- ghc-commits - Haskell.org

[Git][ghc/ghc][wip/marge_bot_batch_merge_job] 23 commits: Check that shift values are valid
by Marge Bot (＠marge-bot) 25 Mar '26

25 Mar '26

Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC Commits: 9c893d01 by Sylvain Henry at 2026-03-25T12:44:19-04:00 Check that shift values are valid In GHC's codebase in non-DEBUG builds we silently substitute shiftL/R with unsafeShiftL/R for performance reasons. However we were not checking that the shift value was valid for unsafeShiftL/R, leading to wrong computations, but only in non-DEBUG builds. This patch adds the necessary checks and reports an error when a wrong shift value is passed. - - - - - 2d066cc7 by Sylvain Henry at 2026-03-25T12:44:19-04:00 Implement basic value range analysis (#25718) Perform basic value range analysis to try to determine at compile time the result of the application of some comparison primops (ltWord#, etc.). This subsumes the built-in rewrite rules used previously to check if one of the comparison argument was a bound (e.g. (x :: Word8) <= 255 is always True). Our analysis is more powerful and handles type conversions: e.g. word8ToWord x <= 255 is now detected as always True too. We also use value range analysis to filter unreachable alternatives in case-expressions. To support this, we had to allow case-expressions for primitive types to not have a DEFAULT alternative (as was assumed before and checked in Core lint). - - - - - 24a785ad by ARATA Mizuki at 2026-03-25T12:44:23-04:00 rts: Align stack to 64-byte boundary in StgRun on x86 When LLVM spills AVX/AVX-512 vector registers to the stack, it requires 32-byte (__m256) or 64-byte (__m512) alignment. If the stack is not sufficiently aligned, LLVM inserts a realignment prologue that reserves %rbp as a frame pointer, conflicting with GHC's use of %rbp as an STG callee-saved register and breaking the tail-call-based calling convention. Previously, GHC worked around this by lying to LLVM about the stack alignment and rewriting aligned vector loads/stores (VMOVDQA, VMOVAPS) to unaligned ones (VMOVDQU, VMOVUPS) in the LLVM Mangler. This had two problems: - It did not extend to AVX-512, which requires 64-byte alignment. (#26595) - When Haskell calls a C function that takes __m256/__m512 arguments on the stack, the callee requires genuine alignment, which could cause a segfault. (#26822) This patch genuinely aligns the stack to 64 bytes in StgRun by saving the original stack pointer before alignment and restoring it in StgReturn. We now unconditionally advertise 64-byte stack alignment to LLVM for all x86 targets, making rewriteAVX in the LLVM Mangler unnecessary. STG_RUN_STACK_FRAME_SIZE is increased from 48 to 56 bytes on non-Windows x86-64 to store the saved stack pointer. Closes #26595 and #26822 Co-Authored-By: Claude Opus 4.5 <noreply(a)anthropic.com> - - - - - 24d83e24 by Teo Camarasu at 2026-03-25T12:44:24-04:00 ghc-internal: Float Generics to near top of module graph We remove GHC.Internal.Generics from the critical path of the `ghc-internal` module graph. GHC.Internal.Generics used to be in the middle of the module graph, but now it is nearer the top (built later). This change thins out the module graph and allows us to get rid of the ByteOrder hs-boot file. We implement this by moving Generics instances from the module where the datatype is defined to the GHC.Internal.Generics module. This trades off increasing the compiled size of GHC.Internal.Generics with reducing the dependency footprint of datatype modules. Not all instances are moved to GHC.Internal.Generics. For instance, `GHC.Internal.Control.Monad.Fix` keeps its instance as it is one of the very last modules compiled in `ghc-internal` and so inverting the relationship here would risk adding GHC.Internal.Generics back onto the critical path. We also don't change modules that are re-exported from the `template-haskell` or `ghc-heap`. This is done to make it easy to eventually move `Generics` to `base` once something like #26657 is implemented. Resolves #26930 Metric Decrease: T21839c - - - - - 8f64bb3e by sheaf at 2026-03-25T12:44:41-04:00 Avoid infinite loop in deep subsumption This commit ensures we only unify after we recur in the deep subsumption code in the FunTy vs non-FunTy case of GHC.Tc.Utils.Unify.tc_sub_type_deep, to avoid falling into an infinite loop. See the new Wrinkle [Avoiding a loop in tc_sub_type_deep] in Note [FunTy vs non-FunTy case in tc_sub_type_deep] in GHC.Tc.Utils.Unify. Fixes #26823 Co-authored-by: simonpj <simon.peytonjones(a)gmail.com> - - - - - e9a15fe5 by Ian Duncan at 2026-03-25T12:44:45-04:00 AArch64: fix MOVK regUsageOfInstr to mark dst as both read and written MOVK (move with keep) modifies only a 16-bit slice of the destination register, so the destination is both read and written. The register allocator must know this to avoid clobbering live values. Update regUsageOfInstr to list the destination in both src and dst sets. No regression test: triggering the misallocation requires specific register pressure around a MOVK sequence, which is difficult to reliably provoke from Haskell source. - - - - - 3d069672 by Simon Jakobi at 2026-03-25T12:44:45-04:00 Add regression test for #12002 Closes #12002. - - - - - 3a415e32 by Simon Jakobi at 2026-03-25T12:44:46-04:00 Add regression test for #12046 Closes #12046. Co-authored-by: Andreas Klebinger <klebinger.andreas(a)gmx.at> - - - - - d09855b5 by Simon Jakobi at 2026-03-25T12:44:46-04:00 Add regression test for #13180 Closes #13180. - - - - - f68cf1ab by Simon Jakobi at 2026-03-25T12:44:46-04:00 Add regression test for #11141 Closes #11141. - - - - - 2512a033 by Simon Jakobi at 2026-03-25T12:44:46-04:00 Add regression test for #11505 Closes #11505. - - - - - be367209 by Simon Jakobi at 2026-03-25T12:44:46-04:00 Add regression perf test for #13820 Closes #13820. - - - - - bde1a9a3 by Simon Jakobi at 2026-03-25T12:44:46-04:00 Add regression test for #10381 Closes #10381. - - - - - d6877173 by Matthew Pickering at 2026-03-25T12:44:47-04:00 rts: forward clone-stack messages after TSO migration MSG_CLONE_STACK assumed that the target TSO was still owned by the capability that received the message. This is not always true: the TSO can migrate before the inbox entry is handled. When that happened, handleCloneStackMessage could clone a live stack from the wrong capability and use the wrong capability for allocation and performTryPutMVar, leading to stack sanity failures such as checkStackFrame: weird activation record found on stack. Fix this by passing the current capability into handleCloneStackMessage, rechecking msg->tso->cap at handling time, and forwarding the message if the TSO has migrated. Once ownership matches, use the executing capability consistently for cloneStack, rts_apply, and performTryPutMVar. Fixes #27008 - - - - - a8795d04 by fendor at 2026-03-25T12:44:49-04:00 Extract Binary instances to `GHC.ByteCode.Binary` - - - - - b5f55256 by fendor at 2026-03-25T12:44:49-04:00 Add `seqNonEmpty` for evaluating `NonEmpty a` - - - - - be29dad6 by fendor at 2026-03-25T12:44:49-04:00 Record `LinkableUsage` instead of `Linkable` in `LoaderState` Retaining a ByteCode `Linkable` after it has been loaded retains its `UnlinkedBCO`, keeping it alive for the remainder of the program. This starts accumulating a lot of `UnlinkedBCO` and memory over time. However, the `Linkable` is merely used to later record its usage in `mkObjectUsage`, which is used for recompilation checking. However, this is incorrect, as the interface file and bytecode objects could be in different states, e.g. the interface changes, but the bytecode library hasn't changed so we don't need to recompile and vice versa. By computing a `Fingerprint` for the `ModuleByteCode`, and recording it in the `LinkableUsage`, we know precisely whether the `ByteCode` object on disk is outdated. Thus, parts of this commit just makes sure that we efficiently compute a `Fingerprint` for `ModuleByteCode` and store it in the on-disk representation of `ModuleByteCode`. We change the `LoaderState` to retain `LinkableUsage`, which is smaller representation of a `Linkable`. This allows us to free the unneeded fields of `Linkable` after linking them. We declare the following memory invariants that this commit implements: * No `LinkablePart` should be retained from `LoaderState`. * `Linkable`s should be unloaded after they have been loaded. These invariants are unfortunately tricky to automatically uphold, so we are simply documenting our assumptions for now. We introduce the `linkable-space` test which makes sure that after loading, no `DotGBC` or `UnlinkedBCO` is retained. ------------------------- Metric Increase: MultiLayerModulesTH_OneShot ------------------------- We allocate a bit more, but the peak number of bytes doesn't change. While a bit unfortunate, accepting the metric increase. We add multiple new performance measurements where we were able to observe the desired memory invariants. Further, we add regression tests to validate that the recompilation checker behaves more correct than before. - - - - - 92a183e5 by mangoiv at 2026-03-25T12:44:50-04:00 release tracking: adopt release tracking ticket from #16816 - - - - - 676b5fa6 by mangoiv at 2026-03-25T12:44:50-04:00 release tracking: add a release tracking ticket Brings the information in the release tracking ticket up to date with https://gitlab.haskell.org/ghc/ghc-hq/-/blob/main/release-management.mkd Resolves #26691 - - - - - 4f97c2c5 by Teo Camarasu