-
ea4c2cbd
by Brandon Chinn at 2026-02-27T16:22:38-08:00
Implement QualifiedStrings (#26503)
See Note [Implementation of QualifiedStrings]
-
08bc245b
by sheaf at 2026-03-01T11:11:54-05:00
Clean up join points, casts & ticks
This commit shores up the logic dealing with casts and ticks occurring
in between a join point binding and a jump.
Fixes #26642 #26929 #26693
Makes progress on #14610 #26157 #26422
Changes:
- Remove 'GHC.Types.Tickish.TickishScoping' in favour of simpler
predicates 'tickishHasNoScope'/'tickishHasSoftScope', as things were
before commit 993975d3. This makes the code easier to read and
document (fewer indirections).
- Introduce 'canCollectArgsThroughTick' for consistent handling of
ticks around PrimOps and other 'Id's that cannot be eta-reduced.
See overhauled Note [Ticks and mandatory eta expansion].
- New Note [JoinId vs TailCallInfo] in GHC.Core.SimpleOpt that explains
robustness of JoinId vs fragility of TailCallInfo.
- Allow casts/non-soft-scoped ticks to occur in between a join point
binder and a jump, but only in Core Prep.
See Note [Join points, casts, and ticks] and
Note [Join points, casts, and ticks... in Core Prep]
in GHC.Core.Opt.Simplify.Iteration.
Also update Core Lint to account for this.
See Note [Linting join points with casts or ticks] in GHC.Core.Lint.
- Update 'GHC.Core.Utils.mergeCaseAlts' to avoid pushing a cast in
between a join point binding and its jumps. This fixes #26642.
See the new (MC5) and (MC6) in Note [Merge Nested Cases].
- Update float out to properly handle source note ticks. They are now
properly floated out instead of being discarded.
This increases the number of ticks in certain tests with -g.
Test cases: T26642 and TrickyJoins.
Metric increase due to more source note ticks with -g:
-------------------------
Metric Increase:
libdir
size_hello_artifact
size_hello_unicode
-------------------------
-
476c4cdf
by Sean D. Gillespie at 2026-03-02T10:14:37-05:00
Add SIMD absolute value on x86 and LLVM
On x86, absolute value of 32 bits or less is implemented with
PABSB/PABSW/PABSD if SSSE3 is available. Otherwise, there is a fallback
for SSE2. For 64 bit integers it uses VPABSQ, required by AVX-512VL,
with fallbacks for SSE4.2 and SSE2.
There is no dedicated instruction for floating point absolute value on
x86, so it is simulated using bitwise AND.
Absolute value for signed integers and floats are implemented by the
"llvm.abs/llvm.fabs" standard library intrinsics. This implementation
uses MachOps constructors, unlike non-vector floating point absolute
value, which uses CallishMachOps.
-
709448c0
by Sean D. Gillespie at 2026-03-02T10:14:46-05:00
Add SIMD floating point square root
On x86, this is implemented with the SQRTPS and SQRTPD instructions. On
LLVM, it uses the sqrt library intrinstic.
-
0deadf66
by Sean D. Gillespie at 2026-03-02T10:14:47-05:00
Improve error message for SIMD on aarch64
When encountering vector literals on aarch64, previously it would
throw:
<no location info>: error:
panic! (the 'impossible' happened)
GHC version 9.15.20251219:
getRegister' (CmmLit:CmmVec):
Now it is more consistent with the other vector operations:
<no location info>: error:
sorry! (unimplemented feature or known bug)
GHC version 9.15.20251219:
SIMD operations on AArch64 currently require the LLVM backend
-
7d64031b
by Vladislav Zavialov at 2026-03-03T11:09:28-05:00
Replace maybeAddSpace with spaceIfSingleQuote
Simplify pretty-printing of HsTypes by using spaceIfSingleQuote.
This allows us to drop the unwieldy lhsTypeHasLeadingPromotionQuote
helper function.
Follow-up to 178c1fd830c78377ef5d338406a41e1d8eb5f0da
-
598db847
by Wolfgang Jeltsch at 2026-03-06T06:25:25-05:00
Correct `hIsReadable` and `hIsWritable` for duplex handles
This contribution implements CLC proposal #371. It changes `hIsReadable`
and `hIsWritable` such that they always throw a respective exception
when encountering a closed or semi-closed handle, not just in the case
of a file handle.
-
b90201e5
by Wolfgang Jeltsch at 2026-03-06T06:25:25-05:00
Document `SemiClosedHandle`
-
c9df72b5
by Wolfgang Jeltsch at 2026-03-06T06:25:25-05:00
Tell users what “semi-closed” means for duplex handles
-
a8aa1868
by Ilias Tsitsimpis at 2026-03-06T06:26:29-05:00
Fix determinism of linker arguments
The switch from Data.Map to UniqMap in 3b5be05ac29 introduced
non-determinism in the order of packages passed to the linker.
This resulted in non-reproducible builds where the DT_NEEDED entries in
dynamic libraries were ordered differently across builds.
Fix the regression by explicitly sorting the package list derived from
UniqMap.
Fixes #26838
-
9b64ad3a
by Matthew Pickering at 2026-03-06T06:27:16-05:00
determinism: Use a deterministic renaming when writing bytecode files
Now when writing the bytecode file, a counter and substitution are used
to provide deterministic keys to local variables (rather than relying on
uniques). This change ensures that `.gbc` are produced
deterministically.
Fixes #26499
-
d29800e0
by Teo Camarasu at 2026-03-06T06:28:46-05:00
ghc-internal: delete Version hs-boot loop
Version has a Read instance which needs Unicode but part of the Unicode interface is the unicode version. This is easy to resolve. We simply don't re-export the version from the Unicode module.
Resolves #26940
-
ad25af90
by Sylvain Henry at 2026-03-06T06:30:33-05:00
Linker: implement support for COMMON symbols (#6107)
Add some support for COMMON symbols. We don't support common symbols
having different sizes where the larger one is allocated after the
smaller one. The linker will fail with an appropriate error message if
it happens.
-
3b59f158
by Cheng Shao at 2026-03-06T06:31:16-05:00
compiler: fix redundant import of GHC.Hs.Lit
This patch removes a redundant import of `GHC.Hs.Lit` which causes a
ghc build failure with validate flavours when bootstrapping from 9.14.
Fixes #26972.
-
148d36f3
by Cheng Shao at 2026-03-06T06:32:01-05:00
compiler: avoid unneeded traversals in GHC.Unit.State
Following !15591, this patch avoids unneeded traversals in
`reportCycles`/`reportUnusable` when log verbosity is below given
threshold. Also applies `logVerbAtLeast` when appropriate.
Co-authored-by: Codex <codex@openai.com>
-
7e31367c
by Cheng Shao at 2026-03-06T06:32:46-05:00
ghc-internal: fix redundant import in GHC.Internal.Event.Windows.ManagedThreadPool
This patch fixes redundant import in
`GHC.Internal.Event.Windows.ManagedThreadPool` that causes a
compilation error when building windows target with validate flavours
and bootstrapping from 9.14. Fixes #26976.
-
fc8b8e27
by sheaf at 2026-03-06T06:33:28-05:00
System.Info.fullCompilerVersion: add 'since' annot
Fixes #26973
-
c8238375
by Sylvain Henry at 2026-03-06T06:34:23-05:00
Hadrian: deprecate --bignum and automatically enable +native_bignum for JS
Deprecate --bignum=... to select the bignum backend. It's only used to
select the native backend, and this can be done with the +native_bignum
flavour transformer.
Additionally, we automatically enable +native_bignum for the JS target
because the GMP backend isn't supported.
-
a3ac7074
by Sylvain Henry at 2026-03-06T06:35:17-05:00
JS: fix putEnum/fromEnum (#24593)
Don't go through Word16 when serializing Enums.
-
0b36e96c
by Andreas Klebinger at 2026-03-06T06:35:58-05:00
Docs: Document -fworker-wrapper-cbv default setting.
Fixes #26841
-
eca445e7
by mangoiv at 2026-03-07T05:02:36-05:00
drop deb9/10 from CI, add deb13
debian 9 and 10 are end of life, hence we drop them
from our CI, but we do add debian 13. Jobs that were
previously run on 9 and 10 run on 13, too, jobs that
were run on 10, are run on 11 now. Jobs that were
previously run on debian 12 are run on debian 13 now.
This MR also updates hadrian's bootstrap plans for that
reason.
Metric Decrease:
T9872d
-
12f8b829
by Luite Stegeman at 2026-03-07T05:03:33-05:00
Fix GHC.Internal.Prim haddock
Haddock used to parse Haskell source to generate documentation,
but switched to using interface files instead. This broke documentation
of the GHC.Internal.Prim module, since it's a wired-in interface that
didn't provide a document structure.
This patch adds the missing document structure and updates genprimopcode
to make the section headers and descriptions available.
fixes #26954
-
f87e5e57
by Luite Stegeman at 2026-03-07T05:03:33-05:00
Remove obsolete --make-haskell-source from genprimopcode
Now that haddock uses the wired-in interface for GHC.Internal.Prim,
the generated Haskell source file is no longer needed. Remove the
--make-haskell-source code generator from genprimopcode and replace
the generated GHC/Internal/Prim.hs with a minimal static source file.
-
4a7ddc7b
by Sylvain Henry at 2026-03-07T05:04:59-05:00
JS: fix linking of exposed but non-preload units (#24886)
Units exposed in the unit database but not explicitly passed on the
command-line were not considered by the JS linker. This isn't an issue
for cabal which passes every unit explicitly but it is an issue when
using GHC directly (cf T24886 test).
-
689aafcd
by mangoiv at 2026-03-07T05:05:52-05:00
testsuite: double foundation timeout multiplier
The runtime timeout in the foundation test was regularly hit by code
generated by the wasm backend - we increase the timout since the high
runtime is expected on the wasm backend for this rather complex test.
Resolves #26938
-
a46a1bb1
by Cheng Shao at 2026-03-09T04:50:30-04:00
compiler: add myCapabilityExpr to GHC.Cmm.Utils
This commit adds `myCapabilityExpr` to `GHC.Cmm.Utils` which is
computed from `BaseReg`. It's convenient for codegen logic where one
needs to pass the current Capability's pointer.
