When doing subtyping (initial declaration, assigning, passing argument as parameter, return type etc) the returned result is a SubTypeResult. The subtyping logic (in types/subtyping.rs) currently collects some of the reasons for mismatches when doing the checking and stores this as a NonEqualityReason.
However this reason is thrown away in many places and does not make it to diagnostics. This can be shown by running rg "SubTypeResult::IsNotSubType\(_".
It would be better to pass this via a field to many of the diagnostics. For example here
There should be something like mismatch: MismatchExplanation. This can be turned into a user readable explanation of why the type error happened. This can be surfaced via adding labels in the Diagnostic::PositionWithAdditionalLabels.
The TypeStringRepresentation enum was designed to cover this, but I don't think it quite fits into this structure.
When doing subtyping (initial declaration, assigning, passing argument as parameter, return type etc) the returned result is a SubTypeResult. The subtyping logic (in
types/subtyping.rs) currently collects some of the reasons for mismatches when doing the checking and stores this as a NonEqualityReason.However this reason is thrown away in many places and does not make it to diagnostics. This can be shown by running
rg "SubTypeResult::IsNotSubType\(_".It would be better to pass this via a field to many of the diagnostics. For example here
https://github.com/kaleidawave/ezno/blob/5191329edd448c364914917297bef2846527b91b/checker/src/context/mod.rs#L1115-L1120
There should be something like
mismatch: MismatchExplanation. This can be turned into a user readable explanation of why the type error happened. This can be surfaced via adding labels in theDiagnostic::PositionWithAdditionalLabels.