at 2026-03-25T12:44:51-04:00 Revert "Set default eventlog-flush-interval to 5s" Flushing the eventlog forces a synchronisation of all the capabilities and there was a worry that this might lead to a performance cost for some highly parallel workloads. This reverts commit 66b96e2a591d8e3d60e74af3671344dfe4061cf2. - - - - - 17aad2d5 by Wen Kokke at 2026-03-25T12:44:56-04:00 Fix race condition between flushEventLog and start/endEventLogging. This commit changes `flushEventLog` to acquire/release the `state_change` mutex to prevent interleaving with `startEventLogging` and `endEventLogging`. In the current RTS, `flushEventLog` _does not_ acquire this mutex, which may lead to eventlog corruption on the following interleaving: - `startEventLogging` writes the new `EventLogWriter` to `event_log_writer`. - `flushEventLog` flushes some events to `event_log_writer`. - `startEventLogging` writes the eventlog header to `event_log_writer`. This causes the eventlog to be written out in an unreadable state, with one or more events preceding the eventlog header. This commit renames the old function to `flushEventLog_` and defines `flushEventLog` simply as: ```c void flushEventLog(Capability **cap USED_IF_THREADS) { ACQUIRE_LOCK(&state_change_mutex); flushEventLog_(cap); RELEASE_LOCK(&state_change_mutex); } ``` The old function is still needed internally within the compilation unit, where it is used in `endEventLogging` in a context where the `state_change` mutex has already been acquired. I've chosen to mark `flushEventLog_` as static and let other uses of `flushEventLog` within the RTS refer to the new version. There is one use in `hs_init_ghc` via `flushTrace`, where the new locking behaviour should be harmless, and one use in `handle_tick`, which I believe was likely vulnerable to the same race condition, so the new locking behaviour is desirable. I have not added a test. The behaviour is highly non-deterministic and requires a program that concurrently calls `flushEventLog` and `startEventLogging`/`endEventLogging`. I encountered the issue while developing `eventlog-socket` and within that context have verified that my patch likely addresses the issue: a test that used to fail within the first dozen or so runs now has been running on repeat for several hours. - - - - - 9a4c68ee by Phil Hazelden at 2026-03-25T12:44:57-04:00 Fix build with werror on glibc 2.43. We've been defining `_XOPEN_SOURCE` and `_POSIX_C_SOURCE` to the same values as defined in glibc prior to 2.43. But in 2.43, glibc changes them to new values, which means we get a warning when redefining them. By `#undef`ing them first, we no longer get a warning. Closes #27076. - - - - - dab85a9a by Tobias Haslop at 2026-03-25T12:45:02-04:00 Fix broken Haddock link to Bifunctor class in description of Functor class - - - - - 136 changed files: - + .gitlab/issue_templates/release_tracking.md - + compiler/GHC/ByteCode/Binary.hs - + compiler/GHC/ByteCode/Recomp/Binary.hs - compiler/GHC/ByteCode/Serialize.hs - compiler/GHC/CmmToAsm/AArch64/Instr.hs - compiler/GHC/CmmToLlvm.hs - compiler/GHC/CmmToLlvm/Config.hs - compiler/GHC/CmmToLlvm/Mangler.hs - compiler/GHC/Core.hs - compiler/GHC/Core/Lint.hs - compiler/GHC/Core/Opt/ConstantFold.hs - + compiler/GHC/Core/Opt/Range.hs - compiler/GHC/Core/Opt/Simplify/Iteration.hs - compiler/GHC/Driver/Config/CmmToLlvm.hs - compiler/GHC/Driver/Hooks.hs - compiler/GHC/Driver/Main.hs - compiler/GHC/Driver/Pipeline.hs - compiler/GHC/Driver/Plugins.hs - compiler/GHC/HsToCore/Usage.hs - compiler/GHC/Iface/Recomp.hs - compiler/GHC/Iface/Recomp/Types.hs - compiler/GHC/Linker/ByteCode.hs - compiler/GHC/Linker/Deps.hs - compiler/GHC/Linker/Loader.hs - compiler/GHC/Linker/Types.hs - compiler/GHC/Prelude/Basic.hs - compiler/GHC/Runtime/Loader.hs - compiler/GHC/StgToCmm/Expr.hs - compiler/GHC/StgToCmm/Utils.hs - compiler/GHC/Tc/Types.hs - compiler/GHC/Tc/Utils/Monad.hs - compiler/GHC/Tc/Utils/Unify.hs - compiler/GHC/Unit/Home/ModInfo.hs - compiler/GHC/Unit/Module/Deps.hs - compiler/GHC/Unit/Module/Status.hs - compiler/GHC/Utils/Misc.hs - compiler/ghc.cabal.in - docs/users_guide/9.16.1-notes.rst - ghc/GHCi/Leak.hs - libraries/ghc-internal/src/GHC/Internal/Base.hs - libraries/ghc-internal/src/GHC/Internal/ByteOrder.hs - − libraries/ghc-internal/src/GHC/Internal/ByteOrder.hs-boot - libraries/ghc-internal/src/GHC/Internal/Char.hs - libraries/ghc-internal/src/GHC/Internal/Data/Foldable.hs - libraries/ghc-internal/src/GHC/Internal/Data/Functor/Const.hs - libraries/ghc-internal/src/GHC/Internal/Data/Functor/Identity.hs - libraries/ghc-internal/src/GHC/Internal/Data/Monoid.hs - libraries/ghc-internal/src/GHC/Internal/Data/Semigroup/Internal.hs - libraries/ghc-internal/src/GHC/Internal/Data/Traversable.hs - libraries/ghc-internal/src/GHC/Internal/Data/Version.hs - libraries/ghc-internal/src/GHC/Internal/Event/Control.hs - libraries/ghc-internal/src/GHC/Internal/Functor/ZipList.hs - libraries/ghc-internal/src/GHC/Internal/Generics.hs - libraries/ghc-internal/src/GHC/Internal/IO/Exception.hs - libraries/ghc-internal/src/GHC/Internal/RTS/Flags.hsc - libraries/ghc-internal/src/GHC/Internal/Read.hs - libraries/ghc-internal/src/GHC/Internal/Unicode/Bits.hs - rts/CloneStack.c - rts/CloneStack.h - rts/Messages.c - rts/RtsFlags.c - rts/StgCRun.c - rts/eventlog/EventLog.c - rts/include/rts/Constants.h - rts/include/rts/PosixSource.h - + testsuite/tests/bytecode/TLinkable/BCOTemplate.hs - + testsuite/tests/bytecode/TLinkable/LinkableUsage01.stderr - + testsuite/tests/bytecode/TLinkable/LinkableUsage02.stderr - + testsuite/tests/bytecode/TLinkable/Makefile - + testsuite/tests/bytecode/TLinkable/all.T - + testsuite/tests/bytecode/TLinkable/genLinkables.sh - + testsuite/tests/bytecode/TLinkable/linkable-space.hs - + testsuite/tests/bytecode/TLinkable/linkable-space.stdout - testsuite/tests/count-deps/CountDepsAst.stdout - testsuite/tests/count-deps/CountDepsParser.stdout - testsuite/tests/driver/multipleHomeUnits/multipleHomeUnits_recomp_th.stdout - + testsuite/tests/driver/recomp022/A1.hs - + testsuite/tests/driver/recomp022/A2.hs - + testsuite/tests/driver/recomp022/A3.hs - + testsuite/tests/driver/recomp022/B.hs - + testsuite/tests/driver/recomp022/C.hs - + testsuite/tests/driver/recomp022/Makefile - + testsuite/tests/driver/recomp022/all.T - + testsuite/tests/driver/recomp022/recomp022a.stdout - + testsuite/tests/driver/recomp022/recomp022b.stdout - testsuite/tests/ghci/scripts/ListTuplePunsPpr.stdout - testsuite/tests/ghci/scripts/T10963.stderr - testsuite/tests/ghci/scripts/ghci064.stdout - testsuite/tests/interface-stability/base-exports.stdout - testsuite/tests/interface-stability/base-exports.stdout-javascript-unknown-ghcjs - testsuite/tests/interface-stability/base-exports.stdout-mingw32 - testsuite/tests/interface-stability/base-exports.stdout-ws-32 - + testsuite/tests/parser/should_compile/T12002.hs - + testsuite/tests/parser/should_compile/T12002.stderr - testsuite/tests/parser/should_compile/all.T - + testsuite/tests/perf/compiler/T13820.hs - testsuite/tests/perf/compiler/all.T - + testsuite/tests/rebindable/T10381.hs - testsuite/tests/rebindable/all.T - testsuite/tests/rts/all.T - + testsuite/tests/rts/cloneThreadStackMigrating.hs - + testsuite/tests/simd/should_run/StackAlignment32.hs - + testsuite/tests/simd/should_run/StackAlignment32.stdout - + testsuite/tests/simd/should_run/StackAlignment32_main.c - + testsuite/tests/simd/should_run/StackAlignment64.hs - + testsuite/tests/simd/should_run/StackAlignment64.stdout - + testsuite/tests/simd/should_run/StackAlignment64_main.c - testsuite/tests/simd/should_run/all.T - + testsuite/tests/simplCore/should_compile/T19166.hs - + testsuite/tests/simplCore/should_compile/T19166.stderr - + testsuite/tests/simplCore/should_compile/T25718.hs - + testsuite/tests/simplCore/should_compile/T25718.stderr - + testsuite/tests/simplCore/should_compile/T25718a.hs - + testsuite/tests/simplCore/should_compile/T25718a.stderr - + testsuite/tests/simplCore/should_compile/T25718b.hs - + testsuite/tests/simplCore/should_compile/T25718b.stderr - + testsuite/tests/simplCore/should_compile/T25718c.hs - + testsuite/tests/simplCore/should_compile/T25718c.stderr-ws-32 - + testsuite/tests/simplCore/should_compile/T25718c.stderr-ws-64 - testsuite/tests/simplCore/should_compile/all.T - + testsuite/tests/typecheck/T13180/T13180.hs - + testsuite/tests/typecheck/T13180/T13180.hs-boot - + testsuite/tests/typecheck/T13180/T13180.stderr - + testsuite/tests/typecheck/T13180/T13180A.hs - + testsuite/tests/typecheck/T13180/all.T - + testsuite/tests/typecheck/should_compile/T11141.hs - + testsuite/tests/typecheck/should_compile/T11141.stderr - + testsuite/tests/typecheck/should_compile/T11505Bar.hs - + testsuite/tests/typecheck/should_compile/T11505Foo.hs - + testsuite/tests/typecheck/should_compile/T11505Foo.hs-boot - + testsuite/tests/typecheck/should_compile/T12046.hs - testsuite/tests/typecheck/should_compile/T26225.hs - testsuite/tests/typecheck/should_compile/all.T - + testsuite/tests/typecheck/should_fail/T26823.hs - + testsuite/tests/typecheck/should_fail/T26823.stderr - testsuite/tests/typecheck/should_fail/all.T The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/75c79c03a0d3f72c86999fda861eb6… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/75c79c03a0d3f72c86999fda861eb6… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/supersven/correctly_propagate_host-build-target] 2 commits: Respect user set CFLAGS, CXXFLAGS and LDFLAGS
by Sven Tennie (＠supersven) 25 Mar '26