-
4afc65b1
by Cheng Shao at 2026-03-09T04:50:30-04:00
compiler: lower tryPutMVar# into a ccall directly
This patch addresses an old TODO of `stg_tryPutMVarzh` by removing it
completely and making the compiler lower `tryPutMVar#` into a ccall to
`performTryPutMVar` directly, without landing into an intermediate C
or Cmm function. `performTryPutMVar` is promoted to a public RTS
function with default visibility, and the compiler lowering logic
takes into account the C ABI of `performTryPutMVar` and converts from
C Bool to primop's `Int#` result properly.
-
9e3d6a58
by Simon Hengel at 2026-03-09T04:51:15-04:00
Don't use #line in haddocks
This confuses the parser. Haddock output is unaffected by this change.
(read: this still produces the same documentation)
-
f4e8fec2
by Wolfgang Jeltsch at 2026-03-09T04:52:01-04:00
Remove in-package dependencies on `GHC.Internal.System.IO`
This contribution eliminates all dependencies on
`GHC.Internal.System.IO` from within `ghc-internal`. It comprises the
following changes:
* Make `GHC.Internal.Fingerprint` independent of I/O support
* Tighten the dependencies of `GHC.Internal.Data.Version`
* Tighten the dependencies of `GHC.Internal.TH.Monad`
* Tighten the dependencies of `GHCi.Helpers`
* Move some code that needs `System.IO` to `template-haskell`
* Move the `GHC.ResponseFile` implementation into `base`
* Move the `System.Exit` implementation into `base`
* Move the `System.IO.OS` implementation into `base`
Metric Decrease:
size_hello_artifact
size_hello_artifact_gzip
size_hello_unicode
size_hello_unicode_gzip
-
91df4c82
by Sylvain Henry at 2026-03-09T04:53:20-04:00
T18832: fix Windows CI failure by dropping removeDirectoryRecursive
On Windows, open file handles prevent deletion. After killThread, the
closer thread may not have called hClose yet, causing removeDirectoryRecursive
to fail with "permission denied". The test harness cleans up the run
directory anyway, so the call is redundant.
-
d7fe9671
by Cheng Shao at 2026-03-09T04:54:04-04:00
compiler: fix redundant import in GHC.StgToJS.Object
This patch fixes a redundant import in GHC.StgToJS.Object that causes
a build failure when compiling head from 9.14 with validate flavours.
Fixes #26991.
-
0bfd29c3
by Cheng Shao at 2026-03-09T04:54:46-04:00
wasm: fix `Illegal foreign declaration` failure when ghci loads modules with JSFFI exports
This patch fixes a wasm ghci error when loading modules with JSFFI
exports; the `backendValidityOfCExport` check in `tcCheckFEType`
should only makes sense and should be performed when not checking the
JavaScript calling convention; otherwise, when the calling convention
is JavaScript, the codegen logic should be trusted to backends that
actually make use of it. Fixes #26998.
-
e659610c
by Duncan Coutts at 2026-03-09T12:08:35-04:00
Apply NOINLINE pragmas to generated Typeable bindings
For context, see the existing Note [Grand plan for Typeable]
and the Note [NOINLINE on generated Typeable bindings] added in the
subsequent commit.
This is about reducing the number of exported top level names and
unfoldings, which reduces interface file sizes and reduces the number of
global/dynamic linker symbols.
Also accept the changed test output and metric decreases.
Tests that record the phase output for type checking or for simplifier
end up with different output: the generated bindings now have an
Inline [~] annotation, and many top level names are now local rather
than module-prefixed for export.
Also accept the numerous metric decreases in compile_time/bytes
allocated, and a few in compile_time/max_bytes_used.
There's also one instance of a decrease in runtime/max_bytes_used but
it's a ghci-way test and so presumably the reason is that it loads
smaller .hi files and/or links fewer symbols.
-------------------------
Metric Decrease:
CoOpt_Singletons
MultiLayerModulesTH_OneShot
MultilineStringsPerf
T10421
T10547
T12150
T12227
T12234
T12425
T13035
T13056
T13253
T13253-spj
T15304
T15703
T16875
T17836b
T17977b
T18140
T18223
T18282
T18304
T18698a
T18698b
T18730
T18923
T20049
T21839c
T24471
T24582
T24984
T3064
T4029
T5030
T5642
T5837
T6048
T9020
T9198
T9961
TcPlugin_RewritePerf
WWRec
hard_hole_fits
mhu-perf
-------------------------
-
67df5161
by Duncan Coutts at 2026-03-09T12:08:35-04:00
Add documentation Note [NOINLINE on generated Typeable bindings]
and refer to it from the code and existing documentation.
-
c4ad6167
by Duncan Coutts at 2026-03-09T12:08:35-04:00
Switch existing note to "named wrinkle" style, (GPT1)..(GPT7)
GPT = Grand plan for Typeable
-
dc84f8e2
by Cheng Shao at 2026-03-09T12:09:21-04:00
ci: only build deb13 for validate pipeline aarch64-linux jobs
This patch drops the redundant aarch64-linux deb12 job from validate pipelines
and only keeps deb13; it's still built in nightly/release pipelines. Closes #27004.
-
23a50772
by Rajkumar Natarajan at 2026-03-10T14:11:37-04:00
chore: Merge GHC.Internal.TH.Quote into GHC.Internal.TH.Monad
Move the QuasiQuoter datatype from GHC.Internal.TH.Quote to
GHC.Internal.TH.Monad and delete the Quote module.
Update submodule template-haskell-quasiquoter to use the merged
upstream version that imports from the correct module.
Co-authored-by: Cursor <cursoragent@cursor.com>
-
a2bb6fc3
by Simon Jakobi at 2026-03-10T14:12:23-04:00
Add regression test for #16122
-
604e1180
by Cheng Shao at 2026-03-11T15:00:42-04:00
hadrian: remove the broken bench flavour
This patch removes the bench flavour from hadrian which has been
broken for years and not used for actual benchmarking (for which
`perf`/`release` is used instead). Closes #26825.
-
c3e64915
by Simon Jakobi at 2026-03-11T15:01:31-04:00
Add regression test for #18186
The original TypeInType language extension is replaced with
DataKinds+PolyKinds for compatibility.
Closes #18186.
-
664996c7
by Andreas Klebinger at 2026-03-11T15:02:16-04:00
Bump nofib submodule.
We accrued a number of nofib fixes we want to have here.
-
517cf64e
by Simon Jakobi at 2026-03-11T15:03:03-04:00
Add regression test for #15907
Closes #15907.
-
fff362cf
by Simon Jakobi at 2026-03-11T15:03:49-04:00
Ensure T14272 is run in optasm way
Closes #16539.
-
ec81ec2c
by Simon Jakobi at 2026-03-11T15:03:49-04:00
Add regression test for #24632
Closes #24632.
-
cefec47b
by Simon Jakobi at 2026-03-11T15:03:50-04:00
Fix module name of T9675: T6975 -> T9675
-
d3690ae8
by Andreas Klebinger at 2026-03-11T15:04:31-04:00
User guide: Clarify phase control on INLINEABLE[foo] pragmas.
Fixes #26851
-
e7054934
by Simon Jakobi at 2026-03-11T15:05:16-04:00
Add regression test for #12694
Closes #12694.
-
4756d9f6
by Simon Jakobi at 2026-03-11T15:05:16-04:00
Add regression test for #16275
Closes #16275.
-
34b7e2c1
by Simon Jakobi at 2026-03-11T15:05:16-04:00
Add regression test for #14908
Closes #14908.
-
4243db3d
by Simon Jakobi at 2026-03-11T15:05:16-04:00
Add regression test for #14151
Closes #14151.
-
0e9f1453
by Simon Jakobi at 2026-03-11T15:05:16-04:00
Add regression test for #12640
Closes #12640.
-
ae606c7f
by Simon Jakobi at 2026-03-11T15:05:16-04:00
Add regression test for #15588
Closes #15588.
-
5a38ce4e
by Simon Jakobi at 2026-03-11T15:05:16-04:00
Add regression test for #9445
Closes #9445.
-
d054b467
by Cheng Shao at 2026-03-11T15:05:59-04:00
compiler: implement string interning logic for BCONPtrFS
This patch adds a `FastStringEnv`-based cache of `MallocStrings`
requests to `Interp`, so that when we load bytecode with many
breakpoints that share the same module names & unit ids, we reuse the
allocated remote pointers instead of issuing duplicte `MallocStrings`
requests and bloating the C heap. Closes #26995.
-
b85a0293
by Simon Jakobi at 2026-03-11T15:06:41-04:00
Add perf test for #1216
Closes #1216.
-
cd7f7420
by Sylvain Henry at 2026-03-11T15:07:58-04:00
JS: check that tuple constructors are linked (#23709)
Test js-mk_tup was failing before because tuple constructors weren't
linked in. It's no longer an issue after the linker fixes.
-
d57f01a4
by Matthew Pickering at 2026-03-11T15:08:40-04:00
testsuite: Add test for foreign import prim with unboxed tuple return
This commit just adds a test that foreign import prim works with unboxed
sums.
-
23d111ce
by Matthew Pickering at 2026-03-11T15:08:41-04:00
Return a valid pointer in advanceStackFrameLocationzh
When there is no next stack chunk, `advanceStackFrameLocationzh` used to
return NULL in the pointer-typed StackSnapshot# result slot.
Even though the caller treats that case as "no next frame", the result is
still materialized in a GC-visible pointer slot. If a GC observes the raw
NULL there, stack decoding can crash.
Fix this by ensuring the dead pointer slot contains a valid closure
pointer. Also make the optional result explicit by returning an unboxed
sum instead of a tuple with a separate tag.
Fixes #27009
-
4c58a3ae
by Cheng Shao at 2026-03-11T15:09:22-04:00
hadrian: build profiled dynamic objects with -dynamic-too
This patch enables hadrian to build profiled dynamic objects with
`-dynamic-too`, addressing a build parallelism bottleneck in release
pipelines. Closes #27010.
-
870243e4
by Zubin Duggal at 2026-03-12T17:33:28+05:30
DmdAnal: Take stable unfoldings into account when determining argument demands
Previously, demand analysis only looked at the RHS to compute argument demands.
