- ghc-devs - Haskell.org

Looking for maintainer: ghc-exactprint
by Alan & Kim Zimmerman 13 May '26

13 May '26

Hi all I have come to the sad realisation that I am not actually part of the Haskell ecosystem any more. I joined at a time of my life when I was doing management work with lots of free time, and needed something to occupy my brain. Haskell and contributing to GHC did that for me in spades. But for the last few years I have been working as an individual contributor, in a non-Haskell ecosystem. So my available time and motivation for Haskell has fallen off. I have always been interested in developer tooling, and refactoring in particular. My efforts to restore the Haskell Refactorer (HaRe) led, after many false starts, to implementing the Exact Print Annotations in GHC (using Trees That Grow), and the companion library to make use of it, ghc-exactprint. This library works in lockstep with GHC releases, so for every GHC (major) release there needs to be a new ghc-exactprint release. I have been doing this every release since the first one in 2015, and it is something that needs to happen for every release. Over the last few years I have got the implementation to work the way I believe it should, and that it is now stable, so any changes going forward should just be tweaks to manage the inevitable changes to the language. I follow a well-defined process[1] for each new GHC version. My plan is to do this one more time, for the impending GHC 10.0, as the branch has now been cut. During/after that I would like to hand over to some other person(s) to take responsibility for this. The sanest approach might be to move the repo fully into GHC, much like Haddock. There is one more piece I have not been able to complete. I presented a built-in CPP at the last HIW at ZuriHac[2]. Afterwards, I completed and polished the branch[3], but ran out of steam to actually land it, as it would take a lot of consistent effort to get all parties to agree on whether it is even needed, and then to land it, deal with unintended consequences, documentation, and all the myriad other things that go with adding a new major feature. The itch I was scratching with it is to be able to cleanly handle CPP in ghc-exactprint, but it has benefits for other users too, in terms of performance, and being focused only on the actual directives, so it does not munge any unrelated code that just happens to look like a CPP directive. I would be happy to do a handover/collaboration with anyone interested in taking that over. Regards Alan [1] https://github.com/alanz/ghc-exactprint/blob/master/NEW-GHC.md [2] https://alanz.github.io/hiw2025/ [3] https://gitlab.haskell.org/ghc/ghc/-/tree/wip/az/ghc-cpp

