Commit Graph

1979 Commits

Author SHA1 Message Date
Allan Shortlidge
034c62cf01 AST: Generalize availability fix-its to support custom availability domains.
Resolves rdar://156118254.
2025-07-18 13:27:08 -07:00
Allan Shortlidge
a3fbe3d530 AST: Introduce queries on Decl for active @available attrs.
Use these queries to replace some duplicated code. Also, move the
`attr_inlinable_available.swift` test to the `Availability` sub-directory since
the test has more to do with availability checking than it has to do
specifically with the `@inlinable` attr.
2025-07-18 13:27:08 -07:00
Anthony Latsis
d7ec3a34bb Merge pull request #83094 from swiftlang/jepa-main
Sema: Make `@concurrent` not imply `async` on closures
2025-07-17 01:54:04 +01:00
Anthony Latsis
58d5059617 Sema: Make @concurrent not imply async on closures
The present approach is not prudent because `@concurrent` synchronous
functions, a natural extension, are a likely-to-happen future direction,
whereas the current inference rule is entirely grounded on `@concurrent`
being exclusive to async functions.

If we were to ship this rule, we would have to keep the promise for
backwards compatibility when implementing the aforementioned future
direction, replacing one inconsistency with another, and possibly
introducing new bug-prone expression checking code.

```swift
func foo(_: () -> Void) {}
func foo(_: () async -> Void) {}

// In a future without this change and  `@concurrent` synchronous
// functions accepted, the first call resolves to the first overload,
// and the second call resolves to the second, despite `@concurrent` no
// longer implying `async`.
foo { }
foo { @concurrent in }
```

This change also drops the fix-it for removing `@concurrent` when used
on a synchronous closure. With the inference rule gone, and the
diagnosis delayed until after solution application, this error raises a
fairly balanced choice between removing the attribute and being
explicit about the effect, where a unilateral suggestion is quite
possibly more harmful than useful.
2025-07-16 20:46:29 +01:00
Allan Shortlidge
d09e117dc0 AST: Canonicalize platform versions for @backDeployed attrs.
Also, diagnose invalid platform versions.

Part of rdar://155558161.
2025-07-14 16:40:16 -07:00
Pavel Yaskevich
d8a12e9b75 Merge pull request #82925 from xedin/remove-concurrency-attr-clash-downgrade
[Concurrency] Don't downgrade explicit isolation attribute clashes
2025-07-11 10:52:06 -07:00
Pavel Yaskevich
a9373c0f3f [Concurrency] Don't downgrade explicit isolation attribute clashes
The compiler shouldn't accept mismatch in explicit isolation attributes
because it could lead to incorrect isolation selection.
2025-07-09 10:59:13 -07:00
Alexis Laferrière
9e953d2f96 Merge pull request #82891 from xymus/cdecl-reserved-names
Sema: Report the use of reserved names by `@cdecl` functions
2025-07-09 09:22:03 -07:00
Alexis Laferrière
a23c658c9d Sema: Report the use of reserved @cdecl names 2025-07-08 13:00:06 -07:00
Pavel Yaskevich
43eec8fede [AST/Sema] SE-0487: Expand @nonexhaustive attribute to support warn argument
The spelling `@nonexhaustive(warn)` replaces `@preEnumExtensibility`
attriubte.
2025-07-04 10:20:25 -07:00
Pavel Yaskevich
fe1ae75711 [AST/Sema] SE-0487: Rename @extensible into @nonexhaustive
This is an accepted spelling for the attribute. This commit
also renames the feature flag from `ExtensibleAttribute` to
`NonexhaustiveAttribute` to match the spelling of the attribute.
2025-07-03 17:50:05 -07:00
Alexis Laferrière
f101bfd969 Merge pull request #82697 from xymus/cdecl-misc
AST: Misc improvements to `@decl` enums: implicit names, code gen fix and more tests
2025-07-03 15:44:34 -07:00
Pavel Yaskevich
a522448e90 [Concurency] Allow declarations with @isolated(any) parameters be marked as @concurrent/nonisolated(nonsending)
It's shouldn't be possible to use these attributes directly on
the function type that is `@isolated(any)` as per SE-0461 proposal
but it shouldn't preclude declarations that have parameters with
`@isolated(any)` from using them.

Resolves: rdar://154754939
2025-07-01 15:07:23 -07:00
Alexis Laferrière
beb980b1f3 AST: Support @cdecl enums without an explicit C name
Accept `@cdecl` enums without an explicit C name. The C name defaults to
the Swift one. It is printed using the `SWIFT_ENUM` macro instead of
`SWIFT_ENUM_NAMED`.
2025-07-01 14:10:53 -07:00
Alexis Laferrière
529cc3ccb6 Merge pull request #82039 from xymus/cdecl-enum
PrintAsClang: Introduce `@cdecl` enums
2025-06-30 09:56:28 -07:00
Alexis Laferrière
81a0f98783 Merge pull request #82194 from xymus/cdecl-parser
Parser: Accept `@cdecl` with an optional identifier for a custom C name
2025-06-27 15:06:11 -07:00
Becca Royal-Gordon
7460ce1c72 Merge pull request #82468 from beccadax/abi-sendable-metatype
[NFC] Test that `@abi` handles `SendableMetatype`
2025-06-26 15:41:03 -07:00
Becca Royal-Gordon
66300b7e6e Test that @abi handles SendableMetatype
The `@abi` checker ignores uses of marker protocols, as they don’t have any impact on call compatibility. Add tests to ensure that this rule encompasses SE-0470’s `SendableMetatype`.