If the optimised RHS discarded uses of an argument that the stable unfolding
still needed, it would be incorrectly marked absent. Worker/wrapper would then
replace it with LitRubbish, and inlining the stable unfolding would use the
rubbish value, causing a segfault.
To fix, we introduce addUnfoldingDemands which analyses the stable unfolding
with dmdAnal and combines its DmdType with the RHS's via the new `maxDmdType`
which combines the demands of the stable unfolding with the rhs, so we can avoid
any situation where we give an absent demand to something which is still used
by the stable unfolding.
Fixes #26416.
-
669d09f9
by Cheng Shao at 2026-03-13T15:06:07-04:00
hadrian: remove redundant library/rts ways definitions from stock flavours
This patch removes redundant library/rts ways definitions from stock
flavours in hadrian; they can be replaced by applying appropriate
filters on `defaultFlavour`.
-
a27dc081
by Teo Camarasu at 2026-03-13T15:06:51-04:00
ghc-internal: move bits Weak of finalizer interface to base
We move parts of the Weak finalizer interface to `base` only the parts
that the RTS needs to know about are kept in `ghc-internal`.
This lets us then prune our imports somewhat and get rid of some SOURCE imports.
Resolves #26985
-
6eef855b
by Sylvain Henry at 2026-03-13T15:08:18-04:00
Stg/Unarise: constant-folding during unarisation (#25650)
When building an unboxed sum from a literal argument, mkUbxSum
previously emitted a runtime cast via `case primop [lit] of var -> ...`.
This wrapper prevented GHC from recognising the result as a static
StgRhsCon, causing top-level closures to be allocated as thunks instead
of being statically allocated.
Fix: try to perform the numeric literal cast at compile time using
mkLitNumberWrap (wrapping semantics). If successful, return the cast
literal directly with an identity wrapper (no case expression). The
runtime cast path is kept as fallback for non-literal arguments.
Test: codeGen/should_compile/T25650
-
905f8723
by Simon Jakobi at 2026-03-13T15:09:09-04:00
Add regression test for #2057
Test that GHC stops after an interface-file error instead of
continuing into the linker.
The test constructs a stale package dependency on purpose. `pkgB` is compiled
against one version of package `A`, then the same unit id is replaced by an
incompatible build of `A`. When `Main` imports `B`, GHC has to read `B.hi`,
finds an unfolding that still mentions the old `A`, and should fail while
loading interfaces.
Closes #2057.
Assisted-by: Codex
-
a13245a9
by Sylvain Henry at 2026-03-13T15:10:06-04:00
JS: fix recompilation avoidance (#23013)
- we were checking the mtime of the *.jsexe directory, not of a file
- we were not computing the PkgsLoaded at all
-
07442653
by Cheng Shao at 2026-03-13T15:10:51-04:00
hadrian: bump index state & bootstrap plans
This patch bumps hadrian index state & bootstrap plans:
- The updated index state allows bootstrapping from 9.14 without cabal
allow-newer hacks
- The updated bootstrap plans all contain shake-0.19.9 containing
important bugfix, allowing a subsequent patch to bump shake bound to
ensure the bugfix is included
- ghc 9.14.1 bootstrap plan is added
-
fdc1dbad
by Cheng Shao at 2026-03-13T15:10:51-04:00
ci: add ghc 9.14.1 to bootstrap matrix
This patch adds ghc 9.14.1 to bootstrap matrix, so that we test
bootstrapping from ghc 9.14.1.
-
91916079
by Sylvain Henry at 2026-03-13T15:11:43-04:00
T17912: wait for opener thread to block before killing it (#24739)
Instead of a fixed 1000ms delay, poll threadStatus until the opener
thread is in BlockedOnForeignCall, ensuring killThread only fires once
the thread is provably inside the blocking open() syscall. This prevents
the test from accidentally passing on Windows due to scheduling races.
-
baa4ebb4
by Cheng Shao at 2026-03-13T15:12:26-04:00
template-haskell: fix redundant import in Language.Haskell.TH.Quote
This patch fixes a redundant import in `Language.Haskell.TH.Quote`
that causes a ghc build failure when bootstrapping from 9.14 with
validate flavours. Fixes #27014.
-
02e68a86
by Brandon Simmons at 2026-03-13T15:13:19-04:00
Add a cumulative gc_sync_elapsed_ns counter to GHC.Internal.Stats
This makes it possible to get an accurate view of time spent in sync
phase when using prometheus-style sampling. Previously this was only
available for the most recent GC.
This intentionally leaves GHC.Stats API unchanged since it is marked as
deprecated, and API changes there require CLC approval.
Fixes #26944
-
a18fa3c1
by Cheng Shao at 2026-03-14T05:12:14-04:00
configure: make $LLVMAS default to $CC when $CcLlvmBackend is YES
This patch changes the $LLVMAS detection logic in configure so that
when it's not manually specified by the user, it defaults to $CC if
$CcLlvmBackend is YES. It's a more sensible default than auto-detected
clang from the environment, especially when cross-compiling, $CC as
the cross target's LLVM assembler is more compatible with the use case
than the system-wide clang. Fixes #26769.
-
3774086e
by Matthew Pickering at 2026-03-14T05:13:00-04:00
exceptions: annotate onException continuation with WhileHandling
Before this patch, an exception thrown in the `onException` handler
would loose track of where the original exception was thrown.
```
import Control.Exception
main :: IO ()
main = failingAction `onException` failingCleanup
where
failingAction = throwIO (ErrorCall "outer failure")
failingCleanup = throwIO (ErrorCall "cleanup failure")
```
would report
```
T28399: Uncaught exception ghc-internal:GHC.Internal.Exception.ErrorCall:
cleanup failure
HasCallStack backtrace:
throwIO, called at T28399.hs:<line>:<column> in <package-id>:Main
```
notice that the "outer failure" exception is not present in the error
message.
With this patch, any exception thrown is in the handler is annotated
with WhileHandling. The resulting message looks like
```
T28399: Uncaught exception ghc-internal:GHC.Internal.Exception.ErrorCall:
cleanup failure
While handling outer failure
HasCallStack backtrace:
throwIO, called at T28399.hs:7:22 in main:Main
```
CLC Proposal: https://github.com/haskell/core-libraries-committee/issues/397
Fixes #26759
-
63ae8eb3
by Andreas Klebinger at 2026-03-14T05:13:43-04:00
Fix missing profiling header for origin_thunk frame.
Fixes #27007
-
213d2c0e
by Cheng Shao at 2026-03-14T05:14:28-04:00
ci: fix ci-images revision
The current ci-images revision was a commit on the WIP branch of
https://gitlab.haskell.org/ghc/ci-images/-/merge_requests/183, and
it's not on the current ci-images master branch. This patch fixes the
image revision to use the current tip of ci-images master.
-
fc2b083f
by Andreas Klebinger at 2026-03-14T05:15:14-04:00
Revert "hadrian/build-cabal: Better respect and utilize -j"
This reverts commit eab3dbba79650e6046efca79133b4c0a5257613d.
While it's neat this currently isn't well supported on all platforms.
It's time will come, but for now I'm reverting this to avoid issues for
users on slightly unconvential platforms.
This will be tracked at #26977.
-
12a706cf
by Cheng Shao at 2026-03-14T16:37:54-04:00
base: fix redundant imports in GHC.Internal.Weak.Finalize
This patch fixes redundant imports in GHC.Internal.Weak.Finalize that
causes a regression in bootstrapping head from 9.14 with validate
flavours. Fixes #27026.
-
b5d39cad
by Matthew Pickering at 2026-03-14T16:38:37-04:00
Use explicit syntax rather than pure
-
43638643
by Andreas Klebinger at 2026-03-15T18:15:48-04:00
Configure: Fix check for --target support in stage0 CC
The check FP_PROG_CC_LINKER_TARGET used $CC unconditionally to check for
--target support. However this fails for the stage0 config where the C
compiler used is not $CC but $CC_STAGE0.
Since we already pass the compiler under test into the macro I simply
changed it to use that instead.
Fixes #26999
-
18fd0df6
by Simon Hengel at 2026-03-15T18:16:33-04:00
Fix typo in recursive_do.rst
-
86bd9bfc
by fendor at 2026-03-17T23:46:09-04:00
Introduce `-fimport-loaded-targets` GHCi flag
This new flag automatically adds all loaded targets to the GHCi session
by adding an `InteractiveImport` for the loaded targets.
By default, this flag is disabled, as it potentially increases memory-usage.
This interacts with the flag `-fno-load-initial-targets` as follows:
* If no module is loaded, no module is added as an interactive import.
* If a reload loads up to a module, all loaded modules are added as
interactive imports.
* Unloading modules removes them from the interactive context.
Fixes #26866 by rendering the use of a `-ghci-script` to achieve the
same thing redundant.
-
e3d4c1bb
by mniip at 2026-03-17T23:47:03-04:00
ghc-internal: Remove GHC.Internal.Data.Eq
It served no purpose other than being a re-export.
-
6f4f6cf0
by mniip at 2026-03-17T23:47:03-04:00
ghc-internal: Refine GHC.Internal.Base imports
Removed re-exports from GHC.Internal.Base. This reveals some modules
that don't actually use anything *defined* in GHC.Internal.Base, and
that can be pushed down a little in the import graph.
Replaced most imports of GHC.Internal.Base with non-wildcard imports
from modules where the identifiers are actually defined.
Part of #26834
Metric Decrease:
T5321FD
-
7fb51f54
by mangoiv at 2026-03-17T23:48:00-04:00
ci: clone, don't copy when creating the cabal cache
Also removed WINDOWS_HOST variable detected via uname - we now just
check whether the CI job has windows in its name. This works because we
only ever care about it if the respective job is not a cross job. We
also statically detect darwin cross jobs in the same way. We only ever have
darwin -> darwin cross jobs so this is enough to detect the host
reliably.
-
f8817879
by mangoiv at 2026-03-17T23:48:44-04:00
ci: mark size_hello_artifact fragile on darwin x86
The size of the x86_64 hello artifact is not stable which results in flaky testruns.
