October 2022 - ghc-devs

Reducing all type family applications as far as possible
by ÉRDI Gergő 28 Oct '22

28 Oct '22

Hi, Is there a way to have GHC reduce all (reducable) type familiy applications? For example, if I have ``` type family Id x where Id x = x foo :: Bool -> Id Bool foo x = x ``` then I would like the `Id` of `foo` to have type `Bool -> Bool`, and the generated Core should be either `foo = \x -> x` or, if that's asking for too much, at least something like `foo = \x -> x |> (somePossiblyOvercomplicatedCoercion :: Bool ~ Bool)`. Is there already some functionality for this in GHC? If not, how messy would it be to implement it as a separate `CoreProgram` transformation just after desugaring? It'd need to be immediately after desugaring because I would like e.g. the specialiser to already specialise in terms of these reduced types. Thanks, Gergo

1 0

Exact Print Annotations Next Steps
by Alan & Kim Zimmerman 20 Oct '22

20 Oct '22

I have put down some thoughts about simplifying the exact print annotations, prior to extending them to support printing the `RenamedSource`. See https://gitlab.haskell.org/ghc/ghc/-/issues/22319 Alan

1 0

GHC.Pack deprecation process: Step 1
by Hécate 18 Oct '22

18 Oct '22

Hi everyone, As discussed earlier this year¹, the deprecation process for GHC.Pack starts with the 9.6.1 milestone². ¹ https://gitlab.haskell.org/ghc/ghc/-/issues/21461 ² https://gitlab.haskell.org/ghc/ghc/-/issues/21541 Cheers, Hécate -- Hécate ✨ 🐦: @TechnoEmpress IRC: Hecate WWW: https://glitchbra.in RUN: BSD

1 0

Release Timelines
by Matthew Pickering 18 Oct '22

18 Oct '22

Hi all, Three releases are currently cooking. * 9.2.5 (End of October) - Fixes to performance regressions introduced in 9.2.4. - Fix to interface file determinism - Critical correctness fix for AArch64 code generation * 9.4.3 (End of October) - Critical correctness fix for AArch64 code generation - Fixes to performance regressions introduced in 9.4.2. - Fix to interface file determinism * 9.6.1 (Early 2023) - Branch will be taken in early november. Cheers, Matt

3 2

9.6.1 branch date - Wednesday 7th Nov
by Matthew Pickering 18 Oct '22

18 Oct '22

Hi all, In order to keep up the 6-month release cadence it is nearly time to take the branch for the 9.6.1 release. The date the branch will be taken is Wednesday 7th Nov. Please reply to this email if there are any in-flight patches which should be included in the 9.6 release. Cheers, Matt

