Structs with one field have different type semantics to newtype-declared UDTs

[sezna]

I encountered this one while fixing #2146 -- here's a repro:

https://microsoft.github.io/qsharp/?code=H4sIAAAAAAAAE2WQQW7CQAxF9znF3zWRUrEHsUmrik135QAuGGakYRyNHUJA3L1iAhVNtx6%252F5%252F9HLXUbw4cILthINI42RyMScC2KyL0NLaOhhCXKyUK1KAod%252BYbOU76GmOP0NnVKy4nMS8Qn%252BVhWmGMdveFSAMBsBnOMbw7SI9%252FmlCTBK9TIfNzDHI2BvYJPLW%252FylH7xrg2M3puDRAYHPnC0%252FBrYcMLyBpeWOq4WD0jl6WxLqqz%252FiIzU1a31nVpJz0dONSRsX9WGwFi%252Ffym2El8Mjo4Mc17hVTv%252B42soPRIUz73zJvbCCuppgN%252Fl8f2XHSkOkm5WirndznPYTsznMSju%252FusPD6bz6%252BMBAAA%253D&profile=unrestricted

I think we are missing a tuple-unwrap in the case of single-arity struct decls.

[sezna]

The solution here might be to disallow using old UDT syntax with new struct syntax, if this is an intentional structural differentiation. Investing more into the old UDT experience (newtype) is likely not worth it at this point, given that its on its way out.