- ghc-commits - Haskell.org

[Git][ghc/ghc] Pushed new branch wip/sjakobi/T27547
by Simon Jakobi (＠sjakobi) 29 Jul '26

29 Jul '26

Simon Jakobi pushed new branch wip/sjakobi/T27547 at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/sjakobi/T27547 You're receiving this email because of your account on gitlab.haskell.org. Manage all notifications: https://gitlab.haskell.org/-/profile/notifications | Help: https://gitlab.haskell.org/help

1 0

[Git][ghc/ghc][wip/spj-reinstallable-base2] Allow GHC.Essentials to be hidden
by sheaf (＠sheaf) 29 Jul '26

29 Jul '26

sheaf pushed to branch wip/spj-reinstallable-base2 at Glasgow Haskell Compiler / GHC Commits: 9471e797 by sheaf at 2026-07-29T02:23:10+02:00 Allow GHC.Essentials to be hidden - - - - - 22 changed files: - compiler/GHC/Builtin.hs - compiler/GHC/Driver/Downsweep.hs - compiler/GHC/Driver/MakeFile.hs - compiler/GHC/HsToCore.hs - compiler/GHC/Iface/Errors/Ppr.hs - compiler/GHC/Iface/Errors/Types.hs - compiler/GHC/Iface/Load.hs - compiler/GHC/Iface/Make.hs - compiler/GHC/Iface/Recomp.hs - compiler/GHC/Unit/External.hs - compiler/GHC/Unit/Finder.hs - compiler/GHC/Unit/Finder/Types.hs - compiler/GHC/Unit/State.hs - libraries/base/base.cabal.in - testsuite/tests/cabal/T12485/Makefile - testsuite/tests/driver/T27013e/T27013e.stderr - testsuite/tests/driver/T27013f/T27013f.stderr - testsuite/tests/driver/multipleHomeUnits/multipleHomeUnitsModuleVisibility.stderr - testsuite/tests/ghci/scripts/all.T - 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 The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/9471e797f4b30fb4bd49116ee7e49e8… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/9471e797f4b30fb4bd49116ee7e49e8… You're receiving this email because of your account on gitlab.haskell.org. Manage all notifications: https://gitlab.haskell.org/-/profile/notifications | Help: https://gitlab.haskell.org/help

1 0

[Git][ghc/ghc][wip/sjakobi/T27368-ppr] Cmm: print unreachable blocks under -dppr-debug (#27368)
by Simon Jakobi (＠sjakobi) 28 Jul '26

28 Jul '26

Simon Jakobi pushed to branch wip/sjakobi/T27368-ppr at Glasgow Haskell Compiler / GHC Commits: 5a644f29 by Simon Jakobi at 2026-07-29T00:17:39+02:00 Cmm: print unreachable blocks under -dppr-debug (#27368) Unreachable blocks linger in a CmmGraph's block map for most of the Cmm pipeline, but pprCmmGraph only ever printed the blocks reachable from the entry, so dumps looked consistent while the graph was not. Bugs like #27368 were hard to debug due to this. pprCmmGraph now appends the stored-but-unreachable blocks under a "// unreachable blocks:" heading when -dppr-debug is on. Existing dumps are unchanged. See Note [unreachable blocks] in GHC.Cmm.Pipeline. Assisted-by: Claude Opus 5 - - - - - 7 changed files: - + changelog.d/T27368-ppr-unreachable-cmm-blocks.md - compiler/GHC/Cmm.hs - compiler/GHC/Cmm/Pipeline.hs - docs/users_guide/debugging.rst - + testsuite/tests/cmm/should_compile/T27368-ppr-debug.cmm - + testsuite/tests/cmm/should_compile/T27368-ppr-debug.stderr - testsuite/tests/cmm/should_compile/all.T Changes: ===================================== changelog.d/T27368-ppr-unreachable-cmm-blocks.md ===================================== @@ -0,0 +1,10 @@ +section: cmm +issues: #27368 +mrs: !16417 +synopsis: + Cmm dumps now show unreachable blocks under ``-dppr-debug`` +description: + Unreachable blocks stay in a Cmm graph's block map for most of the Cmm + pipeline, but ``-ddump-cmm-*`` only ever printed the blocks reachable from + the graph's entry. Adding ``-dppr-debug`` now appends the stored but + unreachable blocks, which makes bugs like #27368 visible in the dumps. ===================================== compiler/GHC/Cmm.hs ===================================== @@ -151,15 +151,25 @@ instance OutputableP Platform CmmGraph where toBlockMap :: CmmGraph -> LabelMap CmmBlock toBlockMap (CmmGraph {g_graph=GMany NothingO body NothingO}) = body +-- | Print the blocks reachable from the entry, in reverse postorder. +-- +-- Under @-dppr-debug@ the unreachable blocks stored in the graph are appended +-- too. See Note [unreachable blocks] in "GHC.Cmm.Pipeline". pprCmmGraph :: Platform -> CmmGraph -> SDoc pprCmmGraph platform g = text "{" <> text "offset" - $$ nest 2 (vcat $ map (pdoc platform) blocks) + $$ nest 2 (vcat (map (pdoc platform) blocks) $$ unreachable) $$ text "}" - where blocks = revPostorder g - -- revPostorder has the side-effect of discarding unreachable code, - -- so pretty-printed Cmm will omit any unreachable blocks. This can - -- sometimes be confusing. + where + blocks = revPostorder g + + unreachable = getPprDebug $ \debug -> + if not debug || mapNull dead_blocks + then empty + else text "// unreachable blocks:" + $$ nest 2 (vcat $ map (pdoc platform) (mapElems dead_blocks)) + + dead_blocks = foldl' (\bs b -> mapDelete (entryLabel b) bs) (toBlockMap g) blocks revPostorder :: CmmGraph -> [CmmBlock] revPostorder g = {-# SCC "revPostorder" #-} ===================================== compiler/GHC/Cmm/Pipeline.hs ===================================== @@ -357,6 +357,7 @@ containing junk code. These aren't necessarily a problem, but removing them is good because it might save time in the native code generator later. +To make unreachable blocks visible in -ddump-cmm-* output, add -dppr-debug. -} dumpGraph :: Logger -> Platform -> Bool -> DumpFlag -> String -> CmmGraph -> IO () ===================================== docs/users_guide/debugging.rst ===================================== @@ -564,6 +564,11 @@ C-\- representation These flags dump various phases of GHC's C-\- pipeline. +Dumps of Cmm graphs print the blocks reachable from the entry, in reverse +post-order. To also show unreachable blocks, which can linger in the graph, +add :ghc-flag:`-dppr-debug`. These blocks are then listed under a +``// unreachable blocks:`` heading. + .. ghc-flag:: -ddump-cmm-verbose-by-proc :shortdesc: Show output from main C-\- pipeline passes (grouped by proc) :type: dynamic @@ -574,9 +579,6 @@ These flags dump various phases of GHC's C-\- pipeline. the chosen backend. Currently only the NCG backends runs additional passes ( :ghc-flag:`-ddump-opt-cmm` ). - Cmm dumps don't include unreachable blocks since we print - blocks in reverse post-order. - .. ghc-flag:: -ddump-cmm-verbose :shortdesc: Write output from main C-\- pipeline passes to files :type: dynamic ===================================== testsuite/tests/cmm/should_compile/T27368-ppr-debug.cmm ===================================== @@ -0,0 +1,19 @@ +#include "Cmm.h" + +// The block "dead" is stored in the graph but no block branches to it, so it +// only shows up in Cmm dumps under -dppr-debug. +testUnreachable (W_ x) +{ + if (x > 0) { + goto live; + } + return (x); + +dead: + x = x + 42; + return (x); + +live: + x = x - 1; + return (x); +} ===================================== testsuite/tests/cmm/should_compile/T27368-ppr-debug.stderr ===================================== @@ -0,0 +1 @@ +// unreachable blocks: ===================================== testsuite/tests/cmm/should_compile/all.T ===================================== @@ -13,6 +13,9 @@ test('T20725', normal, compile, ['-package ghc']) test('T23610', normal, makefile_test, ['T23610']) test('T24224', [cmm_src, grep_errmsg(r'(F64.*);', [1]), only_ways(['normal'])], compile, ['-no-hs-main -ddump-cmm -dsuppress-all -dsuppress-uniques']) test('T24474', cmm_src, compile, ['-optc-g3']) +# -dppr-debug makes stored-but-unreachable blocks visible in Cmm dumps (#27368) +test('T27368-ppr-debug', [cmm_src, grep_errmsg(r'(// unreachable blocks:)', [1]), only_ways(['normal'])], + compile, ['-no-hs-main -ddump-cmm-verbose-by-proc -dppr-debug -dsuppress-uniques']) test('T24474-cmm-gets-c-opts', cmm_src, compile, ['-optc-DFOO']) test('T24474-cmm-opt-order', cmm_src, compile, ['-optc-DFOO ' '-optCmmP-UFOO ' View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/5a644f2981e120e011f0f0f7dd7c133… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/5a644f2981e120e011f0f0f7dd7c133… You're receiving this email because of your account on gitlab.haskell.org. Manage all notifications: https://gitlab.haskell.org/-/profile/notifications | Help: https://gitlab.haskell.org/help

