Currently the definition of the case(s) for unions in <:< don't implement a particularly sane subtype relation. Nonetheless the current semantics are relied on, eg. in fragmentApplies. Probably unions should only relate as subtypes to other unions and we should introduce a separate relation for fragment applicability.
Currently the definition of the case(s) for unions in
<:<don't implement a particularly sane subtype relation. Nonetheless the current semantics are relied on, eg. infragmentApplies. Probably unions should only relate as subtypes to other unions and we should introduce a separate relation for fragment applicability.