July 2018 - ghc-tickets

[GHC] #14741: High-memory usage during compilation using Template Haskell
by GHC 09 Mar '19

09 Mar '19

#14741: High-memory usage during compilation using Template Haskell -------------------------------------+------------------------------------- Reporter: donatello | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.2.2 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: -------------------------------------+------------------------------------- When trying to embed some files into an executable using Template Haskell, I find that memory usage during compilation exceeds 4GB and often crashes my laptop. The files I am trying to embed are only about 25MB in size (totally 35MB in size). I made a somewhat minimal example to demonstrate this problem. To embed the files, I am using the [https://hackage.haskell.org/package/file-embed `file-embed`] package (the issue persists when using the alternative [https://hackage.haskell.org/package/wai-app-static `wai-app-static`] package too. The code to demonstrate runs in Linux and is available here - https://github.com/donatello/file-embed-exp. To try it out, just clone the repository and run Make (it uses the Haskell Stack tool and the Linux dd utility). This appear to be an issue in GHC. Is there anyway to mitigate the issue in the current version? Related discussion: https://github.com/snoyberg/file-embed/issues/24 -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14741> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 21

[GHC] #10069: CPR related performance issue
by GHC 09 Mar '19

09 Mar '19

#10069: CPR related performance issue -------------------------------------+------------------------------------- Reporter: pacak | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.1-rc2 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Runtime Unknown/Multiple | performance bug Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- By default CRP analysis can be too aggressive in trying to pass as much as possible in unboxed tuples, in general it's not a problem but when one big datatype is passed to several consumers it might end up pushed to stack several times instead of once - to heap, things are getting worse when there are sufficient fields to cause stack overflow which otherwise is possible to avoid - in our codebase adding one field with ExistentialQuantification (unused, but that prevents ghc from doing CRP transformation) reduces number of stack overflow by a factor of 1000 and increases overall performance by 10%. In provided example performance for both A and B should be identical and yet B is consistently faster by 3-5% It's possible to increase this performance gap by adding more and more fields. I was able to replicate this issue in ghc 7.8.3 and 7.10,1rc2 {{{#!hs {-# LANGUAGE ExistentialQuantification #-} module Blah where import Criterion import Criterion.Main import Data.Typeable data A = A () !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int data B = forall rep. (Typeable rep) => B rep !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int !Int a :: A a = A () 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 b :: B b = B () 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 {-# NOINLINE a1 #-} a1 :: A -> Int a1 (A _ f1 f2 f3 f4 f5 f6 f7 f8 g1 g2 g3 g4 g5 g6 g7 g8 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _) = f1 {-# NOINLINE a2 #-} a2 :: A -> Int a2 (A _ f1 f2 f3 f4 f5 f6 f7 f8 g1 g2 g3 g4 g5 g6 g7 g8 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _) = f2 {-# NOINLINE a3 #-} a3 :: A -> Int a3 (A _ f1 f2 f3 f4 f5 f6 f7 f8 g1 g2 g3 g4 g5 g6 g7 g8 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _) = f3 {-# NOINLINE a4 #-} a4 :: A -> Int a4 (A _ f1 f2 f3 f4 f5 f6 f7 f8 g1 g2 g3 g4 g5 g6 g7 g8 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _) = f4 {-# NOINLINE b1 #-} b1 :: B -> Int b1 (B _ f1 f2 f3 f4 f5 f6 f7 f8 g1 g2 g3 g4 g5 g6 g7 g8 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _) = f1 {-# NOINLINE b2 #-} b2 :: B -> Int b2 (B _ f1 f2 f3 f4 f5 f6 f7 f8 g1 g2 g3 g4 g5 g6 g7 g8 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _) = f2 {-# NOINLINE b3 #-} b3 :: B -> Int b3 (B _ f1 f2 f3 f4 f5 f6 f7 f8 g1 g2 g3 g4 g5 g6 g7 g8 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _) = f3 {-# NOINLINE b4 #-} b4 :: B -> Int b4 (B _ f1 f2 f3 f4 f5 f6 f7 f8 g1 g2 g3 g4 g5 g6 g7 g8 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _) = f4 {-# NOINLINE fa #-} fa :: A -> Int fa a = a1 a + a2 a + a3 a + a4 a {-# NOINLINE fb #-} fb :: B -> Int fb b = b1 b + b2 b + b3 b + b4 b main :: IO () main = defaultMain [ bgroup "single call" [ bench "A" $ whnf fa a , bench "B" $ whnf fb b ] ] }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10069> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 39

[GHC] #11441: RFC: Inline intermediate languages (Core, STG, Cmm, even StrictCore)
by GHC 08 Mar '19