25 Mar '26

Sven Tennie pushed to branch wip/supersven/correctly_propagate_host-build-target at Glasgow Haskell Compiler / GHC Commits: 7bffb567 by Sven Tennie at 2026-03-25T17:30:29+01:00 Respect user set CFLAGS, CXXFLAGS and LDFLAGS These were ignored on installs. - - - - - 312deb98 by Sven Tennie at 2026-03-25T17:30:29+01:00 Only add flags if they are new This prevents duplicated when the user already set the arguments in CXXFLAGS or CFLAGS. - - - - - 3 changed files: - distrib/configure.ac.in - utils/ghc-toolchain/src/GHC/Toolchain/Tools/Cpp.hs - utils/ghc-toolchain/src/GHC/Toolchain/Tools/Link.hs Changes: ===================================== distrib/configure.ac.in ===================================== @@ -159,6 +159,24 @@ AC_PROG_CXX([c++ g++ clang++]) MOVE_TO_FLAGS([CC],[CFLAGS]) MOVE_TO_FLAGS([CXX],[CXXFLAGS]) +dnl Save the original user-provided environment variables immediately +dnl before any m4 macros modify them. These will be passed to ghc-toolchain +dnl so it can make independent decisions and reach the same conclusions as autoconf. +USER_CONF_CC_OPTS_STAGE2="$CFLAGS" +USER_CONF_CXX_OPTS_STAGE2="$CXXFLAGS" +USER_CONF_GCC_LINKER_OPTS_STAGE2="$LDFLAGS" +USER_CPP_ARGS="$CPPFLAGS" +USER_HS_CPP_ARGS="$HaskellCPPArgs" +USER_JS_CPP_ARGS="$JavaScriptCPPArgs" +USER_CMM_CPP_ARGS="$CmmCPPArgs" + +dnl Initialize CONF_* variables with user-provided environment variables +dnl so that subsequent configure checks modify them and both autoconf and +dnl ghc-toolchain see the same values +CONF_CC_OPTS_STAGE2="$CFLAGS" +CONF_CXX_OPTS_STAGE2="$CXXFLAGS" +CONF_GCC_LINKER_OPTS_STAGE2="$LDFLAGS" + dnl ** figure out how to invoke the C preprocessor (i.e. `gcc -E`) AC_PROG_CPP ===================================== utils/ghc-toolchain/src/GHC/Toolchain/Tools/Cpp.hs ===================================== @@ -61,7 +61,7 @@ findHsCpp progOpt cc = checking "for Haskell C preprocessor" $ do let rawHsCppProgram = addFlagIfNew "-E" foundHsCppProg -- Always try to add the Haskell-specific CPP flags, regardless of the user options hppArgs <- findHsCppArgs rawHsCppProgram - let hsCppProgram = over _prgFlags (++hppArgs) rawHsCppProgram + let hsCppProgram = foldr addFlagIfNew rawHsCppProgram (reverse hppArgs) return HsCpp{hsCppProgram} -- | Given a C preprocessor, figure out how it should be invoked to preprocess ===================================== utils/ghc-toolchain/src/GHC/Toolchain/Tools/Link.hs ===================================== @@ -345,30 +345,30 @@ addPlatformDepLinkFlags archOs cc ccLink0 = do -- -- On OpenSolaris uses gnu ld whereas SmartOS appears to use the Solaris -- implementation, which rather uses the -64 flag. - return $ ccLink2 & _prgFlags %++ "-m64" + return $ addFlagIfNew "-m64" ccLink2 ArchOS ArchAlpha _ -> -- For now, to suppress the gcc warning "call-clobbered -- register used for global register variable", we simply -- disable all warnings altogether using the -w flag. Oh well. - return $ ccLink2 & over _prgFlags (++["-w","-mieee","-D_REENTRANT"]) + return $ foldr addFlagIfNew ccLink2 (reverse ["-w","-mieee","-D_REENTRANT"]) -- ArchOS ArchHPPA? _ -> ArchOS ArchARM{} OSFreeBSD -> -- On arm/freebsd, tell gcc to generate Arm -- instructions (ie not Thumb). - return $ ccLink2 & _prgFlags %++ "-Wl,-z,noexecstack" + return $ addFlagIfNew "-Wl,-z,noexecstack" ccLink2 ArchOS ArchARM{} OSLinux -> -- On arm/linux and arm/android, tell gcc to generate Arm -- instructions (ie not Thumb). - return $ ccLink2 & _prgFlags %++ "-Wl,-z,noexecstack" + return $ addFlagIfNew "-Wl,-z,noexecstack" ccLink2 ArchOS ArchAArch64 OSFreeBSD -> - return $ ccLink2 & _prgFlags %++ "-Wl,-z,noexecstack" + return $ addFlagIfNew "-Wl,-z,noexecstack" ccLink2 ArchOS ArchAArch64 OSLinux -> - return $ ccLink2 & _prgFlags %++ "-Wl,-z,noexecstack" + return $ addFlagIfNew "-Wl,-z,noexecstack" ccLink2 ArchOS ArchAArch64 OSNetBSD -> - return $ ccLink2 & _prgFlags %++ "-Wl,-z,noexecstack" + return $ addFlagIfNew "-Wl,-z,noexecstack" ccLink2 ArchOS ArchPPC OSAIX -> -- We need `-D_THREAD_SAFE` to unlock the thread-local `errno`. - return $ ccLink2 & over _prgFlags (++["-D_THREAD_SAFE","-Wl,-bnotextro"]) + return $ foldr addFlagIfNew ccLink2 (reverse ["-D_THREAD_SAFE","-Wl,-bnotextro"]) _ -> return ccLink2 @@ -383,7 +383,7 @@ addOSMinGW32CcFlags archOs cc link -- See Note [Windows stack allocations]. checkFStackCheck :: Cc -> Program -> M Program checkFStackCheck cc link = checking "that -fstack-check works" $ do - let link' = link & _prgFlags %++ "-fstack-check" + let link' = addFlagIfNew "-fstack-check" link checkLinkWorks cc link' return link' @@ -393,7 +393,7 @@ addNoAsNeeded archOs cc ccLink | os <- archOS_OS archOs , osElfTarget os = checking "that --no-as-needed works" $ do - let ccLink' = over _prgFlags (++["-Wl,--no-as-needed"]) ccLink + let ccLink' = addFlagIfNew "-Wl,--no-as-needed" ccLink checkLinkWorks cc ccLink' return ccLink' | otherwise = return ccLink View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/9ce6620f4254da92b848af3e1a3c75… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/9ce6620f4254da92b848af3e1a3c75… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/romes/26039] Bump default language edition to GHC2024
by Rodrigo Mesquita (＠alt-romes) 25 Mar '26