1 0

[Git][ghc/ghc][wip/sjakobi/T27368-ppr] Cmm: print unreachable blocks under -dppr-debug (#27368)
by Simon Jakobi (＠sjakobi) 28 Jul '26

28 Jul '26

Simon Jakobi pushed to branch wip/sjakobi/T27368-ppr at Glasgow Haskell Compiler / GHC Commits: 7496cace by Simon Jakobi at 2026-07-28T23:59:34+02:00 Cmm: print unreachable blocks under -dppr-debug (#27368) Unreachable blocks linger in a CmmGraph's block map for most of the Cmm pipeline, but pprCmmGraph only ever printed the blocks reachable from the entry, so dumps looked consistent while the graph was not. Bugs like #27368 were hard to debug due to this. pprCmmGraph now appends the stored-but-unreachable blocks under a "// unreachable blocks:" heading when -dppr-debug is on. Existing dumps are unchanged. See Note [unreachable blocks] in GHC.Cmm.Pipeline. Assisted-by: Claude Opus 5 - - - - - 7 changed files: - + changelog.d/T27368-ppr-unreachable-cmm-blocks.md - compiler/GHC/Cmm.hs - compiler/GHC/Cmm/Pipeline.hs - docs/users_guide/debugging.rst - + testsuite/tests/cmm/should_compile/T27368-ppr-debug.cmm - + testsuite/tests/cmm/should_compile/T27368-ppr-debug.stderr - testsuite/tests/cmm/should_compile/all.T Changes: ===================================== changelog.d/T27368-ppr-unreachable-cmm-blocks.md ===================================== @@ -0,0 +1,10 @@ +section: cmm +issues: #27368 +mrs: !16417 +synopsis: + Cmm dumps now show unreachable blocks under ``-dppr-debug`` +description: + Unreachable blocks stay in a Cmm graph's block map for most of the Cmm + pipeline, but ``-ddump-cmm-*`` only ever printed the blocks reachable from + the graph's entry. Adding ``-dppr-debug`` now appends the stored but + unreachable blocks, which makes bugs like #27368 visible in the dumps. ===================================== compiler/GHC/Cmm.hs ===================================== @@ -151,15 +151,24 @@ instance OutputableP Platform CmmGraph where toBlockMap :: CmmGraph -> LabelMap CmmBlock toBlockMap (CmmGraph {g_graph=GMany NothingO body NothingO}) = body +-- | Print the blocks reachable from the entry, in reverse postorder. Under +-- @-dppr-debug@ the blocks that are stored in the graph but unreachable are +-- appended too; see Note [unreachable blocks] in "GHC.Cmm.Pipeline". pprCmmGraph :: Platform -> CmmGraph -> SDoc pprCmmGraph platform g = text "{" <> text "offset" - $$ nest 2 (vcat $ map (pdoc platform) blocks) + $$ nest 2 (vcat (map (pdoc platform) blocks) $$ unreachable) $$ text "}" - where blocks = revPostorder g - -- revPostorder has the side-effect of discarding unreachable code, - -- so pretty-printed Cmm will omit any unreachable blocks. This can - -- sometimes be confusing. + where + blocks = revPostorder g + + unreachable = getPprDebug $ \debug -> + if not debug || mapNull dead_blocks + then empty + else text "// unreachable blocks:" + $$ nest 2 (vcat $ map (pdoc platform) (mapElems dead_blocks)) + + dead_blocks = foldl' (\bs b -> mapDelete (entryLabel b) bs) (toBlockMap g) blocks revPostorder :: CmmGraph -> [CmmBlock] revPostorder g = {-# SCC "revPostorder" #-} ===================================== compiler/GHC/Cmm/Pipeline.hs ===================================== @@ -357,6 +357,7 @@ containing junk code. These aren't necessarily a problem, but removing them is good because it might save time in the native code generator later. +To make unreachable blocks visible in -ddump-cmm-* output, add -dppr-debug. -} dumpGraph :: Logger -> Platform -> Bool -> DumpFlag -> String -> CmmGraph -> IO () ===================================== docs/users_guide/debugging.rst ===================================== @@ -564,6 +564,11 @@ C-\- representation These flags dump various phases of GHC's C-\- pipeline. +Dumps of Cmm graphs print the blocks reachable from the entry, in reverse +post-order. To also show unreachable blocks, which can linger in the graph, +add :ghc-flag:`-dppr-debug`. These blocks are then listed under a +``// unreachable blocks:`` heading. + .. ghc-flag:: -ddump-cmm-verbose-by-proc :shortdesc: Show output from main C-\- pipeline passes (grouped by proc) :type: dynamic @@ -574,9 +579,6 @@ These flags dump various phases of GHC's C-\- pipeline. the chosen backend. Currently only the NCG backends runs additional passes ( :ghc-flag:`-ddump-opt-cmm` ). - Cmm dumps don't include unreachable blocks since we print - blocks in reverse post-order. - .. ghc-flag:: -ddump-cmm-verbose :shortdesc: Write output from main C-\- pipeline passes to files :type: dynamic ===================================== testsuite/tests/cmm/should_compile/T27368-ppr-debug.cmm ===================================== @@ -0,0 +1,19 @@ +#include "Cmm.h" + +// The block "dead" is stored in the graph but no block branches to it, so it +// only shows up in Cmm dumps under -dppr-debug. +testUnreachable (W_ x) +{ + if (x > 0) { + goto live; + } + return (x); + +dead: + x = x + 42; + return (x); + +live: + x = x - 1; + return (x); +} ===================================== testsuite/tests/cmm/should_compile/T27368-ppr-debug.stderr ===================================== @@ -0,0 +1 @@ +// unreachable blocks: ===================================== testsuite/tests/cmm/should_compile/all.T ===================================== @@ -13,6 +13,9 @@ test('T20725', normal, compile, ['-package ghc']) test('T23610', normal, makefile_test, ['T23610']) test('T24224', [cmm_src, grep_errmsg(r'(F64.*);', [1]), only_ways(['normal'])], compile, ['-no-hs-main -ddump-cmm -dsuppress-all -dsuppress-uniques']) test('T24474', cmm_src, compile, ['-optc-g3']) +# -dppr-debug makes stored-but-unreachable blocks visible in Cmm dumps (#27368) +test('T27368-ppr-debug', [cmm_src, grep_errmsg(r'(// unreachable blocks:)', [1]), only_ways(['normal'])], + compile, ['-no-hs-main -ddump-cmm-verbose-by-proc -dppr-debug -dsuppress-uniques']) test('T24474-cmm-gets-c-opts', cmm_src, compile, ['-optc-DFOO']) test('T24474-cmm-opt-order', cmm_src, compile, ['-optc-DFOO ' '-optCmmP-UFOO ' View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/7496cace78470908179ebcfc5833228… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/7496cace78470908179ebcfc5833228… You're receiving this email because of your account on gitlab.haskell.org. Manage all notifications: https://gitlab.haskell.org/-/profile/notifications | Help: https://gitlab.haskell.org/help

1 0

[Git][ghc/ghc] Pushed new branch wip/sjakobi/ghc-config-ignore
by Simon Jakobi (＠sjakobi) 28 Jul '26

28 Jul '26

Simon Jakobi pushed new branch wip/sjakobi/ghc-config-ignore at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/sjakobi/ghc-config-ignore You're receiving this email because of your account on gitlab.haskell.org. Manage all notifications: https://gitlab.haskell.org/-/profile/notifications | Help: https://gitlab.haskell.org/help

1 0

[Git][ghc/ghc][wip/9.14.2-backports] 2 commits: hadrian: add text and parsec to stage0Packages
by Zubin (＠wz1000) 28 Jul '26

