#2956. Update and rename existing tests for type void

[sgrekhov]

No description provided.

[eernstg]

Looks good!

I'm tempted to suggest that each of the situations where void is tested to get a specific treatment is followed up by a test which is identical, except that it uses typedef Void = void;. The purely syntactic cases should then allow Void to be used, but other places (like the implements clause of a class / ... / an extension type) would need to expand the type alias, find void, and report the error for that.

This could be done in these tests, or it could be a separate set of test libraries, in this PR or in another one, whatever is more convenient.

I commented on several cases where a compile-time error is expected, and it is a syntax error, and nothing says "syntax". I think it's useful (for future readers in particular) to explicitly mention that they are syntax errors, because syntax errors behave so differently from other errors. E.g., we would never even attempt to answer questions like "how many syntax errors does this email have, considered as a Dart program?". Also, we can't put several syntax errors together in the same test library, for the same reason.

In short, developers who read (not to mention maintain) tests involving syntax errors should be aware of the fact that they are syntax errors.

[sgrekhov]

Added notes that some compile errors are actually syntax errors and also added tests for type alias of void. PTAL.

[eernstg]

LGTM