1 0

Hadrian has broken my workflow
by Simon Peyton Jones 02 May '26

02 May '26

Friends Something extremely tiresome has happened to Hadrian. When I'm compiling GHC, and I get a type error, Hadrian helpfully tells me the command line that broke. I often copy/paste that command line to re-run it, for two reasons: - It is much much faster than running Hadrian from scratch, as I repeatedly try to compile the same source file. - More important, when building libraries or the stage2 compiler, I often want to add lots of debug or tracing flags to the command line, so I can find what is wrong with the stage1 compiler. This is utterly essential. - Indeed I often make a deliberate syntax error in (say) a library module, run Hadrian, watch it fall over with that syntax error, so I get the command line. Then I fix the deliberate error and repeatedly compile that one file with lots of tracing. But now something has broken. If I re-run the command I get simonpj@PF55P26D:~/code/HEAD-31$ /home/simonpj/.ghcup/bin/ghc -Wall -hisuf hi -osuf o -hcsuf hc -static -hide-all-packages -no-user-package-db '-package-env -' '-package-db _build/stage0/inplace/package.conf.d' '-this-unit-id ghc-10.1-inplace' '-this-package-name ghc' '-package-id array-0.5.8.0-9044' '-package-id base-4.21.0.0-ae91' '-package-id binary-0.8.9.3-25e5' '-package-id bytestring-0.12.2.0-7dd0' '-package-id containers-0.7-5a8f' '-package-id deepseq-1.5.1.0-6837' '-package-id directory-1.3.10.1-inplace' '-package-id exceptions-0.10.9-4d2e' '-package-id filepath-1.5.4.0-inplace' '-package-id ghc-boot-10.1-inplace' '-package-id ghc-boot-th-next-10.1-inplace' '-package-id ghc-heap-9.12.2-4363' '-package-id ghc-toolchain-0.1.0.0-inplace' '-package-id ghci-10.1-inplace' '-package-id hpc-0.7.0.2-inplace' '-package-id os-string-2.0.8-inplace' '-package-id process-1.6.26.1-inplace' '-package-id rts-1.0.2' '-package-id semaphore-compat-1.0.0-inplace' '-package-id stm-2.5.3.1-d35b' '-package-id text-2.1.2-ecdb' '-package-id time-1.15-inplace' '-package-id transformers-0.6.1.2-38b1' '-package-id unix-2.8.8.0-inplace' -i -i/home/simonpj/code/HEAD-31/_build/stage0/compiler/build -i/home/simonpj/code/HEAD-31/_build/stage0/compiler/build/autogen -i/home/simonpj/code/HEAD-31/compiler/ -I_build/stage0/compiler/build -I_build/stage0/compiler/build/. -Icompiler/. -I/home/simonpj/code/HEAD-31/libraries/process/include -I/home/simonpj/code/HEAD-31/_build/stage0/libraries/process/build/include -I/home/simonpj/code/HEAD-31/libraries/directory -I/home/simonpj/code/HEAD-31/_build/stage0/libraries/directory/build -I/home/simonpj/code/HEAD-31/libraries/unix/include -I/home/simonpj/code/HEAD-31/_build/stage0/libraries/unix/build/include -I/home/simonpj/code/HEAD-31/libraries/time/lib/include -I/home/simonpj/code/HEAD-31/_build/stage0/libraries/time/build/lib/include -I/home/simonpj/.ghcup/ghc/9.12.2/lib/ghc-9.12.2/lib/x86_64-linux-ghc-9.12.2-a828/bytestring-0.12.2.0-7dd0/include -I/home/simonpj/.ghcup/ghc/9.12.2/lib/ghc-9.12.2/lib/x86_64-linux-ghc-9.12.2-a828/ghc-internal-9.1202.0-1465/include -I/home/simonpj/.ghcup/ghc/9.12.2/lib/ghc-9.12.2/lib/x86_64-linux-ghc-9.12.2-a828/ghc-bignum-1.3-6db4/include -I/home/simonpj/.ghcup/ghc/9.12.2/lib/ghc-9.12.2/lib/x86_64-linux-ghc-9.12.2-a828/rts-1.0.2/include -optP-include -optP_build/stage0/compiler/build/autogen/cabal_macros.h -optP-DCAN_LOAD_DLL -outputdir _build/stage0/compiler/build -fdiagnostics-color=always -Wall -Wno-name-shadowing -Wnoncanonical-monad-instances -Wnoncanonical-monoid-instances -XGHC2024 -XNoImplicitPrelude -optc-Wno-error=inline -c -o _build/stage0/compiler/build/GHC/Core.o -DDEBUG -haddock -fno-warn-deprecated-flags -O +RTS -O64M -RTS -O2 @/tmp/extra-file-9558430904709-115211-0 ghc-9.12.2: panic! (the 'impossible' happened) GHC version 9.12.2: processArgs: no arg yet It's that pesky "@" thing. And of course GHC has deleted that tmp file. Super frustrating. What can I do? Do we have to use the pesky @ thing? It's just complicating things! ALSO, *all the error messages come out twice*, which is disconcerting and tiresome. See below. Thanks Simon simonpj@PF55P26D:~/code/HEAD-31$ hadrian/build -j6 --flavour=perf+assertions_stage1+no_profiled_libs+ticky_ghc+lint Warning: Specifying an absolute path to the project file is deprecated. Use --project-dir to set the project's directory. Configuration is affected by cabal.project at '/home/simonpj/code/HEAD-31/hadrian'. Up to date Warning: Specifying an absolute path to the project file is deprecated. Use --project-dir to set the project's directory. Configuration is affected by cabal.project at '/home/simonpj/code/HEAD-31/hadrian'. ]0;Starting...