Commit Graph

70 Commits

Author SHA1 Message Date
Robert Widmann
6df3fcf3ad Add Missing Definitions of canDerive{En,De}codable 2021-02-19 16:03:32 -08:00
Doug Gregor
ab3c5dee3e [Concurrency] Introduce Actor protocol to which actor classes all conform.
Introduce a new Actor protocol, which is a class-bound protocol with only
one requirement:

    func enqueue(partialTask: PartialAsyncTask)

All actor classes implicitly conform to this protocol, and will synthesize
a (currently empty) definition of `enqueue(partialTask:)` unless a suitable
one is provided explicitly.
2020-09-28 16:59:21 -07:00
Robert Widmann
3d6c6b4a59 Revert "Sema: Don't need to derive CaseIterable's AllCases associated type"
This reverts commit 3ae31d5173.
2020-08-18 17:37:04 -07:00
Slava Pestov
3ae31d5173 Sema: Don't need to derive CaseIterable's AllCases associated type
Just declaring a default in the standard library works fine.
2020-08-07 16:48:40 -04:00
Suyash Srijan
6ea2034e33 [NFC] Refactor DerivedConformances code to reduce code duplication for diagnostics (#32884) 2020-07-15 00:49:36 +01:00
Suyash Srijan
0aabee59d6 [Typechecker] Emit specialized diagnostic notes on automatic synthesis failure to Comparable (#32797) 2020-07-14 22:03:08 +01:00
Slava Pestov
dfbb958ced Sema: Try to derive type witnesses before running inference
Sema can infer type witnesses for a small set of known conformances, to
RawRepresentable, CaseIterable, and Differentiable.

Previously, we would try to compute the type witness in this order:

1) First, via name lookup, to find an explicit nested type with the
   same name as an associated type.

2) Second, we would attempt inference.

3) Third, we would attempt derivation.

Instead, let's do 3) before 2). This avoids circularity errors in
situations where the witness can be derived, but inference fails.

This breaks source compatibility with enum declarations where the raw
type in the inheritance clause is a lie, and the user defines their
own witnesses with mismatched types. However, I suspect this does not
come up in practice, because if you don't synthesize witnesses, there
is no way to access the actual raw literal values of the enum cases.
2020-06-26 19:46:52 -04:00
Dan Zheng
f9c5d7ae6c [AutoDiff] Derive Differentiable.zeroTangentVectorInitializer. (#31823)
`Differentiable` conformance derivation now supports
`Differentiable.zeroTangentVectorInitializer`.

There are two potential cases:
1. Memberwise derivation: done when `TangentVector` can be initialized memberwise.
2. `{ TangentVector.zero }` derivation: done as a fallback.

`zeroTangentVectorInitializer` is a closure that produces a zero tangent vector,
capturing minimal necessary information from `self`.

It is an instance property, unlike the static property `AdditiveArithmetic.zero`,
and should be used by the differentiation transform for correctness.

Remove `Differentiable.zeroTangentVectorInitializer` dummy default implementation.

Update stdlib `Differentiable` conformances and tests.
Clean up DerivedConformanceDifferentiable.cpp cruft.

Resolves TF-1007.
Progress towards TF-1008: differentiation correctness for projection operations.
2020-05-29 01:59:52 -07:00
Dan Zheng
bbe86e908d [AutoDiff upstream] Add Differentiable protocol derived conformances. (#30671)
Add `AdditiveArithmetic` derived conformances for structs and classes, gated by
the `-enable-experimental-differentiable-programming` flag.

Structs and classes whose stored properties all conform to `Differentiable` can
derive `Differentiable`:

- `associatedtype TangentVector: Differentiable & AdditiveArithmetic`
  - Member `TangentVector` structs are synthesized whose stored properties are
    all `var` stored properties that conform to `Differentiable` and that are
    not `@noDerivative`.
- `mutating func move(along: TangentVector)`

The `@noDerivative` attribute may be declared on stored properties to opt out of
inclusion in synthesized `TangentVector` structs.

Some stored properties cannot be used in `TangentVector` struct synthesis and
are implicitly marked as `@noDerivative`, with a warning:

- `let` stored properties.
  - These cannot be updated by `mutating func move(along: TangentVector)`.
- Non-`Differentiable`-conforming stored properties.

`@noDerivative` also implies `@_semantics("autodiff.nonvarying")`, which is
relevant for differentiable activity analysis.

Add type-checking and SILGen tests.

Resolves TF-845.
2020-03-27 06:40:04 -07:00
Dan Zheng
e5cb871428 [AutoDiff upstream] Add flag-gated AdditiveArithmetic derivation. (#30628)
Add `AdditiveArithmetic` derived conformances for structs, gated by the
`-enable-experimential-additive-arithmetic-derivation` flag.

