January 2015 - ghc-tickets

[GHC] #10017: signal handlers are invoked multiple times when the threaded rts is used
by GHC 29 Jun '15

29 Jun '15

#10017: signal handlers are invoked multiple times when the threaded rts is used -------------------------------------+------------------------------------- Reporter: redneb | Owner: simonmar Type: bug | Status: new Priority: normal | Milestone: Component: Runtime | Version: 7.10.1-rc1 System | Operating System: Linux Keywords: | Type of failure: Incorrect result Architecture: | at runtime Unknown/Multiple | Blocked By: Test Case: | Related Tickets: #9423 Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- When you install a custom signal handler and the threaded rts is being used, then the signal handler will be invoked multiple times. Here's a program that the demonstrates this: {{{#!hs import Control.Concurrent import System.Posix.Signals main :: IO () main = do _ <- flip (installHandler sig) Nothing $ Catch $ putStrLn $ "Received signal " ++ show sig raiseSignal sig threadDelay 100000 where sig = sigUSR2 }}} If you compile this with the `-threaded` flag and then run it with say `+RTS -N4` then it produces the following output: {{{ Received signal 12 Received signal 12 Received signal 12 Received signal 12 Received signal 12 }}} In general the signal handler is invoked `n_capabilities+1` times. This also happens with all other signals. This regression was introduced by f9f89b7884ccc8ee5047cf4fffdf2b36df6832df (which was later [changeset:4748f5936fe72d96edfa17b153dbfd84f2c4c053 reverted] but then [changeset:7e658bc14e2dd6baf208deebbdab9e1285ce4c72 re- added]), a commit addressing #9423. The cause of the problem is [source:/rts/posix/Signals.c@f44bbc83bab62f9a2d25e69d87c2b4af25318d52#L256 this] loop. I don't understand why we need to write an event about the signal received in the per capability control pipe introduced by the aforementioned commit. Aren't these control pipes supposed only to be used to shutdown the capabilities (which happens [source:/rts/posix/Signals.c@f44bbc83bab62f9a2d25e69d87c2b4af25318d52#L183 here])? Removing the loop seems to solve the issue, but I don't know if it makes #9423 reappear. I cannot test this on a Mac OS X right now. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10017> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #8144: Interface hashes include time stamp of included .h file when CPP is used
by GHC 25 Jun '15

25 Jun '15

#8144: Interface hashes include time stamp of included .h file when CPP is used -------------------------+------------------------------------------------- Reporter: nh2 | Owner: Type: bug | Status: new Priority: | Milestone: normal | Version: 7.6.3 Component: | Operating System: Unknown/Multiple Compiler | Type of failure: Compile-time performance bug Keywords: | Test Case: Architecture: | Blocking: Unknown/Multiple | Difficulty: | Unknown | Blocked By: | Related Tickets: | -------------------------+------------------------------------------------- See https://github.com/nh2/ghc-bug-time-dependent-interface-hashes for a fast test case. Having '''Test.hs''': {{{ {-# LANGUAGE CPP #-} module Test () where }}} and an empty '''cabal_macros.h''', {{{ touch cabal_macros.h ghc -c Test.hs -optP-include -optPcabal_macros.h }}} will generate you a different interface hash ''most of the time'' in '''ghc --show-iface Test.hi'''. It looks like the modification time of the .h file makes it into the interface hash, but only at second resolution, and you get ''compilation IS NOT required'' if you manage to run it twice per second. I think the interface hash should be independent from the time stamp of the header file. This bug can trigger unnecessary recompilation and makes compilations with identical inputs non-deterministic. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8144> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 34

[GHC] #8785: Replace hooks API in the RTS with something better
by GHC 16 Jun '15

16 Jun '15

#8785: Replace hooks API in the RTS with something better ------------------------------------+------------------------------------- Reporter: simonmar | Owner: simonmar Type: task | Status: new Priority: high | Milestone: 7.10.1 Component: Runtime System | 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: | ------------------------------------+------------------------------------- Hooks rely on static linking behaviour which doesn't always work: we have to disable -Bsymbolic for the RTS on Linux (see `compiler/main/SysTools.lhs`) and it apparently doesn't work at all on Mac (#8754). So instead of hooks we should be passing in information when we initialize the RTS, like we already do for some other things (`-rtsopts` etc.). -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8785> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #9991: runghc shows the value produced by main when its type is a non-() Show instance
by GHC 14 Jun '15

