[Git][ghc/ghc][wip/T23309] 160 commits: base: Export GHC.Conc.Sync.fromThreadId
Ryan Scott (@RyanGlScott)
gitlab at gitlab.haskell.org
Mon Jun 5 11:27:25 UTC 2023
Ryan Scott pushed to branch wip/T23309 at Glasgow Haskell Compiler / GHC
Commits:
0a513952 by Ben Gamari at 2023-05-11T04:10:17-04:00
base: Export GHC.Conc.Sync.fromThreadId
Closes #22706.
- - - - -
29be39ba by Matthew Pickering at 2023-05-11T04:10:54-04:00
Build vanilla alpine bindists
We currently attempt to build and distribute fully static alpine
bindists (ones which could be used on any linux platform) but most
people who use the alpine bindists want to use alpine to build their own
static applications (for which a fully static bindist is not necessary).
We should build and distribute these bindists for these users whilst the
fully-static bindist is still unusable.
Fixes #23349
- - - - -
40c7daed by Simon Peyton Jones at 2023-05-11T04:11:30-04:00
Look both ways when looking for quantified equalities
When looking up (t1 ~# t2) in the quantified constraints,
check both orientations. Forgetting this led to #23333.
- - - - -
c17bb82f by Rodrigo Mesquita at 2023-05-11T04:12:07-04:00
Move "target has RTS linker" out of settings
We move the "target has RTS linker" information out of configure into a
predicate in GHC, and remove this option from the settings file where it
is unnecessary -- it's information statically known from the platform.
Note that previously we would consider `powerpc`s and `s390x`s other
than `powerpc-ibm-aix*` and `s390x-ibm-linux` to have an RTS linker,
but the RTS linker supports neither platform.
Closes #23361
- - - - -
bd0b056e by Krzysztof Gogolewski at 2023-05-11T04:12:44-04:00
Add a test for #17284
Since !10123 we now reject this program.
- - - - -
630b1fea by Bodigrim at 2023-05-11T04:13:24-04:00
Document unlawfulness of instance Num Fixed
Fixes #22712
- - - - -
87eebf98 by sheaf at 2023-05-11T11:55:22-04:00
Add fused multiply-add instructions
This patch adds eight new primops that fuse a multiplication and an
addition or subtraction:
- `{fmadd,fmsub,fnmadd,fnmsub}{Float,Double}#`
fmadd x y z is x * y + z, computed with a single rounding step.
This patch implements code generation for these primops in the following
backends:
- X86, AArch64 and PowerPC NCG,
- LLVM
- C
WASM uses the C implementation. The primops are unsupported in the
JavaScript backend.
The following constant folding rules are also provided:
- compute a * b + c when a, b, c are all literals,
- x * y + 0 ==> x * y,
- ±1 * y + z ==> z ± y and x * ±1 + z ==> z ± x.
NB: the constant folding rules incorrectly handle signed zero.
This is a known limitation with GHC's floating-point constant folding
rules (#21227), which we hope to resolve in the future.
- - - - -
ad16a066 by Krzysztof Gogolewski at 2023-05-11T11:55:59-04:00
Add a test for #21278
- - - - -
05cea68c by Matthew Pickering at 2023-05-11T11:56:36-04:00
rts: Refine memory retention behaviour to account for pinned/compacted objects
When using the copying collector there is still a lot of data which
isn't copied (such as pinned, compacted, large objects etc). The logic
to decide how much memory to retain didn't take into account that these
wouldn't be copied. Therefore we pessimistically retained 2* the amount
of memory for these blocks even though they wouldn't be copied by the
collector.
The solution is to split up the heap into two parts, the parts which
will be copied and the parts which won't be copied. Then the appropiate
factor is applied to each part individually (2 * for copying and 1.2 *
for not copying).
The T23221 test demonstrates this improvement with a program which first
allocates many unpinned ByteArray# followed by many pinned ByteArray#
and observes the difference in the ultimate memory baseline between the
two.
There are some charts on #23221.
Fixes #23221
- - - - -
1bb24432 by Cheng Shao at 2023-05-11T11:57:15-04:00
hadrian: fix no_dynamic_libs flavour transformer
This patch fixes the no_dynamic_libs flavour transformer and make
fully_static reuse it. Previously building with no_dynamic_libs fails
since ghc program is still dynamic and transitively brings in dyn ways
of rts which are produced by no rules.
- - - - -
0ed493a3 by Josh Meredith at 2023-05-11T23:08:27-04:00
JS: refactor jsSaturate to return a saturated JStat (#23328)
- - - - -
a856d98e by Pierre Le Marre at 2023-05-11T23:09:08-04:00
Doc: Fix out-of-sync using-optimisation page
- Make explicit that default flag values correspond to their -O0 value.
- Fix -fignore-interface-pragmas, -fstg-cse, -fdo-eta-reduction,
-fcross-module-specialise, -fsolve-constant-dicts, -fworker-wrapper.
- - - - -
c176ad18 by sheaf at 2023-05-12T06:10:57-04:00
Don't panic in mkNewTyConRhs
This function could come across invalid newtype constructors, as we
only perform validity checking of newtypes once we are outside the
knot-tied typechecking loop.
This patch changes this function to fake up a stub type in the case of
an invalid newtype, instead of panicking.
This patch also changes "checkNewDataCon" so that it reports as many
errors as possible at once.
Fixes #23308
- - - - -
ab63daac by Krzysztof Gogolewski at 2023-05-12T06:11:38-04:00
Allow Core optimizations when interpreting bytecode
Tracking ticket: #23056
MR: !10399
This adds the flag `-funoptimized-core-for-interpreter`, permitting use
of the `-O` flag to enable optimizations when compiling with the
interpreter backend, like in ghci.
- - - - -
c6cf9433 by Ben Gamari at 2023-05-12T06:12:14-04:00
hadrian: Fix mention of non-existent removeFiles function
Previously Hadrian's bindist Makefile referred to a `removeFiles`
function that was previously defined by the `make` build system. Since
the `make` build system is no longer around, this function is now
undefined. Naturally, make being make, this appears to be silently
ignored instead of producing an error.
Fix this by rewriting it to `rm -f`.
Closes #23373.
- - - - -
eb60ec18 by Bodigrim at 2023-05-12T06:12:54-04:00
Mention new implementation of GHC.IORef.atomicSwapIORef in the changelog
- - - - -
aa84cff4 by Teo Camarasu at 2023-05-12T19:27:23-04:00
rts: Ensure non-moving gc is not running when pausing
- - - - -
5ad776ab by Teo Camarasu at 2023-05-12T19:27:23-04:00
rts: Teach listAllBlocks about nonmoving heap
List all blocks on the non-moving heap.
Resolves #22627
- - - - -
d683b2e5 by Krzysztof Gogolewski at 2023-05-12T19:28:00-04:00
Fix coercion optimisation for SelCo (#23362)
setNominalRole_maybe is supposed to output a nominal coercion.
In the SelCo case, it was not updating the stored role to Nominal,
causing #23362.
- - - - -
59aa4676 by Alexis King at 2023-05-12T19:28:47-04:00
hadrian: Fix linker script flag for MergeObjects builder
This fixes what appears to have been a typo in !9530. The `-t` flag just
enables tracing on all versions of `ld` I’ve looked at, while `-T` is
used to specify a linker script. It seems that this worked anyway for
some reason on some `ld` implementations (perhaps because they
automatically detect linker scripts), but the missing `-T` argument
causes `gold` to complain.
- - - - -
4bf9fa0f by Adam Gundry at 2023-05-12T23:49:49-04:00
Less coercion optimization for non-newtype axioms
See Note [Push transitivity inside newtype axioms only] for an explanation
of the change here. This change substantially improves the performance of
coercion optimization for programs involving transitive type family reductions.
-------------------------
Metric Decrease:
CoOpt_Singletons
LargeRecord
T12227
T12545
T13386
T15703
T5030
T8095
-------------------------
- - - - -
dc0c9574 by Adam Gundry at 2023-05-12T23:49:49-04:00
Move checkAxInstCo to GHC.Core.Lint
A consequence of the previous change is that checkAxInstCo is no longer
called during coercion optimization, so it can be moved back where it belongs.
