| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
Finishes what !7467 (closed) started.
Progress towards #17957
|
|
|
|
|
| |
We have the length already, so we might as well use that rather than
O(n) recomputing it.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As proposed in
https://gitlab.haskell.org/ghc/ghc/-/merge_requests/7508#note_432877 and
https://gitlab.haskell.org/ghc/ghc/-/merge_requests/7508#note_434676,
`GHC.HsToCore.Ticks` is about ticks, breakpoints are separate and
backend-specific (only for the bytecode interpreter), and mix entry
writing is just for HPC.
With this split we separate out those interpreter- and HPC-specific
its, and keep the main `GHC.HsToCore.Ticks` agnostic.
Also, instead of passing the reversed list and count around, we use
`SizedSeq` which abstracts over the algorithm. This is much nicer to
avoid noise and prevents bugs.
(The bugs are not just hypothetical! I missed up the reverses on an
earlier draft of this commit.)
|
|
|
|
|
|
| |
The old name made it confusing why disabling HPC didn't disable the
entire pass. The name makes it clear --- there are other reasons to add
ticks in addition.
|
|
|
|
|
|
| |
No need to inline traversing a maybe for `mkModBreaks`. And better to
make each function do one thing and let the caller deside when than
scatter the decision making and make the caller seem more imperative.
|
| |
|
|
|
|
|
|
|
|
| |
This is a follow-up to !7247 (closed) making the inclusion of compact unwinding
sections the default.
Also a slight refactoring/simplification of the flag handling to add
-fno-compact-unwind.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit redefines the structure of Splices in the AST.
We get rid of `HsSplice` which used to represent typed and untyped
splices, quasi quotes, and the result of splicing either an expression,
a type or a pattern.
Instead we have `HsUntypedSplice` which models an untyped splice or a
quasi quoter, which works in practice just like untyped splices.
The `HsExpr` constructor `HsSpliceE` which used to be constructed with
an `HsSplice` is split into `HsTypedSplice` and `HsUntypedSplice`. The
former is directly constructed with an `HsExpr` and the latter now takes
an `HsUntypedSplice`.
Both `HsType` and `Pat` constructors `HsSpliceTy` and `SplicePat` now
take an `HsUntypedSplice` instead of a `HsSplice` (remember only
/untyped splices/ can be spliced as types or patterns).
The result of splicing an expression, type, or pattern is now
comfortably stored in the extension fields `XSpliceTy`, `XSplicePat`,
`XUntypedSplice` as, respectively, `HsUntypedSpliceResult (HsType
GhcRn)`, `HsUntypedSpliceResult (Pat GhcRn)`, and `HsUntypedSpliceResult
(HsExpr GhcRn)`
Overall the TTG extension points are now better used to
make invalid states unrepresentable and model the progression between
stages better.
See Note [Lifecycle of an untyped splice, and PendingRnSplice]
and Note [Lifecycle of an typed splice, and PendingTcSplice] for more
details.
Updates haddock submodule
Fixes #21263
-------------------------
Metric Decrease:
hard_hole_fits
-------------------------
|
|
|
|
|
| |
UntypedSpliceFlavour was only used in the client-specific `GHC.Hs.Expr`
but was defined in the client-independent L.H.S.Expr.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
too hard
Progress towards #17957
Because of `CoreM`, I did not move the `DynFlags` and `HscEnv` to other
modules as thoroughly as I usually do. This does mean that risk of
`DynFlags` "creeping back in" is higher than it usually is.
After we do the same process to the other Core passes, and then figure
out what we want to do about `CoreM`, we can finish the job started
here.
That is a good deal more work, however, so it certainly makes sense to
land this now.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This fixes two bugs which were adding dependencies on alex/happy when
building from a source dist.
* When we try to pass `--with-alex` and `--with-happy` to cabal when
configuring but the builders are not set. This is fixed by making them
optional.
