Well, the idea was mostly that it would allow writing some extent of RDEF validation alongside the basic HLSL compilation tests when implementing HLSL features.
Eh, fair enough.
The amount of work we'd have to do to parse is daunting, and adding new tests to the existing shaders / tables seems simple enough, but if you feel strongly about it I guess I can rewrite it...
Oh, I wasn't suggesting adding that to this MR, or replacing the existing test. It just seemed like a potentially useful feature for some indeterminate point in the future.