Resolves #26814
-
e34cb6da
by Adam Gundry at 2026-03-20T12:20:00-04:00
ghci: Mention active language edition in startup banner
Per GHC proposal 632, this makes the GHCi startup banner include
the active language edition, plus an indication of whether this
was the default (as opposed to being explicitly selected via an
option such as `-XGHC2024`). For example:
```
$ ghci
GHCi, version 9.14.1: https://www.haskell.org/ghc/ :? for help
Using default language edition: GHC2024
ghci>
```
Fixes #26037.
-
52c3e6ba
by sheaf at 2026-03-20T12:21:09-04:00
Improve incomplete record selector warnings
This commit stops GHC from emitting spurious incomplete record selector
warnings for bare selectors/projections such as .fld
There are two places we currently emit incomplete record selector
warnings:
1. In the desugarer, when we see a record selector or an occurrence
of 'getField'. Here, we can use pattern matching information to
ensure we don't give false positives.
2. In the typechecker, which might sometimes give false positives but
can emit warnings in cases that the pattern match checker would
otherwise miss.
This is explained in Note [Detecting incomplete record selectors]
in GHC.HsToCore.Pmc.
Now, we obviously don't want to emit the same error twice, and generally
we prefer (1), as those messages contain fewer false positives. So we
suppress (2) when we are sure we are going to emit (1); the logic for
doing so is in GHC.Tc.Instance.Class.warnIncompleteRecSel,
and works by looking at the CtOrigin.
Now, the issue was that this logic handled explicit record selectors as
well as overloaded record field selectors such as "x.r" (which turns
into a simple GetFieldOrigin CtOrigin), but it didn't properly handle
record projectors like ".fld" or ".fld1.fld2" (which result in other
CtOrigins such as 'RecordFieldProjectionOrigin').
To solve this problem, we re-use the 'isHasFieldOrigin' introduced in
fbdc623a (slightly adjusted).
On the way, we also had to update the desugarer with special handling
for the 'ExpandedThingTc' case in 'ds_app', to make sure that
'ds_app_var' sees all the type arguments to 'getField' in order for it
to indeed emit warnings like in (1).
Fixes #26686
-
309d7e87
by Cheng Shao at 2026-03-20T12:21:53-04:00
rts: opportunistically grow the MutableByteArray# in-place in resizeMutableByteArray#
Following !15234, this patch improves `resizeMutableByteArray#` memory
efficiency by growing the `MutableByteArray#` in-place if possible,
addressing an old todo comment here. Also adds a new test case
`resizeMutableByteArrayInPlace` that stresses this behavior.
-
7d4ef162
by Matthew Craven at 2026-03-20T12:22:47-04:00
Change representation of floating point literals
This commit changes the representation of floating point literals
throughough the compiler, in particular in Core and Cmm.
The Rational type is deficient for this purpose, dealing poorly
with NaN, +/-Infinity, and negative zero. Instead, the new module
GHC.Types.Literal.Floating uses the host Float/Double type to represent
NaNs, infinities and negative zero. It also contains a Rational
constructor, for the benefit of -fexcess-precision.
Other changes:
- Remove Note [negative zero] and related code
This also removes the restrictions on constant-folding of division
by zero, and should make any problems with NaN/Infinity more obvious.
- Use -0.0 as the additive identity for Core constant folding rules
for floating-point addition, fixing #21227.
- Manual worker-wrapper for GHC.Float.rationalToDouble. This is
intended to prevent the compiler's WW on this function from
interfering with constant-folding. This change means that we now
avoid allocating a box for the result of a 'realToFrac' call in
T10359.
- Combine floatDecodeOp and doubleDecodeOp.
This change also fixes a bug in doubleDecodeOp wherein it
would incorrectly produce an Int# instead of an Int64#
literal for the mantissa component with 64-bit targets.
- Use Float/Double for assembly immediates, and update the X86 and
PowerPC backends to properly handle special values such as NaN and
infinity.
- Allow 'rational_to' to handle zero denominators, fixing a
TODO in GHC.Core.Opt.ConstantFold.
Fixes #8364 #9811 #18897 #21227
Progress towards #26919
Metric Decrease:
T10359
Co-authored-by: sheaf <sam.derbyshire@gmail.com>
-------------------------
Metric Decrease:
T1969
T5321FD
-------------------------
-
80e2dd4f
by Zubin Duggal at 2026-03-20T12:23:33-04:00
compiler/ffi: Collapse void pointer chains in capi wrappers
New gcc/clang treat -Wincompatible-pointer-types as an error by
default. Since C only allows implicit conversion from void*, not void**,
capi wrappers for functions taking e.g. abstract** would fail to compile
when the Haskell type Ptr (Ptr Abstract) was naively translated to void**.
Collapse nested void pointers to a single void* when the pointee type
has no known C representation.
Fixes #26852
-
1c50bd7b
by Luite Stegeman at 2026-03-20T12:24:37-04:00
Move some functions related to pointer tagging to a separate module
-
bfd7aafd
by Luite Stegeman at 2026-03-20T12:24:37-04:00
Branchless unpacking for enumeration types
Change unpacking for enumeration types to go to Word8#/Word16#/Word#
directly instead of going through an intermediate unboxed sum. This
allows us to do a branchless conversion using DataToTag and TagToEnum.
Fixes #26970
-
72b20fc0
by Luite Stegeman at 2026-03-20T12:25:30-04:00
bytecode: Carefully SLIDE off the end of a stack chunk
The SLIDE bytecode instruction was not checking for stack chunk
boundaries and could corrupt the stack underflow frame, leading
to crashes.
We add a check to use safe writes if we cross the chunk boundary
and also handle stack underflow if Sp is advanced past the underflow
frame.
fix #27001
-
2e22b43c
by Cheng Shao at 2026-03-20T12:26:14-04:00
ghci: serialize BCOByteArray buffer directly when possible
This patch changes the `Binary` instances of `BCOByteArray` to
directly serialize the underlying buffer when possible, while also
taking into account the issue of host-dependent `Word` width. See
added comments and amended `Note [BCOByteArray serialization]` for
detailed explanation. Closes #27020.
-
89d9ba37
by Sylvain Henry at 2026-03-20T12:27:34-04:00
JS: replace BigInt with Number arithmetic for 32/64-bit quot/rem (#23597)
Replace BigInt-based implementations of quotWord32, remWord32,
quotRemWord32, quotRem2Word32, quotWord64, remWord64, quotInt64, and
remInt64 with pure Number (double/integer) arithmetic to avoid the
overhead of BigInt promotion.
-
ae4ddd60
by Sylvain Henry at 2026-03-20T12:28:28-04:00
Core: add constant-folding rules for Addr# eq/ne (#18032)
-
3e767f98
by Matthew Pickering at 2026-03-20T12:29:11-04:00
Use OsPath rather than FilePath in Downsweep cache
This gets us one step closure to uniformly using `OsPath` in the
compiler.
-
2c57de29
by Cheng Shao at 2026-03-20T12:29:55-04:00
hadrian: fix ghc-in-ghci flavour stage0 shared libraries
This patch fixes missing stage0 shared libraries in hadrian
ghc-in-ghci flavour, which was accidentally dropped in
669d09f950a6e88b903d9fd8a7571531774d4d5d and resulted in a regression
in HLS support on linux/macos. Fixes #27057.
-
5b1be555
by Sylvain Henry at 2026-03-20T12:30:48-04:00
JS: install rts/Types.h header file (#27033)
It was an omission, making HsFFI.h not usable with GHC using the JS
backend.
-
b883f08f
by Cheng Shao at 2026-03-20T12:31:33-04:00
hadrian: don't compile RTS with -Winline
This patch removes `-Winline` from cflags when compiling the RTS,
given that:
1. It generates a huge pile of spam and hurts developer experience
2. Whether inlining happens is highly dependent on toolchains,
flavours, etc, and it's not really an issue to fix if inlining
doesn't happen; it's a hint to the C compiler anyway.
Fixes #27060.
-
333387d6
by Cheng Shao at 2026-03-20T12:31:33-04:00
hadrian: compile libffi-clib with -Wno-deprecated-declarations
This patch adds `-Wno-deprecated-declarations` to cflags of
`libffi-clib`, given that it produces noise at compile-time that
aren't really our issue to fix anyway, it's from vendored libffi
source code.
-
67c47771
by Rodrigo Mesquita at 2026-03-20T12:32:17-04:00
Expose decodeStackWithIpe from ghc-experimental
This decoding is useful to the debugger and it wasn't originally
exported as an oversight.
-
18513365
by Matthew Pickering at 2026-03-21T04:43:26-04:00
Add support for custom external interpreter commands
It can be useful for GHC API clients to implement their own external
interpreter commands.
For example, the debugger may want an efficient way to inspect the
stacks of the running threads in the external interpreter.
-
4636d906
by mangoiv at 2026-03-21T04:44:10-04:00
ci: remove obsolete fallback for old debian and ubuntu versions
-
2e3a2805
by mangoiv at 2026-03-21T04:44:10-04:00
ci: drop ubuntu 18 and 20
Ubuntu 18 EOL: May 2023
Ubuntu 20 EOL: May 2025
We should probably not make another major release supporting these platforms.
Also updates the generator script.
Resolves #25876
-
de54e264
by Cheng Shao at 2026-03-21T17:52:08+01:00
rts: fix -Wcompare-distinct-pointer-types errors
This commit fixes `-Wcompare-distinct-pointer-types` errors in the RTS
which should have been caught by the `validate` flavour but was
warnings in CI due to the recent `+werror` regression.
-
b9bd73de
by Cheng Shao at 2026-03-21T17:52:08+01:00
ghc-internal: fix unused imports
This commit fixes unused imports in `ghc-internal` which should have
been caught by the `validate` flavour but was warnings in CI due to
the recent `+werror` regression. Fixes #26987 #27059.
-
da946a16
by Cheng Shao at 2026-03-21T17:03:51+00:00
ghci: fix unused imports
This commit fixes unused imports in `ghci` which should have been
caught by the `validate` flavour but was warnings in CI due to the
recent `+werror` regression. Fixes #26987 #27059.
