Implement a query that returns the `AvailabilityContext` for a given
`SourceLoc` and `DeclContext`. Re-implement the existing type checker query
that just returns an `AvailabilityRange` on top of the new query.
Make the pointer to uniqued storage an implementation detail of an
`AvailabilityContext` value. This way clients of `AvailabilityContext` don't
need to think about pointers and can have access to mutating operations on a
context when appropriate.
Begin using `AvailabilityContext` as the availability representation in
`TypeRefinementContext`, instead of only storing platform introduction
availability ranges.
There should be no functional changes since this just changes the
representation of the existing information stored by `TypeRefinementContext`.
However, in the future `AvailabilityContext` will be expanded to represent
additional availability constraints.
Since string literals are only a single token from the perspective of the
parser, the source range for a pattern binding decl that is initialized with a
string literal expression only extends to the beginning of the string literal.
To ensure the TypeRefinementContext for a pattern binding decl includes the
entire contents of the init expression we must calculate the character source
range of the decl instead.
Resolves rdar://110952225 and https://github.com/swiftlang/swift/issues/77050.
This field was complicating initialization of `TypeRefinementContext` and
bloating its storage just to make it slightly more convenient to look up
information that is pretty easy to derive on-demand when it is actually needed.
Effects checking for `async` requires actor isolation to have been
computed. This change removes effects checking out of the contextualize
initializers request, because effects checking isn't needed for
secondary files. Instead, effects checking is done in the primary decl
checker after actor isolation checking.
Instead of doing the pattern parsing in both the
C++ parser and ASTGen, factor out the parsing into
a request that returns the pattern to emit, regex
type, and version. This can then be lazily run
during type-checking.
While returning the actual child vector from
`ExpandChildTypeRefinementContextsRequest` is a nice idea, it is both
inefficient (the vector gets copied in and out) and kind of inaccurate, since
the vector remains mutable after the initial expansion and may gain additional
children as macros are lazily expanded.
Also rename it to `getExplicitReturnStmts` for clarity and have it
take a `SmallVector` out parameter instead as a small optimization and
to discourage use of this new method as an alternative to
`AnyFunctionRef::bodyHasExplicitReturnStmt`.
If there is only one connected component, we don't actually have
to form the array of components at all. Re-organize the code so
that we can short-circuit this decision as soon as possible.
Changing the members of these structs are still API and ABI breaking
changes but they are not as fragile as C++ classes where adding/removing
virtual functions or doing other changes can also result in breaking the
ABI.
rdar://119319825
Tuples in general do have conversions but an empty tuple or `Void`
doesn't, which means that if a type variable has a subtype binding
to `Void` it should be safe to prioritize.