28 Jul '26

Zubin pushed to branch wip/9.14.2-backports at Glasgow Haskell Compiler / GHC Commits: b0fc1bd4 by Zubin Duggal at 2026-07-29T03:06:33+05:30 hadrian: add text and parsec to stage0Packages text depends on template-haskell-lift, and template-haskell-lift needs to be built in tree rather than using the boot package, so `text` and `parsec` (which depends on `text`) must also be built in tree rather than copied from the boot pkgdb. Also need to disable -simdutf for text when building the stage0 version, wasm backend doesn't like -simdutf. - - - - - eddff3ab by Zubin Duggal at 2026-07-29T03:06:33+05:30 Prepare 9.14.2 Bump filpath submodule to 1.5.5.0 Bump Win32 submodule to 2.14.2.2 Metric Increase: T4029 T13379 haddock.base libdir Metric Decrease: T5030 T24602_perf_size T25046_perf_size_gzip T25046_perf_size_unicode T25046_perf_size_unicode_gzip - - - - - 47 changed files: - − changelog.d/26616 - − changelog.d/T26978 - − changelog.d/T26979 - − changelog.d/T27046 - − changelog.d/T27047 - − changelog.d/T27123.md - − changelog.d/T27124.md - − changelog.d/T27131 - − changelog.d/T27182.md - − changelog.d/T27202 - − changelog.d/T27225 - − changelog.d/T27386 - − changelog.d/bump-process - − changelog.d/deterministic-usage-order - − changelog.d/fix-absent-dict-projection - − changelog.d/fix-blackhole-handling - − changelog.d/fix-cmm-atomic-load-store - − changelog.d/fix-compacting-gc-ap-27434 - − changelog.d/fix-exponential-case-desugar-27383 - − changelog.d/fix-finalizers-27072 - − changelog.d/fix-layout-stack-fcall - − changelog.d/fix-peekitbl-no-tntc - − changelog.d/fix-use-std-ap-thunk - − changelog.d/ghc-pkg-faster-closure - − changelog.d/hadrian-stale-package-confs-26661 - − changelog.d/jobserver-leak-fix - − changelog.d/more-efficient-home-unit-imports-finding - − changelog.d/reexported-module-errors - − changelog.d/semaphore-v2 - − changelog.d/tag-inference-27005 - − changelog.d/unused-type - − changelog.d/wasm-fix-serviceworker - − changelog.d/windows-rethrow-overlapped-exception - configure.ac - docs/users_guide/9.14.2-notes.rst - hadrian/bootstrap/generate_bootstrap_plans - + hadrian/bootstrap/plan-9_10_3.json - + hadrian/bootstrap/plan-9_12_4.json - + hadrian/bootstrap/plan-bootstrap-9_10_3.json - + hadrian/bootstrap/plan-bootstrap-9_12_4.json - hadrian/src/Rules/ToolArgs.hs - hadrian/src/Settings/Default.hs - hadrian/src/Settings/Packages.hs - libraries/Win32 - libraries/base/base.cabal.in - libraries/base/changelog.md - libraries/filepath The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/b72a3cb735cf32a1e511379e408d7c… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/b72a3cb735cf32a1e511379e408d7c… You're receiving this email because of your account on gitlab.haskell.org. Manage all notifications: https://gitlab.haskell.org/-/profile/notifications | Help: https://gitlab.haskell.org/help

1 0

[Git][ghc/ghc][wip/marge_bot_batch_merge_job] 5 commits: EPA: ClsInstDecl with decls as [LHsDecl GhcPs] in GhcPs
by Marge Bot (＠marge-bot) 28 Jul '26

28 Jul '26

Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC Commits: bee1913d by Alan Zimmerman at 2026-07-28T16:42:29-04:00 EPA: ClsInstDecl with decls as [LHsDecl GhcPs] in GhcPs Similar to 4fdfe75731e01dad7d7fa474c2703d0d3965afb1, this commit changes the as-parsed representation of class instance declarations to [LHsDecl GhcPs], and only separates them by type from the renamer onward. This also allows us to remove all the AnnSortKey machinery for exact printing, as it is now no longer needed. - - - - - 72c55eee by Cheng Shao at 2026-07-28T16:43:11-04:00 hadrian: implement and use writeFileAtomic to fix race condition This patch implements `writeFileAtomic` in hadrian and change all invocations of shake non-atomic `writeFile'` to use `writeFileAtomic`, to avoid multiple hadrian concurrent invocations overwriting the same in-tree generated file not in the build root directory. Fixes #27536. Additional notes: - `writeFileChanged`/`writeFileChangedBS` cannot be made atomic since it involves reading the file's older