25 Mar '26

Rodrigo Mesquita pushed to branch wip/romes/26039 at Glasgow Haskell Compiler / GHC Commits: 9fb7c2cb by Rodrigo Mesquita at 2026-03-25T16:02:12+00:00 Bump default language edition to GHC2024 As per the accepted ghc-proposal#632 Fixes #26039 - - - - - 122 changed files: - .gitlab/ci.sh - compiler/GHC/Driver/Flags.hs - docs/users_guide/exts/control.rst - testsuite/tests/ado/ado004.hs - testsuite/tests/annotations/should_fail/annfail02.hs - testsuite/tests/annotations/should_fail/annfail02.stderr - testsuite/tests/array/should_run/arr020.hs - testsuite/tests/core-to-stg/T19700.hs - testsuite/tests/deSugar/should_fail/DsStrictFail.hs - testsuite/tests/deriving/should_compile/T15798b.hs - testsuite/tests/deriving/should_compile/T15798c.hs - testsuite/tests/deriving/should_compile/T15798c.stderr - testsuite/tests/deriving/should_compile/T24955a.hs - testsuite/tests/deriving/should_compile/T24955a.stderr - testsuite/tests/deriving/should_compile/T24955b.hs - testsuite/tests/deriving/should_compile/T24955c.hs - testsuite/tests/deriving/should_fail/T10598_fail4.hs - testsuite/tests/deriving/should_fail/T10598_fail4.stderr - testsuite/tests/deriving/should_fail/T10598_fail5.hs - testsuite/tests/deriving/should_fail/T10598_fail5.stderr - testsuite/tests/dmdanal/sigs/T22241.hs - testsuite/tests/gadt/T20485.hs - testsuite/tests/ghci.debugger/scripts/all.T - testsuite/tests/ghci.debugger/scripts/break012.hs - testsuite/tests/ghci.debugger/scripts/break012.stdout - testsuite/tests/ghci/prog-mhu002/all.T - testsuite/tests/ghci/scripts/Makefile - testsuite/tests/ghci/should_run/all.T - testsuite/tests/indexed-types/should_compile/T15322.hs - testsuite/tests/indexed-types/should_compile/T15322.stderr - testsuite/tests/linear/should_fail/T18888.hs - testsuite/tests/module/T20007.hs - testsuite/tests/module/T20007.stderr - testsuite/tests/module/mod90.hs - testsuite/tests/module/mod90.stderr - testsuite/tests/overloadedrecflds/should_fail/NoFieldSelectorsFail.hs - testsuite/tests/overloadedrecflds/should_fail/T18999_NoDisambiguateRecordFields.hs - testsuite/tests/overloadedrecflds/should_fail/all.T - testsuite/tests/parser/should_fail/ParserNoLambdaCase.hs - testsuite/tests/parser/should_fail/ParserNoLambdaCase.stderr - testsuite/tests/parser/should_fail/T16270h.hs - testsuite/tests/parser/should_fail/T16270h.stderr - testsuite/tests/parser/should_fail/readFail001.hs - testsuite/tests/parser/should_fail/readFail001.stderr - testsuite/tests/partial-sigs/should_compile/SomethingShowable.hs - testsuite/tests/plugins/late-plugin/LatePlugin.hs - testsuite/tests/polykinds/T7151.hs - testsuite/tests/polykinds/T7151.stderr - testsuite/tests/polykinds/T7433.hs - testsuite/tests/polykinds/T7433.stderr - testsuite/tests/programs/andy_cherry/test.T - testsuite/tests/rename/should_fail/T10668.hs - testsuite/tests/rename/should_fail/T10668.stderr - testsuite/tests/rename/should_fail/T12681.hs - testsuite/tests/rename/should_fail/T12681.stderr - testsuite/tests/rename/should_fail/T13568.hs - testsuite/tests/rename/should_fail/T13568.stderr - testsuite/tests/rename/should_fail/T13644.hs - testsuite/tests/rename/should_fail/T13644.stderr - testsuite/tests/rename/should_fail/T13847.hs - testsuite/tests/rename/should_fail/T13847.stderr - testsuite/tests/rename/should_fail/T14032c.hs - testsuite/tests/rename/should_fail/T19843l.hs - testsuite/tests/rename/should_fail/T19843l.stderr - testsuite/tests/rename/should_fail/T25901_imp_hq_fail_5.stderr - testsuite/tests/rename/should_fail/T25901_imp_sq_fail_2.stderr - testsuite/tests/rename/should_fail/T5385.hs - testsuite/tests/rename/should_fail/T5385.stderr - testsuite/tests/roles/should_fail/Roles5.hs - testsuite/tests/roles/should_fail/Roles5.stderr - testsuite/tests/showIface/DocsInHiFile.hs - testsuite/tests/showIface/DocsInHiFile1.stdout - testsuite/tests/showIface/DocsInHiFileTH.hs - testsuite/tests/showIface/DocsInHiFileTH.stdout - testsuite/tests/showIface/DocsInHiFileTHExternal.hs - testsuite/tests/showIface/HaddockIssue849.hs - testsuite/tests/showIface/HaddockIssue849.stdout - testsuite/tests/showIface/HaddockOpts.hs - testsuite/tests/showIface/HaddockOpts.stdout - testsuite/tests/showIface/HaddockSpanIssueT24378.hs - testsuite/tests/showIface/HaddockSpanIssueT24378.stdout - testsuite/tests/showIface/MagicHashInHaddocks.hs - testsuite/tests/showIface/MagicHashInHaddocks.stdout - testsuite/tests/showIface/Makefile - testsuite/tests/showIface/NoExportList.hs - testsuite/tests/showIface/NoExportList.stdout - testsuite/tests/showIface/PragmaDocs.stdout - testsuite/tests/showIface/ReExports.stdout - testsuite/tests/simplCore/T9646/test.T - testsuite/tests/simplCore/should_compile/T21960.hs - testsuite/tests/simplCore/should_compile/T26709.stderr - testsuite/tests/th/TH_Promoted1Tuple.hs - testsuite/tests/th/TH_Roles1.hs - testsuite/tests/typecheck/should_compile/MutRec.hs - testsuite/tests/typecheck/should_compile/T10770a.hs - testsuite/tests/typecheck/should_compile/T11339.hs - testsuite/tests/typecheck/should_compile/T11397.hs - testsuite/tests/typecheck/should_compile/T13526.hs - testsuite/tests/typecheck/should_compile/T18467.hs - testsuite/tests/typecheck/should_compile/T18467.stderr - testsuite/tests/typecheck/should_compile/tc081.hs - testsuite/tests/typecheck/should_compile/tc141.hs - testsuite/tests/typecheck/should_fail/T23427.hs - testsuite/tests/typecheck/should_fail/T6078.hs - testsuite/tests/typecheck/should_fail/T7453.hs - testsuite/tests/typecheck/should_fail/T7453.stderr - testsuite/tests/typecheck/should_fail/T8570.hs - testsuite/tests/typecheck/should_fail/T8570.stderr - testsuite/tests/typecheck/should_fail/tcfail083.hs - testsuite/tests/typecheck/should_fail/tcfail083.stderr - testsuite/tests/typecheck/should_fail/tcfail084.hs - testsuite/tests/typecheck/should_fail/tcfail084.stderr - testsuite/tests/typecheck/should_fail/tcfail094.hs - testsuite/tests/typecheck/should_fail/tcfail094.stderr - testsuite/tests/typecheck/should_run/T1735.hs - testsuite/tests/typecheck/should_run/T1735_Help/Basics.hs - testsuite/tests/typecheck/should_run/T3731.hs - testsuite/tests/vdq-rta/should_fail/T24159_type_syntax_th_fail.script - testsuite/tests/warnings/should_fail/CaretDiagnostics1.hs - testsuite/tests/warnings/should_fail/CaretDiagnostics1.stderr - testsuite/tests/warnings/should_fail/T24396c.hs - testsuite/tests/warnings/should_fail/T24396c.stderr The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/9fb7c2cb95e3fed41edc57cd5dd9dbe… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/9fb7c2cb95e3fed41edc57cd5dd9dbe… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/supersven/correctly_propagate_host-build-target] Only add flags if they are new
by Sven Tennie (＠supersven) 25 Mar '26

25 Mar '26

Sven Tennie pushed to branch wip/supersven/correctly_propagate_host-build-target at Glasgow Haskell Compiler / GHC Commits: 9ce6620f by Sven Tennie at 2026-03-25T17:01:40+01:00 Only add flags if they are new This prevents duplicated when the user already set the arguments in CXXFLAGS or CFLAGS. - - - - - 2 changed files: - utils/ghc-toolchain/src/GHC/Toolchain/Tools/Cpp.hs - utils/ghc-toolchain/src/GHC/Toolchain/Tools/Link.hs Changes: ===================================== utils/ghc-toolchain/src/GHC/Toolchain/Tools/Cpp.hs ===================================== @@ -61,7 +61,7 @@ findHsCpp progOpt cc = checking "for Haskell C preprocessor" $ do let rawHsCppProgram = addFlagIfNew "-E" foundHsCppProg -- Always try to add the Haskell-specific CPP flags, regardless of the user options hppArgs <- findHsCppArgs rawHsCppProgram - let hsCppProgram = over _prgFlags (++hppArgs) rawHsCppProgram + let hsCppProgram = foldr addFlagIfNew rawHsCppProgram (reverse hppArgs) return HsCpp{hsCppProgram} -- | Given a C preprocessor, figure out how it should be invoked to preprocess ===================================== utils/ghc-toolchain/src/GHC/Toolchain/Tools/Link.hs ===================================== @@ -345,30 +345,30 @@ addPlatformDepLinkFlags archOs cc ccLink0 = do -- -- On OpenSolaris uses gnu ld whereas SmartOS appears to use the Solaris -- implementation, which rather uses the -64 flag. - return $ ccLink2 & _prgFlags %++ "-m64" + return $ addFlagIfNew "-m64" ccLink2 ArchOS ArchAlpha _ -> -- For now, to suppress the gcc warning "call-clobbered -- register used for global register variable", we simply -- disable all warnings altogether using the -w flag. Oh well. - return $ ccLink2 & over _prgFlags (++["-w","-mieee","-D_REENTRANT"]) + return $ foldr addFlagIfNew ccLink2 (reverse ["-w","-mieee","-D_REENTRANT"]) -- ArchOS ArchHPPA? _ -> ArchOS ArchARM{} OSFreeBSD -> -- On arm/freebsd, tell gcc to generate Arm -- instructions (ie not Thumb). - return $ ccLink2 & _prgFlags %++ "-Wl,-z,noexecstack" + return $ addFlagIfNew "-Wl,-z,noexecstack" ccLink2 ArchOS ArchARM{} OSLinux -> -- On arm/linux and arm/android, tell gcc to generate Arm -- instructions (ie not Thumb). - return $ ccLink2 & _prgFlags %++ "-Wl,-z,noexecstack" + return $ addFlagIfNew "-Wl,-z,noexecstack" ccLink2 ArchOS ArchAArch64 OSFreeBSD -> - return $ ccLink2 & _prgFlags %++ "-Wl,-z,noexecstack" + return $ addFlagIfNew "-Wl,-z,noexecstack" ccLink2 ArchOS ArchAArch64 OSLinux -> - return $ ccLink2 & _prgFlags %++ "-Wl,-z,noexecstack" + return $ addFlagIfNew "-Wl,-z,noexecstack" ccLink2 ArchOS ArchAArch64 OSNetBSD -> - return $ ccLink2 & _prgFlags %++ "-Wl,-z,noexecstack" + return $ addFlagIfNew "-Wl,-z,noexecstack" ccLink2 ArchOS ArchPPC OSAIX -> -- We need `-D_THREAD_SAFE` to unlock the thread-local `errno`. - return $ ccLink2 & over _prgFlags (++["-D_THREAD_SAFE","-Wl,-bnotextro"]) + return $ foldr addFlagIfNew ccLink2 (reverse ["-D_THREAD_SAFE","-Wl,-bnotextro"]) _ -> return ccLink2 @@ -383,7 +383,7 @@ addOSMinGW32CcFlags archOs cc link -- See Note [Windows stack allocations]. checkFStackCheck :: Cc -> Program -> M Program checkFStackCheck cc link = checking "that -fstack-check works" $ do - let link' = link & _prgFlags %++ "-fstack-check" + let link' = addFlagIfNew "-fstack-check" link checkLinkWorks cc link' return link' @@ -393,7 +393,7 @@ addNoAsNeeded archOs cc ccLink | os <- archOS_OS archOs , osElfTarget os = checking "that --no-as-needed works" $ do - let ccLink' = over _prgFlags (++["-Wl,--no-as-needed"]) ccLink + let ccLink' = addFlagIfNew "-Wl,--no-as-needed" ccLink checkLinkWorks cc ccLink' return ccLink' | otherwise = return ccLink View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/9ce6620f4254da92b848af3e1a3c750… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/9ce6620f4254da92b848af3e1a3c750… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/sjakobi/T27097
by Simon Jakobi (＠sjakobi2) 25 Mar '26