-
955b1cf8
by Cheng Shao at 2026-03-21T17:03:51+00:00
compiler: fix unused imports in GHC.Tc.Types.Origin
This commit fixes unused imports in `GHC.Tc.Types.Origin` which should
have been caught by the `validate` flavour but was warnings in CI due
to the recent `+werror` regression. Fixes #27059.
-
3b1aeb50
by Cheng Shao at 2026-03-21T17:03:51+00:00
hadrian: fix missing +werror in validate flavour
This patch fixes missing `+werror` in validate flavour, which was an
oversight in bb3a2ba1eefadf0b2ef4f39b31337a23eec67f29. Fixes #27066.
-
44f118f0
by Cheng Shao at 2026-03-22T04:54:01-04:00
ci: bump CACHE_REV and add the missing reminder
This patch bumps `CACHE_REV` to address recent `[Cabal-7159]` CI
errors due to stale cabal cache on some runners, and also adds a
reminder to remind future maintainers. Fixes #27075.
-
2a218737
by ARATA Mizuki at 2026-03-23T11:11:39-04:00
Add 128-bit SIMD support to AArch64 NCG
Changes:
- Add `Format` field to vector-capable instructions.
These instructions will emit `vN.4s` (for example) as a operand.
- Additional constructors for `Operand`:
`OpVecLane` represents a vector lane and will be emitted as `vN.<width>[<index>]` (`vN.s[3]` for example).
`OpScalarAsVec` represents a scalar, but printed as a vector lane like `vN.<width>[0]` (`vN.s[0]` for example).
- Integer quot/rem are implemented in C, like x86.
Closes #26536
Metric Increase:
T3294
-
5d6e2be9
by ARATA Mizuki at 2026-03-23T11:11:39-04:00
AArch64 NCG: Improve code generation for floating-point and vector constants
Some floating-point constants can be directly encoded using the FMOV instruction.
Similarly, a class of vectors with same values can be encoded using FMOV, MOVI, or MVNI.
-
c6d262aa
by Simon Jakobi at 2026-03-23T11:12:22-04:00
Add regression test for #13729
Closes #13729.
-
aa5dfe67
by Sylvain Henry at 2026-03-26T03:48:56-04:00
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.
-
c8a7b588
by Sylvain Henry at 2026-03-26T03:48:56-04:00
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).
-
a5ec467e
by ARATA Mizuki at 2026-03-26T03:49:49-04:00
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@anthropic.com>
-
661da815
by Teo Camarasu at 2026-03-26T03:50:33-04:00
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
-
45428f88
by sheaf at 2026-03-26T03:51:31-04:00
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@gmail.com>
-
2823b039
by Ian Duncan at 2026-03-26T03:52:21-04:00
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.
-
57b7878d
by Simon Jakobi at 2026-03-26T03:53:07-04:00
Add regression test for #12002
Closes #12002.
-
c8f9df2d
by Simon Jakobi at 2026-03-26T03:53:07-04:00
Add regression test for #12046
Closes #12046.
Co-authored-by: Andreas Klebinger <klebinger.andreas@gmx.at>
-
615d72ac
by Simon Jakobi at 2026-03-26T03:53:07-04:00
Add regression test for #13180
Closes #13180.
-
423eebcf
by Simon Jakobi at 2026-03-26T03:53:07-04:00
Add regression test for #11141
Closes #11141.
-
286849a4
by Simon Jakobi at 2026-03-26T03:53:07-04:00
Add regression test for #11505
Closes #11505.
-
7db149d9
by Simon Jakobi at 2026-03-26T03:53:07-04:00
Add regression perf test for #13820
Closes #13820.
-
e73c4adb
by Simon Jakobi at 2026-03-26T03:53:07-04:00
Add regression test for #10381
Closes #10381.
-
5ebcfb57
by Benjamin Maurer at 2026-03-26T03:54:02-04:00
Generate assembly on x86 for word2float (#22252)
We used to emit C function call for MO_UF_Conv primitive.
Now emits direct assembly instead.
Co-Authored-By: Sylvain Henry <sylvain@haskus.fr>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
-
5b550754
by Matthew Pickering at 2026-03-26T03:54:51-04:00
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
-
ef0a1bd2
by mangoiv at 2026-03-26T03:55:34-04:00
release tracking: adopt release tracking ticket from #16816
-
a7f40fd9
by mangoiv at 2026-03-26T03:55:34-04:00
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
-
161d3285
by Teo Camarasu at 2026-03-26T03:56:18-04:00
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.
-
36eed985
by Cheng Shao at 2026-03-26T03:57:03-04:00
ghc-boot: move GHC.Data.SmallArray to ghc-boot
This commit moves `GHC.Data.SmallArray` from the `ghc` library to
`ghc-boot`, so that it can be used by `ghci` as well:
- The `Binary` (from `ghc`) instance of `SmallArray` is moved to
`GHC.Utils.Binary`
- Util functions `replicateSmallArrayIO`, `mapSmallArrayIO`,
`mapSmallArrayM_`, `imapSmallArrayM_` , `smallArrayFromList` and
`smallArrayToList` are added
- The `Show` instance is added
- The `Binary` (from `binary`) instance is added
-
fdf828ae
by Cheng Shao at 2026-03-26T03:57:03-04:00
compiler: use `Binary` instance of `BCOByteArray` for bytecode objects
This commit defines `Binary` (from `compiler`) instance of
`BCOByteArray` which serializes the underlying buffer directly, and
uses it directly in bytecode object serialization. Previously we reuse
the `Binary` (from `binary`) instance, and this change allows us to
avoid double-copying via an intermediate `ByteString` when using
`put`/`get` in `binnary`. Also see added comment for explanation.
-
3bf62d0a
by Cheng Shao at 2026-03-26T03:57:03-04:00
ghci: use SmallArray directly in ResolvedBCO
This patch makes ghci use `SmallArray` directly in `ResolvedBCO` when
applicable, making the memory representation more compact and reducing
marshaling overhead. Closes #27058.
-
3d6492ce
by Wen Kokke at 2026-03-26T03:57:53-04:00
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.
-
7b9a75f0
by Phil Hazelden at 2026-03-26T03:58:37-04:00
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.
-
fe6e76c5
by Tobias Haslop at 2026-03-26T03:59:30-04:00
Fix broken Haddock link to Bifunctor class in description of Functor class
-
404b71c1
by Luite Stegeman at 2026-03-27T04:40:49-04:00
Fix assert in Interpreter.c
If we skip exactly the number of words on the stack we end up on
the first word in the next chunk.
-
a85bd503
by Luite Stegeman at 2026-03-27T04:40:49-04:00
Support arbitrary size unboxed tuples in bytecode
This stores the size (number of words on the stack) of the next
expected tuple in the TSO, ctoi_spill_size field, eliminating
the need of stg_ctoi_tN frames for each size.
Note: On 32 bit platform there is still a bytecode tuple size
limit of 255 words on the stack.
Fixes #26946
-
e2209031
by Luite Stegeman at 2026-03-27T04:40:49-04:00
Add specialized frames for small tuples
Small tuples are now returned more efficiently to the interpreter.
They use one less word of stack space and don't need manipulation
of the TSO anymore.
-
b26bb2ea
by VeryMilkyJoe at 2026-03-27T04:41:38-04:00
Remove backwards compatibility pattern synonym `ModLocation`
Fixes #24932
-
66e5e324
by Vladislav Zavialov at 2026-03-27T04:42:25-04:00
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
-
f8de456f
by Sylvain Henry at 2026-03-27T04:43:22-04:00
STM: don't create a transaction in the rhs of catchRetry# (#26028)
We don't need to create a transaction for the rhs of (catchRetry#)
because contrary to the lhs we don't need to abort it on retry. Moreover
it is particularly harmful if we have code such as (#26028):
let cN = readTVar vN >> retry
tree = c1 `orElse` (c2 `orElse` (c3 `orElse` ...))
atomically tree
Because it will stack transactions for the rhss and the read-sets of all
the transactions will be iteratively merged in O(n^2) after the
execution of the most nested retry.
This is the second attempt at implementing this. The first attempt
triggered segfaults (#26291) and has been reverted.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
-
fcf092dd
by Luite Stegeman at 2026-03-27T04:44:17-04:00
Windows: remove StgAsyncIOResult and fix crash/leaks
In stg_block_async{_void}, a stack slot was reserved for
an StgAsyncIOResult. This slot would be filled by the IO
manager upon completion of the async call.
However, if the blocked thread was interrupted by an async
exception, we would end up in an invalid state:
- If the blocked computation was never re-entered, the
StgAsyncIOResult would never be freed.
- If the blocked computation was re-entered, the thread would
find an unitialized stack slot for the StgAsyncIOResult,
leading to a crash reading its fields, or freeing the pointer.
We fix this by removing the StgAsyncIOResult altogether and writing
the result directly to the stack.
Fixes #26341
-
05094993
by Luite Stegeman at 2026-03-27T04:45:12-04:00
Don't refine DEFAULT alt for unary typeclasses
A non-DEFAULT data alt for a unary typeclass dictionary would
interfere with Unary Class Magic, leading to segfaults.
fixes #27071
-
4ee260cf
by sheaf at 2026-03-27T04:46:06-04:00
Fix several oversights in hsExprType
This commit fixes several oversights in GHC.Hs.Syn.Type.hsExprType:
- The 'RecordCon' case was returning the type of the constructor,
instead of the constructor application. This is fixed by using
'splitFunTys'.
- The 'ExplicitTuple' case failed to take into account tuple sections,
and was also incorrectly handling 1-tuples (e.g. 'Solo') which can
be constructed using Template Haskell.
- The 'NegApp' case was returning the type of the negation operator,
again failing to apply it to the argument. Fixed by using
'funResultTy'.
- The 'HsProc' case was computing the result type of the arrow proc
block, without taking into account the argument type. Fix that by
adding a new field to 'CmdTopTc' that stores the arrow type, so that
we can construct the correct result type `arr a b` for
`proc (pat :: a) -> (cmd :: b)`.
- The 'ArithSeq' and 'NegApp' cases were failing to take into account
the result 'HsWrapper', which could e.g. silently drop casts.
This is fixed by introducing 'syntaxExpr_wrappedFunResTy' which, on
top of taking the result type, applies the result 'HsWrapper'.
These fixes are validated by the new GHC API test T26910.
