Description

Currently Test.Case.ID only attempts to provide an identifier for arguments (or combinations of arguments, when there are multiple parameters) which are encodable, using one of the heuristics described in Run selected test cases. It does not attempt to model or represent arguments which don't conform to one of those protocols, and this is limiting because it means there's no way to distinguish those arguments in reporting.

Even though arguments which don't conform to one of those protocols are not considered stable, and thus cannot be selectively re-run, it's nevertheless helpful to have a way to uniquely identify and refer to such test cases within the scope of a single run. This issue tracks expanding Test.Case.ID and its associated types to model this, including providing a way to distinguish stable vs. non-stable arguments.

This is a prerequisite to #671.

Tracked by rdar://119522099

Additional information

No response