25 Mar '26

Simon Jakobi pushed new branch wip/sjakobi/T27097 at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/sjakobi/T27097 You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/supersven/correctly_propagate_host-build-target] Only add flags if they are new
by Sven Tennie (＠supersven) 25 Mar '26

25 Mar '26

Sven Tennie pushed to branch wip/supersven/correctly_propagate_host-build-target at Glasgow Haskell Compiler / GHC Commits: b9784bd6 by Sven Tennie at 2026-03-25T14:54:14+01:00 Only add flags if they are new This prevents duplicated when the user already set the arguments in CXXFLAGS or CFLAGS. - - - - - 2 changed files: - utils/ghc-toolchain/src/GHC/Toolchain/Tools/Cpp.hs - utils/ghc-toolchain/src/GHC/Toolchain/Tools/Link.hs Changes: ===================================== utils/ghc-toolchain/src/GHC/Toolchain/Tools/Cpp.hs ===================================== @@ -61,7 +61,7 @@ findHsCpp progOpt cc = checking "for Haskell C preprocessor" $ do let rawHsCppProgram = addFlagIfNew "-E" foundHsCppProg -- Always try to add the Haskell-specific CPP flags, regardless of the user options hppArgs <- findHsCppArgs rawHsCppProgram - let hsCppProgram = over _prgFlags (++hppArgs) rawHsCppProgram + let hsCppProgram = foldr addFlagIfNew rawHsCppProgram (reverse hppArgs) return HsCpp{hsCppProgram} -- | Given a C preprocessor, figure out how it should be invoked to preprocess ===================================== utils/ghc-toolchain/src/GHC/Toolchain/Tools/Link.hs ===================================== @@ -383,7 +383,7 @@ addOSMinGW32CcFlags archOs cc link -- See Note [Windows stack allocations]. checkFStackCheck :: Cc -> Program -> M Program checkFStackCheck cc link = checking "that -fstack-check works" $ do - let link' = link & _prgFlags %++ "-fstack-check" + let link' = addFlagIfNew "-fstack-check" link checkLinkWorks cc link' return link' @@ -393,7 +393,7 @@ addNoAsNeeded archOs cc ccLink | os <- archOS_OS archOs , osElfTarget os = checking "that --no-as-needed works" $ do - let ccLink' = over _prgFlags (++["-Wl,--no-as-needed"]) ccLink + let ccLink' = addFlagIfNew "-Wl,--no-as-needed" ccLink checkLinkWorks cc ccLink' return ccLink' | otherwise = return ccLink View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/b9784bd6dd7f13982ddcbbaa477d17a… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/b9784bd6dd7f13982ddcbbaa477d17a… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/supersven/correctly_propagate_host-build-target] 2 commits: Respect user set CFLAGS, CXXFLAGS and LDFLAGS
by Sven Tennie (＠supersven) 25 Mar '26

25 Mar '26

Sven Tennie pushed to branch wip/supersven/correctly_propagate_host-build-target at Glasgow Haskell Compiler / GHC Commits: 75ce1385 by Sven Tennie at 2026-03-25T14:44:01+01:00 Respect user set CFLAGS, CXXFLAGS and LDFLAGS These were ignored on installs. - - - - - 34f32edd by Sven Tennie at 2026-03-25T14:44:01+01:00 Only add flags if they are new This prevents duplicated when the user already set the arguments in CXXFLAGS or CFLAGS. - - - - - 3 changed files: - distrib/configure.ac.in - utils/ghc-toolchain/src/GHC/Toolchain/Tools/Cpp.hs - utils/ghc-toolchain/src/GHC/Toolchain/Tools/Link.hs Changes: ===================================== distrib/configure.ac.in ===================================== @@ -159,6 +159,13 @@ AC_PROG_CXX([c++ g++ clang++]) MOVE_TO_FLAGS([CC],[CFLAGS]) MOVE_TO_FLAGS([CXX],[CXXFLAGS]) +dnl Initialize CONF_* variables with user-provided environment variables +dnl so that subsequent configure checks modify them and both autoconf and +dnl ghc-toolchain see the same values +CONF_CC_OPTS_STAGE2="$CFLAGS" +CONF_CXX_OPTS_STAGE2="$CXXFLAGS" +CONF_GCC_LINKER_OPTS_STAGE2="$LDFLAGS" + dnl ** figure out how to invoke the C preprocessor (i.e. `gcc -E`) AC_PROG_CPP @@ -424,6 +431,16 @@ checkMake380() { checkMake380 make checkMake380 gmake +# Pass the CONF_* variables (which now include user environment variables, +# autoconf defaults, and all configure-computed flags) to ghc-toolchain +USER_CONF_CC_OPTS_STAGE2="$CONF_CC_OPTS_STAGE2" +USER_CONF_CXX_OPTS_STAGE2="$CONF_CXX_OPTS_STAGE2" +USER_CONF_GCC_LINKER_OPTS_STAGE2="$CONF_GCC_LINKER_OPTS_STAGE2" +USER_CPP_ARGS="$CPPFLAGS" +USER_HS_CPP_ARGS="$HaskellCPPArgs" +USER_JS_CPP_ARGS="$JavaScriptCPPArgs" +USER_CMM_CPP_ARGS="$CmmCPPArgs" + # Toolchain target files FIND_GHC_TOOLCHAIN_BIN([YES]) PREP_TARGET_FILE ===================================== utils/ghc-toolchain/src/GHC/Toolchain/Tools/Cpp.hs ===================================== @@ -61,7 +61,7 @@ findHsCpp progOpt cc = checking "for Haskell C preprocessor" $ do let rawHsCppProgram = addFlagIfNew "-E" foundHsCppProg -- Always try to add the Haskell-specific CPP flags, regardless of the user options hppArgs <- findHsCppArgs rawHsCppProgram - let hsCppProgram = over _prgFlags (++hppArgs) rawHsCppProgram + let hsCppProgram = foldr addFlagIfNew rawHsCppProgram (reverse hppArgs) return HsCpp{hsCppProgram} -- | Given a C preprocessor, figure out how it should be invoked to preprocess ===================================== utils/ghc-toolchain/src/GHC/Toolchain/Tools/Link.hs ===================================== @@ -393,7 +393,7 @@ addNoAsNeeded archOs cc ccLink | os <- archOS_OS archOs , osElfTarget os = checking "that --no-as-needed works" $ do - let ccLink' = over _prgFlags (++["-Wl,--no-as-needed"]) ccLink + let ccLink' = addFlagIfNew "-Wl,--no-as-needed" ccLink checkLinkWorks cc ccLink' return ccLink' | otherwise = return ccLink View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/e60c93b63e96f942e67f705570fe69… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/e60c93b63e96f942e67f705570fe69… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/int-index/star-exp] Extend HsExpr with the StarIsType syntax (#26587, #26967)
by Vladislav Zavialov (＠int-index) 25 Mar '26

