[Version 9.0] Feature support for pattern-matching enhancements

[BillWagner]

This includes all commits from #1026

There are a number of conversations that haven't been addressed from #1026:

• Add support for pattern additions #1026 (comment) (overall review)
• Add support for pattern additions #1026 (comment)
• Add support for pattern additions #1026 (comment)
• Add support for pattern additions #1026 (comment)
• Add support for pattern additions #1026 (comment)

[BillWagner]

The proposal has several open issues. We need to ensure this PR matches the final decision on those questions.

I've looked through the LDM notes, and here's my initial analysis:

Are and, or, and not some kind of contextual keyword?

Yes, and, or and not are contextual keywords in a pattern. Even if a type or identifier is in scope, it's a pattern keyword. when used as a pattern. However, you can write this (confusing though it is):

public class or { }

or or = new or();

if (or is or)
{ }

Semantic type for relational operators: We have proposed that when the input type is already a comparable primitive, that is the type of the comparison. However, when the input is not a comparable primitive, we treat the relational as including an implicit type test to the type of the constant on the right-hand-side of the relational. If the programmer intends to support more than one input type, that must be done explicitly:

That is the decision. Those rules are in this PR.

Most of the other open questions in the speclet have resolutions documented. I haven't yet verified that this PR includes those decisions. Now verified, or added in the 8th commit.

The only remaining issue is the diagnostic rules for relational and logical patterns.