1 0

Missing class op rules when using API
by Erdi, Gergo 18 Oct '22

18 Oct '22

PUBLIC Hi, I'm trying to compile the following two modules: ``` {-# LANGUAGE NoImplicitPrelude #-} module MiniMonad where class Functor f where fmap :: (a -> b) -> f a -> f b class (Functor f) => Applicative f where pure :: a -> f a class (Applicative m) => Monad m where return :: a -> m a return = pure (>>=) :: m a -> (a -> m b) -> m b infixl 1 >>= ``` ``` {-# LANGUAGE NoImplicitPrelude #-} module SpecMain where import MiniMonad data Identity a = MkIdentity a instance Functor Identity where fmap f (MkIdentity x) = MkIdentity (f x) instance Applicative Identity where pure x = MkIdentity x instance Monad Identity where run :: Identity a -> a run (MkIdentity x) = x foo :: () foo = run (return ()) ``` If I compile these from the CLI, I see three rules firing: ``` ./_build/stage1/bin/ghc -debug -fforce-recomp -dno-typeable-binds input/SpecMain.hs -iinput -O0 -ddump-rule-rewrites -dsuppress-all ``` ``` [1 of 2] Compiling MiniMonad ( input/MiniMonad.hs, input/MiniMonad.o ) [2 of 2] Compiling SpecMain ( input/SpecMain.hs, input/SpecMain.o ) tc_iface_decl/2 Monad input/SpecMain.hs:14:10: warning: [-Wmissing-methods] * No explicit implementation for '>>=' * In the instance declaration for 'Monad Identity' | 14 | instance Monad Identity where | ^^^^^^^^^^^^^^ Rule fired Rule: Class op $p1Monad Module: (BUILTIN) Before: MiniMonad.$p1Monad TyArg SpecMain.Identity ValArg SpecMain.$fMonadIdentity After: SpecMain.$fApplicativeIdentity Cont: StrictArg MiniMonad.pure ApplyToTy a_avS Stop[RhsCtxt(Recursive)] a_avS -> SpecMain.Identity a_avS Rule fired Rule: Class op pure Module: (BUILTIN) Before: MiniMonad.pure TyArg SpecMain.Identity ValArg SpecMain.$fApplicativeIdentity After: $cpure_aw8 Cont: ApplyToTy a_avS Stop[RhsCtxt(Recursive)] a_avS -> SpecMain.Identity a_avS Rule fired Rule: Class op return Module: (BUILTIN) Before: MiniMonad.return TyArg SpecMain.Identity ValArg SpecMain.$fMonadIdentity After: $creturn_avP Cont: ApplyToTy () ApplyToVal nodup hole () -> SpecMain.Identity () GHC.Tuple.Prim.() Select nodup wild_X1 Stop[RhsCtxt(NonRecursive)] () ``` However, if I try this via the API (see attached `main.hs`), then only `Class op return` fires: ``` Processing module MiniMonad Processing module SpecMain input/SpecMain.hs:14:10: warning: [-Wmissing-methods] * No explicit implementation for '>>=' * In the instance declaration for 'Monad Identity' | 14 | instance Monad Identity where | ^^^^^^^^^^^^^^ Rule fired Rule: Class op return Module: (BUILTIN) Before: MiniMonad.return TyArg SpecMain.Identity ValArg SpecMain.$fMonadIdentity After: $creturn_aqm Cont: ApplyToTy () ApplyToVal nodup hole () -> SpecMain.Identity () GHC.Tuple.Prim.() Select nodup wild_X1 Stop[RhsCtxt(NonRecursive)] () All done. ``` This of course leads to much worse code further downstream (and is catastrophic for my use case where I rely on extensive specialisation). Why is that? What is missing from my `main.hs` that only one of the three method rules is correctly used when compiling the second module? Thanks, Gergo This email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please delete all copies and notify the sender immediately. You may wish to refer to the incorporation details of Standard Chartered PLC, Standard Chartered Bank and their subsidiaries at https: //www.sc.com/en/our-locations Where you have a Financial Markets relationship with Standard Chartered PLC, Standard Chartered Bank and their subsidiaries (the "Group"), information on the regulatory standards we adhere to and how it may affect you can be found in our Regulatory Compliance Statement at https: //www.sc.com/rcs/ and Regulatory Compliance Disclosures at http: //www.sc.com/rcs/fm Insofar as this communication is not sent by the Global Research team and contains any market commentary, the market commentary has been prepared by the sales and/or trading desk of Standard Chartered Bank or its affiliate. It is not and does not constitute research material, independent research, recommendation or financial advice. Any market commentary is for information purpose only and shall not be relied on for any other purpose and is subject to the relevant disclaimers available at https: //www.sc.com/en/regulatory-disclosures/#market-disclaimer. Insofar as this communication is sent by the Global Research team and contains any research materials prepared by members of the team, the research material is for information purpose only and shall not be relied on for any other purpose, and is subject to the relevant disclaimers available at https: //research.sc.com/research/api/application/static/terms-and-conditions. Insofar as this e-mail contains the term sheet for a proposed transaction, by responding affirmatively to this e-mail, you agree that you have understood the terms and conditions in the attached term sheet and evaluated the merits and risks of the transaction. We may at times also request you to sign the term sheet to acknowledge the same. Please visit https: //www.sc.com/en/regulatory-disclosures/dodd-frank/ for important information with respect to derivative products.