Also includes some edits to Note [Conflict checking with AxiomInstCo] as
suggested by @simonpj.
- - - - -
8b9b7dbc by Simon Peyton Jones at 2023-05-12T23:50:25-04:00
Use the eager unifier in the constraint solver
This patch continues the refactoring of the constraint solver
described in #23070.
The Big Deal in this patch is to call the regular, eager unifier from the
constraint solver, when we want to create new equalities. This
replaces the existing, unifyWanted which amounted to
yet-another-unifier, so it reduces duplication of a rather subtle
piece of technology. See
* Note [The eager unifier] in GHC.Tc.Utils.Unify
* GHC.Tc.Solver.Monad.wrapUnifierTcS
I did lots of other refactoring along the way
* I simplified the treatment of right hand sides that contain CoercionHoles.
Now, a constraint that contains a hetero-kind CoercionHole is non-canonical,
and cannot be used for rewriting or unification alike. This required me
to add the ch_hertero_kind flag to CoercionHole, with consequent knock-on
effects. See wrinkle (2) of `Note [Equalities with incompatible kinds]` in
GHC.Tc.Solver.Equality.
* I refactored the StopOrContinue type to add StartAgain, so that after a
fundep improvement (for example) we can simply start the pipeline again.
* I got rid of the unpleasant (and inefficient) rewriterSetFromType/Co functions.
With Richard I concluded that they are never needed.
* I discovered Wrinkle (W1) in Note [Wanteds rewrite Wanteds] in
GHC.Tc.Types.Constraint, and therefore now prioritise non-rewritten equalities.
Quite a few error messages change, I think always for the better.
Compiler runtime stays about the same, with one outlier: a 17% improvement in T17836
Metric Decrease:
T17836
T18223
- - - - -
5cad28e7 by Bartłomiej Cieślar at 2023-05-12T23:51:06-04:00
Cleanup of dynflags override in export renaming
The deprecation warnings are normally emitted whenever the name's GRE is being looked up, which calls the GHC.Rename.Env.addUsedGRE function. We do not want those warnings to be emitted when renaming export lists, so they are artificially turned off by removing all warning categories from DynFlags at the beginning of GHC.Tc.Gen.Export.rnExports. This commit removes that dependency by unifying the function used for GRE lookup in lookup_ie to lookupGreAvailRn and disabling the call to addUsedGRE in said function (the warnings are also disabled in a call to lookupSubBndrOcc_helper in lookupChildrenExport), as per #17957. This commit also changes the setting for whether to warn about deprecated names in addUsedGREs to be an explicit enum instead of a boolean.
- - - - -
d85ed900 by Alexis King at 2023-05-13T08:45:18-04:00
Use a uniform return convention in bytecode for unary results
fixes #22958
- - - - -
8a0d45f7 by Bodigrim at 2023-05-13T08:45:58-04:00
Add more instances for Compose: Enum, Bounded, Num, Real, Integral
See https://github.com/haskell/core-libraries-committee/issues/160 for discussion
- - - - -
902f0730 by Simon Peyton Jones at 2023-05-13T14:58:34-04:00
Make GHC.Types.Id.Make.shouldUnpackTy a bit more clever
As #23307, GHC.Types.Id.Make.shouldUnpackTy was leaving money on the
table, failing to unpack arguments that are perfectly unpackable.
The fix is pretty easy; see Note [Recursive unboxing]
- - - - -
a5451438 by sheaf at 2023-05-13T14:59:13-04:00
Fix bad multiplicity role in tyConAppFunCo_maybe
The function tyConAppFunCo_maybe produces a multiplicity coercion
for the multiplicity argument of the function arrow, except that
it could be at the wrong role if asked to produce a representational
coercion. We fix this by using the 'funRole' function, which computes
the right roles for arguments to the function arrow TyCon.
Fixes #23386
- - - - -
5b9e9300 by sheaf at 2023-05-15T11:26:59-04:00
Turn "ambiguous import" error into a panic
This error should never occur, as a lookup of a type or data constructor
should never be ambiguous. This is because a single module cannot export
multiple Names with the same OccName, as per item (1) of
Note [Exporting duplicate declarations] in GHC.Tc.Gen.Export.
This code path was intended to handle duplicate record fields, but the
rest of the code had since been refactored to handle those in a
different way.
We also remove the AmbiguousImport constructor of IELookupError, as
it is no longer used.
Fixes #23302
- - - - -
e305e60c by M Farkas-Dyck at 2023-05-15T11:27:41-04:00
Unbreak some tests with latest GNU grep, which now warns about stray '\'.
Confusingly, the testsuite mangled the error to say "stray /".
We also migrate some tests from grep to grep -E, as it seems the author actually wanted an "POSIX extended" (a.k.a. sane) regex.
Background: POSIX specifies 2 "regex" syntaxen: "basic" and "extended". Of these, only "extended" syntax is actually a regular expression. Furthermore, "basic" syntax is inconsistent in its use of the '\' character — sometimes it escapes a regex metacharacter, but sometimes it unescapes it, i.e. it makes an otherwise normal character become a metacharacter. This baffles me and it seems also the authors of these tests. Also, the regex(7) man page (at least on Linux) says "basic" syntax is obsolete. Nearly all modern tools and libraries are consistent in this use of the '\' character (of which many use "extended" syntax by default).
- - - - -
5ae81842 by sheaf at 2023-05-15T14:49:17-04:00
Improve "ambiguous occurrence" error messages
This error was sometimes a bit confusing, especially when data families
were involved. This commit improves the general presentation of the
"ambiguous occurrence" error, and adds a bit of extra context in the
case of data families.
Fixes #23301
- - - - -
2f571afe by Sylvain Henry at 2023-05-15T14:50:07-04:00
Fix GHCJS OS platform (fix #23346)
- - - - -
86aae570 by Oleg Grenrus at 2023-05-15T14:50:43-04:00
Split DynFlags structure into own module
This will allow to make command line parsing to depend on
diagnostic system (which depends on dynflags)
- - - - -
fbe3fe00 by Josh Meredith at 2023-05-15T18:01:43-04:00
Replace the implementation of CodeBuffers with unboxed types
- - - - -
21f3aae7 by Josh Meredith at 2023-05-15T18:01:43-04:00
Use unboxed codebuffers in base
Metric Decrease:
encodingAllocations
- - - - -
18ea2295 by Ben Gamari at 2023-05-15T18:02:20-04:00
rts: Weak pointer cleanups
Various stylistic cleanups. No functional changes.
- - - - -
c343112f by Ben Gamari at 2023-05-15T18:02:20-04:00
rts: Don't force debug output to stderr
Previously `+RTS -Dw -l` would emit debug output to the eventlog while
`+RTS -l -Dw` would emit it to stderr. This was because the parser for
`-D` would unconditionally override the debug output target. Now we
instead only do so if no it is currently `TRACE_NONE`.
- - - - -
a5f5f067 by Ben Gamari at 2023-05-15T18:02:20-04:00
rts: Forcibly flush eventlog on barf
Previously we would attempt to flush via `endEventLogging` which can
easily deadlock, e.g., if `barf` fails during GC.
Using `flushEventLog` directly may result in slightly less consistent
eventlog output (since we don't take all capabilities before flushing)
but avoids deadlocking.
- - - - -
73b1e87c by Ben Gamari at 2023-05-15T18:02:20-04:00
rts: Assert that pointers aren't cleared by -DZ
This turns many segmentation faults into much easier-to-debug assertion
failures by ensuring that LOOKS_LIKE_*_PTR checks recognize bit-patterns
produced by `+RTS -DZ` clearing as invalid pointers.
This is a bit ad-hoc but this is the debug runtime.
- - - - -
37fb61d8 by Ben Gamari at 2023-05-15T18:02:20-04:00
rts: Introduce printGlobalThreads
- - - - -
451d65a6 by Ben Gamari at 2023-05-15T18:02:20-04:00
rts: Don't sanity-check StgTSO.global_link
See Note [Avoid dangling global_link pointers].