Structs whose stored properties all conform to `AdditiveArithmetic` can derive
`AdditiveArithmetic`:
- `static var zero: Self`
- `static func +(lhs: Self, rhs: Self) -> Self`
- `static func -(lhs: Self, rhs: Self) -> Self`
- An "effective memberwise initializer":
  - Either a synthesized memberwise initializer or a user-defined initializer
    with the same type.

Effective memberwise initializers are used only by derived conformances for
`Self`-returning protocol requirements like `AdditiveArithmetic.+`, which
require memberwise initialization.

Resolves TF-844.
Unblocks TF-845: upstream `Differentiable` derived conformances.
2020-03-25 10:31:50 -07:00
Robert Widmann
4be0e4a2bd Strip Forward-Declararations of TypeChecker 2020-03-16 19:19:02 -07:00
kelvin13
d71b3dfec8 small vector sizes 2020-01-25 13:50:28 -06:00
kelvin13
884584546a clang format 2020-01-23 22:53:23 -06:00
kelvin13
1715ebcf9d address more reviewed issues 2020-01-23 22:46:30 -06:00
kelvin13
2701012434 fix some reviewed issues 2020-01-22 21:56:47 -06:00
kelvin13
fb4bd68653 Merge branch 'master' into comparable-enums 2020-01-22 18:38:31 -06:00
Brent Royal-Gordon
0c478b6be6 Revert "Merge pull request #28665 from CodaFi/the-phantom-menace"
This reverts commit 43a3ab7e35, reversing
changes made to 4f39d9c749.

# Conflicts:
#	include/swift/AST/Attr.def
#	lib/AST/Attr.cpp
#	lib/Serialization/Deserialization.cpp
#	lib/Serialization/ModuleFormat.h
#	lib/Serialization/Serialization.cpp
2020-01-15 15:28:42 -08:00
Slava Pestov
1dd527ef4e Sema: Fix crash when synthesizing RawRepresentable conformance with non-Equatable raw type
The error recovery logic around derived conformances is a little bit
tricky. Make sure we don't crash if a type explicitly provides a
RawValue type witness that is not equatable, but omits the witnesses
for init(rawValue:) and the rawValue property.

Fixes <rdar://problem/58127114>.
2020-01-13 19:07:06 -05:00
kelvin13
ed2552a694 fix merge conflict and update caseless stdlib enums to use synthesized Comparable 2020-01-02 18:40:50 -06:00
Robert Widmann
fbdd907a1f Add a way to just synthesize CodingKeys 2019-12-10 16:28:51 -08:00
Slava Pestov
2cada3bce1 Revert "Eagerly TypeCheck Synthesized Decls"
This reverts commit cab4e9f011.
2019-11-22 17:41:34 -05:00
kelvin13
e4dc295d64 fix conflicts 2019-11-21 12:42:10 -06:00
Robert Widmann
0775808db1 Reject invalid derived Hashable members explicitly
Hashable doesn't quite have the know-how to reject invalid derivation contexts before hand.  Give it a little help by adding a way to retrieve if a decl added to the conformance context was invalid after type checking completes.  Otherwise we'll emit "Hashable is broken".
2019-11-11 11:58:33 -08:00
Robert Widmann
cab4e9f011 Eagerly TypeCheck Synthesized Decls 2019-11-11 11:34:52 -08:00
Robert Widmann
7993f2ffeb Downgrade the TypeChecker in DerivedConformance to an ASTContext 2019-10-30 12:55:42 -07:00
taylorswift
c915fb6fc3 support associated values 2019-09-18 20:08:21 -05:00
taylorswift
210ec69a1b comparable enums 2019-09-01 13:41:16 -05:00
Slava Pestov
4c499fd4ac AST: Stop passing around LazyResolvers in various places 2019-07-06 00:43:22 -04:00
Owen Voorhees
e273d05460 [Sema] Add notes to explain why Equatable/Hashable conformance couldn't be synthesized
If a struct/enum cannot have Equatable/Hashable conformance automatically synthesized because a member's type is not Equatable/Hashable, add a note to the existing 'does not conform' diagnostic pointing out the type that blocked synthesis.
2019-06-08 13:02:49 -07:00
Doug Gregor
2756377367 [Type checker] Separate more functionality from the TypeChecker instance.
Use the usual bag of tricks to eliminating dependence on the
TypeChecker instance: static functions, LazyResolver callbacks, and
emitting diagnostics on decls/ASTContext.
2018-08-20 21:39:44 -07:00
Doug Gregor
32b9b0f6d2 [Type checker] Eliminate some unnecessary TypeChecker uses. NFC 2018-07-27 09:59:04 -07:00
Huon Wilson
47e643af0e [Sema] Fix synthesis of Equatable and Hashable conformances in extensions of generic types.
https://bugs.swift.org/browse/SR-6803
2018-05-07 09:41:25 +10:00
Huon Wilson
3488a3a53b [Sema] Allow synthesis of protocol conformances in extensions.
This works for all protocols except for Decodable on non-final classes, because
the init requirement has to be 'required' and thus in the type's declaration.