08 Mar '19

#11441: RFC: Inline intermediate languages (Core, STG, Cmm, even StrictCore) -------------------------------------+------------------------------------- Reporter: Iceland_jack | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 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: -------------------------------------+------------------------------------- Allow intermediate code inline. == Inline ''Core'' == Typeable/WithTypeable: > {{{#!hs > withTypeable :: TTypeRep a -> (Typeable a => r) -> r > }}} > > Currently, `withTypeable` cannot be written in Haskell, but it is straightforward to write in Core. Can we fix this? [http://www.seas.upenn.edu/~sweirich/papers/wadlerfest2016.pdf A reflection on types] (''2016''): > We cannot implement ''withTypeable'' in Haskell source. But we ''can'' implement it in GHC's statically-typed intermediate language, System FC. The definition is simple, roughly like this: > > {{{#!hs > withTypeable tr k = k tr -- Not quite right > }}} > > Its second argument ''k'' expects a ''Typeable'' dictionary as its value argument. But since > a ''Typeable'' dictionary is represented by a ''TypeRep'', we can simply pass ''tr'' to ''k''. > When written in System FC there is a type-safe coercion to move from ''TypeRep a'' to ''Typeable a'', > but that coercion is erased at runtime. Since the definition can be statically type checked, > ''withTypeable'' does not form part of the trusted code base. == Inline ''Cmm'' == Edkward Kmett writes his own `.cmm` [https://github.com/ekmett/concurrent/blob/master/cbits/hashmap.cmm hither], [https://github.com/ekmett/concurrent/blob/master/cbits/array.cmm yon]. == Inline ''Strict Core'' == Should [http://research.microsoft.com/en-us/um/people/simonpj/papers /strict-core/tacc-hs09.pdf Types are calling conventions] (''2009'') be implemented: > === '''4.4 The ''`seq`'' function''' === > > A nice feature of Strict Core,,ANF,, is that it is possible to give a straightforward definition of the primitive ''seq'' function of Haskell: > {{{#!hs > seq : {a:* -> b:* -> {a} -> {b} -> b} > = { \a:*. \b:*. \x:{a}. \y:{b}. let _:a = x <> in y <> } > }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11441> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #11315: GHC doesn't restore split objects
by GHC 08 Mar '19

08 Mar '19

#11315: GHC doesn't restore split objects -------------------------------------+------------------------------------- Reporter: snowleopard | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 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: -------------------------------------+------------------------------------- While working on the new Shake-based build system I came across an annoying problem related to missing split objects ([#note see note]). I still don't fully understand the reasons behind the problem, but I found that GHC doesn't restore missing split objects, as explained in the example below. I think this is a bug, but maybe this is actually a missing feature. I think it shouldn't be difficult to fix. {{{#!hs module Test (test) where test :: Int test = 0 }}} I compile the above with `ghc -split-objs Test.hs` and get two split object files in `Test_o_split`. Then if I delete any of these files, or even the whole folder, and rerun `ghc -split-objs Test.hs` the compiler doesn't bother to restore split objects, pretending that everything is up- to-date. However, if I delete `Test.o` all object files are restored. Another related issue reported by Neil Mitchell (which may be worth a separate ticket) is that even if we corrupt `Test.o` and rerun `ghc -split-objs Test.hs`, the compiler doesn't restore the object file (nor missing split objects) to the consistent state. ---- [=#note Note:] The problem I have is described here: https://github.com/snowleopard/shaking-up-ghc/issues/30. It's not necessary to follow the link to understand this ticket, however I'd appreciate any insight you might have about the reasons behind the problem and/or a possible solution. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11315> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #9832: Get rid of PERL dependency of `ghc-split`
by GHC 08 Mar '19

08 Mar '19

#9832: Get rid of PERL dependency of `ghc-split` -------------------------------------+------------------------------------- Reporter: hvr | Owner: Type: task | Status: new Priority: normal | Milestone: 7.12.1 Component: Compiler | Version: Keywords: | Operating System: Windows Architecture: Unknown/Multiple | Type of failure: Difficulty: Unknown | None/Unknown Blocked By: | Test Case: Related Tickets: | Blocking: | Differential Revisions: -------------------------------------+------------------------------------- (writeme) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9832> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 37

[GHC] #8964: split_marker_entry assert breaks -split-objs and -ddump-opt-cmm
by GHC 08 Mar '19

08 Mar '19

#8964: split_marker_entry assert breaks -split-objs and -ddump-opt-cmm -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: lowest | Milestone: ⊥ Component: Compiler | Version: 7.9 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: Compile-time Difficulty: Easy (less than 1 | crash hour) | Test Case: Blocked By: | Blocking: Related Tickets: | -------------------------------------+------------------------------------- ...since the pretty-printer attempts to force the blocks list of the split_marker_entry proc. I'm not really sure what the right way to fix this is: just provide an empty block list, or hack the pretty-printer to treat split markers specially. But should be easy to fix. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8964> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #8685: -split-objs doesn't work for executables
by GHC 08 Mar '19