Fixes rdar://152602566.
2025-06-24 15:40:01 -07:00
Slava Pestov
7f8175b3da RequirementMachine: Add two more completion termination checks for concrete type requirements
The concrete nesting limit, which defaults to 30, catches
things like A == G<A>. However, with something like
A == (A, A), you end up with an exponential problem size
before you hit the limit.

Add two new limits.

The first is the total size of the concrete type, counting
all leaves, which defaults to 4000. It can be set with the
-requirement-machine-max-concrete-size= frontend flag.

The second avoids an assertion in addTypeDifference() which
can be hit if a certain counter overflows before any other
limit is breached. This also defaults to 4000 and can be set
with the -requirement-machine-max-type-differences= frontend flag.
2025-06-17 17:51:25 -04:00
Alexis Laferrière
89f7ebfb18 Sema: @cdecl functions C name defaults to the Swift name 2025-06-11 12:42:38 -07:00
Alexis Laferrière
2601ff44d4 Parser: Accept @cdecl with an indentifier for the C name
Begin accepting the attribute in the form of `@cdecl(cName)`, using an
identifier instead of a string.

For ease of landing this change we still accept the string form. We
should stop accepting it before making this feature available in
production.
2025-06-11 12:42:38 -07:00
Allan Shortlidge
e16c638fc3 AST: Warn for non-existent platform versions in @available attributes. 2025-06-10 22:11:05 -07:00
Allan Shortlidge
e358cd6309 Tests: Correct invalid platform versions in @available attributes. 2025-06-10 22:11:05 -07:00
Allan Shortlidge
3f305a48c8 AST: Re-map macOS 16 aligned availability versions to 26.
- watchOS 12 -> 26
- visionOS 3 -> 26
- macos 16 -> 26
- iOS 19 -> 26
- tvOS 19 -> 26

The version numbers for `if #available(...)` queries are intentionally not
re-mapped.
2025-06-10 22:11:05 -07:00
Meghana Gupta
7d454533a0 Merge pull request #82067 from meg-gupta/lifetimeunderscored
Update spelling for representing lifetime dependencies to @_lifetime
2025-06-09 19:01:31 -07:00
Alexis Laferrière
6d091123f9 Sema: Type-check @cdecl enums
@cdecl enums are Swift enums representable in C. These enums must have
an integer raw type. They can be referenced from @cdecl functions and
@objc methods. @objc enums are still rejected from @cdecl functions.
2025-06-09 11:54:13 -07:00
Meghana Gupta
44e05fa858 [NFC] Update tests and diagnostics 2025-06-07 12:49:01 -07:00
Allan Shortlidge
85444fd1e5 AST: Diagnose unrecognized platforms as errors with CustomAvailability enabled.
When the CustomAvailability experimental feature is enabled, make it an error
to specify an unrecognized availability domain name. Also, add these
diagnostics to a diagnostic group so that developers can control their behavior
when they are warnings.

Resolves rdar://152741624.
2025-06-06 17:30:01 -07:00
Allan Shortlidge
924eb4898b AST: Fix iOS -> visionOS version remap for @backDeployed attrs.
The version remapping for `@backDeployed` regressed due to a bug introduced by
https://github.com/swiftlang/swift/pull/81922.

Also, fix some visionOS tests that have gotten out of date because we don't
seem to be running them in CI.

Resolves rdar://152542983.
2025-06-05 19:19:07 -07:00
Alexis Laferrière
5df615d3e0 Sema: Reject enums marked with both @cdecl and @objc 2025-06-05 13:46:36 -07:00
Allan Shortlidge
f4b4dc9889 AST/Sema: Fix remapping of iOS availability in diagnostics for visionOS.
When compiling for visionOS, iOS availability attributes are remapped into the
visionOS availability domain automatically. While the version remapping was
being performed correctly, there was a regression that caused the platform name
to be printed incorrectly in many diagnostics. Whenever an iOS version is
remapped to a visionOS version, availability diagnostics will now present
those versions as visionOS versions instead of iOS versions.

