Checking for staticness is not enough because top-level decls count as
non-static yet do not have a sufficient curry level to satisfy this
check. Make sure we're looking for decls that reside inside of types.
rdar://75299825, SR-14327
While it is very convenient to default the ExtInfo state when creating
new function types, it also make the intent unclear to those looking to
extend ExtInfo state. For example, did a given call site intend to have
the default ExtInfo state or does it just happen to work? This matters a
lot because function types are regularly unpacked and rebuilt and it's
really easy to accidentally drop ExtInfo state.
By changing the ExtInfo state to an optional, we can track when it is
actually needed.
`visitTopLevelCodeDecl` ignored the `Stmt` visit returning a nullptr.
This caused the `walkToDeclPost` to run for the `TopLevelCodeDecl` and
thus an imbalance in the `RangeResolver` pre and posts (since none of
the children would have their `walkTo*Post` called).
This was originally incorrectly fixed while assuming that the
`walkTo*Post` are called regardless of whether the children were visited
or not. Those changes have been reverted - fixing an imbalance in
`ExtDecls` in `SemaAnnotator`.
Added a test case for the `ExtDecls` imbalance which can occur while an
extension is being added.
Resolves rdar://74820040
At the every usage of 'collectPossibleReturnTypesFromContext()', the
closure must have been typechecked earlier. If the closure has no
valid type at this point, there's no chance to be type-checked by
re-typechecking.
rdar://problem/74430478
Default arguments were still being visited when converting the call,
adding extra commas to the converted call. Skip over them.
Resolves rdar://74248990
Still convert the call if it was requested directly - only check the name
when converting a whole function. Once we have an attribute, we should
use that instead.
The replacement range was `FunctionDecl::getRange`, which does not
include attributes. This would cause attributes to be duplicated when
converting a function to async. Override the start with the attribute
start instead so that they are replaced as well.
Resolves rdar://74063741
`SourceEntityWalker` had an unbalanced `walkToDeclPre` and
`walkToDeclPost`, ie. `walkToDeclPost` could be called even though
`walkToDeclPre` was not. Specifically, this would occur for both
`OperatorDecl` and `PrecedenceGroupDecl` declarations.
These could both be added to the `if` in `walkToDeclPost`, but this
seems fairly errorprone in general - especially as new decls are added.
Indeed, there's already declarations that are being skipped because they
aren't explicitly tested for in `walkToDeclPre`, ie.
`PatternBindingDecl`.
Instead of skipping if not explcitly handled, only skip running the
`SEWalker` walk methods if the declaration is implicit (and not a
constructor decl, see TODO). This should probably also always visit
children, with various decls changed to become implicit (eg.
TopLevelCodeDecl), but we can do that later - breaks too many tests for
now.
This change exposed a few parameter declarations that were missing their
implicit flag, as well as unbalanced walk methods in `RangeResolver`.
When generating a module interface, emit `#if` around any declarations
that are tied to specific, named language features. This allows module
interfaces to be processed by older Swift compilers that do not
support these newer features, such as async/await or actors.
The amount of effort required to correctly handle a new kind of
feature varies somewhat drastically based on the feature itself. The
"simple" case is where a particular declaration can only exist if a
feature is available. For example, and `async` declaration is fairly
easy to handle; a `@_marker` protocol's conformances are not.
Fixes rdar://73326633.
Always parse `async` and `await`, allowing the definition and use of
asynchronous functions without the "experimental concurrency" flag.
Note that, at present, use of asynchronous functions requires one to
explicitly import or link against the `_Concurrency` library. We'll
sort this out in a follow-up change.
Tracked by rdar://73455330.
Adds three refactorings intended to help users migrate their existing
code to use the new async language features:
1. Convert call to use async alternative
2. Convert function to async
3. Add async alternative function
A function is considered to have an async alternative if it has a void
return type and has a void returning closure as its last parameter. A
method to explicitly mark functions as having an async alternative may
be added to make this more accurate in the future (required for eg.
a warning about a call to the non-async version of a function in an
async context).
(1) converts a call to use the new `await` async language syntax. If the
async alternative throws, it will also add `try`. The closure itself is
hoisted out of the call, see the comments on
`AsyncConversionStringBuilder` for specifics.
(2) converts a whole function to `async`, using (1) to convert any calls
in the function to their async alternatives. (3) is similar to (2), but
instead *adds* a function and replaces calls to its
completion/handler/callback closure parameter with `return` or `throws`.
Resolves rdar://68254700
Current modes are:
1. output the whole rewritten buffer, with RUN and CHECK lines removed
2. output each replacement in JSON
To make each refactoring (rather than the whole file) easier to test,
add a plain text output option that can be easily checked with
FileCheck.
Implementation-only imports are unnecessary in generated module interfaces, since they aren't exported to the module's dependencies, and the module's public API cannot refer to symbols imported as implementation-only.
Generated Swift interfaces for modules with overlays, like Foundation or Dispatch, currently contain `import Foundation`/`import Dispatch` statements.
These imports are redundant, and this change removes them.
Consolidate ThrowsKeyword, RethrowsKeyword, and AsyncKeyword to
EffectsSpecifierKeyword.
Abolish 'key.throwsoffset' and 'key.throwslength' as they aren't used.
In derivatives of loops, no longer allocate boxes for indirect case payloads. Instead, use a custom pullback context in the runtime which contains a bump-pointer allocator.
When a function contains a differentiated loop, the closure context is a `Builtin.NativeObject`, which contains a `swift::AutoDiffLinearMapContext` and a tail-allocated top-level linear map struct (which represents the linear map struct that was previously directly partial-applied into the pullback). In branching trace enums, the payloads of previously indirect cases will be allocated by `swift::AutoDiffLinearMapContext::allocate` and stored as a `Builtin.RawPointer`.
