| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
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
|
|
|
|
|
|
| |
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.
|
|
|
|
|
|
|
|
| |
`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
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
| |
Implements #22702
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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).
|
|
|
|
|
|
|
| |
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.
|
|
|
|
|
| |
This will allow to make command line parsing to depend on
diagnostic system (which depends on dynflags)
|
| |
|
|
|
|
|
|
|
|
|
| |
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
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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
|
|
|
|
|
|
|
|
|
|
| |
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
|
|
|
|
|
|
|
| |
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]
|
|
|
|
| |
fixes #22958
|
|
|
|
| |
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.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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
-------------------------
|
|
|
|
|
|
| |
setNominalRole_maybe is supposed to output a nominal coercion.
In the SelCo case, it was not updating the stored role to Nominal,
causing #23362.
|
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
| |
Since !10123 we now reject this program.
|
|
|
|
|
|
|
|
|
|
|
|
| |
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
|
|
|
|
|
| |
When looking up (t1 ~# t2) in the quantified constraints,
check both orientations. Forgetting this led to #23333.
|
|
|
|
| |
Closes #23163.
|
|
|
|
|
| |
The prologue of each stack frame are the saved LR and FP registers, 8
byte each. I.e. the size of the stack frame header is 2 * 8 byte.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This MR changes some simple optimizations and is a first step in re-architecting
the JS backend pipeline to add the optimizer. In particular it:
- removes simple peep hole optimizations from `GHC.StgToJS.Printer` and removes that module
- adds module `GHC.JS.Optimizer`
- defines the same peep hole opts that were removed only now they are `Syntax -> Syntax` transformations rather than `Syntax -> JS code` optimizations
- hooks the optimizer into code gen
- adds FuncStat and ForStat constructors to the backend.
Working Ticket:
- #22736
Related MRs:
- MR !10142
- MR !10000
-------------------------
Metric Decrease:
CoOpt_Read
ManyAlternatives
PmSeriesS
PmSeriesT
PmSeriesV
T10421
T12707
T13253
T13253-spj
T15164
T17516
T18140
T18282
T18698a
T18698b
T18923
T1969
T19695
T20049
T3064
T5321FD
T5321Fun
T783
T9198
T9233
T9630
-------------------------
|
| |
|
|
|
|
| |
As requested by @treeowl in CLC#139.
|
|
|
|
|
|
|
|
|
| |
Tracking ticket: #20114
MR: !10390
This converts uses of `mkTcRnUnknownMessage` to newly added constructors
of `TcRnMessage`.
|
|
|
|
|
|
|
|
|
|
|
| |
Tracking ticket: #20115
MR: !10361
This converts uses of `mkTcRnUnknownMessage` to newly added constructors
of `TcRnMessage`.
Only addresses the single warning missing from the previous MR.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In general this patch pushes plugin initialisation points to earlier in
the pipeline. As plugins can modify the `HscEnv`, it's imperative that
the plugins are initialised as soon as possible and used thereafter.
For example, there are some new tests which modify hsc_logger and other
hooks which failed to fire before (and now do)
One consequence of this change is that the error for specifying the
usage of a HPT plugin from the command line has changed, because it's
now attempted to be loaded at initialisation rather than causing a
cyclic module import.
Closes #21279
Co-authored-by: Matthew Pickering <matthewtpickering@gmail.com>
|
|
|
|
|
|
| |
This fixes a bug in oneshot mode where hooks modified in a plugin
wouldn't be used in oneshot mode because we neglected to use the right
hsc_env. This was observed by @csabahruska.
|
|
|
|
|
|
|
|
|
| |
Tracking ticket: #20115
MR: !10350
This converts uses of `mkTcRnUnknownMessage` to newly added constructors
of `TcRnMessage`.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously, `gen_Newtype_fam_insts` was substituting the type variable binders
of a type family instance using `substTyVars`, which failed to take type
variable dependencies into account. There is similar code in
`GHC.Tc.TyCl.Class.tcATDefault` that _does_ perform this substitution properly,
so this patch:
1. Factors out this code into a top-level `substATBndrs` function, and
2. Uses `substATBndrs` in `gen_Newtype_fam_insts`.
Fixes #23329.
|
|
|
|
| |
This has no observable effect, but avoids storing useless data.
|
|
|
|
| |
fixes #23278
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
While fixing these I've also changed the way we store addresses into
ByteArray#. Addr# are composed of two parts: a JavaScript array and an
offset (32-bit number).