version, so their uses are left alone. It doesn't affect #27536 given their outputs are contained in the build root directory. - It's possible to shrink this patch by only making writes outside the build root directory atomic. But I think it's not worth the effort for fine grained distinction here, and atomic writes within the build root directory should also improve robustness of a hadrian build. - In the longer term we do want to make a ghc build only generate files within the build root directory, though that's a lot of work and outside the scope of this particular bugfix. Co-authored-by: Codex <codex(a)openai.com> - - - - - 540db46e by Sylvain Henry at 2026-07-28T17:15:37-04:00 RTS: correctly mark slop bytes when shrinking large arrays (#19048) Correctly mark slop bytes even when profiling is off so that heap census doesn't traverse garbage-collected closures. - - - - - 586e8666 by Simon Jakobi at 2026-07-28T17:15:39-04:00 testsuite: Relax T24471 residency tolerance T24471 peak residency fluctuates enough on i386 to cause spurious failures. Use the standard residency tolerance while retaining the existing allocation threshold. See https://gitlab.haskell.org/ghc/ghc/-/work_items/24471#note_682303. Assisted-by: gpt-5.6-sol via Codex CLI - - - - - a95f73da by Cheng Shao at 2026-07-28T17:15:40-04:00 compiler: fix missing top-level procedure labels in cmm dumps This patch fixes missing top-level procedure labels in some intermediate Cmm pass dumps. Fixes #27553. Co-authored-by: Codex <codex(a)openai.com> - - - - - 65 changed files: - + changelog.d/fix-cmm-dump-labels - + changelog.d/fix-heap-census-large-arrays-19048 - compiler/GHC/Cmm/Pipeline.hs - compiler/GHC/Hs/Decls.hs - compiler/GHC/Hs/Instances.hs - compiler/GHC/Hs/Stats.hs - compiler/GHC/Hs/Utils.hs - compiler/GHC/HsToCore/Docs.hs - compiler/GHC/HsToCore/Quote.hs - compiler/GHC/Iface/Ext/Ast.hs - compiler/GHC/Parser.y - compiler/GHC/Parser/Annotation.hs - compiler/GHC/Parser/PostProcess.hs - compiler/GHC/Rename/Module.hs - compiler/GHC/Rename/Names.hs - compiler/GHC/Tc/TyCl.hs - compiler/GHC/Tc/TyCl/Class.hs - compiler/GHC/Tc/TyCl/Instance.hs - compiler/GHC/Tc/Utils/Env.hs - compiler/GHC/ThToHs.hs - compiler/Language/Haskell/Syntax/Decls.hs - hadrian/src/Hadrian/Builder/Ar.hs - hadrian/src/Hadrian/Utilities.hs - hadrian/src/Rules/BinaryDist.hs - hadrian/src/Rules/CabalReinstall.hs - hadrian/src/Rules/Documentation.hs - hadrian/src/Rules/Generate.hs - hadrian/src/Rules/Library.hs - hadrian/src/Rules/Nofib.hs - hadrian/src/Rules/Program.hs - hadrian/src/Rules/Register.hs - hadrian/src/Rules/Rts.hs - hadrian/src/Rules/SourceDist.hs - hadrian/src/Rules/Test.hs - hadrian/src/Rules/ToolArgs.hs - rts/Apply.cmm - rts/ZeroSlop.c → rts/MarkSlop.c - rts/PrimOps.cmm - rts/Printer.c - rts/ProfHeap.c - rts/RtsFlags.c - rts/ThreadPaused.c - rts/include/Cmm.h - rts/include/rts/storage/ClosureMacros.h - rts/rts.cabal - rts/sm/NonMovingMark.c - rts/sm/Sanity.c - rts/sm/Storage.c - testsuite/tests/haddock/should_compile_flag_haddock/T17544.stderr - testsuite/tests/parser/should_compile/DumpRenamedAst.stderr - testsuite/tests/perf/compiler/all.T - testsuite/tests/printer/Test24533.stdout - + testsuite/tests/rts/T19048.hs - + testsuite/tests/rts/T19048.stdout - testsuite/tests/rts/all.T - utils/check-exact/ExactPrint.hs - utils/check-exact/Utils.hs - utils/haddock/haddock-api/src/Haddock/Backends/Hoogle.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/Convert.hs - utils/haddock/haddock-api/src/Haddock/GhcUtils.hs - utils/haddock/haddock-api/src/Haddock/Interface/Create.hs - utils/haddock/haddock-api/src/Haddock/Interface/Rename.hs - utils/haddock/haddock-api/src/Haddock/Types.hs The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/a2e94c6bb1b28a05f8004ce51f03f3… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/a2e94c6bb1b28a05f8004ce51f03f3… You're receiving this email because of your account on gitlab.haskell.org. Manage all notifications: https://gitlab.haskell.org/-/profile/notifications | Help: https://gitlab.haskell.org/help

