- ghc-devs - Haskell.org

Re: Deriving mechanisms in GHC
by Ryan Scott 04 Jul '24

04 Jul '24

(Disclaimer: I have never used hs-to-coq before, so everything below is merely an educated guess.) Looking at the source code for hs-to-coq, I see this [1]: addDerivedInstances :: GhcMonad m => TypecheckedModule -> m > TypecheckedModule > addDerivedInstances tcm = do > let Just (hsgroup, a, b, c) = tm_renamed_source tcm > > (_gbl_env, inst_infos, _rn_binds) <- initTcHack tcm $ do > let tcg_env = fst (tm_internals_ tcm) > tcg_env_hack = tcg_env { tcg_mod = fakeDerivingMod, > tcg_semantic_mod = fakeDerivingMod } > -- Set the module to make it look like we are in GHCi > -- so that GHC will allow us to re-typecheck existing > instances > setGblEnv tcg_env_hack $ > #if __GLASGOW_HASKELL__ >= 810 > tcInstDeclsDeriv [] (hs_derivds hsgroup) > #else > tcInstDeclsDeriv [] (hs_tyclds hsgroup >>= group_tyclds) > (hs_derivds hsgroup) > #endif If I understand this code correctly, it is using the GHC API to produce deriving-generated code and typecheck it, which seems reasonable. The part that confuses me is the #if __GLASGOW_HASKELL__ >= 810 bit. Prior to GHC 8.10, this code would obtain deriving-related information from two places: 1. The deriving clauses of data type declarations (i.e., the hs_tyclds) 2. Standalone deriving declarations (i.e., the hs_derivds) In GHC 8.10 and later, however, the code only obtains deriving-related information from standalone deriving declarations. This means that you'll completely skip over any derived instances that arise from deriving clauses, which is likely the source of the trouble you're encountering. In order to give advice on what you *should* be doing, let me briefly describe why the type of tcInstDeclsDeriv changed in GHC 8.10. The commit that changed tcInstDeclsDeriv is [2]. Prior to that commit, tcInstDeclsDeriv would obtain information related to deriving clauses via its first two arguments: 1. The first argument (of type [DerivInfo]) contained all of the deriving clauses for data family instances. (Note that hs-to-coq uses an empty list, so this means that hs-to-coq will always skip over data family instances, before or after GHC 8.10. I'm not sure if this should be considered as a separate bug.) 2. The second argument (of type [LTyClDecl GhcRn]) contained all of the data type definitions, which might have deriving clauses attached to them. The linked commit changes things so that *all* deriving clause–related information (both for ordinary data types as well as data family instances) is passed as a [DerivInfo] list. This means that hs-to-coq needs to produce those DerivInfo values somehow. As inspiration, you might want to look at how GHC calls tcInstDeclsDeriv here [3]. GHC first calls the tcTyAndClassDecls function to produce the DerivInfo values, and then it passes the DerivInfo values to tcInstDeclsDeriv. I would hope that something similar would work for hs-to-coq. Best, Ryan ----- [1] https://github.com/plclub/hs-to-coq/blob/03e823972fc7c5f85a300e554c691563f8… [2] https://gitlab.haskell.org/ghc/ghc/-/commit/679427f878e50ba5a9981bac4c2f9c7… [3] https://gitlab.haskell.org/ghc/ghc/-/blob/bc1d435e399d8376b4e33d5d936424ff7…

3 3

Deriving mechanisms in GHC
by Yao Li 03 Jul '24

03 Jul '24

Dear All, My name is Yao Li. I am an assistant professor at Portland State University. I am currently working with my Ph.D. student Nicholas Coltharp on updating a tool called hs-to-coq (https://github.com/plclub/hs-to-coq) to support GHC 8.10 and the corresponding base library. However, we are running into some trouble involving the "deriving" mechanisms in GHC. It seems the derived type class instances are no longer generated by hs-to-coq after we updated GHC. I wonder what changes happened between GHC 8.4 and 8.10 that could possibly lead to this issue? Or if we hope to dig into this issue ourselves, what are some GHC resources that we should look into? Any pointers would be much appreciated! For context, hs-to-coq uses GHC's parser to parse Haskell code. Our translation is based on the source code but on the GhcRn pass. Thanks, Yao Li, Ph.D., (He/Him) Assistant Professor, Computer Science Department, Portland State University