* When we configure, cabal requires alex/happy because of the
build-tool-depends fields. These are now made optional with a cabal
flag (build-tool-depends) for compiler/hpc-bin/genprimopcode.
Fixes #21627
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a large collection of changes all relating to eta
reduction, originally triggered by #18993, but there followed
a long saga.
Specifics:
* Move state-hack stuff from GHC.Types.Id (where it never belonged)
to GHC.Core.Opt.Arity (which seems much more appropriate).
* Add a crucial mkCast in the Cast case of
GHC.Core.Opt.Arity.eta_expand; helps with T18223
* Add clarifying notes about eta-reducing to PAPs.
See Note [Do not eta reduce PAPs]
* I moved tryEtaReduce from GHC.Core.Utils to GHC.Core.Opt.Arity,
where it properly belongs. See Note [Eta reduce PAPs]
* In GHC.Core.Opt.Simplify.Utils.tryEtaExpandRhs, pull out the code for
when eta-expansion is wanted, to make wantEtaExpansion, and all that
same function in GHC.Core.Opt.Simplify.simplStableUnfolding. It was
previously inconsistent, but it's doing the same thing.
* I did a substantial refactor of ArityType; see Note [ArityType].
This allowed me to do away with the somewhat mysterious takeOneShots;
more generally it allows arityType to describe the function, leaving
its clients to decide how to use that information.
I made ArityType abstract, so that clients have to use functions
to access it.
* Make GHC.Core.Opt.Simplify.Utils.rebuildLam (was stupidly called
mkLam before) aware of the floats that the simplifier builds up, so
that it can still do eta-reduction even if there are some floats.
(Previously that would not happen.) That means passing the floats
to rebuildLam, and an extra check when eta-reducting (etaFloatOk).
* In GHC.Core.Opt.Simplify.Utils.tryEtaExpandRhs, make use of call-info
in the idDemandInfo of the binder, as well as the CallArity info. The
occurrence analyser did this but we were failing to take advantage here.
In the end I moved the heavy lifting to GHC.Core.Opt.Arity.findRhsArity;
see Note [Combining arityType with demand info], and functions
idDemandOneShots and combineWithDemandOneShots.
(These changes partly drove my refactoring of ArityType.)
* In GHC.Core.Opt.Arity.findRhsArity
* I'm now taking account of the demand on the binder to give
extra one-shot info. E.g. if the fn is always called with two
args, we can give better one-shot info on the binders
than if we just look at the RHS.
* Don't do any fixpointing in the non-recursive
case -- simple short cut.
* Trim arity inside the loop. See Note [Trim arity inside the loop]
* Make SimpleOpt respect the eta-reduction flag
(Some associated refactoring here.)
* I made the CallCtxt which the Simplifier uses distinguish between
recursive and non-recursive right-hand sides.
data CallCtxt = ... | RhsCtxt RecFlag | ...
It affects only one thing:
- We call an RHS context interesting only if it is non-recursive
see Note [RHS of lets] in GHC.Core.Unfold
* Remove eta-reduction in GHC.CoreToStg.Prep, a welcome simplification.
See Note [No eta reduction needed in rhsToBody] in GHC.CoreToStg.Prep.
Other incidental changes
* Fix a fairly long-standing outright bug in the ApplyToVal case of
GHC.Core.Opt.Simplify.mkDupableContWithDmds. I was failing to take the
tail of 'dmds' in the recursive call, which meant the demands were All
Wrong. I have no idea why this has not caused problems before now.