1 0

[Git][ghc/ghc][wip/sjakobi/nonmoving-gc-test-fixes] 2 commits: testsuite: Omit T22859 in nonmoving threaded ways
by Simon Jakobi (＠sjakobi) 28 Jul '26

28 Jul '26

Simon Jakobi pushed to branch wip/sjakobi/nonmoving-gc-test-fixes at Glasgow Haskell Compiler / GHC Commits: 84b93041 by Simon Jakobi at 2026-07-28T23:12:27+02:00 testsuite: Omit T22859 in nonmoving threaded ways T22859 checks allocation-limit handlers with output that depends on precise allocation behaviour. The nonmoving threaded ways change where these limits are reached, just as the already-omitted LLVM ways do. Omit these ways instead of treating their incidental output differences as test failures. Assisted-by: gpt-5.6-sol via Codex CLI - - - - - 514e985c by Simon Jakobi at 2026-07-28T23:12:27+02:00 testsuite: Make listThreads1 insensitive to the RTS's own threads listThreads1 expected `listThreads` to return exactly `[ThreadId 1]`. That holds only under a non-threaded RTS: a threaded RTS forks the per-capability `IOManager on cap N` threads and the `TimerManager` from `hs_init_ghc` before `main` runs, so the test fails in every threaded way. The test was added in d442ac053f ("JS: fix thread-related primops") as a regression test for the JS backend's ListThreadsOp, which returned the `h$threads` run queue and thereby omitted the *running* thread. Checking that `myThreadId` is among the listed threads states that property directly, and still catches the original bug. Checking for the presence of `ThreadId 1` would not do: under a threaded RTS the main thread is created last, and `ThreadId 1` is a startup thread that has already finished and disappears from the list at the next major GC. This mirrors 8519af6080 ("testsuite: Make listThreads more robust"). Assisted-by: Claude Opus 4.5 - - - - - 3 changed files: - libraries/base/tests/listThreads1.hs - libraries/base/tests/listThreads1.stdout - testsuite/tests/rts/all.T Changes: ===================================== libraries/base/tests/listThreads1.hs ===================================== @@ -2,5 +2,10 @@ module Main where import GHC.Conc.Sync +-- Regression test for the JS backend's ListThreadsOp, which used to omit the +-- running thread. Whatever other threads the RTS has is irrelevant here. main :: IO () -main = listThreads >>= print +main = do + tid <- myThreadId + ts <- listThreads + print (tid `elem` ts) ===================================== libraries/base/tests/listThreads1.stdout ===================================== @@ -1 +1 @@ -[ThreadId 1] +True ===================================== testsuite/tests/rts/all.T ===================================== @@ -679,7 +679,7 @@ test('T22859', [js_skip, # This test is vulnerable to changes in allocation behaviour, so we disable it in some ways when(arch('wasm32'), skip), - omit_ways(llvm_ways)], + omit_ways(llvm_ways + ['nonmoving_thr', 'nonmoving_thr_sanity'])], compile_and_run, ['-with-rtsopts -A8K']) # These tests need access to the internal RTS headers. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/ac137b50d30c9511a7994ac59e6d95… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/ac137b50d30c9511a7994ac59e6d95… You're receiving this email because of your account on gitlab.haskell.org. Manage all notifications: https://gitlab.haskell.org/-/profile/notifications | Help: https://gitlab.haskell.org/help