(/home/simonpj/code/HEAD-31) | Run Ghc CompileHs (Stage0 InTreeLibs): compiler/GHC/Core.hs => _build/stage0/compiler/build/GHC/Core.o ]0;Running for 5s [9912/37741], predicted 4m26s (1%)(/home/simonpj/code/HEAD-31) compiler/GHC/Core.hs:1680:41: error: [GHC-88464] Variable not in scope: mkTick :: CoreTickish -> CoreExpr -> CoreExpr Suggested fix: Perhaps use data constructor ‘Tick’ (line 270) | 1680 | wrapFloat (FloatTick t) body = mkTick t body | ^^^^^^ compiler/GHC/Core.hs:1682:41: error: [GHC-88464] Variable not in scope: mkSingleAltCase :: CoreExpr -> Id -> AltCon -> [Var] -> CoreExpr -> CoreExpr | 1682 | wrapFloat (FloatCase e b con bs) body = mkSingleAltCase e b con bs body | ^^^^^^^^^^^^^^^ ]0;Finished in 5.74s(/home/simonpj/code/HEAD-31) Development.Shake.cmd, system command failed Command line: /home/simonpj/.ghcup/bin/ghc -Wall -hisuf hi -osuf o -hcsuf hc -static -hide-all-packages -no-user-package-db '-package-env -' '-package-db _build/stage0/inplace/package.conf.d' '-this-unit-id ghc-10.1-inplace' '-this-package-name ghc' '-package-id array-0.5.8.0-9044' '-package-id base-4.21.0.0-ae91' '-package-id binary-0.8.9.3-25e5' '-package-id bytestring-0.12.2.0-7dd0' '-package-id containers-0.7-5a8f' '-package-id deepseq-1.5.1.0-6837' '-package-id directory-1.3.10.1-inplace' '-package-id exceptions-0.10.9-4d2e' '-package-id filepath-1.5.4.0-inplace' '-package-id ghc-boot-10.1-inplace' '-package-id ghc-boot-th-next-10.1-inplace' '-package-id ghc-heap-9.12.2-4363' '-package-id ghc-toolchain-0.1.0.0-inplace' '-package-id ghci-10.1-inplace' '-package-id hpc-0.7.0.2-inplace' '-package-id os-string-2.0.8-inplace' '-package-id process-1.6.26.1-inplace' '-package-id rts-1.0.2' '-package-id semaphore-compat-1.0.0-inplace' '-package-id stm-2.5.3.1-d35b' '-package-id text-2.1.2-ecdb' '-package-id time-1.15-inplace' '-package-id transformers-0.6.1.2-38b1' '-package-id unix-2.8.8.0-inplace' -i -i/home/simonpj/code/HEAD-31/_build/stage0/compiler/build -i/home/simonpj/code/HEAD-31/_build/stage0/compiler/build/autogen -i/home/simonpj/code/HEAD-31/compiler/ -I_build/stage0/compiler/build -I_build/stage0/compiler/build/. -Icompiler/. -I/home/simonpj/code/HEAD-31/libraries/process/include -I/home/simonpj/code/HEAD-31/_build/stage0/libraries/process/build/include -I/home/simonpj/code/HEAD-31/libraries/directory -I/home/simonpj/code/HEAD-31/_build/stage0/libraries/directory/build -I/home/simonpj/code/HEAD-31/libraries/unix/include -I/home/simonpj/code/HEAD-31/_build/stage0/libraries/unix/build/include -I/home/simonpj/code/HEAD-31/libraries/time/lib/include -I/home/simonpj/code/HEAD-31/_build/stage0/libraries/time/build/lib/include -I/home/simonpj/.ghcup/ghc/9.12.2/lib/ghc-9.12.2/lib/x86_64-linux-ghc-9.12.2-a828/bytestring-0.12.2.0-7dd0/include -I/home/simonpj/.ghcup/ghc/9.12.2/lib/ghc-9.12.2/lib/x86_64-linux-ghc-9.12.2-a828/ghc-internal-9.1202.0-1465/include -I/home/simonpj/.ghcup/ghc/9.12.2/lib/ghc-9.12.2/lib/x86_64-linux-ghc-9.12.2-a828/ghc-bignum-1.3-6db4/include -I/home/simonpj/.ghcup/ghc/9.12.2/lib/ghc-9.12.2/lib/x86_64-linux-ghc-9.12.2-a828/rts-1.0.2/include -optP-include -optP_build/stage0/compiler/build/autogen/cabal_macros.h -optP-DCAN_LOAD_DLL -outputdir _build/stage0/compiler/build -fdiagnostics-color=always -Wall -Wno-name-shadowing -Wnoncanonical-monad-instances -Wnoncanonical-monoid-instances -XGHC2024 -XNoImplicitPrelude -optc-Wno-error=inline -c -o _build/stage0/compiler/build/GHC/Core.o -DDEBUG -haddock -fno-warn-deprecated-flags -O +RTS -O64M -RTS -O2 @/tmp/extra-file-80099975488219-115449-0 Exit code: 1 Stderr: compiler/GHC/Core.hs:1680:41: error: [GHC-88464] Variable not in scope: mkTick :: CoreTickish -> CoreExpr -> CoreExpr Suggested fix: Perhaps use data constructor ‘Tick’ (line 270) | 1680 | wrapFloat (FloatTick t) body = mkTick t body | ^^^^^^ compiler/GHC/Core.hs:1682:41: error: [GHC-88464] Variable not in scope: mkSingleAltCase :: CoreExpr -> Id -> AltCon -> [Var] -> CoreExpr -> CoreExpr | 1682 | wrapFloat (FloatCase e b con bs) body = mkSingleAltCase e b con bs body | ^^^^^^^^^^^^^^^ Build failed.