25 Mar '26

Vladislav Zavialov pushed to branch wip/int-index/star-exp at Glasgow Haskell Compiler / GHC Commits: 7f73c68b by Vladislav Zavialov at 2026-03-25T16:16:23+03:00 Extend HsExpr with the StarIsType syntax (#26587, #26967) This patch allows kinds of the form `k -> *` and `* -> k` to occur in expression syntax, i.e. to be used as required type arguments. For example: {-# LANGUAGE RequiredTypeArguments, StarIsType #-} x1 = f (* -> * -> *) x2 = f (forall k. k -> *) x3 = f ((* -> *) -> Constraint) Summary of the changes: * Introduce the HsStar constructor of HsExpr and its extension field XStar. It is analogous to HsStarTy in HsType. * Refactor HsStarTy to store the unicode flag as TokStar, defined as type TokStar = EpUniToken "*" "★" -- similar to TokForall, TokRArrow, etc. The token is stored in the extension field and replaces the Bool field. * Extend the `infixexp2` nonterminal to parse `*` as a direct argument of `->`. This is more limited than the full StarIsType syntax, but has the nice property of not conflicting with the multiplication operator `a * b`. Test case: T26967 T26967_tyop - - - - - 37 changed files: - compiler/GHC/Hs/Expr.hs - compiler/GHC/Hs/Syn/Type.hs - compiler/GHC/Hs/Type.hs - compiler/GHC/HsToCore/Expr.hs - compiler/GHC/HsToCore/Quote.hs - compiler/GHC/HsToCore/Ticks.hs - compiler/GHC/Iface/Ext/Ast.hs - compiler/GHC/Iface/Type.hs - compiler/GHC/Parser.y - compiler/GHC/Parser/Annotation.hs - compiler/GHC/Parser/Errors/Types.hs - compiler/GHC/Parser/PostProcess.hs - compiler/GHC/Rename/Expr.hs - compiler/GHC/Rename/HsType.hs - compiler/GHC/Rename/Pat.hs - compiler/GHC/Tc/Errors/Ppr.hs - compiler/GHC/Tc/Errors/Types.hs - compiler/GHC/Tc/Gen/App.hs - compiler/GHC/Tc/Gen/Expr.hs - compiler/GHC/Tc/Gen/HsType.hs - compiler/GHC/Tc/Types/Origin.hs - compiler/GHC/Tc/Zonk/Type.hs - compiler/GHC/Utils/Outputable.hs - compiler/Language/Haskell/Syntax/Expr.hs - compiler/Language/Haskell/Syntax/Extension.hs - compiler/Language/Haskell/Syntax/Type.hs - docs/users_guide/exts/required_type_arguments.rst - + testsuite/tests/vdq-rta/should_compile/T26967.hs - + testsuite/tests/vdq-rta/should_compile/T26967.stderr - + testsuite/tests/vdq-rta/should_compile/T26967_tyop.hs - + testsuite/tests/vdq-rta/should_compile/T26967_tyop.stderr - testsuite/tests/vdq-rta/should_compile/all.T - utils/check-exact/ExactPrint.hs - utils/haddock/haddock-api/src/Haddock/Backends/LaTeX.hs - utils/haddock/haddock-api/src/Haddock/Backends/Xhtml/Decl.hs - utils/haddock/haddock-api/src/Haddock/Interface/Rename.hs - utils/haddock/haddock-api/src/Haddock/Interface/RenameType.hs The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/7f73c68bf808053f49cfbd77050e894… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/7f73c68bf808053f49cfbd77050e894… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/spj-reinstallable-base] Add imports in ghc-internal for known-key names [skip ci]
by Simon Peyton Jones (＠simonpj) 25 Mar '26

25 Mar '26

Simon Peyton Jones pushed to branch wip/spj-reinstallable-base at Glasgow Haskell Compiler / GHC Commits: c8f6fc8a by Simon Peyton Jones at 2026-03-25T12:37:50+00:00 Add imports in ghc-internal for known-key names [skip ci] - - - - - 36 changed files: - compiler/GHC/Types/Name.hs - compiler/GHC/Types/Name/Cache.hs - libraries/ghc-internal/src/GHC/Internal/ByteOrder.hs - libraries/ghc-internal/src/GHC/Internal/Clock.hsc - libraries/ghc-internal/src/GHC/Internal/ClosureTypes.hs - libraries/ghc-internal/src/GHC/Internal/Control/Monad/Fail.hs - libraries/ghc-internal/src/GHC/Internal/Data/Bits.hs - libraries/ghc-internal/src/GHC/Internal/Data/Dynamic.hs - libraries/ghc-internal/src/GHC/Internal/Data/Either.hs - libraries/ghc-internal/src/GHC/Internal/Data/OldList.hs - libraries/ghc-internal/src/GHC/Internal/Data/Proxy.hs - libraries/ghc-internal/src/GHC/Internal/Data/STRef.hs - libraries/ghc-internal/src/GHC/Internal/Data/Type/Coercion.hs - libraries/ghc-internal/src/GHC/Internal/Data/Type/Equality.hs - libraries/ghc-internal/src/GHC/Internal/Data/Typeable/Internal.hs - libraries/ghc-internal/src/GHC/Internal/Data/Unique.hs - libraries/ghc-internal/src/GHC/Internal/Exception.hs - libraries/ghc-internal/src/GHC/Internal/Exception/Type.hs - libraries/ghc-internal/src/GHC/Internal/Fingerprint/Type.hs - libraries/ghc-internal/src/GHC/Internal/Foreign/C/ConstPtr.hs - libraries/ghc-internal/src/GHC/Internal/Foreign/Storable.hs - libraries/ghc-internal/src/GHC/Internal/ForeignSrcLang.hs - libraries/ghc-internal/src/GHC/Internal/Generics.hs - libraries/ghc-internal/src/GHC/Internal/Heap/ProfInfo/Types.hs - libraries/ghc-internal/src/GHC/Internal/IO.hs - libraries/ghc-internal/src/GHC/Internal/IO/Device.hs - libraries/ghc-internal/src/GHC/Internal/IO/IOMode.hs - libraries/ghc-internal/src/GHC/Internal/IOArray.hs - libraries/ghc-internal/src/GHC/Internal/MVar.hs - libraries/ghc-internal/src/GHC/Internal/Numeric.hs - libraries/ghc-internal/src/GHC/Internal/Ptr.hs - libraries/ghc-internal/src/GHC/Internal/Read.hs - libraries/ghc-internal/src/GHC/Internal/Text/Read.hs - libraries/ghc-internal/src/GHC/Internal/TypeNats.hs - libraries/ghc-internal/src/GHC/Internal/Unicode/Char/DerivedCoreProperties.hs - libraries/ghc-internal/src/GHC/Internal/Unicode/Char/UnicodeData/GeneralCategory.hs The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/c8f6fc8aa65ec564b77029576d185e3… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/c8f6fc8aa65ec564b77029576d185e3… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/T26878] Simplify mkTick
by sheaf (＠sheaf) 25 Mar '26