* Delete dead function GHC.Core.Opt.Simplify.Utils.contIsRhsOrArg
Metrics: compile_time/bytes allocated
Test Metric Baseline New value Change
---------------------------------------------------------------------------------------
MultiLayerModulesTH_OneShot(normal) ghc/alloc 2,743,297,692 2,619,762,992 -4.5% GOOD
T18223(normal) ghc/alloc 1,103,161,360 972,415,992 -11.9% GOOD
T3064(normal) ghc/alloc 201,222,500 184,085,360 -8.5% GOOD
T8095(normal) ghc/alloc 3,216,292,528 3,254,416,960 +1.2%
T9630(normal) ghc/alloc 1,514,131,032 1,557,719,312 +2.9% BAD
parsing001(normal) ghc/alloc 530,409,812 525,077,696 -1.0%
geo. mean -0.1%
Nofib:
Program Size Allocs Runtime Elapsed TotalMem
--------------------------------------------------------------------------------
banner +0.0% +0.4% -8.9% -8.7% 0.0%
exact-reals +0.0% -7.4% -36.3% -37.4% 0.0%
fannkuch-redux +0.0% -0.1% -1.0% -1.0% 0.0%
fft2 -0.1% -0.2% -17.8% -19.2% 0.0%
fluid +0.0% -1.3% -2.1% -2.1% 0.0%
gg -0.0% +2.2% -0.2% -0.1% 0.0%
spectral-norm +0.1% -0.2% 0.0% 0.0% 0.0%
tak +0.0% -0.3% -9.8% -9.8% 0.0%
x2n1 +0.0% -0.2% -3.2% -3.2% 0.0%
--------------------------------------------------------------------------------
Min -3.5% -7.4% -58.7% -59.9% 0.0%
Max +0.1% +2.2% +32.9% +32.9% 0.0%
Geometric Mean -0.0% -0.1% -14.2% -14.8% -0.0%
Metric Decrease:
MultiLayerModulesTH_OneShot
T18223
T3064
T15185
T14766
Metric Increase:
T9630
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch concerns #20155, part (1)
The general idea is that since primops have curried bindings
(currently in PrimOpWrappers.hs) we don't need to eta-expand
them. But we /do/ need to eta-expand the levity-polymorphic ones,
because they /don't/ have bindings.
This patch makes a start in that direction, by identifying the
levity-polymophic primops in the PrimOpId IdDetails constructor.
For the moment, I'm still eta-expanding all primops (by saying
that hasNoBinding returns True for all primops), because of the
bug reported in #20155. But I hope that before long we can
tidy that up too, and remove the TEMPORARILY stuff in hasNoBinding.
|
| |
|
|
|
|
|
| |
This avoids a fixpoint iteration for the common case of
non-recursive bindings.
|
|
|
|
|
| |
Sometimes there are very large casts, and coercionRKind
can be slow.
|
|
|
|
|
|
|
| |
We want to be able to eta-reduce
\x y. ((f x) |> co) y
by pushing 'co' inwards. A very small change accommodates this
See Note [Eta reduction with casted function]
|
|
|
|
|
|
| |
This is preliminary work for JavaScript support. It's better to put the
code handling the desugaring of Prim, C and JavaScript declarations into
separate modules.
|
|
|
|
|
|
|
| |
It was previously disabled because of:
- a confusion about "SRT inlining" (see removed comment in this commit)
- a linker bug (overflow) in the handling of ARM64_RELOC_SUBTRACTOR
relocation: fixed by a previous commit.
|
|
|
|
|
|
|
|
|
| |
- reordered the 3 SRT implementation cases from the most general to the
most specific one:
USE_SRT_POINTER -> USE_SRT_OFFSET -> USE_INLINE_SRT_FIELD
- added requirements for each
- found and documented a confusion about "SRT inlining" not supported
with MachO. (It is fixed in the following commit)
|
|
|
|
| |
it is still re-exported from GHC.Exts
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit acb188e0 introduced a regression in the computation of free
variables in mdo statements, as the logic in
GHC.Rename.Expr.segmentRecStmts was slightly different depending on
whether the recursive do block corresponded to an mdo statement or
a rec statment.
This patch restores the previous computation for mdo blocks.