1 1

gitlab downtime this morning
by Magnus Viernickel 27 Apr '26

27 Apr '26

Dear GHC Devs, as you may have noticed, the gitlab went down this morning. Properly this time, so that's new, hurray! Since I am kind of running out of options, I have reinstated Anubis again. Anubis is a a service that sits in front of our gitlab and lets users solve a proof of work challenge to obtain a cookie. It doesn't really solve the issue but it adds some throttling on the client side and results look promising for now; if performance degrades again, I'll have to look for another solution. In the meantime I will also start looking into more long term solutions, with the Haskell infra team, the current state of affairs is not sustainable, as you might agree. If you have any painpoints with Anubis, please share these. I know that this is a further UX degradation and if possible I would like to keep it as small as possible. Best Magnus

1 0

GitLab
by Simon Peyton Jones 27 Apr '26

27 Apr '26

Friends Is it just me or is GHC's GitLab down? I'm getting 502 (bad gateway). Simon

2 1

Gitlab currently throwing a lot of 502s
by Magnus Viernickel 23 Apr '26

23 Apr '26

Hello GHC devs, we're seeing a lot of 502s at the moment, we're getting twice the rate of requests than normal, I am investigating what is going on and trying to fix this. Best Magnus

4 6

Short emergency gitlab downtime to apply some fixes
by Magnus Viernickel 20 Apr '26

20 Apr '26

Hi, due to the increased number of 502s caused by out of memory kills, I had to do a quick adjustment that should hopefully improve the situation. This causes the usual 5 minutes downtime of the gitlab container rebooting. Thank you for your patience. Best Magnus