Suppose we want to store an Addr# in a ByteArray# foo at offset i.
Before this patch, we were storing both fields as a tuple in the "arr"
array field:
foo.arr[i] = [addr_arr, addr_offset];
Now we only store the array part in the "arr" field and the offset
directly in the array:
foo.dv.setInt32(i, addr_offset):
foo.arr[i] = addr_arr;
It avoids wasting space for the tuple.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* For ByteArray-based bounds-checking, the JavaScript backend must use the
`len` field, instead of the inbuild JavaScript `length` field.
* Range-based operations must also check both the start and end of the range
for bounds
* All indicies are valid for ranges of size zero, since they are essentially no-ops
* For cases of ByteArray accesses (e.g. read as Int), the end index is
(i * sizeof(type) + sizeof(type) - 1), while the previous implementation
uses (i + sizeof(type) - 1). In the Int32 example, this is (i * 4 + 3)
* IndexByteArrayOp_Word8As* primitives use byte array indicies (unlike
the previous point), but now check both start and end indicies
* Byte array copies now check if the arrays are the same by identity and
then if the ranges overlap.
|
|
|
|
|
|
|
| |
Ensure, when loading plugins, that the ghc the plugin depends on is the
ghc loading the plugin -- otherwise fail to load the plugin.
Progress towards #20742.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously, the unit-id of ghc-the-library was fixed as `ghc`.
This was done primarily because the compiler must know the unit-id of
some packages (including ghc) a-priori to define wired-in names.
However, as seen in #20742, a reinstallable `ghc` whose unit-id is fixed
to `ghc` might result in subtle bugs when different ghc's interact.
A good example of this is having GHC_A load a plugin compiled by GHC_B,
where GHC_A and GHC_B are linked to ghc-libraries that are ABI
incompatible. Without a distinction between the unit-id of the ghc library
GHC_A is linked against and the ghc library the plugin it is loading was
compiled against, we can't check compatibility.
This patch gives a slightly better unit-id to ghc (ghc-version) by
(1) Not setting -this-unit-id to ghc, but rather to the new unit-id (modulo stage0)
(2) Adding a definition to `GHC.Settings.Config` whose value is the new unit-id.
(2.1) `GHC.Settings.Config` is generated by Hadrian
(2.2) and also by cabal through `compiler/Setup.hs`
This unit-id definition is imported by `GHC.Unit.Types` and used to
set the wired-in unit-id of "ghc", which was previously fixed to "ghc"
The commits following this one will improve the unit-id with a
cabal-style package hash and check compatibility when loading plugins.
Note that we also ensure that ghc's unit key matches unit id both when
hadrian or cabal builds ghc, and in this way we no longer need to add
`ghc` to the WiringMap.
|
|
|
|
|
|
|
|
| |
Commit 3f374399 introduced a bug which caused us to forget to include
the parent of an export item of the form T(..) (that is, IEThingAll)
when checking for duplicate exports.
Fixes #23318
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adds a new LANGUAGE pragma ExtendedLiterals, which enables defining
unboxed numeric literals such as `0xFF#Word8 :: Word8#`.
Implements GHC proposal 0451:
https://github.com/ghc-proposals/ghc-proposals/blob/b384a538b34f79d18a0201455b7b3c473bc8c936/proposals/0451-sized-literals.rst
Fixes #21422.
Bumps haddock submodule.
Co-authored-by: Krzysztof Gogolewski <krzysztof.gogolewski@tweag.io>
|
|
|
|
|
|
|
|
|
| |
Tracking ticket: #20115
MR: !10336
This converts uses of `mkTcRnUnknownMessage` to newly added constructors
of `TcRnMessage`.
|
|
|
|
|
|
|
|
|
| |
This commit implements GHC proposal #433, adding the Unsatisfiable
class to the GHC.TypeError module. This provides an alternative to
TypeError for which error reporting is more predictable: we report it
when we are reporting unsolved Wanted constraints.
Fixes #14983 #16249 #16906 #18310 #20835
|
| |
|
|
|
|
|
|
|
|
|
| |
!9018 brought in exact print annotations in LayoutInfo for open and
close braces at the top level.
But it retained them in the HsModule annotations too.
Remove the originals, so exact printing uses LayoutInfo
|