1 0

[Git][ghc/ghc][master] hadrian: implement and use writeFileAtomic to fix race condition
by Marge Bot (＠marge-bot) 28 Jul '26

28 Jul '26

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: 72c55eee by Cheng Shao at 2026-07-28T16:43:11-04:00 hadrian: implement and use writeFileAtomic to fix race condition This patch implements `writeFileAtomic` in hadrian and change all invocations of shake non-atomic `writeFile'` to use `writeFileAtomic`, to avoid multiple hadrian concurrent invocations overwriting the same in-tree generated file not in the build root directory. Fixes #27536. Additional notes: - `writeFileChanged`/`writeFileChangedBS` cannot be made atomic since it involves reading the file's older version, so their uses are left alone. It doesn't affect #27536 given their outputs are contained in the build root directory. - It's possible to shrink this patch by only making writes outside the build root directory atomic. But I think it's not worth the effort for fine grained distinction here, and atomic writes within the build root directory should also improve robustness of a hadrian build. - In the longer term we do want to make a ghc build only generate files within the build root directory, though that's a lot of work and outside the scope of this particular bugfix. Co-authored-by: Codex <codex(a)openai.com> - - - - - 14 changed files: - hadrian/src/Hadrian/Builder/Ar.hs - hadrian/src/Hadrian/Utilities.hs - hadrian/src/Rules/BinaryDist.hs - hadrian/src/Rules/CabalReinstall.hs - hadrian/src/Rules/Documentation.hs - hadrian/src/Rules/Generate.hs - hadrian/src/Rules/Library.hs - hadrian/src/Rules/Nofib.hs - hadrian/src/Rules/Program.hs - hadrian/src/Rules/Register.hs - hadrian/src/Rules/Rts.hs - hadrian/src/Rules/SourceDist.hs - hadrian/src/Rules/Test.hs - hadrian/src/Rules/ToolArgs.hs Changes: ===================================== hadrian/src/Hadrian/Builder/Ar.hs ===================================== @@ -43,7 +43,7 @@ runAr :: FilePath -- ^ base name to use for response files -> Action () runAr outputFilePath arPath flagArgs fileArgs buildOptions = do rspFile <- responseFilePath outputFilePath - writeFile' rspFile $ unwords fileArgs + writeFileAtomic rspFile $ unwords fileArgs cmd [arPath] flagArgs ('@' : rspFile) buildOptions -- | Invoke @ar@ given a path to it and a list of arguments. Note that @ar@ ===================================== hadrian/src/Hadrian/Utilities.hs ===================================== @@ -21,7 +21,9 @@ module Hadrian.Utilities ( -- * File system operations copyFile, copyFileUntracked, createFileLink, fixFile, makeExecutable, moveFile, removeFile, createDirectory, copyDirectory, - moveDirectory, removeDirectory, removeFile_, writeFileChangedBS, + moveDirectory, removeDirectory, removeFile_, + writeFileAtomic, writeFileLinesAtomic, + writeFileChangedBS, -- * Diagnostic info Colour (..), ANSIColour (..), putColoured, shouldUseColor, @@ -48,6 +50,7 @@ import Data.List.Extra import Data.Maybe import Data.Typeable (TypeRep, typeOf) import Development.Shake hiding (Normal) +import qualified Development.Shake as Shake import Development.Shake.Classes import Development.Shake.Command (CmdArgument (..), IsCmdArgument (toCmdArgument)) import Development.Shake.FilePath @@ -341,7 +344,7 @@ withResponseFileIfLongCmd outputFilePath argsPre argsResp argsPost = do then cmd argsPre argsResp argsPost else do rspFile <- responseFilePath outputFilePath - writeFile' rspFile (escapeArgs argsResp) + writeFileAtomic rspFile (escapeArgs argsResp) cmd argsPre ['@' : rspFile] argsPost -- | Convert a command's output file path to a response file path to be used for that command. @@ -400,6 +403,19 @@ copyFileUntracked source target = do putProgressInfo =<< renderAction "Copy file (untracked)" source target liftIO $ IO.copyFile source target +-- | An atomic version of Shake's 'Shake.writeFile''. +writeFileAtomic :: FilePath -> String -> Action () +writeFileAtomic file contents = do + let dir = takeDirectory file + liftIO $ IO.createDirectoryIfMissing True dir + withTempFileWithin dir $ \temp -> do + Shake.writeFile' temp contents + liftIO $ IO.renameFile temp file + +-- | An atomic version of Shake's 'Shake.writeFileLines'. +writeFileLinesAtomic :: FilePath -> [String] -> Action () +writeFileLinesAtomic file = writeFileAtomic file . unlines + -- | Transform a given file by applying a function to its contents. fixFile :: FilePath -> (String -> String) -> Action () fixFile file f = do @@ -409,7 +425,7 @@ fixFile file f = do let new = f old IO.evaluate $ rnf new return new - liftIO $ writeFile file contents + writeFileAtomic file contents -- | Make a given file executable by running the @chmod +x@ command. makeExecutable :: FilePath -> Action () ===================================== hadrian/src/Rules/BinaryDist.hs ===================================== @@ -234,7 +234,7 @@ buildBinDistDir root conf@BindistConfig{..} = do bindistContext = vanillaContext library_stage compiler bindistSettingsContent <- interpretInContext bindistContext $ generateSettings bindistSettings False "package.conf.d" - writeFile' bindistSettings bindistSettingsContent + writeFileAtomic bindistSettings bindistSettingsContent copyDirectory rtsIncludeDir bindistFilesDir when windowsHost $ createGhcii (bindistFilesDir -/- "bin") @@ -296,7 +296,7 @@ buildBinDistDir root conf@BindistConfig{..} = do need $ map (bindistFilesDir -/-) (["configure", "Makefile"] ++ bindistInstallFiles) copyFile ("hadrian" -/- "bindist" -/- "config.mk.in") (bindistFilesDir -/- "config.mk.in") - generateBuildMk conf >>= writeFile' (bindistFilesDir -/- "build.mk") + generateBuildMk conf >>= writeFileAtomic (bindistFilesDir -/- "build.mk") copyFile ("hadrian" -/- "cfg" -/- "default.target.in") (bindistFilesDir -/- "default.target.in") copyFile ("hadrian" -/- "cfg" -/- "default.host.target.in") (bindistFilesDir -/- "default.host.target.in") @@ -314,7 +314,7 @@ buildBinDistDir root conf@BindistConfig{..