Fixes #19146.
- - - - -
d69cbd78 by sheaf at 2023-05-15T18:03:00-04:00
Split up tyThingToIfaceDecl from GHC.Iface.Make
This commit moves tyThingToIfaceDecl and coAxiomToIfaceDecl
from GHC.Iface.Make into GHC.Iface.Decl.
This avoids GHC.Types.TyThing.Ppr, which needs tyThingToIfaceDecl,
transitively depending on e.g. GHC.Iface.Load and GHC.Tc.Utils.Monad.
- - - - -
4d29ecdf by sheaf at 2023-05-15T18:03:00-04:00
Migrate errors to diagnostics in GHC.Tc.Module
This commit migrates the errors in GHC.Tc.Module to use the new
diagnostic infrastructure.
It required a significant overhaul of the compatibility checks between
an hs-boot or signature module and its implementation; we now use
a Writer monad to accumulate errors; see the BootMismatch datatype
in GHC.Tc.Errors.Types, with its panoply of subtypes.
For the sake of readability, several local functions inside the
'checkBootTyCon' function were split off into top-level functions.
We split off GHC.Types.HscSource into a "boot or sig" vs "normal hs file"
datatype, as this mirrors the logic in several other places where we
want to treat hs-boot and hsig files in a similar fashion.
This commit also refactors the Backpack checks for type synonyms
implementing abstract data, to correctly reject implementations that
contain qualified or quantified types (this fixes #23342 and #23344).
- - - - -
d986c98e by Rodrigo Mesquita at 2023-05-16T00:14:04-04:00
configure: Drop unused AC_PROG_CPP
In configure, we were calling `AC_PROG_CPP` but never making use of the
$CPP variable it sets or reads.
The issue is $CPP will show up in the --help output of configure,
falsely advertising a configuration option that does nothing.
The reason we don't use the $CPP variable is because HS_CPP_CMD is
expected to be a single command (without flags), but AC_PROG_CPP, when
CPP is unset, will set said variable to something like `/usr/bin/gcc -E`.
Instead, we configure HS_CPP_CMD through $CC.
- - - - -
a8f0435f by Cheng Shao at 2023-05-16T00:14:42-04:00
rts: fix --disable-large-address-space
This patch moves
ACQUIRE_ALLOC_BLOCK_SPIN_LOCK/RELEASE_ALLOC_BLOCK_SPIN_LOCK from
Storage.h to HeapAlloc.h. When --disable-large-address-space is passed
to configure, the code in HeapAlloc.h makes use of these two macros.
Fixes #23385.
- - - - -
bdb93cd2 by Oleg Grenrus at 2023-05-16T07:59:21+03:00
Add -Wmissing-role-annotations
Implements #22702
- - - - -
41ecfc34 by Ben Gamari at 2023-05-16T07:28:15-04:00
base: Export {get,set}ExceptionFinalizer from System.Mem.Weak
As proposed in CLC Proposal #126 [1].
[1]: https://github.com/haskell/core-libraries-committee/issues/126
- - - - -
67330303 by Ben Gamari at 2023-05-16T07:28:16-04:00
base: Introduce printToHandleFinalizerExceptionHandler
- - - - -
5e3f9bb5 by Josh Meredith at 2023-05-16T13:59:22-04:00
JS: Implement h$clock_gettime in the JavaScript RTS (#23360)
- - - - -
90e69d5d by Zubin Duggal at 2023-05-16T14:00:00-04:00
compiler: Use compact representation for SourceText
SourceText is serialized along with INLINE pragmas into interface files. Many of
these SourceTexts are identical, for example "{-# INLINE#". When deserialized,
each such SourceText was previously expanded out into a [Char], which is highly
wasteful of memory, and each such instance of the text would allocate an
independent list with its contents as deserializing breaks any sharing that might
have existed.
Instead, we use a `FastString` to represent these, so that each instance unique
text will be interned and stored in a memory efficient manner.
- - - - -
b70bc690 by Zubin Duggal at 2023-05-16T14:00:00-04:00
compiler: Use compact representation/FastStrings for `SourceNote`s
`SourceNote`s should not be stored as [Char] as this is highly wasteful
and in certain scenarios can be highly duplicated.
Metric Decrease:
hard_hole_fits
- - - - -
6231a126 by Zubin Duggal at 2023-05-16T14:00:00-04:00
compiler: Use compact representation for UsageFile (#22744)
Use FastString to store filepaths in interface files, as this data is
highly redundant so we want to share all instances of filepaths in the
compiler session.
- - - - -
47a58150 by Zubin Duggal at 2023-05-16T14:00:00-04:00
testsuite: add test for T22744
This test checks for #22744 by compiling 100 modules which each have
a dependency on 1000 distinct external files.
Previously, when loading these interfaces from disk, each individual instance
of a filepath in the interface will would be allocated as an individual object
on the heap, meaning we have heap objects for 100*1000 files, when there are
only 1000 distinct files we care about.
This test checks this by first compiling the module normally, then measuring
the peak memory usage in a no-op recompile, as the recompilation checking will
force the allocation of all these filepaths.
- - - - -
0451bdc9 by Ben Gamari at 2023-05-16T21:31:40-04:00
users guide: Add glossary
Currently this merely explains the meaning of "technology preview" in
the context of released features.
- - - - -
0ba52e4e by Ben Gamari at 2023-05-16T21:31:40-04:00
Update glossary.rst
- - - - -
3d23060c by Ben Gamari at 2023-05-16T21:31:40-04:00
Use glossary directive
- - - - -
2972fd66 by Sylvain Henry at 2023-05-16T21:32:20-04:00
JS: fix getpid (fix #23399)
- - - - -
5fe1d3e6 by Matthew Pickering at 2023-05-17T21:42:00-04:00
Use setSrcSpan rather than setLclEnv in solveForAll
In subsequent MRs (#23409) we want to remove the TcLclEnv argument from
a CtLoc. This MR prepares us for that by removing the one place where
the entire TcLclEnv is used, by using it more precisely to just set the
contexts source location.
Fixes #23390
- - - - -
385edb65 by Torsten Schmits at 2023-05-17T21:42:40-04:00
Update the users guide paragraph on -O in GHCi
In relation to #23056
- - - - -
87626ef0 by Ben Gamari at 2023-05-18T15:18:53-04:00
base: Add test for #13660
- - - - -
9eef53b1 by Ben Gamari at 2023-05-18T15:18:53-04:00
base: Move implementation of GHC.Foreign to GHC.Internal
- - - - -
174ea2fa by Ben Gamari at 2023-05-18T15:18:53-04:00
base: Introduce {new,with}CStringLen0
These are useful helpers for implementing the internal-NUL code unit
check needed to fix #13660.
- - - - -
a46ced16 by Ben Gamari at 2023-05-18T15:18:53-04:00
base: Clean up documentation
- - - - -
b98d99cc by Ben Gamari at 2023-05-18T15:18:53-04:00
base: Ensure that FilePaths don't contain NULs
POSIX filepaths may not contain the NUL octet but previously we did not
reject such paths. This could be exploited by untrusted input to cause
discrepancies between various `FilePath` queries and the opened
filename. For instance, `readFile "hello.so\x00.txt"` would open the
file `"hello.so"` yet `takeFileExtension` would return `".txt"`.
The same argument applies to Windows FilePaths
Fixes #13660.
- - - - -
7ae45459 by Simon Peyton Jones at 2023-05-18T15:19:29-04:00
Allow the demand analyser to unpack tuple and equality dictionaries
Addresses #23398. The demand analyser usually does not unpack class
dictionaries: see Note [Do not unbox class dictionaries] in
GHC.Core.Opt.DmdAnal.
This patch makes an exception for tuple dictionaries and equality
dictionaries, for reasons explained in wrinkles (DNB1) and (DNB2) of
the above Note.
Compile times fall by 0.1% for some reason (max 0.7% on T18698b).