Fixes #21654
|
|
|
|
|
|
|
|
|
|
|
|
| |
This moves handling of the magic 'withDict' function from the desugarer
to the typechecker. Details in Note [withDict].
I've extracted a part of T16646Fail to a separate file T16646Fail2,
because the new error in 'reify' hides the errors from 'f' and 'g'.
WithDict now works with casts, this fixes #21328.
Part of #19915
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The function breakTyVarCycle_maybe has been installed
in a dark corner of GHC to catch some gremlins (a.k.a.
occurs-check failures) who lurk
there. But it previously only caught gremlins of the
form (a ~ ... F a ...), where some of our intrepid users
have spawned gremlins of the form (G a ~ ... F (G a) ...).
This commit improves breakTyVarCycle_maybe (and renames
it to breakTyEqCycle_maybe) to catch the new gremlins.
Happily, the change is remarkably small.
The gory details are in Note [Type equality cycles].
Test cases: typecheck/should_compile/{T21515,T21473}.
|
| |
|
|
|
|
| |
We want `DynFlags` only mentioned in `GHC.Driver`.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The functions `pushCoValArg` and `pushCoercionIntoLambda` could
introduce bad representation-polymorphism. Example:
type RR :: RuntimeRep
type family RR where { RR = IntRep }
type F :: TYPE RR
type family F where { F = Int# }
co = GRefl F (TYPE RR[0])
:: (F :: TYPE RR)
~# (F |> TYPE RR[0] :: TYPE IntRep)
f :: F -> ()
`pushCoValArg` would transform the unproblematic application
(f |> (co -> <()>)) (arg :: F |> TYPE RR[0])
into an application in which the argument does not have a fixed
`RuntimeRep`:
f ((arg |> sym co) :: (F :: TYPE RR))
|
|
|
|
| |
Progress towards #17957
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch typechecks record updates by desugaring them inside
the typechecker using the HsExpansion mechanism, and then typechecking
this desugared result.
Example:
data T p q = T1 { x :: Int, y :: Bool, z :: Char }
| T2 { v :: Char }
| T3 { x :: Int }
| T4 { p :: Float, y :: Bool, x :: Int }
| T5
The record update `e { x=e1, y=e2 }` desugars as follows
e { x=e1, y=e2 }
===>
let { x' = e1; y' = e2 } in
case e of
T1 _ _ z -> T1 x' y' z
T4 p _ _ -> T4 p y' x'
The desugared expression is put into an HsExpansion, and we typecheck
that.
The full details are given in Note [Record Updates] in GHC.Tc.Gen.Expr.
Fixes #2595 #3632 #10808 #10856 #16501 #18311 #18802 #21158 #21289
Updates haddock submodule
|
|
|
|
|
|
|
|
|
|
|
|
| |
The simple optimiser would sometimes fail to
beta-reduce a lambda when there were casts
in between the lambda and its arguments.
This can cause problems because we rely on
representation-polymorphic lambdas getting
beta-reduced away (for example, those
that arise from newtype constructors with
representation-polymorphic arguments, with
UnliftedNewtypes).
|
|
|
|
|
| |
Metric Decrease:
T16875
|
|
|
|
|
|
|
|
|
|
|
| |
-haddock on GHC < 9.0 is quite fragile and can result in obtuse parse errors
when it encounters invalid haddock syntax.
This has started to affect users since 297156e0b8053a28a860e7a18e1816207a59547b
enabled -haddock by default on many flavours.
Furthermore, since we don't test bootstrapping with 8.10 on CI, this problem
managed to slip throught the cracks.
|
|
|
|
|
|
| |
We use the 64bit shifts only on 64bit platforms. But we
compile the code always so compiling it on 32bit caused a
lint error. So use Word64 instead.
|
|
|
|
| |
This should get rid of most, if not all "Overlong lists" errors and fix #20016
|
|
|
|
|
|
|
| |
Make sure comments captured in the exact print annotations are in
order of increasing location
Closes #20718
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The code
data instance Method PGMigration = MigrationQuery Query
-- ^ Run a query against the database
| MigrationCode (Connection -> IO (Either String ()))
-- ^ Run any arbitrary IO code
Resulted in two instances of the "-- ^ Run a query against the database"
comment appearing in the Exact Print Annotations when it was parsed.
