[Variadic Generics] Add a new type syntax node and parsing support for explicit pack references. #1134
Conversation
|
@swift-ci please test |
There was a problem hiding this comment.
Could you add a couple of test cases? Interesting test cases that come to my mind are
- the example in your PR description
each MyTypeas a parameter / return type (I assume that this should be invalid but we can/should defer diagnostic generation for this case to ASTGen since we cannot forbideach MyTypein the current SwiftSyntax tree)
| @@ -123,6 +123,14 @@ | |||
| Child('Ellipsis', kind='EllipsisToken') | |||
| ]), | |||
|
|
|||
| # pack-reference-type -> 'each' type | |||
| Node('PackReferenceType', name_for_diagnostics='pack reference', kind='Type', | |||
There was a problem hiding this comment.
This node is identical to the ConstrainedSugarType. What do you think about using the same node type for it instead of introducing a new node? I'm open to naming suggestions.
There was a problem hiding this comment.
Hmm. You're right that the syntactic structure is the same (though it's parsed slightly differently), but conceptually a pack element type is very different from a constrained sugar type. Do you think that most tooling operating on these type syntax nodes would want to treat pack elements the same as constraint types? The only reason why I added a new syntax node is because I added a new TypeRepr on the Swift side to make type resolution easier, since the resolution logic is completely different.
There was a problem hiding this comment.
It was an open question, maybe I should have phrased that more clearly. It was motivated because I’m merging the nodes for #available and #unavailable in #1125 and that simplified quite a few things. If you think it’s better to have different nodes for some/any and each, I’m totally fine with it.
I have test cases in my PR to the Swift repo here: https://github.com/apple/swift/pull/62509/files#diff-b4b3fbedbd5cb01a2a671c3e6e039e0bdc87f29ae3744e6aa6a713b32bb9ad04. These should exercise this code through parser validation. Do I need to add test cases here in addition to those? |
hborla
force-pushed
the
explicit-pack-reference-keyword
branch
from
91035df to
227fe88
Compare
|
@swift-ci please test |
1 similar comment
|
@swift-ci please test |
…rd applied to type parameter pack references.
hborla
force-pushed
the
explicit-pack-reference-keyword
branch
from
227fe88 to
ded777e
Compare
|
@swift-ci please test |
PackReferenceType, for pack references spelled with a contextualeachkeyword.PackReferenceTypeSyntaxwhen the contextualeachkeyword is applied to a type.