Resolves rdar://146293165.
2025-06-02 17:55:38 -07:00
Alexis Laferrière
acdc9514c2 AST: Begin accepting @cdecl on enums (but not @_cdecl) 2025-06-02 13:19:36 -07:00
Alexis Laferrière
7d6447be20 Merge pull request #81519 from xymus/cdecl-diags
Sema: Update more diagnostics for `@cdecl` and representability in the C language
2025-05-29 13:37:29 -07:00
Arnold Schwaighofer
7ac551636b Merge pull request #81714 from aschwaighofer/se0460
SE-0460: Introduce @specialized attribute
2025-05-28 12:03:48 -07:00
Arnold Schwaighofer
13ff5abdb8 Introduce @specialized attribute
Implements SE-0460 -- the non-underscored version of @specialized.

It allows to specify "internal" (not abi affecting) specializations.

rdar://150033316
2025-05-23 13:12:47 -07:00
Michael Gottesman
3ed4059a60 [sema] Change non-sendable -> non-Sendable in diagnostics.
This matches send non sendable but importantly also makes it clear that we are
talking about something that doesn't conform to the Sendable protocol which is
capitalized.

rdar://151802975
2025-05-22 11:37:58 -07:00
Pavel Yaskevich
cda9866b26 [CSGen] Prevent @concurrent on closures from skipping throws inference
Introduction of `@concurrent` attribute caused an unintended
side-effect in `ClosureEffectsRequest` since the attribute
could only be used on `async` types setting `async` too early
prevented body analysis for `throws` from running.

Resolves: rdar://151421590
2025-05-15 17:17:02 -07:00
Pavel Yaskevich
17b8f7ef12 [Sema] Validate that @_inheritActorContext is used only on @Sendable/sending and async/@isolated(any) parameters 2025-05-14 20:08:00 -07:00
Pavel Yaskevich
04d46760bb [AST] Extend @_inheritActorContext attribute to support optional always modifier
By default (currently) the closure passed to a parameter with `@_inheritActorContext`
would only inherit isolation from `nonisolated`, global actor isolated or actor
context when "self" is captured by the closure. `always` changes this behavior to
always inherit actor isolation from context regardless of whether it's captured
or not.
2025-05-14 20:07:57 -07:00
Alexis Laferrière
ef84f0e1aa Sema: Update more diagnostics for @cdecl vs @objc 2025-05-14 13:13:03 -07:00
Hamish Knight
3c67271869 [AST] Handle a few more cases in getStartLoc()
Handle PatternBindingDecls with missing var locations, which can
happen for loop iterator vars, and FuncDecls with missing name and
func locations, which can happen for `defer`. Also while here make
sure we set the source location of a parser-produced ErrorExpr.
2025-05-14 11:15:42 +01:00
Alexis Laferrière
9c7f0c7550 Merge pull request #80917 from xymus/cdecl-print
PrintAsClang: Print a C block in the compatibility header for `@cdecl` functions
2025-05-13 09:21:30 -07:00
Becca Royal-Gordon
602a1168a1 Merge pull request #81421 from beccadax/just-not-my-type 2025-05-12 16:12:40 -07:00
Becca Royal-Gordon
7e7f9d024b Fix @_implements type resolution diagnostics
If a type in an `@_implements` attribute failed to resolve, Sema would assume it was because the type existed but wasn’t a protocol, even if there was another reason for the problem (such as the type not existing). Explicitly resolve the TypeRepr again through a path that will produce diagnostics.
2025-05-09 14:41:10 -07:00
Ben Barham
6667471b3d Merge pull request #81115 from beccadax/abi-let-you-graduate
Finish implementing `@abi` (SE-0476)
2025-05-08 12:52:00 -07:00
Alexis Laferrière
f6dd432450 PrintAsObjC Test: Test pointer types with @cdecl 2025-05-08 11:31:51 -07:00
Anthony Latsis
4f4141fea8 Frontend: Obsolete -fixit-all and -emit-fixits-path
With `ARCMigrate` and `arcmt-test` removed from clang in
https://github.com/llvm/llvm-project/pull/119269 and the new code
migration experience under way (see
https://github.com/swiftlang/swift-evolution/pull/2673), these options
are no longer relevant nor known to be in use. They were introduced
long ago to support fix-it application in Xcode.

For now, turn them into a no-op and emit a obsoletion warning.
2025-05-07 02:30:30 +01:00
Becca Royal-Gordon
01431b87b2 Make @abi non-experimental
This includes changing the feature name so that compilers with the experimental feature don’t accidentally pick up content that only works in the final version.

Resolves rdar://150065196.
2025-05-05 13:50:51 -07:00
Becca Royal-Gordon
0c508a8b78 Ban @abi on multi-var PBDs
Per a comment in the review thread that the comma in a multi-variable pattern binding makes it look like the `@abi` attribute has several arguments.
2025-05-05 13:50:17 -07:00
Pavel Yaskevich
498430afaf [AST/Sema] Add @preEnumExtensibility attribute to downgrade warnings for extensible enums
Just like `@preconcurrency` for concurrency, this attribute is going
to allow exhaustiveness error downgrades for enums that were retroactively
marked as `@extensible`.
2025-05-02 09:07:22 -07:00