Fixes #26910
-
e97232ce
by Hai at 2026-03-27T04:47:04-04:00
Parser.y: avoid looking at token with QualifiedDo
This changes the behavior of 'hintQualifiedDo' so that the supplied
token is not inspected when the QualifiedDo language extension bit is
set.
-
9831385b
by Vladislav Zavialov at 2026-03-27T17:22:30-04:00
Infix holes in types (#11107)
This patch introduces several improvements that follow naturally from
refactoring HsOpTy to represent the operator as an HsType, aligning it
with the approach taken by OpApp and HsExpr.
User-facing changes:
1. Infix holes (t1 `_` t2) are now permitted in types, following the
precedent set by term-level expressions.
Test case: T11107
2. Error messages for illegal promotion ticks are now reported at more
precise source locations.
Test case: T17865
Internal changes:
* The definition of HsOpTy now mirrors that of OpApp:
| HsOpTy (XOpTy p) (LHsType p) (LHsType p) (LHsType p)
| OpApp (XOpApp p) (LHsExpr p) (LHsExpr p) (LHsExpr p)
This moves us one step closer to unifying HsType and HsExpr.
* Ignoring locations,
the old pattern match (HsOpTy x prom lhs op rhs)
is now written as (HsOpTy x lhs (HsTyVar x' prom op) rhs)
but we also handle (HsOpTy x lhs (HsWildCardTy x') rhs)
Constructors other than HsTyVar and HsWildCardTy never appear
in the operator position.
* The various definitions across the compiler have been updated to work
with the new representation, drawing inspiration from the term-level
pipeline where appropriate. For example,
ppr_infix_ty <=> ppr_infix_expr
get_tyop <=> get_op
lookupTypeFixityRn <=> lookupExprFixityRn
(the latter is factored out from rnExpr)
Test cases: T11107 T17865
-
5b6757d7
by mangoiv at 2026-03-27T17:23:19-04:00
ci: build i386 non-validate for deb12
This is a small fix that will unlock ghcup metadata to run, i386 debian
12 was missing as a job.
-
cf942119
by Cheng Shao at 2026-03-30T15:24:37-04:00
ghc-boot: remove unused SizedSeq instances and functions
This commit removes unused `SizedSeq` instances and functions, only
keeping the bits we need for hpc tick sequence for now.
-
22c5b7cc
by Cheng Shao at 2026-03-30T15:24:38-04:00
ghci: remove unused GHCi.BinaryArray
This patch removes the unused `GHCi.BinaryArray` module from `ghci`.
Closes #27108.
-
77abb4ab
by Cheng Shao at 2026-03-30T15:25:21-04:00
testsuite: mark T17912 as fragile on Windows
T17912 is still fragile on Windows, it sometimes unexpectedly pass in
CI. This especially strains our already scarce Windows CI runner
resources. Mark it as fragile on Windows for the time being.
-
d741a6cc
by Andreas Klebinger at 2026-03-31T04:39:33-04:00
Bump minimum shake version for hadrian.
We also add the shake version we want to stack.yaml
Fixes #26884
-
5e556f9e
by Vladislav Zavialov at 2026-03-31T04:40:16-04:00
Status check for the HsType~HsExpr refactoring (#25121)
Add a test case to track the status of a refactoring project within GHC
whose goal is to arrive at the following declaration:
type HsType = HsExpr
The rationale for this is to increase code reuse between the term- and
type-level code in the compiler front-end (AST, parser, renamer, type checker).
The status report is saved to testsuite/tests/ghc-api/T25121_status.stdout
and provides useful insights into what needs to happen to make progress on
the ticket.
-
acffb1b1
by fendor at 2026-03-31T04:41:02-04:00
Extract Binary instances to `GHC.ByteCode.Binary`
-
e2ea8e25
by fendor at 2026-03-31T04:41:02-04:00
Add `seqNonEmpty` for evaluating `NonEmpty a`
-
048b00b7
by fendor at 2026-03-31T04:41:02-04:00
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.
-
2d1c1997
by Simon Jakobi at 2026-03-31T04:41:46-04:00
Eliminate dictionary-passing in ListMap operations
Mark the ListMap helpers 'INLINABLE' so importing modules can specialise
the 'TrieMap (ListMap m)' methods and avoid recursive dictionary-passing.
See Note [Making ListMap operations specialisable].
Fixes #27097
-
ed2c6570
by Cheng Shao at 2026-03-31T04:42:33-04:00
testsuite: fix testdir cleanup logic on Windows
testdir cleanup is unreliable on Windows (#13162) and despite existing
hacks in the driver, new failure mode has occurred. This patch makes
it print the warning and carry on when failed to clean up a testdir,
instead of reporting a spurious framework failure. See added comment
for detailed explanation.
-
d9388e29
by Simon Jakobi at 2026-03-31T13:14:59-04:00
Add regression test for #18177
Closes #18177.
Assisted-by: Codex
-
6a10045c
by mangoiv at 2026-03-31T13:15:43-04:00
ci: allow metric decrease for two tests on i386
There has been a nightly failure on i386 due to a compiler runtime
improvement on i386 debian 12. We allow that.
Metric Decrease (test_env='i386-linux-deb12'):
T12707 T8095
-
7fbb4fcb
by Rodrigo Mesquita at 2026-04-01T12:16:33+00:00
Bump default language edition to GHC2024
As per the accepted ghc-proposal#632
Fixes #26039
-
5ae43275
by Peng Fan at 2026-04-01T19:01:06-04:00
NCG/LA64: add cmpxchg and xchg primops
And append some new instructions for LA664 uarch.
Apply fix to cmpxchg-prim by Andreas Klebinger.
Suggestions in https://gitlab.haskell.org/ghc/ghc/-/merge_requests/15515
-
8f95534a
by Duncan Coutts at 2026-04-01T19:01:52-04:00
Remove signal-based ticker implementations
Fixes issue #27073
All supported platforms should work with the pthreads + nanosleep based
ticker implementation. This avoids all the problems with using signals.
In practice, all supported platforms were probably using the non-signal
tickers already, which is probably why we do not get lots of reports
about deadlocks and other weirdness: we were definately using functions
that are not async signal safe in the tick handler (such as fflush to
flussh the eventlog).
Only Solaris was explicitly using the timer_create ticker impl, and even
Solaris could probably use the pthreads one (if anyone cared: Solaris is
no longer a Teir 3 supported platform).
Plausibly the only supported platform that this will change will be AIX,
which should now use the pthreads impl.
-
51b32b0d
by Duncan Coutts at 2026-04-01T19:01:52-04:00
Tidy up some timer/ticker comments elsewhere
-
7562bcd7
by Duncan Coutts at 2026-04-01T19:01:52-04:00
Remove now-unused install_vtalrm_handler
Support function used by both of the signal-based ticker
implementations.
-
6da127c7
by Duncan Coutts at 2026-04-01T19:01:52-04:00
No longer probe for timer_create in rts/configure
It was only used by the TimerCreate.c ticker impl.
-
3fd490fa
by Duncan Coutts at 2026-04-01T19:01:53-04:00
Note that rtsTimerSignal is deprecated.
-
63099b0f
by Simon Jakobi at 2026-04-01T19:02:39-04:00
Add perf test for #13960
Closes #13960.
-
58009c14
by Apoorv Ingle at 2026-04-02T09:51:24+01:00
Streamline expansions using HsExpansion (#25001)
Notes added [Error Context Stack] [Typechecking by expansion: overview]
Notes updated Note [Expanding HsDo with XXExprGhcRn] [tcApp: typechecking applications]
-------------------------
Metric Decrease:
T9020
-------------------------
There are 2 key changes:
1. `HsExpand` datatype mediates between expansions
2. Replace `ErrCtxtM` to a simpler `HsCtxt` that does not depend on a `TidyEnv`
This has some consequences detailed below:
1. `HsExpand` datatype mediates between expansions
* Simplifies the implementations of `tcExpr` to work on `XExpr`
* Removes `VACtxt` (and its associated `VAExpansion` and `VACall`) datatype, it is subsumed by simply a `SrcSpan`.
* Removes the function `addHeadCtxt` as it is now mearly setting a location
* The function `tcValArgs` does its own argument number management
* move `splitHsTypes` out of `tcApp`
* Removes special case of tcBody from `tcLambdaMatches`
* Removes special case of `dsExpr` for `ExpandedThingTc`
* Renames `tcMonoExpr` -> `tcMonoLExpr`, `tcMonoExprNC` -> `tcMonoLExpr`
* Renames `EValArg`, `EValArgQL` fields: `ea_ctxt` -> `ea_loc_span` and `eaql_ctx` -> `eaql_loc_span`
* Remove `PopErrCtxt` from `XXExprGhcRn`
* `fun_orig` in tcInstFun depends on the SrcSpan of the head of the application chain (similar to addArgCtxt)
- it references the application chain head if it is user located, or
uses the error context stack as a fallback if it's a generated
location
* Make a new variant `GeneratedSrcSpan` in `SrcSpan` for HIEAst Nodes
- Expressions wrapped around `GeneratedSrcSpan` are ignored and never added to the error context stack
- In Explicit list expansion `fromListN` is wrapped with a `GeneratedSrcSpan` with `GeneratedSrcSpanDetails` field to store the original srcspan
2. Replace `ErrCtxtM` to a simpler `HsCtxt` that does not depend on a `TidyEnv`
* Merge `HsThingRn` to `HsCtxt`
* Landmark Error messages are now just computed on the fly
* Make HsExpandedRn and HsExpandedTc payload a located HsExpr GhcRn
* `HsCtxt` are tidied and zonked at the end right before printing
Co-authored-by: simonpj <simon.peytonjones@gmail.com>
-
bc4b4487
by Zubin Duggal at 2026-04-03T14:22:27-04:00
driver: recognise .dyn_o as a valid object file to link if passed on the command line.
This allows plugins compiled with this suffix to run.
Fixes #24486
-
5ebb9121
by Simon Jakobi at 2026-04-03T14:23:11-04:00
Add regression test for #16145
Closes #16145.
-
c1fc1c44
by Simon Peyton Jones at 2026-04-03T19:56:07-04:00
Refactor eta-expansion in Prep
The Prep pass does eta-expansion but I found cases where it was
doing bad things. So I refactored and simplified it quite a bit.