} = do versioned_wrapper = wrapper_name ++ "-" ++ suffix versioned_wrapper_path = bindistFilesDir -/- "wrappers" -/- versioned_wrapper -- Write the wrapper to the versioned path - writeFile' versioned_wrapper_path wrapper_content + writeFileAtomic versioned_wrapper_path wrapper_content -- Create a symlink from the non-versioned to the versioned. liftIO $ do IO.removeFile unversioned_wrapper_path <|> return () ===================================== hadrian/src/Rules/CabalReinstall.hs ===================================== @@ -38,7 +38,7 @@ cabalBuildRules = do withVerbosity Diagnostic $ buildWithCmdOptions [] $ target (vanillaContext Stage2 pkg) (Cabal Install Stage2) [] [] - liftIO $ writeFile outpath "done" + writeFileAtomic outpath "done" phony "build-cabal" $ need [root -/- "stage-cabal" -/- "bin" -/- ".stamp"] @@ -86,8 +86,8 @@ cabalBuildRules = do ] output_file = outputDir -/- wrapper_name wrapper_content <- wrapper Stage2 wrapper_name - writeFile' output_file (wrapper_prefix ++ wrapper_content) + writeFileAtomic output_file (wrapper_prefix ++ wrapper_content) makeExecutable output_file pure () - writeFile' stamp "OK" + writeFileAtomic stamp "OK" ===================================== hadrian/src/Rules/Documentation.hs ===================================== @@ -268,7 +268,7 @@ buildPackageDocumentation = do syn <- pkgSynopsis (Context.package ctx) desc <- pkgDescription (Context.package ctx) let prologue = if null desc then syn else desc - liftIO $ writeFile file prologue + writeFileAtomic file prologue root -/- htmlRoot -/- "libraries/*/*.haddock" %> \file -> do context <- pkgDocContext =<< getPkgDocTarget root file ===================================== hadrian/src/Rules/Generate.hs ===================================== @@ -353,7 +353,7 @@ templateRuleFrom inPath outPath interps = do outPath %> \_ -> do s <- readFile' inPath result <- runInterpolations interps s - writeFile' outPath result + writeFileAtomic outPath result putSuccess ("| Successfully generated " ++ outPath ++ " from its template") templateRule :: FilePath -> Interpolations -> Rules () ===================================== hadrian/src/Rules/Library.hs ===================================== @@ -140,7 +140,7 @@ buildPackage root fp = do -- Write the current time into the file so the file always changes if -- we restamp it because a dependency changes. time <- liftIO $ getCurrentTime - liftIO $ writeFile fp (show time) + writeFileAtomic fp (show time) ways <- interpretInContext ctx getLibraryWays let hasVanilla = elem vanilla ways hasDynamic = elem dynamic ways @@ -148,7 +148,7 @@ buildPackage root fp = do when ((hasVanilla && hasDynamic) && support && way == vanilla) $ do stamp <- (pkgStampFile (ctx { way = dynamic })) - liftIO $ writeFile stamp (show time) + writeFileAtomic stamp (show time) ===================================== hadrian/src/Rules/Nofib.hs ===================================== @@ -42,7 +42,7 @@ nofibRules = do unit $ cmd (Cwd "nofib") [makePath] ["clean"] unit $ cmd (Cwd "nofib") [makePath] (nofibArgs ++ ["boot"]) (Exit e, Stdouterr log) <- cmd (Cwd "nofib") [makePath] nofibArgs - writeFile' fp log + writeFileAtomic fp log if e == ExitSuccess then putVerbose $ "nofib log available at " ++ fp else error $ "nofib failed, full log available at " ++ fp ===================================== hadrian/src/Rules/Program.hs ===================================== @@ -32,7 +32,7 @@ buildProgramRules rs = do top <- topDirectory need [ top -/- "configure" ] copyDirectory (top -/- "inplace" -/- "mingw") root - writeFile' stampPath "OK" + writeFileAtomic stampPath "OK" -- Rules for programs that are actually built by hadrian. forM_ allStages $ \stage -> ===================================== hadrian/src/Rules/Register.hs ===================================== @@ -94,7 +94,7 @@ registerPackageRules rs stage iplace = do -- leads to errors in GHC). buildWithResources rs $ target (Context stage compiler vanilla iplace) (GhcPkg Recache stage) [] [] - writeFileLines stamp [] + writeFileLinesAtomic stamp [] -- Special rule for registering system-cxx-std-lib root -/- relativePackageDbPath (PackageDbLoc stage iplace) -/- systemCxxStdLibConf %> \file -> do ===================================== hadrian/src/Rules/Rts.hs ===================================== @@ -21,7 +21,7 @@ buildGhcInternalImportDef target = do templateIn <- readFile' "rts/win32/libHSghc-internal.def.in" let dllName = takeFileName target -<.> "dll" templateOut = replace "@GhcInternalDll@" dllName templateIn - writeFile' target templateOut + writeFileAtomic target templateOut buildGhcInternalImportLib :: Stage -> FilePath -> Action () buildGhcInternalImportLib stg target = do ===================================== hadrian/src/Rules/SourceDist.hs ===================================== @@ -36,7 +36,7 @@ sourceDistRules = alternatives $ do need [mingw_tarballs_stamp] mingw_tarballs_stamp %> \stamp -> do build (target (vanillaContext Stage1 compiler) (Win32Tarballs DownloadTarballs) [] []) - writeFile' stamp "OK" + writeFileAtomic stamp "OK" archiveSourceTree :: (FilePath -> Action ()) -> FilePath -> Action () ===================================== hadrian/src/Rules/Test.hs ===================================== @@ -328,7 +328,7 @@ timeoutProgBuilder = do let script = unlines [ "#!/bin/sh" , "exec " ++ python ++ " $0.py \"$@\"" ] - writeFile' (root -/- timeoutPath) script + writeFileAtomic (root -/- timeoutPath) script makeExecutable (root -/- timeoutPath) -- | Build extra programs and libraries required by testsuite ===================================== hadrian/src/Rules/ToolArgs.hs ===================================== @@ -63,7 +63,7 @@ multiSetup pkg_s = do -- Get the arguments for all the targets pargs <- mapM one_args tool_targets -- Build any other dependencies (such as generated files) - liftIO $ writeOutput (concatMap (\x -> ["-unit", x]) (map ( "@" <>) pargs)) + writeOutput (concatMap (\x -> ["-unit", x]) (map ( "@" <>) pargs)) where resp_file root p = root </> "multi" </> pkgName p @@ -96,7 +96,7 @@ multiSetup pkg_s = do normalisePackageIds (x:xs) = x : normalisePackageIds xs normalisePackageIds [] = [] - writeFile' (resp_file root p) (intercalate "\n" (normalise_ghc arg_list + writeFileAtomic (resp_file root p) (intercalate "\n" (normalise_ghc arg_list ++ modules cd ++ concatMap rexp (reexportModules cd) ++ ["-outputdir", hidir, @@ -112,11 +112,12 @@ toolRuleBody fp = do Just (_, (p, extra)) -> mkToolTarget extra p Nothing -> fail $ "No prefixes matched " ++ show fp ++ " IN\n " ++ show mm -writeOutput :: [String] -> IO () +writeOutput :: [String] -> Action () writeOutput args = do - liftIO $ lookupEnv "TOOL_OUTPUT" >>= \case - Nothing -> putStrLn (intercalate "\n" args) - Just out -> writeFile out (intercalate "\n" args) + output <- liftIO $ lookupEnv "TOOL_OUTPUT" + case output of + Nothing -> liftIO $ putStrLn (intercalate "\n" args) + Just out -> writeFileAtomic out (intercalate "\n" args) mkToolTarget :: [String] -> Package -> Action () mkToolTarget es p = do @@ -138,7 +139,7 @@ mkToolTarget es p = do need (gens ++ srcs ++ dep_confs) arg_list <- interpret fake_target getArgs - liftIO $ writeOutput (arg_list ++ es) + writeOutput (arg_list ++ es) -- This list is quite a lot like stage0packages but doesn't include -- critically the `exe:ghc` component as that depends on the GHC library @@ -200,4 +201,3 @@ dirMap = do cd <- readContextData c ids <- liftIO $ mapM canonicalizePath [pkgPath p </> i | i <- srcDirs cd] return $ map (,(p, modules cd ++ otherModules cd)) ids - View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/72c55eee0160af1d408c45c551051f9… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/72c55eee0160af1d408c45c551051f9… You're receiving this email because of your account on gitlab.haskell.org. Manage all notifications: https://gitlab.haskell.org/-/profile/notifications | Help: https://gitlab.haskell.org/help