25 Mar '26

sheaf pushed to branch wip/T26878 at Glasgow Haskell Compiler / GHC Commits: 2c7a6e8f by sheaf at 2026-03-25T12:44:43+01:00 Simplify mkTick This commit simplifies 'GHC.Core.Utils.mkTick', removing the accumulating parameter 'rest' which was suspiciously treating a bunch of different ticks as a group, and moving the group as a whole around the AST, ignoring that the ticks in the group might have different placement properties. Also adds Note [Pushing SCCs inwards] which clarifies the logic for pushing SCCs into lambdas, constructor applications, and dropping SCCs around non-function variables (in particular the treatment of splittable ticks). A few other changes are also implemented: - simplify 'can_split' predicate (no functional change) - drop profiling ticks around coercions, fixing #26941 - combine profiling ticks into one when possible Fixes #26878 and #26941 Co-authored-by: simonpj <simon.peytonjones(a)gmail.com> - - - - - 5 changed files: - compiler/GHC/Core/Utils.hs - compiler/GHC/Types/Tickish.hs - + testsuite/tests/simplCore/should_compile/T26941.hs - + testsuite/tests/simplCore/should_compile/T26941_aux.hs - testsuite/tests/simplCore/should_compile/all.T Changes: ===================================== compiler/GHC/Core/Utils.hs ===================================== @@ -303,101 +303,229 @@ mkCast expr co * * ********************************************************************* -} --- | Wraps the given expression in the source annotation, dropping the --- annotation if possible. +-- | Wraps the given expression in a Tick, floating the tick as far into +-- the AST as possible in order to try to satisfy the tick's desired placement +-- properties (as per Note [Tickish placement] in GHC.Types.Tickish). +-- +-- Prefer using 'mkTick' over explicit use of the 'Tick' constructor. +-- +-- Also performs small on-the-fly optimisations: +-- +-- * Eliminate unnecessary ticks by either absorbing them into existing ones +-- or dropping them if that is valid (e.g. dropping profiling ticks around +-- types, coercions and literals). +-- * Split profiling ticks into counting/scoping parts so that the two parts +-- can be placed independently into the AST. mkTick :: CoreTickish -> CoreExpr -> CoreExpr -mkTick t orig_expr = mkTick' id orig_expr +mkTick t orig_expr = mkTick' orig_expr where -- Some ticks (cost-centres) can be split in two, with the -- non-counting part having laxer placement properties. - canSplit = tickishCanSplit t && tickishPlace (mkNoCount t) /= tickishPlace t - - -- mkTick' handles floating of ticks *into* the expression. - mkTick' :: (CoreExpr -> CoreExpr) -- Apply before adding tick (float with) - -- Always a composition of (Tick t) wrappers - -> CoreExpr -- Current expression - -> CoreExpr - -- So in the call (mkTick' rest e), the expression - -- (rest e) - -- has the same type as e - -- Returns an expression equivalent to (Tick t (rest e)) - mkTick' rest expr = case expr of - -- Float ticks into unsafe coerce the same way we would do with a cast. - Case scrut bndr ty alts@[Alt ac abs _rhs] - | Just rhs <- isUnsafeEqualityCase scrut bndr alts - -> Case scrut bndr ty [Alt ac abs (mkTick' rest rhs)] + -- See Note [Scoping ticks and counting ticks] in GHC.Types.Tickish. + can_split = tickishCanSplit t - -- Cost centre ticks should never be reordered relative to each - -- other. Therefore we can stop whenever two collide. + -- mkTick' handles floating of tick `t` *into* the expression. + mkTick' :: CoreExpr -> CoreExpr + mkTick' expr = case expr of Tick t2 e - | ProfNote{} <- t2, ProfNote{} <- t -> Tick t $ rest expr - - -- Otherwise we assume that ticks of different placements float - -- through each other. - | tickishPlace t2 /= tickishPlace t -> Tick t2 $ mkTick' rest e - - -- For annotations this is where we make sure to not introduce - -- redundant ticks. - | tickishContains t t2 -> mkTick' rest e -- Drop t2 - | tickishContains t2 t -> rest e -- Drop t - | otherwise -> mkTick' (rest . Tick t2) e - - -- Ticks don't care about types, so we just float all ticks - -- through them. Note that it's not enough to check for these - -- cases top-level. While mkTick will never produce Core with type - -- expressions below ticks, such constructs can be the result of - -- unfoldings. We therefore make an effort to put everything into - -- the right place no matter what we start with. - Cast e co -> mkCast (mkTick' rest e) co - Coercion co -> Tick t $ rest (Coercion co) + + -- Common up ticks when possible, including profiling ticks that + -- share a cost centre and source notes that subsume one another. + | Just t' <- combineTickish_maybe t t2 + -> mkTick t' e + + -- Profiling ticks for different cost centres should never be reordered + -- relative to each other. Therefore, we stop whenever two collide. + | ProfNote {} <- t + , ProfNote {} <- t2 + -> Tick t e + + -- Ticks of different placements float through each other, so that each + -- tick can be floated into its expected position in the AST. + -- See Note [Tickish placement] in GHC.Types.Tickish. + | tickishPlace t2 /= tickishPlace t + -> Tick t2 $ mkTick' e Lam x e -- Always float through type lambdas. Even for non-type lambdas, -- floating is allowed for all but the most strict placement rule. | not (isRuntimeVar x) || tickishPlace t /= PlaceRuntime - -> Lam x $ mkTick' rest e + -> Lam x $ mkTick' e - -- If it is both counting and scoped, we split the tick into its - -- two components, often allowing us to keep the counting tick on - -- the outside of the lambda and push the scoped tick inside. - -- The point of this is that the counting tick can probably be - -- floated, and the lambda may then be in a position to be - -- beta-reduced. - | canSplit - -> Tick (mkNoScope t) $ rest $ Lam x $ mkTick (mkNoCount t) e + -- Push SCCs into lambdas. + -- See (PSCC2) in Note [Pushing SCCs inwards]. + | can_split + -> Tick (mkNoScope t) $ Lam x $ mkTick (mkNoCount t) e App f arg - -- Always float through type applications. + -- All ticks float inwards through non-runtime arguments, as per + -- Note [Tickish placement] in GHC.Types.Tickish. | not (isRuntimeArg arg) - -> App (mkTick' rest f) arg + -> App (mkTick' f) arg - -- We can also float through constructor applications, placement - -- permitting. Again we can split. - | isSaturatedConApp expr && (tickishPlace t==PlaceCostCentre || canSplit) + -- Push SCCs into saturated constructor applications. + -- See (PSCC3) in Note [Pushing SCCs inwards]. + | isSaturatedConApp expr + , tickishPlace t == PlaceCostCentre || can_split -> if tickishPlace t == PlaceCostCentre - then rest $ tickHNFArgs t expr - else Tick (mkNoScope t) $ rest $ tickHNFArgs (mkNoCount t) expr + then tickHNFArgs t expr + else Tick (mkNoScope t) $ tickHNFArgs (mkNoCount t) expr + + -- All ticks can be floated through casts, as per Note [Tickish placement]. + Cast e co -> mkCast (mkTick' e) co + + -- Treat 'unsafeCoerce' as if it was a cast: float all ticks inwards. + -- See Note [Push ticks into unsafeCoerce] + Case scrut bndr ty alts@[Alt ac abs _rhs] + | Just rhs <- isUnsafeEqualityCase scrut bndr alts + -> Case scrut bndr ty [Alt ac abs (mkTick' rhs)] Var x - | notFunction && tickishPlace t == PlaceCostCentre - -> rest expr -- Drop t - | notFunction && canSplit - -> Tick (mkNoScope t) $ rest expr - where - -- SCCs can be eliminated on variables provided the variable - -- is not a function. In these cases the SCC makes no difference: - -- the cost of evaluating the variable will be attributed to its - -- definition site. When the variable refers to a function, however, - -- an SCC annotation on the variable affects the cost-centre stack - -- when the function is called, so we must retain those. - notFunction = not (isFunTy (idType x)) - - Lit{} + -- Don't drop any ticks around a function + | isFunTy (idType x) + -> Tick t expr + -- Drop SCCs around non-function variables. + -- See (PSCC1) in Note [Pushing SCCs inwards]. | tickishPlace t == PlaceCostCentre - -> rest expr -- Drop t + -> -- Drop pure SCC ticks: scc<foo> (x :: Int) ==> x + expr + | can_split + -> -- Drop the scoping part of the tick, but keep the counting part. + Tick (mkNoScope t) expr + + -- It doesn't make sense to wrap static data (such as coercions, types and literals) + -- in a tick which compiles to code, as the code will never be run. + -- + -- It is in fact actively harmful, because Core Lint will fail on a + -- coercion binding such as let co = <scc> (...), see #26941. + -- It makes more sense to discard the cost centre tick rather than weakening + -- Core Lint. + e@(Coercion {}) | tickishIsCode t -> e + e@(Type {}) | tickishIsCode t -> e + e@(Lit {}) | tickishIsCode t -> e + + -- Catch-all: annotate where we stand. + -- In particular (but not only): Let, most Cases. + _other -> Tick t expr + +{- Note [Pushing SCCs inwards] +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Amongst all ticks, SCCs have the laxest placement properties (PlaceCostCentre, +as described in Note [Tickish placement] GHC.Types.Tickish): + + (PSCC1) SCCs around non-function variables can be eliminated. + The cost of evaluating the variable will be attributed to its definition + site, so the SCC makes no difference. Example: + + scc<foo> (x :: Int) ==> x + + NB: this is only valid when the variable is not a function. For example, in: + + scc<foo> (f :: Int -> Int) + + we must retain the cost centre annotation, as it affects the cost-centre + pointer when the function is called. Discarding the SCC in this case would + defeat the profiling mechanism entirely! + + (PSCC2) SCCs can be pushed into lambdas. + + scc<foo> (\x -> e) ==> \x -> scc<foo> e + + (PSCC3) We can push SCCs into (saturated) constructor applications. + For example, for an arity 2 data constructor 'D': + + scc<foo> (D e1 e2) ==> D (scc<foo> e1) (scc<foo> e2) + +Now, two kinds of ticks contain SCCs: + + - bare SCCs (i.e. ProfNote with profNoteCounts = False, profNoteScopes = True) + - profiling ticks that both count and scope + +The above explanation deals with bare SCCs. When handling profiling ticks that +both count and scope, we can split tick into two, so that the scoping part can +be pushed inwards (or even discarded). Specifically, we perform the following +transformations: - -- Catch-all: Annotate where we stand - _any -> Tick t $ rest expr + (PSCC1) Drop the SCC around non-function variables, keeping only the counting + part: + + scctick<foo> (x :: Int) ==> tick<foo> x + + (PSCC2) Push the SCC inside lambdas: + + scctick<foo> (\x. e) ==> tick<foo> (\x. scc<foo> e) + + NB: we must keep the counting part outside the lambda, in order to preserve + tick counter tallies – it would not be sound to push the counting part inside. + + (PSCC3) Push the SCC inside saturated contructor applications. + + scctick<foo> (D e1 e2) ==> tick<foo> (D (scc<foo> e1) (scc<foo> e2)) + +The benefit of these transformation is that the counting part, tick<foo>, can +likely be floated out of the way, which may expose additional optimisation +opportunities. For example: + + (scctick<foo> (\x. e)) arg + + ==>{PSCC2} + + (tick<foo> (\x. scc<foo> e)) arg + + ==>{GHC.Core.Opt.FloatOut.floatExpr, because 