- - - - -
b53a9086 by Greg Steuck at 2023-05-18T15:20:08-04:00
Use a simpler and more portable construct in ld.ldd check
printf '%q\n' is a bash extension which led to incorrectly
failing an ld.lld test on OpenBSD which uses pdksh as /bin/sh
- - - - -
dd5710af by Torsten Schmits at 2023-05-18T15:20:50-04:00
Update the warning about interpreter optimizations
to reflect that they're not incompatible anymore, but guarded by a flag
- - - - -
4f6dd999 by Matthew Pickering at 2023-05-18T15:21:26-04:00
Remove stray dump flags in GHC.Rename.Names
- - - - -
4bca0486 by Oleg Grenrus at 2023-05-19T11:51:33+03:00
Make Warn = Located DriverMessage
This change makes command line argument parsing use diagnostic
framework for producing warnings.
- - - - -
525ed554 by Simon Peyton Jones at 2023-05-19T10:09:15-04:00
Type inference for data family newtype instances
This patch addresses #23408, a tricky case with data family
newtype instances. Consider
type family TF a where TF Char = Bool
data family DF a
newtype instance DF Bool = MkDF Int
and [W] Int ~R# DF (TF a), with a Given (a ~# Char). We must fully
rewrite the Wanted so the tpye family can fire; that wasn't happening.
- - - - -
c6fb6690 by Peter Trommler at 2023-05-20T03:16:08-04:00
testsuite: fix predicate on rdynamic test
Test rdynamic requires dynamic linking support, which is
orthogonal to RTS linker support. Change the predicate accordingly.
Fixes #23316
- - - - -
735d504e by Matthew Pickering at 2023-05-20T03:16:44-04:00
docs: Use ghc-ticket directive where appropiate in users guide
Using the directive automatically formats and links the ticket
appropiately.
- - - - -
b56d7379 by Sylvain Henry at 2023-05-22T14:21:22-04:00
NCG: remove useless .align directive (#20758)
- - - - -
15b93d2f by Simon Peyton Jones at 2023-05-22T14:21:58-04:00
Add test for #23156
This program had exponential typechecking time in GHC 9.4 and 9.6
- - - - -
2b53f206 by Greg Steuck at 2023-05-22T20:23:11-04:00
Revert "Change hostSupportsRPaths to report False on OpenBSD"
This reverts commit 1e0d8fdb55a38ece34fa6cf214e1d2d46f5f5bf2.
- - - - -
882e43b7 by Greg Steuck at 2023-05-22T20:23:11-04:00
Disable T17414 on OpenBSD
Like on other systems it's not guaranteed that there's sufficient
space in /tmp to write 2G out.
- - - - -
9d531f9a by Greg Steuck at 2023-05-22T20:23:11-04:00
Bring back getExecutablePath to getBaseDir on OpenBSD
Fix #18173
- - - - -
9db0eadd by Krzysztof Gogolewski at 2023-05-22T20:23:47-04:00
Add an error origin for impedance matching (#23427)
- - - - -
33cf4659 by Ben Gamari at 2023-05-23T03:46:20-04:00
testsuite: Add tests for #23146
Both lifted and unlifted variants.
- - - - -
76727617 by Ben Gamari at 2023-05-23T03:46:21-04:00
codeGen: Fix some Haddocks
- - - - -
33a8c348 by Ben Gamari at 2023-05-23T03:46:21-04:00
codeGen: Give proper LFInfo to datacon wrappers
As noted in `Note [Conveying CAF-info and LFInfo between modules]`,
when importing a binding from another module we must ensure that it gets
the appropriate `LambdaFormInfo` if it is in WHNF to ensure that
references to it are tagged correctly.
However, the implementation responsible for doing this,
`GHC.StgToCmm.Closure.mkLFImported`, only dealt with datacon workers and
not wrappers. This lead to the crash of this program in #23146:
module B where
type NP :: [UnliftedType] -> UnliftedType
data NP xs where
UNil :: NP '[]
module A where
import B
fieldsSam :: NP xs -> NP xs -> Bool
fieldsSam UNil UNil = True
x = fieldsSam UNil UNil
Due to its GADT nature, `UNil` produces a trivial wrapper
$WUNil :: NP '[]
$WUNil = UNil @'[] @~(<co:1>)
which is referenced in the RHS of `A.x`. Due to the above-mentioned bug
in `mkLFImported`, the references to `$WUNil` passed to `fieldsSam` were
not tagged. This is problematic as `fieldsSam` expected its arguments to
be tagged as they are unlifted.
The fix is straightforward: extend the logic in `mkLFImported` to cover
(nullary) datacon wrappers as well as workers. This is safe because we
know that the wrapper of a nullary datacon will be in WHNF, even if it
includes equalities evidence (since such equalities are not runtime
relevant).
Thanks to @MangoIV for the great ticket and @alt-romes for his
minimization and help debugging.
Fixes #23146.
- - - - -
2fc18e9e by Rodrigo Mesquita at 2023-05-23T03:46:21-04:00
codeGen: Fix LFInfo of imported datacon wrappers
As noted in #23231 and in the previous commit, we were failing to give a
an LFInfo of LFCon to a nullary datacon wrapper from another module,
failing to properly tag pointers which ultimately led to the
segmentation fault in #23146.
On top of the previous commit which now considers wrappers where we
previously only considered workers, we change the order of the guards so
that we check for the arity of the binding before we check whether it is
a constructor. This allows us to
(1) Correctly assign `LFReEntrant` to imported wrappers whose worker was
nullary, which we previously would fail to do
(2) Remove the `isNullaryRepDataCon` predicate:
(a) which was previously wrong, since it considered wrappers whose
workers had zero-width arguments to be non-nullary and would fail to
give `LFCon` to them
(b) is now unnecessary, since arity == 0 guarantees
- that the worker takes no arguments at all
- and the wrapper takes no arguments and its RHS must be an
application of the worker to zero-width-args only.
- we lint these two items with an assertion that the datacon
`hasNoNonZeroWidthArgs`
We also update `isTagged` to use the new logic in determining the
LFInfos of imported Ids.
The creation of LFInfos for imported Ids and this detail are explained
in Note [The LFInfo of Imported Ids].
Note that before the patch to those issues we would already consider these
nullary wrappers to have `LFCon` lambda form info; but failed to re-construct
that information in `mkLFImported`
Closes #23231, #23146
(I've additionally batched some fixes to documentation I found while
investigating this issue)
- - - - -
0598f7f0 by Rodrigo Mesquita at 2023-05-23T03:46:21-04:00
Make LFInfos for DataCons on construction
As a result of the discussion in !10165, we decided to amend the
previous commit which fixed the logic of `mkLFImported` with regard to
datacon workers and wrappers.
Instead of having the logic for the LFInfo of datacons be in
`mkLFImported`, we now construct an LFInfo for all data constructors on
GHC.Types.Id.Make and store it in the `lfInfo` field.
See the new Note [LFInfo of DataCon workers and wrappers] and
ammendments to Note [The LFInfo of Imported Ids]
- - - - -
12294b22 by Rodrigo Mesquita at 2023-05-23T03:46:21-04:00
Update Note [Core letrec invariant]
Authored by @simonpj
- - - - -
e93ab972 by Rodrigo Mesquita at 2023-05-23T03:46:21-04:00
Rename mkLFImported to importedIdLFInfo
The `mkLFImported` sounded too much like a constructor of sorts, when
really it got the `LFInfo` of an imported Id from its `lf_info` field
when this existed, and otherwise returned a conservative estimate of
that imported Id's LFInfo. This in contrast to functions such as
`mkLFReEntrant` which really are about constructing an `LFInfo`.
- - - - -
e54d9259 by Rodrigo Mesquita at 2023-05-23T03:46:21-04:00
Enforce invariant on typePrimRepArgs in the types
As part of the documentation effort in !10165 I came across this
invariant on 'typePrimRepArgs' which is easily expressed at the
type-level through a NonEmpty list.
It allowed us to remove one panic.
- - - - -
b8fe6a0c by Rodrigo Mesquita at 2023-05-23T03:46:21-04:00
Merge outdated Note [Data con representation] into Note [Data constructor representation]
Introduce new Note [Constructor applications in STG] to better support
the merge, and reference it from the relevant bits in the STG syntax.