14 Jun '15

#9991: runghc shows the value produced by main when its type is a non-() Show instance -------------------------------------+------------------------------------- Reporter: shachaf | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.4 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Incorrect result Unknown/Multiple | at runtime Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- When running a program like this: {{{#!hs main :: IO Int main = return 5 }}} With `runghc`, it'll print a value, which is contrary to expected behavior and to compiled program behavior: {{{ $ runghc Main.hs 5 $ ghc Main.hs [1 of 1] Compiling Main ( Main.hs, Main.o ) Linking Main ... $ ./Main $ }}} This is the same as `ghci`'s behavior with an IO action, so presumably `runghc` is taking a shortcut here. But a program should behave the same way in both cases. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9991> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #9769: ghc-7.8.3-x86_64-unknown-mingw32.tar.xz missing User's Guide
by GHC 14 Jun '15

14 Jun '15

#9769: ghc-7.8.3-x86_64-unknown-mingw32.tar.xz missing User's Guide -------------------------------------+------------------------------------- Reporter: Dangthrimble | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Documentation | Version: 7.8.3 Keywords: | Operating System: Windows Architecture: x86_64 (amd64) | Type of failure: Difficulty: Unknown | Documentation bug Blocked By: | Test Case: Related Tickets: | Blocking: | Differential Revisions: -------------------------------------+------------------------------------- Context: I am relatively new to Haskell and am trying to build a Haskell IDE around `Sublime Text`. I have been using `Haskell Platform` but have encountered a number of problems using `cabal`. Consequently I have been advised to uninstall `Haskell Platform` and download `ghc`. I have just downloaded and extracted `ghc-7.8.3-x86_64-unknown- mingw32.tar.xz`. Hoping for advice on setting up `ghc`, I have opened `ghc-7.8.3\doc\html\index.html` in Firefox and clicked on "The User's Guide". Unfortunately the `ghc-7.8.3/doc/html/users_guide` folder is missing. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9769> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8591: Concurrent executions of ghc-pkg can cause inconstant package.cache files
by GHC 14 Jun '15

14 Jun '15

#8591: Concurrent executions of ghc-pkg can cause inconstant package.cache files ------------------------------------+---------------------------- Reporter: janm | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Package system | Version: 7.6.3 Keywords: ghc-pkg race | Operating System: FreeBSD Architecture: Unknown/Multiple | Type of failure: Other Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+---------------------------- I am doing 24 way parallel builds of system images, including all packages on a system. This includes ghc and multiple ghc packages. I am seeing intermittent dependency failure from the ghc packaging system. From examining Main.hs in ghc-pkg, I see the function withFileAtomic write to a temporary file in package.conf.d and then atomically rename on top of a target file, package.cache in the case. With parallel execution the last rename would win, leading to lost entries in package.cache. In my case, the following things happened: ("Building" indicates a start, "Built" indicates completion, "Installing" is setup in a separate chroot'd environment and is isolated) The FreeBSD ports system is used to drive the Haskell build system. The process works single threaded and fails intermittently when done in parallel. Building: devel/hs-data-default-instances-base Building: devel/hs-data-default-instances-containers Building: devel/hs-data-default-instances-old-locale Built: devel/hs-dlist Building: devel/hs-data-default-instances-dlist Built: devel/hs-temporary Built: jail-image-full Installing: system-image__jail-image-full Built: devel/hs-base64-bytestring Built: archivers/hs-zlib Building: security/hs-digest Built: devel/hs-syb Building: textproc/hs-hs-bibutils Building: textproc/hs-pandoc-types Built: devel/hs-utf8-string Built: devel/hs-data-default-instances-old-locale Built: devel/hs-data-default-instances-containers Built: devel/hs-data-default-instances-base Built: devel/hs-data-default-instances-dlist Building: devel/hs-data-default Built: devel/hs-random Installed: system-image__lang/ghc Installing: system-image__archivers/hs-zlib Installing: system-image__devel/hs-utf8-string Installing: system-image__devel/hs-syb Installing: system-image__devel/hs-base64-bytestring Installing: system-image__devel/hs-data-default-class Installing: system-image__devel/hs-dlist Installing: system-image__devel/hs-random Installing: system-image__devel/hs-temporary Installing: system-image__devel/hs-extensible-exceptions Built: devel/hs-data-default FAILED The error from the Haskell data-default build was: setup: At least the following dependencies are missing: data-default-instances-base -any Looking in the in the package.conf.d directory shows that the data- default-instances-base-0.0.1-7bdf8678f0d8637e096e397e7910f82a.conf file was present, but running "ghc-pkg list" did not show data-default- instances-base Running /usr/local/lib/cabal/ghc-7.6.3/data-default-instances- base-0.0.1/register.sh (which was also present) caused ghc-pkg to now show data-default-instances-base. To me this looks like a race condition between multiple instances of ghc- pkg causing the cache to become inconsistent. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8591> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 11

