: #11376 Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Running the following code: {{{#!hs {-# LANGUAGE TemplateHaskell #-} {-# LANGUAGE TypeInType #-} module IdStarK where import Data.Kind import Language.Haskell.TH type Id a = a data Proxy (a :: Id k) = Proxy $(return []) main :: IO () main = do putStrLn $(reify ''Proxy >>= stringE . pprint) putStrLn $(reify ''Proxy >>= stringE . show) }}} Gives a result I wouldn't have expected: {{{ $ /opt/ghc/head/bin/runghc IdStarK.hs data IdStarK.Proxy (a_0 :: IdStarK.Id * k_1) = IdStarK.Proxy TyConI (DataD [] IdStarK.Proxy [KindedTV a_1627394516 (AppT (AppT (ConT IdStarK.Id) StarT) (VarT k_1627394515))] Nothing [NormalC IdStarK.Proxy []] []) }}} From the output, it appears that `Id` is being applied to ''two'' arguments, both `*` and `k`! Perhaps this indirectly (or directly) a consequence of #11376? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11463> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6