[ASTGen] Generate Swift key-path expressions #77398
Conversation
rintaro
requested review from
ahoppen,
bnbarham,
CodaFi,
hamishknight,
hborla,
slavapestov and
xedin
as code owners
|
@swift-ci Please smoke test |
|
|
||
| // RUN: %empty-directory(%t) | ||
| // RUN: %target-swift-frontend %s -dump-ast -enable-experimental-feature ParserASTGen -verify > %t/astgen.ast.raw | ||
| // RUN: not %target-swift-frontend %s -dump-ast > %t/cpp-parser.ast.raw |
There was a problem hiding this comment.
Comparing -dump-ast instead of -dump-parse because we generate KeyPathExpr differently.
In libParse the root expression is parsed as a normal expression, but in SwiftParser it's parsed as a TypeSyntax, so we generate it as TypeExpr. But they are normalized in PreCheckTarget and should be end up with the same type-checked AST.
There was a problem hiding this comment.
Would it be worth adding this as a comment to the test?
| var i = \[A].property.nonMutatingProperty | ||
| expect(&i, toHaveType: Exactly<ReferenceWritableKeyPath<[A], B>>.self) | ||
|
|
||
| var j = \[A].[x] |
There was a problem hiding this comment.
Amusingly you can also do \[A][x] 😄
There was a problem hiding this comment.
Hm I thought we require a dot if the first "path" part is a subscript. But apparently not 🤔
There was a problem hiding this comment.
FWIW this cases are pulled from test/expr/unary/keypath/keypath.swift And I will add more test cases from it when I revisit here for adding AST diagnostics (for e.g. empty components)
rintaro
force-pushed
the
astgen-keypathexpr
branch
from
c4d183d to
dc156a6
Compare
|
@swift-ci Please smoke test |
|
@swift-ci Please smoke test Windows |
1 similar comment
|
@swift-ci Please smoke test Windows |
No description provided.