[GHC] #9511: Remove deprecated -fglasgow-exts from NoFib suite
by GHC 14 Jun '15

14 Jun '15

#9511: Remove deprecated -fglasgow-exts from NoFib suite -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: Type: task | Status: new Priority: low | Milestone: Component: NoFib benchmark | Version: 7.8.3 suite | Operating System: Keywords: | Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: Other Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Revisions: -------------------------------------+------------------------------------- Someone on #ghc noted that replacing -fglasgow-exts with only the necessary options in each case may change compile times, which may be undesirable. I think a reasonable workaround would be to just put all the options on the command line, whether necessary or not. This will increase the time needed to parse the command lines, but presumably by a trivial amount. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9511> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

[GHC] #8435: Do not copy stack after stack overflow
by GHC 12 Jun '15

12 Jun '15

#8435: Do not copy stack after stack overflow ------------------------------+-------------------------------------------- Reporter: ezyang | Owner: simonmar Type: bug | Status: new Priority: normal | Milestone: Component: Runtime | Version: 7.7 System | Operating System: Unknown/Multiple Keywords: | Type of failure: Runtime performance bug Architecture: | Test Case: Unknown/Multiple | Blocking: Difficulty: Unknown | Blocked By: | Related Tickets: | ------------------------------+-------------------------------------------- I am not sure, but I was doing a close reading of `threadStackOverflow` and noticed that after we throw a stack overflow exception, we fall through to the code responsible for allocating a new stack and copying the code over. An old iteration of the stack overflow code (removed in the commit cited below) did return after throwing the exception, and I did not see anything in the commit message suggesting the change was intentional. It seems sound to copy the stack; it will just become dead immediately. The relevant commit: {{{ commit f30d527344db528618f64a25250a3be557d9f287 Author: Simon Marlow <marlowsd(a)gmail.com> Date: Wed Dec 15 12:08:43 2010 +0000 Implement stack chunks and separate TSO/STACK objects }}} Suggested (untested) patch: {{{ diff --git a/rts/Threads.c b/rts/Threads.c index 742119d..ccd6b17 100644 --- a/rts/Threads.c +++ b/rts/Threads.c @@ -533,6 +533,7 @@ threadStackOverflow (Capability *cap, StgTSO *tso) // Send this thread the StackOverflow exception throwToSingleThreaded(cap, tso, (StgClosure *)stackOverflow_closure); + return; } }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8435> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 7

[GHC] #9746: tests/ghci/scripts/T8172 is sensitive to drive name on Windows
by GHC 11 Jun '15

11 Jun '15

#9746: tests/ghci/scripts/T8172 is sensitive to drive name on Windows -------------------------------------+------------------------------------- Reporter: gintas | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Test Suite | Version: 7.8.3 Keywords: | Operating System: Windows Architecture: Unknown/Multiple | Type of failure: Difficulty: Unknown | None/Unknown Blocked By: | Test Case: Related Tickets: | Blocking: | Differential Revisions: -------------------------------------+------------------------------------- The test fails when the checkout is not on drive C:, like this: --- ./ghci/scripts/T8172.stdout-mingw32 2014-10-28 23:28:12.264410700 +0100 +++ ./ghci/scripts/T8172.run.stdout 2014-10-30 15:20:13.964595800 +0100 @@ -1,11 +1,11 @@ current working directory: - C:\ + K:\ module import search paths: . current working directory: - C:\ + K:\ module import search paths: none current working directory: - C:\ + K:\ module import search paths: / -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9746> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #9252: Generalize hs-boot files to be more like module signatures
by GHC 11 Jun '15

11 Jun '15

#9252: Generalize hs-boot files to be more like module signatures ------------------------------------+------------------------------------- Reporter: ezyang | Owner: ezyang Type: feature request | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.2 Keywords: backpack | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- Right now, `hs-boot` is used solely to support mutually recursive modules. However, as [wiki:Backpack Backpack] has shown, signatures are an essential part of separate modular development, and since hs-boot files are essentially module signatures, it would be good to move away from "hs- boot as a technical mechanism to make mutual recursion to work" to "hsig as a module signature useful for separate modular development and mutual recursion." Here are some of the things that this would entail: 1. Introduce hsig as a valid alternate file suffix to hs-boot (similarly have hisig for hi-boot; the plan being to eventually deprecate the hs-boot prefix) 2. Given an hsig file corresponding to a module in the current package database (but not from the currently being compiled package), check if the corresponding module implements the signature (this is the reverse of the current check we do for hs-boot files, which occurs when we typecheck the module, as opposed to the hs-boot file.) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9252> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9