Custom attributes like global actors carry crucial semantic information
and should never be suppressed in the ASTPrinter. In particular,
`printQuickHelpDeclaration()` sets `PrintImplicitAttrs` to false,
but it's important for quick help / cursor info to include global
actors.
That way, when the request gets truncated by os_log in sourcekit-lsp, we see most of the request. Most likely the sourcetext and the compiler args wouldn't have made it into the log message completely anyway.
rdar://121322828
SourceKit-LSP currently parses the XML comment to generate Markdown again but round-tripping a (probably markdown) doc comment to XML to Markdown is lossy in many cases and unnecessary work. Include the comment as it is spelled in source in the cursor info response so that sourcekit-lsp can display it.
Part of rdar://120685874
When completing in cases like `bar(arg: foo(|, option: 1)`, we don’t know if `option` belongs to the call to `foo` or `bar`. Be defensive and also suggest the signature.
Rather than doing the transform in the parser, and
then potentially undoing it in Sema, move the
entire transform into Sema. This also lets us
unify the logic between function decls and
closures, and allows ASTGen to benefit from it.
Performing a related identifiers request on `self` returns an empty result because `self` is not an identifier. We thus can’t retrieve the first location to compute the old name.
rdar://121668042
If the file didn’t exist on disk, `fileSystem->getRealPath` returns an empty string, which means we can’t find the `InputBuffer` and thsu won’t run solver-based cursor info. There’s no reason to realpath `PrimaryInputPath`, `getMemoryBuffer` already does that.
rdar://120737244
This removes the distinction between argument completions and postfix expr paren completions, which was meaningless since solver-based completion.
It then determines whether to suggest the entire function call pattern (with all argument labels) or only a single argument based on whether there are any existing arguments in the call.
For this to work properly, we need to improve parser recovery a little bit so that it parsers arguments after the code completion token properly.
This should make call pattern heuristics obsolete.
rdar://84809503
Some of the tests now depend on features provided by Swift Syntax, and
will fail to pass if building without Swift Syntax support.
This commits marks all the tests I know fail to pass when Swift Syntax
is disabled.
- #70281 modified `ASTGen/verify-parse.swift`
- #70287 created `refactoring/SyntacticRename/operator.swift` and `SourceKit/RelatedIdents/operator.swift`
- #70389 modified `SourceKit/Refactoring/basic.swift` and indirectly
modified all `refactoring/ExtractFunction/*` tests.
The `NameMatcher` implemented in swift-syntax as slightly different semantics because it consideres `(callable.|callAsFunction)(x: 78)` as a reference to `callAsFunction` instead of a call with argument labels `(x: 78)`, which means that one test needs to be updated.
This allows building sourcekitd and swift-refactor with `SWIFT_BUILD_SWIFT_SYNTAX=NO`. In these builds, the `relatedidents` and `find-syntactic-rename-ranges` requests will always return an error.