1 0

[Git][ghc/ghc][master] EPA: ClsInstDecl with decls as [LHsDecl GhcPs] in GhcPs
by Marge Bot (＠marge-bot) 28 Jul '26

28 Jul '26

Marge Bot pushed to branch master at Glasgow Haskell Compiler / GHC Commits: bee1913d by Alan Zimmerman at 2026-07-28T16:42:29-04:00 EPA: ClsInstDecl with decls as [LHsDecl GhcPs] in GhcPs Similar to 4fdfe75731e01dad7d7fa474c2703d0d3965afb1, this commit changes the as-parsed representation of class instance declarations to [LHsDecl GhcPs], and only separates them by type from the renamer onward. This also allows us to remove all the AnnSortKey machinery for exact printing, as it is now no longer needed. - - - - - 31 changed files: - compiler/GHC/Hs/Decls.hs - compiler/GHC/Hs/Instances.hs - compiler/GHC/Hs/Stats.hs - compiler/GHC/Hs/Utils.hs - compiler/GHC/HsToCore/Docs.hs - compiler/GHC/HsToCore/Quote.hs - compiler/GHC/Iface/Ext/Ast.hs - compiler/GHC/Parser.y - compiler/GHC/Parser/Annotation.hs - compiler/GHC/Parser/PostProcess.hs - compiler/GHC/Rename/Module.hs - compiler/GHC/Rename/Names.hs - compiler/GHC/Tc/TyCl.hs - compiler/GHC/Tc/TyCl/Class.hs - compiler/GHC/Tc/TyCl/Instance.hs - compiler/GHC/Tc/Utils/Env.hs - compiler/GHC/ThToHs.hs - compiler/Language/Haskell/Syntax/Decls.hs - testsuite/tests/haddock/should_compile_flag_haddock/T17544.stderr - testsuite/tests/parser/should_compile/DumpRenamedAst.stderr - testsuite/tests/printer/Test24533.stdout - utils/check-exact/ExactPrint.hs - utils/check-exact/Utils.hs - utils/haddock/haddock-api/src/Haddock/Backends/Hoogle.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/Convert.hs - utils/haddock/haddock-api/src/Haddock/GhcUtils.hs - utils/haddock/haddock-api/src/Haddock/Interface/Create.hs - utils/haddock/haddock-api/src/Haddock/Interface/Rename.hs - utils/haddock/haddock-api/src/Haddock/Types.hs The diff was not included because it is too large. View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/bee1913d0e0786b585ddc1135461ba5… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/bee1913d0e0786b585ddc1135461ba5… You're receiving this email because of your account on gitlab.haskell.org. Manage all notifications: https://gitlab.haskell.org/-/profile/notifications | Help: https://gitlab.haskell.org/help

1 0