More generic attributes tests #55202
Conversation
| @@ -2108,6 +2108,22 @@ static void Main(string[] args) | |||
| Assert.NotNull(symbolInfo.Symbol); | |||
| } | |||
|
|
|||
| // TODO: it's not clear that this test is meaningful. | |||
There was a problem hiding this comment.
Yeah, test plan suggests including a LookupSymbols test, but I'm genuinely not sure what would be meaningful here. @333fred in case you have any suggestions.
There was a problem hiding this comment.
Have a test with a constructor that takes a parameter of type T, and make sure that you can get constructor information (I can't remember whether you need to use GetSymbolInfo or LookupSymbols for that.
There was a problem hiding this comment.
For LookupSymbols we'd want a test like LookupConstrAndDestr with Attr1</*bind*/string/*bind*/> to confirm we can tell the IDE symbols are in scope there.
|
Marking as draft for now to make room in the review queue. Ping when ready for another look |
| var node = tree.GetRoot().DescendantNodes().OfType<AttributeSyntax>().Single(); | ||
| var symbol = model.GetSymbolInfo(node); | ||
| Assert.Equal("Attr1<System.String>..ctor(System.String t)", symbol.Symbol.ToTestDisplayString()); | ||
| } |
There was a problem hiding this comment.
Curious: have you tried a Attr1<Attr1<string>>?
|
Integration test failures are not related to these changes so just merging. |
Related to test plan #36285
Related to #55190
This adds coverage for everything considered blocking. Note that there is an issue when an attribute constructor parameter is of a type parameter type in its original definition, even when the parameter is a closed type in its usage. Currently I think this might be a runtime bug but we will have to dig in more closely to be sure.