In the new design
* There is no distinction between `rhs` and `body`; in particular,
lambdas can now appear anywhere, rather than just as the RHS of
a let-binding.
* This change led to a significant simplification of Prep, and
a more straightforward explanation of eta-expansion. See the new
Note [Eta expansion]
* The consequences is that CoreToStg needs to handle naked lambdas.
This is very easy; but it does need a unique supply, which forces
some simple refactoring. Having a unique supply to hand is probably
a good thing anyway.
-
21beda2c
by Simon Peyton Jones at 2026-04-03T19:56:07-04:00
Clarify Note [Interesting dictionary arguments]
Ticket #26831 ended up concluding that the code for
GHC.Core.Opt.Specialise.interestingDict was good, but the
commments were a bit inadequate.
This commit improves the comments slightly.
-
3eaac1f2
by Simon Peyton Jones at 2026-04-03T19:56:07-04:00
Make inlining a bit more eager for overloaded functions
If we have
f d = ... (class-op d x y) ...
we should be eager to inline `f`, because that may change the
higher order call (class-op d x y) into a call to a statically
known function.
See the discussion on #26831.
Even though this does a bit /more/ inlining, compile times
decrease by an average of 0.4%.
Compile time changes:
DsIncompleteRecSel3(normal) 431,786,104 -2.2%
ManyAlternatives(normal) 670,883,768 -1.6%
ManyConstructors(normal) 3,758,493,832 -2.6% GOOD
MultilineStringsPerf(normal) 29,900,576 -2.8%
T14052Type(ghci) 1,047,600,848 -1.2%
T17836(normal) 392,852,328 -5.2%
T18478(normal) 442,785,768 -1.4%
T21839c(normal) 341,536,992 -14.1% GOOD
T3064(normal) 174,086,152 +5.3% BAD
T5631(normal) 506,867,800 +1.0%
hard_hole_fits(normal) 209,530,736 -1.3%
info_table_map_perf(normal) 19,523,093,184 -1.2%
parsing001(normal) 377,810,528 -1.1%
pmcOrPats(normal) 60,075,264 -0.5%
geo. mean -0.4%
minimum -14.1%
maximum +5.3%
Runtime changes
haddock.Cabal(normal) 27,351,988,792 -0.7%
haddock.base(normal) 26,997,212,560 -0.6%
haddock.compiler(normal) 219,531,332,960 -1.0%
Metric Decrease:
LinkableUsage01
ManyConstructors
T17949
T21839c
T13035
TcPlugin_RewritePerf
hard_hole_fits
Metric Increase:
T3064
-
5cbc2c82
by Matthew Pickering at 2026-04-03T19:57:02-04:00
bytecode: Add magic header/version to bytecode files
In order to avoid confusing errors when using stale interface files (ie
from an older compiler version), we add a simple header/version check
like the one for interface files.
Fixes #27068
-
d95a1936
by fendor at 2026-04-03T19:57:02-04:00
Add constants for bytecode in-memory buffer size
Introduce a common constant for the default size of the .gbc and
.bytecodelib binary buffer.
The buffer is by default set to 1 MB.
-
b822c30a
by mangoiv at 2026-04-03T19:57:49-04:00
testsuite: filter stderr for static001 on darwin
This reactivates the test on x86_64 darwin as this should have been done
long ago and ignores warnings emitted by ranlib on newer version of the
darwin toolchain since they are benign. (no symbols for stub libraries)
Fixes #27116
-
28ce1f8a
by Andreas Klebinger at 2026-04-03T19:58:44-04:00
Give the Data instance for ModuleName a non-bottom toConstr implementation.
I've also taken the liberty to add Note [Data.Data instances for GHC AST Types]
describing some of the uses of Data.Data I could find.
Fixes #27129
-
8ca41ffe
by mangoiv at 2026-04-03T19:59:30-04:00
issue template: fix add bug label
-
3981db0c
by Sylvain Henry at 2026-04-03T20:00:33-04:00
Add more canned GC functions for common register patterns (#27142)
Based on analysis of heap-check sites across the GHC compiler and Cabal,
the following patterns were not covered by existing canned GC functions
but occurred frequently enough to warrant specialisation:
stg_gc_ppppp -- 5 GC pointers
stg_gc_ip -- unboxed word + GC pointer
stg_gc_pi -- GC pointer + unboxed word
stg_gc_ii -- two unboxed words
stg_gc_bpp -- byte (I8) + two GC pointers
Adding these reduces the fraction of heap-check sites falling back to
the generic GC path from ~1.4% to ~0.4% when compiling GHC itself.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
-
d17d1435
by Matthew Pickering at 2026-04-03T20:01:19-04:00
Make home unit dependencies stored as sets
Co-authored-by: Wolfgang Jeltsch <wolfgang@well-typed.com>
-
92a97015
by Simon Peyton Jones at 2026-04-05T00:58:57+01:00
Add Invariant (NoTypeShadowing) to Core
This commit addresses #26868, by adding
a new invariant (NoTypeShadowing) to Core.
See Note [No type-shadowing in Core] in GHC.Core
-
8b5a5020
by Simon Peyton Jones at 2026-04-05T00:58:57+01:00
Major refactor of free-variable functions
For some time we have had two free-variable mechanims for types:
* The "FV" mechanism, embodied in GHC.Utils.FV, which worked OK, but
was fragile where eta-expansion was concerned.
* The TyCoFolder mechanism, using a one-shot EndoOS accumulator
I finally got tired of this and refactored the whole thing, thereby
addressing #27080. Now we have
* `GHC.Types.Var.FV`, which has a composable free-variable result type,
very much in the spirit of the old `FV`, but much more robust.
(It uses the "one shot trick".)
* GHC.Core.TyCo.FVs now has just one technology for free variables.
All this led to a lot of renaming.
There are couple of error-message changes. The change in T18451
makes an already-poor error message even more mysterious. But
it really needs a separate look.
We also now traverse the AST in a different order leading to a different
but still deterministic order for FVs and test output has been adjusted
accordingly.
-
4bf040c6
by sheaf at 2026-04-05T14:56:29-04:00
Add utility pprTrace_ function
This function is useful for quick debugging, as it can be added to a
where clause to pretty-print debugging information:
fooBar x y
| cond = body1
| otherwise = body2
where
!_ = pprTrace_ "fooBar" $
vcat [ text "x:" <+> ppr x
, text "y:" <+> ppr y
, text "cond:" <+> ppr cond
]
-
502e6ffe
by Andrew Lelechenko at 2026-04-07T04:47:21-04:00
base: improve error message for Data.Char.chr
As per https://github.com/haskell/core-libraries-committee/issues/384
-
b21bd52e
by Simon Peyton Jones at 2026-04-07T04:48:07-04:00
Refactor FunResCtxt a bit
Fixes #27154
-
7fe84ea5
by Zubin Duggal at 2026-04-07T19:11:52+05:30
compiler: Warn when -finfo-table-map is used with -fllvm
These are currently not supported together.
Fixes #26435
-
4a45a7da
by Matthew Pickering at 2026-04-08T04:37:29-04:00
packaging: correctly propagate build/host/target to bindist configure script
At the moment the host and target which we will produce a compiler for
is fixed at the initial configure time. Therefore we need to persist
the choice made at this time into the installation bindist as well so we
look for the right tools, with the right prefixes at install time.
In the future, we want to provide a bit more control about what kind of
bindist we produce so the logic about what the host/target will have to
be written by hadrian rather than persisted by the configure script. In
particular with cross compilers we want to either build a normal stage 2
cross bindist or a stage 3 bindist, which creates a bindist which has a
native compiler for the target platform.
Fixes #21970
Co-authored-by: Sven Tennie <sven.tennie@gmail.com>
-
b0950df6
by Sven Tennie at 2026-04-08T04:37:29-04:00
Cross --host and --target no longer required for cross (#21970)
We set sane defaults in the configure script. Thus, these paramenters
aren't required any longer.
-
fef35216
by Sven Tennie at 2026-04-08T04:37:30-04:00
ci: Define USER_CONF_CC_OPTS_STAGE2 for aarch64/mingw
ghc-toolchain doesn't see $CONF_CC_OPTS_STAGE2 when the bindist gets
configured. So, the hack to override the compiler gets lost.
-
8dd6f453
by Cheng Shao at 2026-04-08T04:38:11-04:00
ghci: use ShortByteString for LookupSymbol/LookupSymbolInDLL/LookupClosure messages
This patch refactors ghci to use `ShortByteString` for
`LookupSymbol`/`LookupSymbolInDLL`/`LookupClosure` messages as the
first part of #27147.
Co-authored-by: Codex <codex@openai.com>
-
371ef200
by Cheng Shao at 2026-04-08T04:38:11-04:00
ghci: use ShortByteString for MkCostCentres message
This patch refactors ghci to use `ShortByteString` for `MkCostCentres`
messages as a first part of #27147. This also considerably lowers the
memory overhead of breakpoints when cost center profiling is enabled.
-------------------------
Metric Decrease:
interpreter_steplocal
-------------------------
Co-authored-by: Codex <codex@openai.com>
-
4a122bb6
by Phil Hazelden at 2026-04-08T20:49:42-04:00
Implement modifiers syntax.
The `%m` syntax of linear types is now accepted in more places, to allow
use by future extensions, though so far linear types is still the only
consumer.
This may break existing code where it
* Uses -XLinearTypes.
* Has code of the form `a %m -> b`, where `m` can't be inferred to be
kind Multiplicity.
The code can be fixed either by adding a kind annotation, or by setting
`-XLinearTypes -XNoModifiers`.
Proposal:
https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0370-modifiers.rst
-
07267f79
by Zubin Duggal at 2026-04-08T20:50:25-04:00
hadrian: Don't include the package hash in the haddock directory
Since GHC 9.8 and hash_unit_ids, haddock urls have looked like`ghc-9.10.3/doc/html/libraries/base-4.20.2.0-39f9/**/*.html`
The inclusion of the hash makes it hard for downstream non-boot packages to properly link to these files, as the hash is not
part of a standard cabal substitution.