- - - - -
e1590ddc by Simon Peyton Jones at 2023-05-23T03:46:57-04:00
Add the SolverStage monad
This refactoring makes a substantial improvement in the
structure of the type-checker's constraint solver: #23070.
Specifically:
* Introduced the SolverStage monad. See GHC.Tc.Solver.Monad
Note [The SolverStage monad]
* Make each solver pipeline (equalities, dictionaries, irreds etc)
deal with updating the inert set, as a separate SolverStage. There
is sometimes special stuff to do, and it means that each full
pipeline can have type SolverStage Void, indicating that they never
return anything.
* Made GHC.Tc.Solver.Equality.zonkEqTypes into a SolverStage. Much nicer.
* Combined the remnants of GHC.Tc.Solver.Canonical and
GHC.Tc.Solver.Interact into a new module GHC.Tc.Solver.Solve.
(Interact and Canonical are removed.)
* Gave the same treatment to dictionary and irred constraints
as I have already done for equality constraints:
* New types (akin to EqCt): IrredCt and DictCt
* Ct is now just a simple sum type
data Ct
= CDictCan DictCt
| CIrredCan IrredCt
| CEqCan EqCt
| CQuantCan QCInst
| CNonCanonical CtEvidence
* inert_dicts can now have the better type DictMap DictCt, instead of
DictMap Ct; and similarly inert_irreds.
* Significantly simplified the treatment of implicit parameters.
Previously we had a number of special cases
* interactGivenIP, an entire function
* special case in maybeKickOut
* special case in findDict, when looking up dictionaries
But actually it's simpler than that. When adding a new Given, implicit
parameter constraint to the InertSet, we just need to kick out any
existing inert constraints that mention that implicit parameter.
The main work is done in GHC.Tc.Solver.InertSet.delIPDict, along with
its auxiliary GHC.Core.Predicate.mentionsIP.
See Note [Shadowing of implicit parameters] in GHC.Tc.Solver.Dict.
* Add a new fast-path in GHC.Tc.Errors.Hole.tcCheckHoleFit.
See Note [Fast path for tcCheckHoleFit]. This is a big win in some cases:
test hard_hole_fits gets nearly 40% faster (at compile time).
* Add a new fast-path for solving /boxed/ equality constraints
(t1 ~ t2). See Note [Solving equality classes] in GHC.Tc.Solver.Dict.
This makes a big difference too: test T17836 compiles 40% faster.
* Implement the PermissivePlan of #23413, which concerns what happens with
insoluble Givens. Our previous treatment was wildly inconsistent as that
ticket pointed out.
A part of this, I simplified GHC.Tc.Validity.checkAmbiguity: now we simply
don't run the ambiguity check at all if -XAllowAmbiguousTypes is on.
Smaller points:
* In `GHC.Tc.Errors.misMatchOrCND` instead of having a special case for
insoluble /occurs/ checks, broaden in to all insouluble constraints.
Just generally better. See Note [Insoluble mis-match] in that module.
As noted above, compile time perf gets better. Here are the changes
over 0.5% on Fedora. (The figures are slightly larger on Windows for
some reason.)
Metrics: compile_time/bytes allocated
-------------------------------------
LargeRecord(normal) -0.9%
MultiLayerModulesTH_OneShot(normal) +0.5%
T11822(normal) -0.6%
T12227(normal) -1.8% GOOD
T12545(normal) -0.5%
T13035(normal) -0.6%
T15703(normal) -1.4% GOOD
T16875(normal) -0.5%
T17836(normal) -40.7% GOOD
T17836b(normal) -12.3% GOOD
T17977b(normal) -0.5%
T5837(normal) -1.1%
T8095(normal) -2.7% GOOD
T9020(optasm) -1.1%
hard_hole_fits(normal) -37.0% GOOD
geo. mean -1.3%
minimum -40.7%
maximum +0.5%
Metric Decrease:
T12227
T15703
T17836
T17836b
T8095
hard_hole_fits
LargeRecord
T9198
T13035
- - - - -
6abf3648 by Simon Peyton Jones at 2023-05-23T03:46:57-04:00
Avoid an assertion failure in abstractFloats
The function GHC.Core.Opt.Simplify.Utils.abstractFloats
was carelessly calling lookupIdSubst_maybe on a CoVar;
but a precondition of the latter is being given an Id.
In fact it's harmless to call it on a CoVar, but still, the
precondition on lookupIdSubst_maybe makes sense, so I added
a test for CoVars.
This avoids a crash in a DEBUG compiler, but otherwise has
no effect. Fixes #23426.
- - - - -
838aaf4b by hainq at 2023-05-24T12:41:19-04:00
Migrate errors in GHC.Tc.Validity
This patch migrates the error messages in GHC.Tc.Validity to use
the new diagnostic infrastructure.
It adds the constructors:
- TcRnSimplifiableConstraint
- TcRnArityMismatch
- TcRnIllegalInstanceDecl, with sub-datatypes for HasField errors
and fundep coverage condition errors.
- - - - -
8539764b by Krzysztof Gogolewski at 2023-05-24T12:41:56-04:00
linear lint: Add missing processing of DEFAULT
In this correct program
f :: a %1 -> a
f x = case x of x { _DEFAULT -> x }
after checking the alternative we weren't popping the case binder 'x'
from the usage environment, which meant that the lambda-bound 'x'
was counted twice: in the scrutinee and (incorrectly) in the alternative.
In fact, we weren't checking the usage of 'x' at all.
Now the code for handling _DEFAULT is similar to the one handling
data constructors.
Fixes #23025.
- - - - -
ae683454 by Matthew Pickering at 2023-05-24T12:42:32-04:00
Remove outdated "Don't check hs-boot type family instances too early" note
This note was introduced in 25b70a29f623 which delayed performing some
consistency checks for type families. However, the change was reverted
later in 6998772043a7f0b0360116eb5ffcbaa5630b21fb but the note was not
removed.
I found it confusing when reading to code to try and work out what
special behaviour there was for hs-boot files (when in-fact there isn't
any).
- - - - -
44af57de by Matthew Pickering at 2023-05-24T12:43:08-04:00
rts: Define ticky macro stubs
These macros have long been undefined which has meant we were missing
reporting these allocations in ticky profiles.
The most critical missing definition was TICK_ALLOC_HEAP_NOCTR which was
missing all the RTS calls to allocate, this leads to a the overall
ALLOC_RTS_tot number to be severaly underreported.
Of particular interest though is the ALLOC_STACK_ctr and ALLOC_STACK_tot
counters which are useful to tracking stack allocations.
Fixes #23421
- - - - -
b2dabe3a by Matthew Pickering at 2023-05-24T12:43:08-04:00
rts: ticky: Rename TICK_ALLOC_HEAP_NOCTR to TICK_ALLOC_RTS
This macro increments the ALLOC_HEAP_tot and ALLOC_HEAP_ctr so it makes
more sense to name it after that rather than the suffix NOCTR, whose
meaning has been lost to the mists of time.
- - - - -
eac4420a by Ben Gamari at 2023-05-24T12:43:45-04:00
users guide: A few small mark-up fixes
- - - - -
a320ca76 by Rodrigo Mesquita at 2023-05-24T12:44:20-04:00
configure: Fix support check for response files.
In failing to escape the '-o' in '-o\nconftest\nconftest.o\n' argument
to printf, the writing of the arguments response file always failed.
The fix is to pass the arguments after `--` so that they are treated
positional arguments rather than flags to printf.
Closes #23435
- - - - -
f21ce0e4 by mangoiv at 2023-05-24T12:45:00-04:00
[feat] add .direnv to the .gitignore file
- - - - -
36d5944d by Bodigrim at 2023-05-24T20:58:34-04:00
Add Data.List.unsnoc
See https://github.com/haskell/core-libraries-committee/issues/165 for discussion
- - - - -
c0f2f9e3 by Bartłomiej Cieślar at 2023-05-24T20:59:14-04:00
Fix crash in backpack signature merging with -ddump-rn-trace
In some cases, backpack signature merging could crash in addUsedGRE
when -ddump-rn-trace was enabled, as pretty-printing the GREInfo would cause
unavailable interfaces to be loaded.