Ensure only one is kept.
Closes #20239
|
|
|
|
|
|
|
| |
We don't need any more resolution than this.
Rename the field to `stgToCmmEmitDebugInfo` to indicate it is no longer
conveying any "level" information.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch addresses a relatively obscure situation that arose
when chasing perf regressions in !7847, which itself is fixing
It does two things:
* SpecConstr can specialise on ($df d1 d2) dictionary arguments
* FloatOut no longer checks argument strictness
See Note [Specialising on dictionaries] in GHC.Core.Opt.SpecConstr.
A test case is difficult to construct, but it makes a big difference
in nofib/real/eff/VSM, at least when we have the patch for #21286
installed. (The latter stops worker/wrapper for dictionary arguments).
There is a spectacular, but slightly illusory, improvement in
runtime perf on T15426. I have documented the specifics in
T15426 itself.
Metric Decrease:
T15426
|
|
|
|
| |
Progress towards #17957
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We were considering all Typeable evidence to be "BuiltinInstance"s which
meant the stage restriction was going unchecked. In-fact, typeable has
evidence and so we need to apply the stage restriction.
This is
complicated by the fact we don't generate typeable evidence and the
corresponding DFunIds until after typechecking is concluded so we
introcue a new `InstanceWhat` constructor, BuiltinTypeableInstance which
records whether the evidence is going to be local or not.
Fixes #21547
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With this change, `Backend` becomes an abstract type
(there are no more exposed value constructors).
Decisions that were formerly made by asking "is the
current back end equal to (or different from) this named value
constructor?" are now made by interrogating the back end about
its properties, which are functions exported by `GHC.Driver.Backend`.
There is a description of how to migrate code using `Backend` in the
user guide.
Clients using the GHC API can find a backdoor to access the Backend
datatype in GHC.Driver.Backend.Internal.
Bumps haddock submodule.
Fixes #20927
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This firstly caused spurious output to be emitted (as evidenced by
#21555) but even worse caused a massive coercion to be attempted to be
printed (> 200k terms) which would invariably eats up all the memory of
your computer.
The good news is that removing this trace allows the program to compile
to completion, the bad news is that the program exhibits a core lint
error (on 9.0.2) but not any other releases it seems.
Fixes #21577 and #21555
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit adds module `GHC.Cmm.Dominators`, which provides a wrapper
around two existing algorithms in GHC: the Lengauer-Tarjan dominator
analysis from the X86 back end and the reverse postorder ordering from
the Cmm Dataflow framework. Issue #20726 proposes that we evaluate
some alternatives for dominator analysis, but for the time being, the
best path forward is simply to use the existing analysis on
`CmmGraph`s.
This commit addresses a bullet in #21200.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
LlvmConfig contains information read from llvm-passes and llvm-targets
files in GHC's top directory. Reading these files is done only when
needed (i.e. when the LLVM backend is used) and cached for the whole
compiler session. This patch changes the way this is done:
- Split LlvmConfig into LlvmConfig and LlvmConfigCache
- Store LlvmConfigCache in HscEnv instead of DynFlags: there is no
good reason to store it in DynFlags. As it is fixed per session, we
store it in the session state instead (HscEnv).
- Initializing LlvmConfigCache required some changes to driver functions
such as newHscEnv. I've used the opportunity to untangle initHscEnv
from initGhcMonad (in top-level GHC module) and to move it to
GHC.Driver.Main, close to newHscEnv.
- I've also made `cmmPipeline` independent of HscEnv in order to remove
the call to newHscEnv in regalloc_unit_tests.
|