Since we only build one version of each package, we don't need the hash to disambiguate anything, we can just remove it.
Fixes #26635
-
0a83b95b
by ARATA Mizuki at 2026-04-08T20:51:18-04:00
testsuite: Allow multiple ways to be run by setting multiple command-line options
This patch allows multiple `--test-way`s to take effect, like:
$ hadrian/build test --test-way=normal --test-way=llvm
Previously, only one way was run if the test speed was 'normal' or 'fast'.
Closes #26926
Co-authored-by: sheaf <sam.derbyshire@gmail.com>
-
e841931c
by Teo Camarasu at 2026-04-08T20:52:00-04:00
doc: improve eventlog-flush-interval flag documentation
We mention the performance cost and how this flag can be turned off.
Resolves #27056
-
e332db25
by Teo Camarasu at 2026-04-08T20:52:01-04:00
docs/user_guide: fix typo
-
5b82080a
by Simon Jakobi at 2026-04-08T20:52:44-04:00
Fix -dsuppress-uniques for free variables in demand signatures
Before: Str=b{sXyZ->S}
With this patch: Str=b{S}
T13143.stderr is updated accordingly.
Fixes #27106.
-
b7a084cc
by Simon Jakobi at 2026-04-08T20:53:27-04:00
Documentation fixes for demand signature notation
Fixes #27115.
-
59391132
by Simon Jakobi at 2026-04-08T20:54:08-04:00
Use upsert for non-deleting map updates
Some compiler functions were using `alter`, despite never removing
any entries: they only update an existing entry or insert a new one.
These functions are converted to using `upsert`:
alter :: (Maybe a -> Maybe a) -> Key -> Map a -> Map a
upsert :: (Maybe a -> a) -> Key -> Map a -> Map a
`upsert` variants are also added to APIs of the various Word64Map
wrapper types.
The precedent for this `upsert` operation is in the containers library:
see https://github.com/haskell/containers/pull/1145
Metrics: compile_time/bytes allocated
-------------------------------------
geo. mean: -0.1%
minimum: -0.5%
maximum: +0.0%
Resolves #27140.
-
da7e82f4
by Cheng Shao at 2026-04-08T20:54:49-04:00
testsuite: fix testsuite run for +ipe again
This patch makes the +ipe flavour transformer pass the entire
testsuite again by dropping stdout/stderr checks of certain tests that
are sensitive to stack layout changes with `+ipe`. Related: #26799.
-
b135a87d
by Zubin Duggal at 2026-04-09T19:36:50+05:30
Bump directory submodule to 1.3.11.0 (unreleased)
-
3a291d07
by Zubin Duggal at 2026-04-09T19:36:50+05:30
Bump file-io submodule to 0.2.0
-
e0ab606d
by Zubin Duggal at 2026-04-10T18:40:20+05:30
Release notes for GHC 10.0
-
e08b9b34
by Zubin Duggal at 2026-04-10T18:40:20+05:30
Bump ghc-prim version to 0.14.0
-
a92aac6e
by Zubin Duggal at 2026-04-10T18:40:20+05:30
Bump template-haskell to 2.25.0.0; update submodule exceptions for TH 2.25
-
f254d9e8
by Zubin Duggal at 2026-04-10T18:40:20+05:30
Bump GHC version to 10.0
-
6ce0368a
by Zubin Duggal at 2026-04-10T18:40:28+05:30
Bump base to 4.23.0.0; update submodules for base 4.24 upper bound
-
702fb8a5
by Zubin Duggal at 2026-04-10T18:40:28+05:30
Bump GHC version to 10.1; update submodules template-haskell-lift and template-haskell-quasiquoter for ghc-internal 10.200
-
75df1ca4
by Zubin Duggal at 2026-04-10T18:40:28+05:30
Use changelog.d for release notes (#26002)
GHC now uses a fragment-based changelog workflow using a custom script adapted from https://codeberg.org/fgaz/changelog-d.
Contributors add a file in changelog.d/ for each user-facing change.
At release time, these are assembled into release notes for sphinx (in RST) format, using
the tool.
New hadrian `changelog` target to generate changelogs
CI job to validate changelog entries for MRs unless skipped with ~"no-changelog" label
Teach sphinx about ghc-mr: extlink to link to MRs
Remove `ghc-package-list` from sphinx, and implement it in changelog-d instead (Fixes #26476).
(cherry picked from commit 989c07249978f418dfde1353abfad453f024d61a)
-
585d7450
by Luite Stegeman at 2026-04-11T02:17:13-04:00
tc: discard warnings in tcUserStmt Plan C
We typecheck let_stmt twice, but we don't want the warnings twice!
see #26233
-
2df604e9
by Sylvain Henry at 2026-04-11T02:19:30-04:00
Introduce TargetInt to represent target's Int (#15973)
GHC was using host 'Int' in several places to represent values that
live in the target machine's 'Int' type. This is silently wrong when
cross-compiling from a 32-bit host to a 64-bit target: the host Int
is 32 bits while the target Int is 64 bits.
See Note [TargetInt] in GHC.Platform.
Also used the opportunity to make DynTag = Word8.
Fixes #15973
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
-
d419e972
by Luite Stegeman at 2026-04-13T15:16:04-04:00
Suppress desugaring warnings in the pattern match checker
Avoid duplicating warnings from the actual desugaring pass.
fixes #25996
-
c5b80dd0
by Phil de Joux at 2026-04-13T15:16:51-04:00
Typo ~/ghc/arch-os-version/environments
-
71462fff
by Luite Stegeman at 2026-04-13T15:17:38-04:00
add changelog entry for #26233
-
d1ddfd4b
by Rodrigo Mesquita at 2026-04-14T18:41:12-04:00
Add test for #25636
The existing test behaviour of "T23146_liftedeq" changed because the
simplifier now does a bit more inlining. We can restore the previous bad
behavior by using an OPAQUE pragma.
This test doubles as a test for #25636 when run in ghci, so we add it as
such.
-
b9df40ee
by Rodrigo Mesquita at 2026-04-14T18:41:12-04:00
refactor: protoBCOName is always a Name
Simplifies the code by removing the unnecessary type argument to
ProtoBCO which was always 'Name'
-
5c2a179e
by Rodrigo Mesquita at 2026-04-14T18:41:12-04:00
Allocate static constructors for bytecode
This commit adds support for static constructors when compiling and
linking ByteCode objects.
Top-level StgRhsCon get lowered to ProtoStaticCons rather than to
ProtoBCOs. A ProtoStaticCon gets allocated directly as a data con
application on the heap (using the new primop newConApp#).
Previously, we would allocate a ProtoBCO which, when evaluated, would
PACK and return the constructor.
A few more details are given in Note [Static constructors in Bytecode].
Secondly, this commit also fixes issue #25636 which was caused by
linking *unlifted* constructors in BCO instructions as
- (1) a thunk indexing the array of BCOs in a module
- (2) which evaluated to a BCO which still had to be evaluated to
return the unlifted constructor proper.
The (2) issue has been resolved by allocating the static constructors
directly. The (1) issue can be resolved by ensuring that we allocate all
unlifted top-level constructors eagerly, and leave the knot-tying for
the lifted BCOs and top-level constructors only.
The top-level unlifted constructors are never mutually recursive, so we
can allocate them all in one go as long as we do it in topological
order. Lifted fields of unlifted constructors can still be filled by the
knot-tied lifted variables since in those fields it is fine to keep
those thunks. See Note [Tying the knot in createBCOs] for more details.
Fixes #25636
-------------------------
Metric Decrease:
LinkableUsage01
-------------------------
-
cde47053
by Rodrigo Mesquita at 2026-04-14T18:41:12-04:00
Revert "StgToByteCode: Assert that PUSH_G'd values are lifted"
This reverts commit ec26c54d818e0cd328276196930313f66b780905.
Ever since f7a22c0f4e9ae0dc767115d4c53fddbd8372b777, we now do support
and will link top-level unlifted constructors into evaluated and
properly tagged values which we can reference with PUSH_G.
This assertion is no longer true and triggered a failure in T25636
-
c7a7e5b8
by Rodrigo Mesquita at 2026-04-14T18:41:12-04:00
refactor: Tag more remote Ptrs as RemotePtr
Pure refactor which improves the API of
- GHC.ByteCode.Linker
- GHC.Runtime.Interpreter
- GHC.Runtime.Interpreter.Types.SymbolCache
by using `RemotePtr` for more functions which used to return `Ptr`s that
could potentially be in a foreign process. E.g. `lookupIE`,
`lookupStaticPtr`, etc...
-
fc59494c
by Rodrigo Mesquita at 2026-04-14T18:41:12-04:00
Add float# and subword tests for #25636
These tests cover that static constructors in bytecode work correctly
for Float# and subword values (Word8#, Word16#)
-
477f521b
by Rodrigo Mesquita at 2026-04-14T18:41:12-04:00
test: Validate topoSort logic in createBCOs
This test validates that the topological sorting and ordering of the
unlifted constructors and lifted constructors in `createBCOs` is
correct.
See `Note [Tying the knot in createBCOs]` for why tying the knot for the
created BCOs is slightly difficult and why the topological sorting is
necessary.
This test fails when `let topoSortedObjs = topSortObjs objs` is
substituted by `let topoSortedObjs = zip [0..] objs`, thus witnessing
the toposort logic is correct and necessary.
The test calls the ghci `createBCOs` directly because it is currently
impossible to construct in Source Haskell a situation where a top-level
static unlifted constructor depends on another (we don't have top-level
unlifted constructors except for nullary constructors like `Leaf ::
(UTree :: UnliftedType)`).
This is another test for fix for #25636
-
2d9c30be
by Simon Jakobi at 2026-04-14T18:42:00-04:00
Improve tests for `elem`
...in order to simplify the work on #27096.
* Improve T17752 by including the Core output in golden files, checking
both -O1 and -O2.
* Add tests for fusion and no-fusion cases.
Fixes #27101.
-
9722d8c3
by Brandon Chinn at 2026-04-15T09:39:33-07:00
Add Data.RealFloat
-
909ee5fc
by Brandon Chinn at 2026-04-15T09:39:34-07:00
Add Infinity/NegInfinity/NaN pattern synonyms (#26961)