Rename -experimental-serialize-external-decls only to
-experimental-skip-non-exportable-decls in preparation for the flag being used
to influence more than just serialization.
Resolves rdar://116771543
This attribute instructs the compiler that this function declaration
should be "import"ed from host environment. It's equivalent of Clang's
`__attribute__((import_module("module"), import_name("field")))`
Add the thrown type into the AST representation of function types,
mapping from function type representations and declarations into the
appropriate thrown type. Add tests for serialization, printing, and
basic equivalence of function types that have thrown errors.
Parse typed throw specifiers as `throws(X)` in every place where there
are effects specified, and record the resulting thrown error type in
the AST except the type system. This includes:
* `FunctionTypeRepr`, for the parsed representation of types
* `AbstractFunctionDecl`, for various function-like declarations
* `ClosureExpr`, for closures
* `ArrowExpr`, for parsing of types within expression context
This also introduces some serialization logic for the thrown error
type of function-like declarations, along with an API to extract the
thrown interface type from one of those declarations, although right
now it will either be `Error` or empty.
This attribute instructs the compiler that this function declaration
should be "export"ed from this .wasm module. It's equivalent of Clang's
`__attribute__((export_name("name")))`
KeyPath's getter/setter/hash/equals functions have their own calling
convention, which receives generic arguments and embedded indices from a
given KeyPath argument buffer.
The convention was previously implemented by:
1. Accepting an argument buffer as an UnsafeRawPointer and casting it to
indices tuple pointer in SIL.
2. Bind generic arguments info from the given argument buffer while emitting
prologue in IRGen by creating a new forwarding thunk.
This 2-phase lowering approach was not ideal, as it blocked KeyPath
projection optimization [^1], and also required having a target arch
specific signature lowering logic in SIL-level [^2].
This patch centralizes the KeyPath accessor calling convention logic to
IRGen, by introducing `@convention(keypath_accessor_XXX)` convention in
SIL and lowering it in IRGen. This change unblocks the KeyPath projection
optimization while capturing subscript indices, and also makes it easier
to support WebAssembly target.
[^1]: https://github.com/apple/swift/pull/28799
[^2]: https://forums.swift.org/t/wasm-support/16087/21
In order to support lazy typechecking during module emission for modules
containing specialized functions, the computation of generic signatures for
`@_specialized` attributes must be requestified.
Resolves rdar://115569606
Extract `ExternallyAccessibleDeclVisitor` from `Serialization.cpp` into its own
header and rename it to `DeclExportabilityVisitor` to better align with
terminology in other parts of the compiler. Ideally, `DeclExportabilityVisitor`
should become the canoncial implementation of for exportability checks,
consolidating logic that is currently spread between serialization, module
interface printing, TBDGen, and type checking. For now, it is only used in
serialization to implement serialization safety checks.
Refactor deserialization safety to use a `DeclVisitor` CRTP instead of ad-hoc
casts. This ensures every kind of decl is handled explicitly.
Resolves rdar://115456536
Previously, fully qualified types would be missing for global vars and
properties in `.swiftinterface` files that were emitted lazily.
Adding the test case also revealed that PatternBindingDecls needed to be
typechecked before lazy module serialization as well.
Previously, unsatisfiable conformances could be omitted from emitted
`.swiftinterface` files in lazy typechecking mode since inherited types might
be unresolved when gathering the conformances.
Adding these test cases also revealed that serialization restrictions needed to
be relaxed in order to accomodate unsatisfiable conformances.
- Add a flag to the serialized module (IsEmbeddedSwiftModule)
- Check on import that the mode matches (don't allow importing non-embedded module in embedded mode and vice versa)
- Drop TBD support, it's not expected to work in embedded Swift for now
- Drop auto-linking backdeploy libraries, it's not expected to backdeploy embedded Swift for now
- Drop prespecializations, not expected to work in embedded Swift for now
- Use CMO to serialize everything when emitting an embedded Swift module
- Change SILLinker to deserialize/import everything when importing an embedded Swift module
- Add an IR test for importing modules
- Add a deserialization validation test
Wrap the `InheritedEntry` array available on both `ExtensionDecl` and
`TypeDecl` in a new `InheritedTypes` class. This class will provide shared
conveniences for working with inherited type clauses. NFC.
When `-experimental-serialize-external-decls-only` is specified, skip
serializing conformances to protocols that should be skipped to avoid
unnecessary typechecking. Also, ensure type and value witnesses are resolved
lazily during serialization by passing `true` for `useResolver`.
Resolves rdar://114799742
In https://github.com/apple/swift/pull/65267 deserialization safety was made
more conservative, allowing deserialization of any protocol and
deserialization of any extension declaring an explicit conformance to any
protocol. We can refine this to only allow deserialization of safe protocols
and deserialization of extensions declaring conformances to safe protocols.
Importantly, though, we must look up all conformances declared by the
extension, not just the explicit ones.
Resolves rdar://114673761
This is motivated by needing to typecheck conformances before serializing them
when lazy typechecking is enabled. It will also provide a centralized funnel
point for filtering out conformances we don't want to serialize.
This option is designed to be used in conjunction with
`-experimental-lazy-typecheck` and `-experimental-skip-all-function-bodies`
when emitting a resilient module. The emitted binary module should contain only
the decls needed by clients and should contain roughly the same contents as it
would if the corresponding swiftinterface were emitted instead and then built.
This functionality is a work in progress. Some parts of the AST may still get
typechecked unnecessarily. Additionally, serialization does not trigger the
appropriate typechecking requests for some ASTs and then fails due to missing
types.
Resolves rdar://114230586
This attribute can be attached to a noncopyable struct to specify that its
storage is raw, meaning the type definition is (with some limitations)
able to do as it pleases with the storage. This provides a basis for
implementing types for things like atomics, locks, and data structures
that use inline storage to store conditionally-initialized values.
The example in `test/Prototypes/UnfairLock.swift` demonstrates the use
of a raw layout type to wrap Darwin's `os_unfair_lock` APIs, allowing
a lock value to be stored inside of classes or other types without
needing a separate allocation, and using the borrow model to enforce
safe access to lock-guarded storage.
This is phase-1 of switching from llvm::Optional to std::optional in the
next rebranch. llvm::Optional was removed from upstream LLVM, so we need
to migrate off rather soon. On Darwin, std::optional, and llvm::Optional
have the same layout, so we don't need to be as concerned about ABI
beyond the name mangling. `llvm::Optional` is only returned from one
function in
```
getStandardTypeSubst(StringRef TypeName,
bool allowConcurrencyManglings);
```
It's the return value, so it should not impact the mangling of the
function, and the layout is the same as `std::optional`, so it should be
mostly okay. This function doesn't appear to have users, and the ABI was
already broken 2 years ago for concurrency and no one seemed to notice
so this should be "okay".
I'm doing the migration incrementally so that folks working on main can
cherry-pick back to the release/5.9 branch. Once 5.9 is done and locked
away, then we can go through and finish the replacement. Since `None`
and `Optional` show up in contexts where they are not `llvm::None` and
`llvm::Optional`, I'm preparing the work now by going through and
removing the namespace unwrapping and making the `llvm` namespace
explicit. This should make it fairly mechanical to go through and
replace llvm::Optional with std::optional, and llvm::None with
std::nullopt. It's also a change that can be brought onto the
release/5.9 with minimal impact. This should be an NFC change.