1 0

CountDeps
by Simon Peyton Jones 02 Jul '24

02 Jul '24

Does anyone know what the CountDeps test does? I'm getting the failure below in my branch. Should I just accept it? I think it's because GHC.Core.FamInstEnv now depends on GHC.Builtin.Types.Literals, a very reasonable dependency Thanks Simon +++ "/builds/ghc/ghc/tmp/ghctest-su6yq239/test spaces/testsuite/tests/count-deps/CountDepsAst.run/CountDepsAst.run.stdout.normalised" 2024-07-01 18:37:33.372548372 +0000 @@ -2,6 +2,7 @@ GHC.Builtin.Names GHC.Builtin.PrimOps GHC.Builtin.Types +GHC.Builtin.Types.Literals GHC.Builtin.Types.Prim GHC.Builtin.Uniques GHC.ByteCode.Types --- "/builds/ghc/ghc/tmp/ghctest-su6yq239/test spaces/testsuite/tests/count-deps/CountDepsParser.run/CountDepsParser.stdout.normalised" 2024-07-01 18:37:33.393548751 +0000 +++ "/builds/ghc/ghc/tmp/ghctest-su6yq239/test spaces/testsuite/tests/count-deps/CountDepsParser.run/CountDepsParser.run.stdout.normalised" 2024-07-01 18:37:33.393548751 +0000 @@ -2,6 +2,7 @@ GHC.Builtin.Names GHC.Builtin.PrimOps GHC.Builtin.Types +GHC.Builtin.Types.Literals GHC.Builtin.Types.Prim GHC.Builtin.Uniques GHC.ByteCode.Types

2 1

[ANNOUNCE] GHC 9.6.6 is now available
by Zubin Duggal 01 Jul '24

01 Jul '24

The GHC developers are happy to announce the availability of GHC 9.6.6. Binary distributions, source distributions, and documentation are available on the [release page](https://www.haskell.org/ghc/download_ghc_9_6_6.html). Blog Post: https://www.haskell.org/ghc/blog/20240701-ghc-9.6.6-released.html This release is primarily a bugfix release addressing some issues found in the 9.6 series. These include: * A fix for a bug in the NCG that could lead to incorrect runtime results due to erroneously removing a jump instruction (#24507). * A fix for a linker error that manifested on certain platform/toolchain combinations, particularly darwin with a brew provisioned toolchain, arising due to a confusion in linker options between GHC and cabal (#22210). * A fix for a compiler panic in the simplifier due to incorrect eta expansion (#24718). * A fix for possible segfaults when using the bytecode interpreter due to incorrect constructor tagging (#24870). * And a few more fixes A full accounting of changes can be found in the [release notes]. As some of the fixed issues do affect correctness users are encouraged to upgrade promptly. We would like to thank Microsoft Azure, GitHub, IOG, the Zw3rk stake pool, Well-Typed, Tweag I/O, Serokell, Equinix, SimSpace, Haskell Foundation, and other anonymous contributors whose on-going financial and in-kind support has facilitated GHC maintenance and release management over the years. Finally, this release would not have been possible without the hundreds of open-source contributors whose work comprise this release. As always, do give this release a try and open a [ticket][] if you see anything amiss. Happy Hacking! -Zubin [ticket]: https://gitlab.haskell.org/ghc/ghc/-/issues/new [release notes]: https://downloads.haskell.org/~ghc/9.6.6/docs/users_guide/9.6.6-notes.html

1 0

Deriving via TH (#12457)
by Sebastian Graf 24 Jun '24

24 Jun '24

Hi Ryan and GHC devs, I'm working on-and-off on a prototype that enables use of TemplateHaskell as a deriving strategy: https://gitlab.haskell.org/ghc/ghc/-/commit/82aea77ed908fe36bed829c9c4a01ea…. My current working example is: ``` {-# LANGUAGE TemplateHaskell #-} {-# LANGUAGE TypeFamilies #-} module T12457A where import GHC.Internal.TH.Lib import GHC.Internal.TH.Syntax class C a where type Assoc a m :: a -> Int n :: a -> Int instance DeriveTH C where deriveTH _p head = do let AppT (ConT t) (VarT a) = head x <- newName "x" x2 <- newName "x" addTopDecls =<< [d| $(varP x) = 12 $(varP x2) = 23 |] [d| instance C $(varT a) => C ($(conT t) $(varT a)) where type Assoc ($(conT t) $(varT a)) = Char m :: Eq a => a -> Int m _ = $(varE x) + 42 {-# INLINE m #-} n :: Show a => a -> Int n _ = $(varE x2) + 13 |] --- {-# LANGUAGE TemplateHaskell #-} {-# LANGUAGE ConstraintKinds #-} {-# LANGUAGE DerivingStrategies #-} module T12457 where import Language.Haskell.TH import GHC.Internal.TH.Lib import T12457A newtype T a = T [a] deriving th C ``` I just managed to implement `GHC.Tc.Deriv.Infer.inferConstraints` for this mechanism (still hacky and broken in many ways) and am now stuck in `Deriv.genFamInst`. I realised I would need to replicate the first half of `tcClsInstDecl` to implement it. Before long, I will probably also need to replicate the other half to check method bindings. That leaves me wondering: Is it a good idea to integrate this new deriving strategy so tightly with the existing deriving framework? I would rather just call `tcClsInstDecl`, do a bit of sanity checking for specified constraints in standalone deriving declarations and call it a day. Given that you are the architect of our current deriving code, I hope you are the right person to ask for input. Thanks, Sebastian

