Remove `BooleanType` as a type for consistent expression grammar

[gregsdennis]

This is an option for fixing the grammar issues explained in #387 around functions and their role in expressions.

• Removes the BooleanType type and makes match() and search() merely produce OptionalValueType. That value is either JSON true or false.
• Removes the special handling for functions that can return a JSON true or false result.

A fallout of this is that $[?match(@.timezone, 'Europe/.*')] is not valid/well-formed.

[gregsdennis]

I think the usability of having to compare the result of match and search is poor

I agree that it's more clunky, but returning a value (thereby requiring the comparison) does have the advantage of providing support for $[?match(@.a, 'a.*')==@.b].

I'm not tied to it though.

[cabo]

The current status of the document has solved the clunkiness. I still don't know why we need this regression.

[gregsdennis]

The current status of the document has solved the clunkiness.

No. I will not debate this here. I've made the case perfectly clear in #387 over multiple comments that the current solution is subpar at best. The document is wrong.

[cabo]

Well, it seems the document is not describing the solution very well. Let's work on that.

[gregsdennis]

Well, it seems the document is not describing the solution very well. Let's work on that.

Make a proposal, but the solution is wrong. The document presents the solution. And that's why the document is wrong.

[cabo]

Well, it seems the document is not describing the solution very well. Let's work on that.

Make a proposal, but the solution is wrong. The document presents the solution. And that's why the document is wrong.

You have worked yourself into this fervor. I'm going to stop commenting on this issue.

[cabo]

This is superseded by #403.