Preserve HKTypeLambdas representing kinds in wildApprox #6240
Conversation
wildApprox should not replace references to type params of HKTypeLambdas which represent kinds with wildcards. Fixes #i6238.
|
test performance please |
|
performance test scheduled: 2 job(s) in queue, 1 running. |
|
Performance test finished successfully: Visit http://dotty-bench.epfl.ch/6240/ to see the changes. Benchmarks is based on merging with master (7143421) |
| tl.paramInfos.forall { | ||
| case TypeBounds(lo, hi) => | ||
| def isKindBound(tp: Type): Boolean = | ||
| Inferencing.isFullyDefined(tp, ForceDegree.none) || (tp match { |
There was a problem hiding this comment.
It would be good to have some explanations about the logic here as a comment. In particular I did not understand why a fully defined bounds represents a kind.
There was a problem hiding this comment.
I am also worried about performance implications. isFullyDefined is not cheap, and wildApprox is a hotspot.
There was a problem hiding this comment.
Doesn't isKindBound also need to return true for AnyKind? If not, we could use a comment explaining why that's the case.
There was a problem hiding this comment.
I'm just finishing up with a different approach ... no need to continue reviewing this.
|
Superseded by #6271. |
wildApproxshould not replace references to type params ofHKTypeLambdaswhich represent kinds with wildcards.Fixes #6238.