08 Mar '19

#8685: -split-objs doesn't work for executables ------------------------------------+------------------------------------- Reporter: refold | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Driver | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- While investigating [this bug report](https://github.com/haskell/cabal/issues/1611), I've noticed that the `-split-objs` option only works for libraries, but not for executables. That is, unlike in the case of libraries, the driver feeds only `A.o B.o ...` files to the linker instead of `A_split_0.o A_split_1.o ... B_split_0.o B_split_1.o ...`. It'd be nice if `-split-objs` could be changed to also work on executables. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8685> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #8629: Option 'split-objs' being ignored when trying to reduce object code size in iOS cross-compilation
by GHC 08 Mar '19

08 Mar '19

#8629: Option 'split-objs' being ignored when trying to reduce object code size in iOS cross-compilation ------------------------------------+--------------------------------- Reporter: f1rstmistake | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: ios | Operating System: MacOS X Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+--------------------------------- I have been trying to statically link Pandoc packages to iOS application following the guide in https://ghc.haskell.org/trac/ghc/wiki/Building/CrossCompiling/iOS and it's great so far! I have finally successfully use Pandoc library in iOS app, but the binary size is very big! 1. More than 50MB when using Pandoc library. 2. Around 5MB for simple haskell code. I try to add “split-objs: True” in .cabal/config file, but when running cabal I got this message: {{{ Preprocessing library pandoc-1.12.2.1... unused terminals: 1 on the commandline: Warning: ignoring -fsplit-objs }}} And if I ignore that, at the end of build, I get another error message that stop the build. {{{ cabal: dist-i386/build/Text/Pandoc_o_split: does not exist }}} I try with another package and got similar error message. It seems that split-objs not only being ignored by cross-compilation ghc, but will also fail the build. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8629> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8095: TypeFamilies painfully slow
by GHC 08 Mar '19

08 Mar '19

#8095: TypeFamilies painfully slow -------------------------------------+------------------------------------- Reporter: MikeIzbicki | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler (Type | Version: 7.6.3 checker) | Operating System: Unknown/Multiple Keywords: | Type of failure: Compile-time Architecture: Unknown/Multiple | performance bug Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: 5321 | -------------------------------------+------------------------------------- I'm using the TypeFamilies extension to generate types that are quite large. GHC can handle these large types fine when they are created manually, but when type families get involved, GHC's performance dies. Unlike in ticket #5321, using tail recursion does not eliminate the problem, and the order of arguments greatly affects compile time. I've attached a file Types.hs that demonstrates the problems. This file generates another Haskell file which has the problems. It takes 3 flags. The first is the size of the type to generate, the second is which type family function to use, and the third is whether to call the type family or just use a manually generated type. Here are my performance results: Using non-tail recursion, I get these results. I have to increase the stack size based on the size of the type I want to generate. {{{ $ ./Types 200 a a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=250 real 0m2.973s $ ./Types 300 a a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=350 real 0m6.018s $ ./Types 400 a a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=450 real 0m9.995s $ ./Types 500 a a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=550 real 0m15.645s }}} Tail recursion generates much slower compile times for some reason, and I still need to adjust the stack size: {{{ $ ./Types 200 b a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=250 real 0m16.120s }}} Changing the order of arguments to the recursive type family greatly changes the run times: {{{ $ ./Types 200 c a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=250 real 0m6.095s }}} Without the type family, I get MUCH better performance: {{{ $ ./Types 10000 a d > test.hs && time ghc test.hs > /dev/null real 0m2.271s }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8095> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 64

[GHC] #11715: There's something fishy going on with Constraint
by GHC 07 Mar '19

07 Mar '19

#11715: There's something fishy going on with Constraint -------------------------------------+------------------------------------- Reporter: bgamari | Owner: Type: bug | Status: new Priority: high | Milestone: 8.0.1 Component: Compiler | Version: 8.0.1-rc1 (Type checker) | 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 first noticed something was a bit odd with `Constraint` when chasing #11334. Consider this testcase, {{{#!hs import Data.Typeable main = do print $ typeRep (Proxy :: Proxy Eq) print $ typeOf (Proxy :: Proxy Eq) }}} With `ghc-8.0` this will produce, {{{ Eq Proxy (* -> *) Eq }}} Notice the second line; GHC seems to be claiming that `Eq :: * -> *`, which is clearly nonsense. I believe this issue may be the cause of some of the testsuite failures that I'm seeing on #11011. Unfortunately I haven't the foggiest where this issue might originate. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11715> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 82