This commit fixes that issue by not pretty-printing the GREInfo in addUsedGRE
when -ddump-rn-trace is enabled.
Fixes #23424
Signed-off-by: Bartłomiej Cieślar <bcieslar2001 at gmail.com>
- - - - -
5a07d94a by Krzysztof Gogolewski at 2023-05-25T03:30:20-04:00
Add a regression test for #13981
The panic was fixed by 6998772043a7f0b. Fixes #13981.
- - - - -
182df90e by Krzysztof Gogolewski at 2023-05-25T03:30:57-04:00
Add a test for #23355
It was fixed by !10061, so I'm adding it in the same group.
- - - - -
1b31b039 by uhbif19 at 2023-05-25T12:08:28+02:00
Migrate errors in GHC.Rename.Splice GHC.Rename.Pat
This commit migrates the errors in GHC.Rename.Splice and GHC.Rename.Pat
to use the new diagnostic infrastructure.
- - - - -
56abe494 by sheaf at 2023-05-25T12:09:55+02:00
Common up Template Haskell errors in TcRnMessage
This commit commons up the various Template Haskell errors into a
single constructor, TcRnTHError, of TcRnMessage.
- - - - -
a487ba9e by Krzysztof Gogolewski at 2023-05-25T14:35:56-04:00
Enable ghci tests for unboxed tuples
The tests were originally skipped because ghci used not to support
unboxed tuples/sums.
- - - - -
dc3422d4 by Matthew Pickering at 2023-05-25T18:57:19-04:00
rts: Build ticky GHC with single-threaded RTS
The threaded RTS allows you to use ticky profiling but only for the
counters in the generated code. The counters used in the C portion of
the RTS are disabled. Updating the counters is also racy using the
threaded RTS which can lead to misleading or incorrect ticky results.
Therefore we change the hadrian flavour to build using the
single-threaded RTS (mainly in order to get accurate C code counter
increments)
Fixes #23430
- - - - -
fbc8e04e by sheaf at 2023-05-25T18:58:00-04:00
Propagate long-distance info in generated code
When desugaring generated pattern matches, we skip pattern match checks.
However, this ended up also discarding long-distance information, which
might be needed for user-written sub-expressions.
Example:
```haskell
okay (GADT di) cd =
let sr_field :: ()
sr_field = case getFooBar di of { Foo -> () }
in case cd of { SomeRec _ -> SomeRec sr_field }
```
With sr_field a generated FunBind, we still want to propagate the outer
long-distance information from the GADT pattern match into the checks
for the user-written RHS of sr_field.
Fixes #23445
- - - - -
f8ced241 by Matthew Pickering at 2023-05-26T15:26:21-04:00
Introduce GHCiMessage to wrap GhcMessage
By introducing a wrapped message type we can control how certain
messages are printed in GHCi (to add extra information for example)
- - - - -
58e554c1 by Matthew Pickering at 2023-05-26T15:26:22-04:00
Generalise UnknownDiagnostic to allow embedded diagnostics to access
parent diagnostic options.
* Split default diagnostic options from Diagnostic class into
HasDefaultDiagnosticOpts class.
* Generalise UnknownDiagnostic to allow embedded diagnostics to access
options.
The principle idea here is that when wrapping an error message (such as
GHCMessage to make GHCiMessage) then we need to also be able to lift the
configuration when overriding how messages are printed (see load' for an
example).
- - - - -
b112546a by Matthew Pickering at 2023-05-26T15:26:22-04:00
Allow API users to wrap error messages created during 'load'
This allows API users to configure how messages are rendered when they
are emitted from the load function. For an example see how
'loadWithCache' is used in GHCi.
- - - - -
2e4cf0ee by Matthew Pickering at 2023-05-26T15:26:22-04:00
Abstract cantFindError and turn Opt_BuildingCabal into a print-time option
* cantFindError is abstracted so that the parts which mention specific
things about ghc/ghci are parameters. The intention being that
GHC/GHCi can specify the right values to put here but otherwise
display the same error message.
* The BuildingCabalPackage argument from GenericMissing is removed and
turned into a print-time option. The reason for the error is not
dependent on whether `-fbuilding-cabal-package` is passed, so we don't
want to store that in the error message.
- - - - -
34b44f7d by Matthew Pickering at 2023-05-26T15:26:22-04:00
error messages: Don't display ghci specific hints for missing packages
Tickets like #22884 suggest that it is confusing that GHC used on the
command line can suggest options which only work in GHCi.
This ticket uses the error message infrastructure to override certain
error messages which displayed GHCi specific information so that this
information is only showed when using GHCi.
The main annoyance is that we mostly want to display errors in the same
way as before, but with some additional information. This means that the
error rendering code has to be exported from the Iface/Errors/Ppr.hs
module.
I am unsure about whether the approach taken here is the best or most
maintainable solution.
Fixes #22884
- - - - -
05a1b626 by Matthew Pickering at 2023-05-26T15:26:58-04:00
ghcup-metadata: Don't override existing metadata if version already exists.
If a nightly pipeline runs twice for some reason for the same version
then we really don't want to override an existing entry with new
bindists. This could cause ABI compatability issues for users or break
ghcup's caching logic.
- - - - -
fcbcb3cc by Matthew Pickering at 2023-05-26T15:26:58-04:00
ghcup-metadata: Use proper API url for bindist download
Previously we were using links from the web interface, but it's more
robust and future-proof to use the documented links to the artifacts.
https://docs.gitlab.com/ee/api/job_artifacts.html
- - - - -
5b59c8fe by Matthew Pickering at 2023-05-26T15:26:58-04:00
ghcup-metadata: Set Nightly and LatestNightly tags
The latest nightly release needs the LatestNightly tag, and all other
nightly releases need the Nightly tag. Therefore when the metadata is
updated we need to replace all LatestNightly with Nightly.`
- - - - -
914e1468 by Matthew Pickering at 2023-05-26T15:26:58-04:00
ghcup-metadata: Download nightly metadata for correct date
The metadata now lives in https://gitlab.haskell.org/ghc/ghcup-metadata
with one metadata file per year. When we update the metadata we download
and update the right file for the current year.
- - - - -
16cf7d2e by Matthew Pickering at 2023-05-26T15:26:58-04:00
ghcup-metadata: Download metadata and update for correct year
something about pipeline date
- - - - -
14792c4b by Matthew Pickering at 2023-05-26T15:26:58-04:00
ghcup-metadata: Don't skip CI
On a push we now have a CI job which updates gitlab pages with the
metadata files.
- - - - -
1121bdd8 by Matthew Pickering at 2023-05-26T15:26:59-04:00
ghcup-metadata: Add --date flag to specify the release date
The ghcup-metadata now has a viReleaseDay field which needs to be
populated with the day of the release.
- - - - -
bc478bee by Matthew Pickering at 2023-05-26T15:26:59-04:00
ghcup-metadata: Add dlOutput field
ghcup now requires us to add this field which specifies where it should
download the bindist to. See
https://gitlab.haskell.org/ghc/ghcup-metadata/-/issues/1 for some more
discussion.
- - - - -
2bdbd9da by Josh Meredith at 2023-05-26T15:27:35-04:00
JS: Convert rendering to use HLine instead of SDoc (#22455)
- - - - -
abd9e37c by Norman Ramsey at 2023-05-26T15:28:12-04:00
testsuite: add WasmControlFlow test
This patch adds the WasmControlFlow test to test the wasm backend's
relooper component.
- - - - -
07f858eb by Sylvain Henry at 2023-05-26T15:28:53-04:00
Factorize getLinkDeps
Prepare reuse of getLinkDeps for TH implementation in the JS backend
(cf #22261 and review of !9779).
- - - - -
fad9d092 by Oleg Grenrus at 2023-05-27T13:38:08-04:00
Change GHC.Driver.Session import to .DynFlags
Also move targetPlatform selector
Plenty of GHC needs just DynFlags.