2 3

HsCmdArrForm Fixity
by Alan & Kim Zimmerman 21 Jun '24

21 Jun '24

We are currently wrapping up https://gitlab.haskell.org/ghc/ghc/-/merge_requests/12842, which moves Fixity declarations into TTG extensions point, and into the haskell syntax, not GHC. But rnCmd for HsCommandArrForm in current master has rnCmd (HsCmdArrForm _ op _ (Just _) [arg1, arg2]) ... It matches on HsCmd GhcPs, which never introduce the `Just` form for fixity. The only place the `Just` form is introduced, is in mkOpFormRn, but this is only ever called from this specific equation. So both can be removed, without changing behaviour. The question is, this code was added for a reason at some point, what has broken in the meantime? Should this be fixed somehow instead of being removed? Alan

2 1

TTG and annotations
by Simon Peyton Jones 20 Jun '24

20 Jun '24

Hi Alan I know that you are continuing to work hard on annotations in the TTG extension fields -- thank you. As you know, I'm very eager to get a big Note to explain how it all works. At the moment I am completely lost, and I am reduced to cargo-culting code by copy/pasting and hoping for the best. Plus we are over-reliant on you, which isn't really fair on you. Several of us (including me and Rodrigo, and doubtless other ghc-devs, cc'd) would be happy to review, edit, and improve -- but it is hard for us to see the "big picture". Could you see your way to drafting something? It need not be that long initially, but having *something* to ask questions about and improve would be super helpful. I suggest drafting it in a Google doc so it was easy to edit, and for others to contribute to, but ultimately it should be in the repo. Thanks! Simon

2 1

embedded hashes in library names
by Shayne Fletcher 14 Jun '24

14 Jun '24

when i use ghcup to install a distribution, the libraries that result have embedded hashes in their names e.g. the -f942 in libHSghc-prim-0.11.0-f942-ghc9.10.1.so. is there some way to disable this so that for example in this case, i'd get libHSghc-prim-0.11.0-ghc9.10.1.so? -- Shayne Fletcher

1 1

A dashboard for fragile tests
by Bryan Richter 10 Jun '24

10 Jun '24

Hi all, This weekend at Zurihac, I wanted to start making it possible to explore fragile tests. Fragile tests are tests that pass nondeterministically. Marking them fragile means that they do not influence the overall success of the testsuite. It's been a way of sweeping problems under the rug. For a few years, the GHC test infrastructure has been recording fragile test results into a database. We record whether the test passed or failed. Now we can start peeking under the rug to see what kind of patterns have developed. I was mostly interested in whether or not fragile tests are truly fragile. It looks like many are not. The dashboard has bucketed results into months. Every cell shows the pass rate for a given month (column) and test (row). Red cells mean 0% success. Green cells, 100% success. Shades of yellow are everything in between. It's remarkable how much green there is. One would assume even fragile tests would fail for legitimate reasons sometimes! At this point, the visualization is mostly good for seeing large trends. Any other observations would be appreciated! The dashboard is interactive at https://grafana.gitlab.haskell.org/goto/XcuO3ZUIg?orgId=2 [image: image.png] -Bryan

1 0

Changes to HIE File due to ApplicativeDo going via HsExpansions
by Ingle, Apoorv N 09 Jun '24

09 Jun '24

Dear ghc-devs, For past few months I’ve been working on Rebindable Syntax<https://gitlab.haskell.org/ghc/ghc/-/wikis/Rebindable-syntax>, specifically on expanding ApplicativeDo (#24406, MR !12066<https://gitlab.haskell.org/ghc/ghc/-/merge_requests/12066>) statements before type checking. I noticed that there is a failing test case related to HIE files (T23540). I have been unsuccessful in debugging why this happens and/or if this has any effect on the users of HIE files. Can anyone who has experience in this area be able to take a look and guide me? Thanks! Apoorv

2 1