1 0

Reinstallable base failure -- help
by Simon Peyton Jones 16 Apr '26

16 Apr '26

Friends I'm working on !15719, which is a big step towards making `base` reinstallable. But in one build I'm getting a strange failure, below, just in one build x86_64-linux-fedora43-release https://gitlab.haskell.org/ghc/ghc/-/jobs/2484333 Now, it's likely that this failure is due to my patch. *But I can't reproduce it locally*. How can I repo? I tried hadrian/build --flavour=release but that didn't fall over. Thanks! Simon * Raised the exception: Development.Shake.cmd, system command failed Command line: _build/stage1/bin/haddock --verbosity=0 -B_build/stage1/lib --lib=_build/stage1/lib --odir=_build/doc/html/libraries/ghc-internal-9.1500.0 --dump-interface=_build/doc/html/libraries/ghc-internal-9.1500.0/ghc-internal.haddock --html '--title=ghc-internal-9.1500.0: Basic libraries' --prologue=_build/doc/html/libraries/ghc-internal-9.1500.0/haddock-prologue.txt --optghc=-D__HADDOCK_VERSION__=2300 --hide=GHC.Internal.IO.Handle.Lock.Common --hide=GHC.Internal.IO.Handle.Lock.Flock --hide=GHC.Internal.IO.Handle.Lock.LinuxOFD --hide=GHC.Internal.IO.Handle.Lock.NoOp --hide=GHC.Internal.IO.Handle.Lock.Windows --hide=GHC.Internal.Event.Arr --hide=GHC.Internal.Event.Array --hide=GHC.Internal.Event.Internal --hide=GHC.Internal.Event.Internal.Types --hide=GHC.Internal.Event.IntTable --hide=GHC.Internal.Event.IntVar --hide=GHC.Internal.Event.PSQ --hide=GHC.Internal.Event.Unique --hide=GHC.Internal.Unicode.Bits --hide=GHC.Internal.Unicode.Char.DerivedCoreProperties --hide=GHC.Internal.Unicode.Char.UnicodeData.GeneralCategory --hide=GHC.Internal.Unicode.Char.UnicodeData.SimpleLowerCaseMapping --hide=GHC.Internal.Unicode.Char.UnicodeData.SimpleTitleCaseMapping --hide=GHC.Internal.Unicode.Char.UnicodeData.SimpleUpperCaseMapping --hide=GHC.Internal.System.Environment.ExecutablePath --hide=GHC.Internal.Bignum.Backend.GMP --hide=GHC.Internal.Event.Control --hide=GHC.Internal.Event.EPoll --hide=GHC.Internal.Event.KQueue --hide=GHC.Internal.Event.Manager --hide=GHC.Internal.Event.Poll --hide=GHC.Internal.Event.Thread --hide=GHC.Internal.Event.TimerManager --optghc=-hisuf --optghc=dyn_hi --optghc=-osuf --optghc=dyn_o --optghc=-hcsuf --optghc=dyn_hc --optghc=-fPIC --optghc=-dynamic --optghc=-hide-all-packages --optghc=-no-user-package-db '--optghc=-package-env -' '--optghc=-this-unit-id ghc-internal-9.1500.0-7cec' '--optghc=-this-package-name ghc-internal' '--optghc=-package-id rts-1.0.3' --optghc=-i --optghc=-i/builds/ghc/ghc/_build/stage1/libraries/ghc-internal/build --optghc=-i/builds/ghc/ghc/_build/stage1/libraries/ghc-internal/build/autogen --optghc=-i/builds/ghc/ghc/libraries/ghc-internal/src --optghc=-Irts/include --optghc=-I_build/stage1/libraries/ghc-internal/build --optghc=-I_build/stage1/libraries/ghc-internal/build/include --optghc=-Ilibraries/ghc-internal/include --optghc=-I/builds/ghc/ghc/rts/include --optghc=-I/builds/ghc/ghc/_build/stage1/rts/build/include --optghc=-I/builds/ghc/ghc/libraries/libffi-clib/src/x86 --optghc=-I/builds/ghc/ghc/_build/stage1/libraries/libffi-clib/build/src/x86 --optghc=-I/builds/ghc/ghc/libraries/libffi-clib/include --optghc=-I/builds/ghc/ghc/_build/stage1/libraries/libffi-clib/build/include --optghc=-optP-include --optghc=-optP_build/stage1/libraries/ghc-internal/build/autogen/cabal_macros.h --optghc=-optP-DBIGNUM_GMP --optghc=-outputdir --optghc=_build/stage1/libraries/ghc-internal/build --optghc=-package-env=- --optghc=-this-unit-id --optghc=ghc-internal --optghc=-frebindable-known-key-names --optghc=-Wcompat --optghc=-Wnoncanonical-monad-instances --optghc=-XHaskell2010 --optghc=-XNoImplicitPrelude --optghc=-no-global-package-db --optghc=-package-db=/builds/ghc/ghc/_build/stage1/inplace/package.conf.d --optghc=-ghcversion-file=rts/include/ghcversion.h --optghc=-ghcversion-file=rts/include/ghcversion.h --optghc=-package-env=- --optghc=-this-unit-id --optghc=ghc-internal --optghc=-frebindable-known-key-names --optghc=-Wcompat --optghc=-Wnoncanonical-monad-instances --optghc=-XHaskell2010 --optghc=-XNoImplicitPrelude --optghc=-no-global-package-db --optghc=-package-db=/builds/ghc/ghc/_build/stage1/inplace/package.conf.d --optghc=-ghcversion-file=rts/include/ghcversion.h --optghc=-ghcversion-file=rts/include/ghcversion.h --optghc=-Wno-deprecated-flags --optghc=-Wno-trustworthy-safe +RTS -t_build/stage1/haddock-timing-files/ghc-internal.t --machine-readable -RTS --hyperlinked-source --hoogle --quickjump @/tmp/extra-file-93748803258994-2958-12008 Exit code: 1 Stderr: libraries/ghc-internal/src/GHC/Internal/Types.hs:1:1: error: [GHC-58427] attempting to use module ‘ghc-internal-9.1500.0:GHC.Internal.Types’ (libraries/ghc-internal/src/GHC/Internal/Types.hs) which is not loaded | 1 | {-# LANGUAGE Trustworthy #-} -- So that we can import in into Safe modules | ^ Build failed.