Even more can be made to use .DynFlags if more selectors is migrated.
This is a low hanging fruit.
- - - - -
69fdbece by Alan Zimmerman at 2023-05-27T13:38:45-04:00
EPA: Better fix for #22919
The original fix for #22919 simply removed the ability to match up
prior comments with the first declaration in the file.
Restore it, but add a check that the comment is on a single line, by
ensuring that it comes immediately prior to the next thing (comment or
start of declaration), and that the token preceding it is not on the
same line.
closes #22919
- - - - -
0350b186 by Josh Meredith at 2023-05-29T12:46:27+00:00
Remove JavaScriptFFI from --supported-extensions for non-JS targets (#11214)
- - - - -
b4816919 by Matthew Pickering at 2023-05-30T17:07:43-04:00
testsuite: Pass -kb16k -kc128k for performance tests
Setting a larger stack chunk size gives a greater protection
from stack thrashing (where the repeated overflow/underflow allocates a lot of
stack chunks which sigificantly impact allocations). This
stabilises some tests against differences cause by more things being
pushed onto the stack.
The performance tests are generally testing work done by the compiler,
using allocation as a proxy, so removing/stabilising the allocations due
to the stack gives us more stable tests which are also more sensitive
to actual changes in compiler performance.
The tests which increase are ones where we compile a lot of modules, and
for each module we spawn a thread to compile the module in. Therefore
increasing these numbers has a multiplying effect on these tests because
there are many more stacks which we can increase in size.
The most significant improvements though are cases such as T8095 which
reduce significantly in allocations (30%). This isn't a performance
improvement really but just helps stabilise the test against this
threshold set by the defaults.
Fixes #23439
-------------------------
Metric Decrease:
InstanceMatching
T14683
T8095
T9872b_defer
T9872d
T9961
hie002
T19695
T3064
Metric Increase:
MultiLayerModules
T13701
T14697
-------------------------
- - - - -
6629f1c5 by Ben Gamari at 2023-05-30T17:08:20-04:00
Move via-C flags into GHC
These were previously hardcoded in configure (with no option for
overriding them) and simply passed onto ghc through the settings file.
Since configure already guarantees gcc supports those flags, we simply
move them into GHC.
- - - - -
981e5e11 by Arnaud Spiwack at 2023-05-31T08:34:33-04:00
Allow CPR on unrestricted constructors
Per the new `Note [Linting linearity]`, we want optimisations over
trying to preserve linearity. This will allow CPR to handle `Ur`, in
particular.
- - - - -
bf9344d2 by Arnaud Spiwack at 2023-05-31T08:34:33-04:00
Push coercions across multiplicity boundaries
Per the new `Note [Linting linearity]`, we want optimisations over
trying to preserve linearity. This will avoid preventing inlinings and
reductions and make linear programs more efficient.
- - - - -
d56dd695 by sheaf at 2023-05-31T11:37:12-04:00
Data.Bag: add INLINEABLE to polymorphic functions
This commit allows polymorphic methods in GHC.Data.Bag to be
specialised, avoiding having to pass explicit dictionaries when they
are instantiated with e.g. a known monad.
- - - - -
5366cd35 by sheaf at 2023-05-31T11:37:12-04:00
Split off TcBinderStack into its own module
This commit splits off TcBinderStack into its own module,
to avoid module cycles: we might want to refer to it without also
pulling in the TcM monad.
- - - - -
09d4d307 by sheaf at 2023-05-31T11:37:12-04:00
Split off TcRef into its own module
This helps avoid pull in the full TcM monad when we just want access
to mutable references in the typechecker. This facilitates later patches
which introduce a slimmed down TcM monad for zonking.
- - - - -
88cc19b3 by sheaf at 2023-05-31T11:37:12-04:00
Introduce Codensity monad
The Codensity monad is useful to write state-passing computations in
continuation-passing style, e.g. to implement a State monad as
continuation-passing style over a Reader monad.
- - - - -
f62d8195 by sheaf at 2023-05-31T11:37:12-04:00
Restructure the zonker
This commit splits up the zonker into a few separate components, described
in Note [The structure of the zonker] in `GHC.Tc.Zonk.Type`.
1. `GHC.Tc.Zonk.Monad` introduces a pared-down `TcM` monad, `ZonkM`,
which has enough information for zonking types.
This allows us to refactor `ErrCtxt` to use `ZonkM` instead of `TcM`,
which guarantees we don't throw an error while reporting an error.
2. `GHC.Tc.Zonk.Env` is the new home of `ZonkEnv`, and also defines
two zonking monad transformers, `ZonkT` and `ZonkBndrT`.
`ZonkT` is a reader monad transformer over `ZonkEnv`.
`ZonkBndrT m` is the codensity monad over `ZonkT m`.
`ZonkBndrT` is used for computations that accumulate binders
in the `ZonkEnv`.
3. `GHC.Tc.Zonk.TcType` contains the code for zonking types, for use
in the typechecker. It uses the `ZonkM` monad.
4. `GHC.Tc.Zonk.Type` contains the code for final zonking to `Type`,
which has been refactored to use `ZonkTcM = ZonkT TcM` and
`ZonkBndrTcM = ZonkBndrT TcM`.
Allocations slightly decrease on the whole due to using
continuation-passing style instead of manual state passing of ZonkEnv
in the final zonking to Type.
-------------------------
Metric Decrease:
T4029
T8095
T14766
T15304
hard_hole_fits
RecordUpdPerf
Metric Increase:
T10421
-------------------------
- - - - -
70526f5b by mimi.vx at 2023-05-31T11:37:53-04:00
Update rdt-theme to latest upstream version
Fixes https://gitlab.haskell.org/ghc/ghc/-/issues/23444
- - - - -
f3556d6c by Finley McIlwaine at 2023-06-01T04:53:12-04:00
Restructure IPE buffer layout
Reference ticket #21766
This commit restructures IPE buffer list entries to not contain
references to their corresponding info tables. IPE buffer list nodes now
point to two lists of equal length, one holding the list of info table
pointers and one holding the corresponding entries for each info table.
This will allow the entry data to be compressed without losing the
references to the info tables.
- - - - -
5d1f2411 by Finley McIlwaine at 2023-06-01T04:53:12-04:00
Add IPE compression to configure
Reference ticket #21766
Adds an `--enable-ipe-data-compreesion` flag to the configure script
which will check for libzstd and set the appropriate flags to allow for
IPE data compression in the compiler
- - - - -
b7a640ac by Finley McIlwaine at 2023-06-01T04:53:12-04:00
IPE data compression
Reference ticket #21766
When IPE data compression is enabled, compress the emitted IPE buffer
entries and decompress them in the RTS.
- - - - -
5aef5658 by Finley McIlwaine at 2023-06-01T04:53:12-04:00
Fix libzstd detection in configure and RTS
Ensure that `HAVE_LIBZSTD` gets defined to either 0 or 1 in all cases
and properly check that before IPE data decompression in the RTS. See
ticket #21766.
- - - - -
69563c97 by Finley McIlwaine at 2023-06-01T04:53:12-04:00
Add note describing IPE data compression
See ticket #21766
- - - - -
7872e2b6 by Finley McIlwaine at 2023-06-01T04:53:12-04:00
Fix byte order of IPE data, fix IPE tests
Make sure byte order of written IPE buffer entries matches target.
Make sure the IPE-related tests properly access the fields of IPE buffer
entry nodes with the new IPE layout.
This commit also introduces checks to avoid importing modules if IPE
compression is not enabled.
See ticket #21766.
- - - - -
0e85099b by Finley McIlwaine at 2023-06-01T04:53:12-04:00
Fix IPE data decompression buffer allocation
Capacity of buffers allocated for decompressed IPE data was
incorrect due to a misuse of the `ZSTD_findFrameCompressedSize`
function. Fix by always storing decompressed size of IPE data in IPE
buffer list nodes and using `ZSTD_findFrameCompressedSize` to determine
the size of the compressed data.