'tick<foo>' has no scope} + + tick<foo> ((\x. scc<foo> e) arg) + + ==>{beta reduction} + + tick<foo> (let x = arg in scc<foo> e) + +Note [Push ticks into unsafeCoerce] +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +In #25212, we had a program of the form: + + data Box = Box Any + asBox :: a -> Box + asBox x = {-# SCC asBox #-} Box (unsafeCoerce x) + +As per Note [Implementing unsafeCoerce] in GHC.Internal.Unsafe.Coerce, the call +to `unsafeCoerce` turns into + + case unsafeEqualityProof @Type @a @Any of + UnsafeRefl (co :: a ~# Any) -> x |> Sub co + +The worker for 'asBox' is then of the form: + + $wasBox = \@a (x :: a) -> + (# case unsafeEqualityProof @Type @a @Any of + UnsafeRefl (co :: a ~# Any) -> x |> Sub co + #) + +When inserting the SCC, we push it into the constructor as per (PSCC3) in +Note [Pushing SCCs inwards], so we get: + + $wasBox = \@a (x :: a) -> + tick<asBox> + (# scc<asBox> + case unsafeEqualityProof @Type @a @Any of + UnsafeRefl (co :: a ~# Any) -> x |> Sub co + #) + +Now, if we don't push the SCC tick into the case statement, Core Prep will +see an expression like 'MkSolo# (scc<asBox> ...)', which it will ANFise to +'let x = scc<asBox> ... in MkSolo# x', creating an unwanted thunk in the process. + +So the strategy is to treat this 'unsafeEqualityProof' case statement as if it +was a cast. We thus push the SCC into the RHS of the pattern match: + + $wasBox = \@a (x :: a) -> + tick<asBox> + (# case unsafeEqualityProof @Type @a @Any of + UnsafeRefl (co :: a ~# Any) -> scc<asBox> x |> Sub co + #) + +Then the SCC completely evaporates, as per (PSCC1) in Note [Pushing SCCs inwards]. +-} mkTicks :: [CoreTickish] -> CoreExpr -> CoreExpr mkTicks ticks expr = foldr mkTick expr ticks @@ -2573,8 +2701,8 @@ exprIsTickedString = isJust . exprIsTickedString_maybe exprIsTickedString_maybe :: CoreExpr -> Maybe ByteString exprIsTickedString_maybe (Lit (LitString bs)) = Just bs exprIsTickedString_maybe (Tick t e) - -- we don't tick literals with CostCentre ticks, compare to mkTick - | tickishPlace t == PlaceCostCentre = Nothing + -- Shortcut: ticks with code never wrap literals (compare with 'mkTick') + | tickishIsCode t = Nothing | otherwise = exprIsTickedString_maybe e exprIsTickedString_maybe _ = Nothing ===================================== compiler/GHC/Types/Tickish.hs ===================================== @@ -17,6 +17,7 @@ module GHC.Types.Tickish ( TickishPlacement(..), tickishPlace, tickishContains, + combineTickish_maybe, -- * Breakpoint tick identifiers BreakpointId(..), BreakTickIndex @@ -261,8 +262,12 @@ Ticks have two independent attributes: See Note [Scoped ticks] +Note that profiling notes which both count and scope can be split into two +separate ticks, one that counts and doesn't scope and one that scopes and doesn't +count; see 'tickishCanSplit', 'mkNoCount' and 'mkNoScope'. + Note [Counting ticks] -~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~~~ The following ticks count: - ProfNote ticks with profNoteCounts = True - HPC ticks @@ -290,7 +295,7 @@ sharing, so in practice the actual number of ticks may vary, except that we never change the value from zero to non-zero or vice-versa. Note [Scoped ticks] -~~~~~~~~~~~~~~~~~~~~ +~~~~~~~~~~~~~~~~~~~ The following ticks are scoped: - ProfNote ticks with profNoteScope = True - Breakpoints @@ -375,6 +380,52 @@ Whether we are allowed to float in additional cost depends on the tick: While these transformations are legal, we want to make a best effort to only make use of them where it exposes transformation opportunities. + +Note [Tickish placement] +~~~~~~~~~~~~~~~~~~~~~~~~ +The placement behaviour of ticks (i.e. which nodes we want the tick to be placed +around in the AST) is governed by 'TickishPlacement'. We generally try to push +ticks inwards until they end up placed around an AST node that is appropriate +for their placement rule, as we are about to describe. + +From most restrictive to least restrictive placement rules: + + - PlaceRuntime: counting ticks. + + Ticks with 'PlaceRuntime' placement want to be placed around run-time + expressions. They can be moved through pure compile-time constructs such as + other type arguments, casts, or type lambdas: + + tick <t> (f @ty) ==> (tick <t> f) @ty + tick <t> (e |> co) ==> (tick <t> e) |> co + tick <t> (/\a. e) ==> /\a. tick <t> e + + This is the most restrictive placement rule for ticks, as all tickishs have + in common that they want to track runtime behaviour. + + Any tick that counts (see Note [Counting ticks]) has 'PlaceRuntime' placement. + + - PlaceNonLam: source notes. + + Like PlaceRuntime, but we can also float the tick through value lambdas: + + tick <t> (\x. e) ==> \x. tick <t> e + + This makes sense where there is little difference between annotating the + lambda and annotating the lambda's code. + + - PlaceCostCentre: non-counting profiling ticks. + + In addition to floating through lambdas, cost-centre style tickishs can be + pushed into (saturated) constructor applications, and can be eliminated when + placed around non-function variables: + + tick <t> (C e1 e2) ==> C (tick <t> e1) (tick <t> e2) + + tick <t> (x :: Int) ==> (x :: Int) + + Neither the constructor application nor the variable 'x' are likely to have + any cost worth mentioning. -} -- | Returns @True@ for ticks that can be floated upwards easily even @@ -441,35 +492,19 @@ isProfTick _ = False -- annotating for example using @mkTick@. If we find that we want to -- put a tickish on an expression ruled out here, we try to float it -- inwards until we find a suitable expression. +-- +-- See Note [Tickish placement]. data TickishPlacement = - -- | Place ticks exactly on run-time expressions. We can still - -- move the tick through pure compile-time constructs such as - -- other ticks, casts or type lambdas. This is the most - -- restrictive placement rule for ticks, as all tickishs have in - -- common that they want to track runtime processes. The only - -- legal placement rule for counting ticks. - -- NB: We generally try to move these as close to the relevant - -- runtime expression as possible. This means they get pushed through - -- tyoe arguments. E.g. we create `(tick f) @Bool` instead of `tick (f @Bool)`. + -- | Place ticks exactly on run-time expressions, moving them through pure + -- compile-time constructs such as other ticks, casts or type lambdas. PlaceRuntime - -- | As @PlaceRuntime@, but we float the tick through all - -- lambdas. This makes sense where there is little difference - -- between annotating the lambda and annotating the lambda's code. + -- | As @PlaceRuntime@, but also allow to float the tick through all lambdas. | PlaceNonLam - -- | In addition to floating through lambdas, cost-centre style - -- tickishs can also be moved from constructors, non-function - -- variables and literals. For example: - -- - -- let x = scc<...> C (scc<...> y) (scc<...> 3) in ... - -- - -- Neither the constructor application, the variable or the - -- literal are likely to have any cost worth mentioning. And even - -- if y names a thunk, the call would not care about the - -- evaluation context. Therefore removing all annotations in the - -- above example is safe. + -- | As 'PlaceNonLam', but also float through constructors, non-function + -- variables and literals. | PlaceCostCentre deriving (Eq,Show) @@ -477,7 +512,9 @@ data TickishPlacement = instance Outputable TickishPlacement where ppr = text . show --- | Placement behaviour we want for the ticks +-- | Placement behaviour we want for the ticks. +-- +-- See Note [Tickish placement]. tickishPlace :: GenTickish pass -> TickishPlacement tickishPlace n@ProfNote{} | profNoteCount n = PlaceRuntime @@ -486,6 +523,43 @@ tickishPlace HpcTick{} = PlaceRuntime tickishPlace Breakpoint{} = PlaceRuntime tickishPlace SourceNote{} = PlaceNonLam +-- | Merge two ticks into one, if that is possible. +-- +-- Examples: +-- +-- - combine two source note ticks if one contains the other, +-- - combine a non-counting profiling tick with a non-scoping profiling tick +-- for the same cost centre +-- - combine two equal breakpoint ticks or HPC ticks +combineTickish_maybe :: Eq (GenTickish pass) + => GenTickish pass -> GenTickish pass -> Maybe (GenTickish pass) +combineTickish_maybe + (ProfNote { profNoteCC = cc1, profNoteCount = cnt1, profNoteScope = scope1 }) + (ProfNote { profNoteCC = cc2, profNoteCount = cnt2, profNoteScope = scope2 }) + | cc1 == cc2 + , not cnt1 || not cnt2 + = Just $ ProfNote { profNoteCC = cc1 + , profNoteCount = cnt1 || cnt2 + , profNoteScope = scope1 || scope2 + } +combineTickish_maybe t1@(SourceNote sp1 n1) t2@(SourceNote sp2 n2) + | n1 == n2 + , sp1 `containsSpan` sp2 + = Just t1 + | n1 == n2 + , sp2 `containsSpan` sp1 + = Just t2 + -- NB: it would be possible to use 'combineRealSrcSpans' instead, + -- but that has the risk of combining many source note ticks into a single + -- tick with a huge source span. +combineTickish_maybe t1@(HpcTick {}) t2@(HpcTick {}) + | t1 == t2 + = Just t1 +combineTickish_maybe t1@(Breakpoint {}) t2@(Breakpoint {}) + | t1 == t2 + = Just t1 +combineTickish_maybe _ _ = Nothing + -- | Returns whether one tick "contains" the other one, therefore -- making the second tick redundant. tickishContains :: Eq (GenTickish pass) ===================================== testsuite/tests/simplCore/should_compile/T26941.hs ===================================== @@ -0,0 +1,14 @@ +{-# LANGUAGE DataKinds #-} +{-# LANGUAGE GADTs #-} +{-# LANGUAGE TypeOperators #-} + +module T26941 where + +import GHC.TypeLits + +import T26941_aux ( SMayNat(SKnown), ListH, shxHead ) + +shsHead :: ListH (Just n : sh) Int -> SNat n +shsHead shx = + case shxHead shx of + SKnown SNat -> SNat ===================================== testsuite/tests/simplCore/should_compile/T26941_aux.hs ===================================== @@ -0,0 +1,20 @@ +{-# LANGUAGE DataKinds #-} +{-# LANGUAGE GADTs #-} +{-# LANGUAGE StandaloneKindSignatures #-} +{-# LANGUAGE TypeOperators #-} + +module T26941_aux where + +import Data.Kind +import GHC.TypeLits + +shxHead :: ListH (n : sh) i -> SMayNat i n +shxHead list = {-# SCC "bad_scc" #-} + ( case list of (i `ConsKnown` _) -> SKnown i ) + +type ListH :: [Maybe Nat] -> Type -> Type +data ListH sh i where + ConsKnown :: SNat n -> ListH sh i -> ListH (Just n : sh) i + +data SMayNat i n where + SKnown :: SNat n -> SMayNat i (Just n) ===================================== testsuite/tests/simplCore/should_compile/all.T ===================================== @@ -576,6 +576,8 @@ test('T26117', [grep_errmsg(r'==')], compile, ['-O -ddump-simpl -dsuppress-uniqu test('T26349', normal, compile, ['-O -ddump-rules']) test('T26681', normal, compile, ['-O']) +test('T26941', [extra_files(['T26941_aux.hs']), req_profiling], multimod_compile, ['T26941', '-v0 -O -prof']) + # T26709: we expect three `case` expressions not four test('T26709', [grep_errmsg(r'case')], multimod_compile, View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/2c7a6e8f983930724f70c9d603b9125… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/2c7a6e8f983930724f70c9d603b9125… You're receiving this email because of your account on gitlab.haskell.org.

1 0