4 6

Inplace libraries
by Simon Peyton Jones 15 Apr '26

15 Apr '26

Dear devs I have thirty or so builds of GHC on my disk. Sometimes I want to use one build to build and install (for that build alone) a bunch of libraries. If I do cabal install hspec -w $HOME/code/HEAD-22/_build/stage1/bin/ghc then Cabal rightly warns me Warning: The libraries were installed by creating a global GHC environment file at: /home/simonpj/.ghc/x86_64-linux-9.15.20260309/environments/default The presence of such an environment file is likely to confuse or break other tools because it changes GHC's behaviour: it changes the default package set in ghc and ghci from its normal value (which is "all boot libraries"). GHC environment files are little-used and often not tested for. Question: how can I install the libraries in the build tree for $HOME/code/HEAD-22? After all, I think ghc-internal, base etc are all in that build-tree. Surely hspec can be too? But how? Thanks! Simon

4 8

Gitlab API and web rate limiting for unauthenticated users
by Magnus Viernickel 14 Apr '26

14 Apr '26

Dear GHC devs, as part of my effort to increase the usability of our gitlab instance, I have started employing rate limiting for unauthenticated API requests. The majority of the traffic on the machine and the reason why you experience slowness is disrespectful crawlers that hit expensive API endpoints (hundreds of requests per second look at random commits). Previously they were allowed to do this multiple times per second for hours and hours. Now they get a more modest but still very generous budget. If you're using the API unauthenticated with scripts and hit rate limits, please consider if you really have to hammer it that hard. If you really can't get around sending so many requests, you may use an API token. Authenticated requests are currently almost not rate limited at all. If this turns out to be to impairing for users, I will increase the limits. Best Magnus

1 0

Rotation of Marge's SSH key
by Magnus Viernickel 13 Apr '26

13 Apr '26

Hello GHC developers, For security reasons we have rotated Marge-Bot's SSH key. If you notice any issues that could have been caused by this change, please do not hesitate to reach out. Best Magnus

1 0