Fixes most of https://bugs.swift.org/browse/SR-6803.
2018-05-07 09:41:25 +10:00
Huon Wilson
5e094bdaa2 [Sema] Factor out "can't derive in extension" logic.
NFC.
2018-05-07 09:11:08 +10:00
Huon Wilson
cfd068de6a [Sema] Migrate almost everything to use DerivedConformance. 2018-05-07 09:11:07 +10:00
Huon Wilson
41942c9276 [Sema] Create a "DerivedConformance" class, to store common values.
Instead of passing around a TypeChecker and three Decls (the nominal type, the
protocol, and the decl declaring the conformance) everywhere, we can just pass
one object.

This should be [NFC].
2018-05-07 09:10:06 +10:00
Robert Widmann
dac06898e9 [SE-0194] Deriving Collections of Enum Cases
Implements the minimum specified by the SE-proposal.

* Add the CaseIterable protocol with AllCases associatedtype and
allCases requirement
* Automatic synthesis occurs for "simple" enums
    - Caveat: Availability attributes suppress synthesis.  This can be
              lifted in the future
    - Caveat: Conformance must be stated on the original type
              declaration (just like synthesizing Equatable/Hashable)
    - Caveat: Synthesis generates an [T].  A more efficient collection
              - possibly even a lazy one - should be put here.
2018-03-09 00:22:55 -05:00
John McCall
7f0f8830cd Split AccessorDecl out from FuncDecl. NFC.
This has three principal advantages:

- It gives some additional type-safety when working
  with known accessors.

- It makes it significantly easier to test whether a declaration
  is an accessor and encourages the use of a common idiom.

- It saves a small amount of memory in both FuncDecl and its
  serialized form.
2018-01-12 14:20:27 -05:00
Doug Gregor
4ae4ab70fa [Conformance checking] Put associated type inference into its own file.
Factor associated type inference out into its own separate source file, so
it's easier to maintain on its own.
2017-12-09 23:06:51 -08:00
Graydon Hoare
6d46fb144f [AST] Move copyFormalAccess to ValueDecl. 2017-11-06 14:31:21 -08:00
Tony Allevato
f2c434a038 Merge branch 'master' into synthesize-equatable-hashable 2017-09-21 22:54:36 -07:00
Slava Pestov
10bd85cb34 Sema: Fix Equatable, Hashable, RawRepresentable derivation for @_versioned enums
The synthesized declarations should inherit the @_versioned attribute
from the type, just like they inherit access control.

Fixes <rdar://problem/34342955>.
2017-09-17 21:35:55 -07:00
Tony Allevato
a3db968057 Move derivesProtocolConformances from AST to Sema 2017-08-08 18:30:22 -07:00
Itai Ferber
e2318392be Add CodingKey and *codable derived conformance
* Allow CodingKey conformance to be automatically derived for enums
  which have no raw type (with no associated values) and which have
  a raw type of String or Int.
* Allow Encodable and Decodable conformance to be automatically derived
  for classes and structs with Encodable/Decodable properties
* Add initial unit tests for verifying derived conformance
2017-04-25 13:13:35 -07:00
practicalswift
6d1ae2a39c [gardening] 2016 → 2017 2017-01-06 16:41:22 +01:00
practicalswift
797b80765f [gardening] Use the correct base URL (https://swift.org) in references to the Swift website
Remove all references to the old non-TLS enabled base URL (http://swift.org)
2016-11-20 17:36:03 +01:00
Doug Gregor
e1545a7d5a [Type checker/runtime] Move default implementation of Error._code into the runtime.
Rather than having Sema provide a default implementation of
Error._code when needed, introduce a runtime function to extract the
default code, so that we can provide a default implementation via a
protocol extension in the standard library.
2016-08-25 16:00:19 -07:00
Doug Gregor
6d2f11a305 [Type checker] Synthesize Error._code witnesses as 'final' within a class.
When synthesizing the witness for Error._code, synthesize it as
final. This isn't meant to be user-visible (and, therefore, isn't
meant to be user-overridable), so it's a minor efficiency
win. Moreover, we weren't making sure this member got synthesized in
in cross-module situations, leading to runtime crashes. Fixes
rdar://problem/27335637.
2016-08-01 21:27:51 -07:00
Doug Gregor
1dcfd337e5 [Cleanup] Eliminate the notion of derived global declarations for DeclContexts.
These were only used for synthesized global operators, which have now
been eliminated.
2016-07-21 12:54:27 -07:00
Doug Gregor
823c24b355 [SE-0112] Rename ErrorProtocol to Error.
This is bullet (5) of the proposed solution in SE-0112, and the last
major piece to be implemented.
2016-07-12 10:53:52 -07:00