See ticket #21766
- - - - -
a0048866 by Finley McIlwaine at 2023-06-01T04:53:12-04:00
Add optional dependencies to ./configure output
Changes the configure script to indicate whether libnuma, libzstd, or
libdw are being used as dependencies due to their optional features
being enabled.
- - - - -
09d93bd0 by Finley McIlwaine at 2023-06-01T04:53:12-04:00
Add IPE-enabled builds to CI
- Adds an IPE job to the CI pipeline which is triggered by the ~IPE label
- Introduces CI logic to enable IPE data compression
- Enables uncompressed IPE data on debug CI job
- Regenerates jobs.yaml
MR https://gitlab.haskell.org/ghc/ci-images/-/merge_requests/112 on the
images repository is meant to ensure that the proper images have
libzstd-dev installed.
- - - - -
3ded9a1c by Finley McIlwaine at 2023-06-01T04:53:12-04:00
Update user's guide and release notes, small fixes
Add mention of IPE data compression to user's guide and the release
notes for 9.8.1. Also note the impact compression has on binary size in
both places.
Change IpeBufferListNode compression check so only the value `1`
indicates compression.
See ticket #21766
- - - - -
41b41577 by Finley McIlwaine at 2023-06-01T04:53:12-04:00
Remove IPE enabled builds from CI
We don't need to explicitly specify the +ipe transformer to test IPE data
since there are tests which manually enable IPE information. This commit does
leave zstd IPE data compression enabled on the debian CI jobs.
- - - - -
982bef3a by Krzysztof Gogolewski at 2023-06-01T04:53:49-04:00
Fix build with 9.2
GHC.Tc.Zonk.Type uses an equality constraint.
ghc.nix currently provides 9.2.
- - - - -
1c96bc3d by Krzysztof Gogolewski at 2023-06-01T10:56:11-04:00
Output Lint errors to stderr instead of stdout
This is a continuation of 7b095b99, which fixed warnings but not errors.
Refs #13342
- - - - -
8e81f140 by sheaf at 2023-06-01T10:56:51-04:00
Refactor lookupExactOrOrig & friends
This refactors the panoply of renamer lookup functions relating to
lookupExactOrOrig to more graciously handle Exact and Orig names.
In particular, we avoid the situation in which we would add Exact/Orig
GREs to the tcg_used_gres field, which could cause a panic in bestImport
like in #23240.
Fixes #23428
- - - - -
5d415bfd by Krzysztof Gogolewski at 2023-06-01T10:57:31-04:00
Use the one-shot trick for UM and RewriteM functors
As described in Note [The one-shot state monad trick],
we shouldn't use derived Functor instances for monads using
one-shot. This was done for most of them, but UM and RewriteM
were missed.
- - - - -
2c38551e by Krzysztof Gogolewski at 2023-06-01T10:58:08-04:00
Fix testsuite skipping Lint
setTestOpts() is used to modify the test options for an entire .T file,
rather than a single test.
If there was a test using collect_compiler_stats, all of the tests
in the same file had lint disabled.
Fixes #21247
- - - - -
00a1e50b by Krzysztof Gogolewski at 2023-06-01T10:58:44-04:00
Add testcases for already fixed #16432
They were fixed by 40c7daed0.
Fixes #16432
- - - - -
f6e060cc by Krzysztof Gogolewski at 2023-06-02T09:07:25-04:00
cleanup: Remove unused field from SelfBoot
It is no longer needed since Note [Extra dependencies from .hs-boot files]
was deleted in 6998772043.
I've also added tildes to Note headers, otherwise they're not detected
by the linter.
- - - - -
82eacab6 by sheaf at 2023-06-02T09:08:01-04:00
Delete GHC.Tc.Utils.Zonk
This module was split up into GHC.Tc.Zonk.Type and GHC.Tc.Zonk.TcType
in commit f62d8195, but I forgot to delete the original module
- - - - -
4a4eb761 by Ben Gamari at 2023-06-02T23:53:21-04:00
base: Add build-order import of GHC.Types in GHC.IO.Handle.Types
For reasons similar to those described in Note [Depend on
GHC.Num.Integer].
Fixes #23411.
- - - - -
f53ac0ae by Sylvain Henry at 2023-06-02T23:54:01-04:00
JS: fix and enhance non-minimized code generation (#22455)
Flag -ddisable-js-minimizer was producing invalid code. Fix that and
also a few other things to generate nicer JS code for debugging.
The added test checks that we don't regress when using the flag.
- - - - -
f7744e8e by Andrey Mokhov at 2023-06-03T16:49:44-04:00
[hadrian] Fix multiline synopsis rendering
- - - - -
b2c745db by Bodigrim at 2023-06-03T16:50:23-04:00
Elaborate on performance properties of Data.List.++
- - - - -
0cc394b4 by Ryan Scott at 2023-06-05T07:26:47-04:00
Restore mingwex dependency on Windows
This partially reverts some of the changes in !9475 to make `base` and
`ghc-prim` depend on the `mingwex` library on Windows. It also restores the
RTS's stubs for `mingwex`-specific symbols such as `_lock_file`.
This is done because the C runtime provides `libmingwex` nowadays, and
moreoever, not linking against `mingwex` requires downstream users to link
against it explicitly in difficult-to-predict circumstances. Better to always
link against `mingwex` and prevent users from having to do the guesswork
themselves.
See https://gitlab.haskell.org/ghc/ghc/-/merge_requests/10360#note_495873 for
the discussion that led to this.
- - - - -
35d7aee2 by Ryan Scott at 2023-06-05T07:26:53-04:00
RtsSymbols.c: Remove mingwex symbol stubs
As of !9475, the RTS now links against `ucrt` instead of `msvcrt` on Windows,
which means that the RTS no longer needs to declare stubs for the `__mingw_*`
family of symbols. Let's remove these stubs to avoid confusion.
Fixes #23309.
- - - - -
30 changed files:
- .gitignore
- .gitlab-ci.yml
- .gitlab/gen_ci.hs
- .gitlab/jobs.yaml
- .gitlab/rel_eng/fetch-gitlab-artifacts/fetch_gitlab.py
- .gitlab/rel_eng/mk-ghcup-metadata/README.mkd
- .gitlab/rel_eng/mk-ghcup-metadata/mk_ghcup_metadata.py
- compiler/GHC.hs
- compiler/GHC/Builtin/Types/Prim.hs
- compiler/GHC/Builtin/primops.txt.pp
- compiler/GHC/ByteCode/Asm.hs
- compiler/GHC/ByteCode/Instr.hs
- compiler/GHC/Cmm/CLabel.hs
- compiler/GHC/Cmm/MachOp.hs
- compiler/GHC/Cmm/Parser.y
- compiler/GHC/CmmToAsm.hs
- compiler/GHC/CmmToAsm/AArch64/CodeGen.hs
- compiler/GHC/CmmToAsm/AArch64/Instr.hs
- compiler/GHC/CmmToAsm/AArch64/Ppr.hs
- compiler/GHC/CmmToAsm/Dwarf.hs
- compiler/GHC/CmmToAsm/PPC/CodeGen.hs
- compiler/GHC/CmmToAsm/PPC/Instr.hs
- compiler/GHC/CmmToAsm/PPC/Ppr.hs
- compiler/GHC/CmmToAsm/Wasm/FromCmm.hs
- compiler/GHC/CmmToAsm/X86/CodeGen.hs
- compiler/GHC/CmmToAsm/X86/Instr.hs
- compiler/GHC/CmmToAsm/X86/Ppr.hs
- compiler/GHC/CmmToC.hs
- compiler/GHC/CmmToLlvm.hs
- compiler/GHC/CmmToLlvm/Base.hs
The diff was not included because it is too large.
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/d2013f4dac3890aa67a8ec6b3b9b7999ca61dca8...35d7aee26a52a36f67ac411a8916bfa9883e305b
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/d2013f4dac3890aa67a8ec6b3b9b7999ca61dca8...35d7aee26a52a36f67ac411a8916bfa9883e305b
You're receiving this email because of your account on gitlab.haskell.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-commits/attachments/20230605/5a5e8fb9/attachment-0001.html>
More information about the ghc-commits
mailing list