1 1

Serious bug in GHC 9.4 on basic math on aarch64
by Ian-Woo Kim 14 Oct '22

14 Oct '22

Hi, We found a reproducible error with a simple math test (div/mod) on aarch64. (on linux, on macos), so we want to call for attention on this. https://gitlab.haskell.org/ghc/ghc/-/issues/22282 This was originally caught by this issue: https://github.com/haskell-foundation/foundation/issues/571 Thank you! Ian-Woo Kim Mercury Technologies

4 4

Milestones page is missing or has inaccurate dates
by George Colpitts 13 Oct '22

13 Oct '22

Hi The Milestones page <https://gitlab.haskell.org/ghc/ghc/-/milestones> is missing many dates or has inaccurate dates: - Most milestones don't have any associated completion date. - As far as I know 9.4.2 and 9.2.4 are released but the milestones pages doesn't seem to show that. Can the page be updated so we have an idea of what is coming when? There has been a lot of great work done and I know a lot is being worked on. It would be great if the milestones page could be updated to better reflect that. Thanks George

3 4

Warnings in RTS
by Simon Peyton Jones 11 Oct '22

11 Oct '22

Whenever I compile the RTS I get these warnings. Are they necessary? Simon In file included from rts/Schedule.c:16:0: error: rts/Schedule.c: In function ‘schedule’: rts/Schedule.h:143:1: error: warning: inlining failed in call to ‘appendToRunQueue’: call is unlikely and code size would grow [-Winline] 143 | appendToRunQueue (Capability *cap, StgTSO *tso) | ^~~~~~~~~~~~~~~~ | 143 | appendToRunQueue (Capability *cap, StgTSO *tso) | ^ rts/Schedule.c:1226:9: error: note: called from here 1226 | appendToRunQueue(cap,t); | ^~~~~~~~~~~~~~~~~~~~~~~ | 1226 | appendToRunQueue(cap,t); | ^ In file included from rts/Schedule.c:16:0: error: rts/Schedule.h:164:1: error: warning: inlining failed in call to ‘pushOnRunQueue’: call is unlikely and code size would grow [-Winline] 164 | pushOnRunQueue (Capability *cap, StgTSO *tso) | ^~~~~~~~~~~~~~ | 164 | pushOnRunQueue (Capability *cap, StgTSO *tso) | ^ rts/Schedule.c:1228:9: error: note: called from here 1228 | pushOnRunQueue(cap,t); | ^~~~~~~~~~~~~~~~~~~~~ | 1228 | pushOnRunQueue(cap,t); | ^ rts/Schedule.c: At top level: cc1: warning: unrecognized command line option ‘-Wno-nonportable-include-path’

2 1

Implementing a compilation server
by Domínguez, Facundo 11 Oct '22

11 Oct '22

Dear ghc devs, I'm using the ghc API to write a compilation server (a.k.a. persistent worker). The idea is to serve requests to compile individual modules. In this fashion, we can compile modules with different compilation flags and yet pay only once for the startup overheads of the compiler. One challenge of this approach is to reuse as much as possible from the ghc API session/environment from one compilation request to the next, so we save the trouble of reconstructing it each time. This message is to ask for advise on how to better accomplish this reuse. I tried reusing the whole environment for multiple requests, but I'm conjecturing that this might cause troubles when the requests require building modules with different values of -this-unit-id. Another problem that streams from this is that recompiling a module which defines a type class instance fails because it encounters in the environment the type class instance from the previous compilation. My work-in-progress implementation is here [1]. There appears to be multiple ways to compile a module in the API, so far I have been trying DriverPipeline.compileFile. My best lead right now is to look for inspiration in how GHCi implements the load command, but this does a sort of --make compilation while I'm going here for the one-shot style. Thanks in advance, Facundo [1] https://github.com/tweag/rules_haskell/blob/16ba422457ea4daa5dbf40d46327ebc…

2 3