Two different Options .should.not.eql each other

[matthew-white]

#1287 removes direct access to the value property of Option objects. That caused Ramda to consider all Options to be equal, so the PR also added an equals() method to Option. However, I'm realizing that there's a similar issue with Should.js. Now that it doesn't have access to value, eql() (AKA deepEqual()) no longer throws for two different Options. For example:

// This passes.
Option.of(1).should.eql(Option.of(2));

We have tests with assertions like that. I noticed this issue while looking at #1291, where you can see this test:

central-backend/test/unit/http/middleware.js

Lines 56 to 63 in ca1652e

	it('should pass through any field key content', (done) => {
	const request = createRequest({ url: '/key/12|45/users/23' });
	fieldKeyParser(request, null, () => {
	request.fieldKey.should.eql(Option.of('12|45'));
	request.url.should.equal('/users/23');
	done();
	});
	});

If I change Option.of('12|45') to something else in the test, the test continues to pass.

It looks like Should.js still distinguishes between empty and non-empty Options, since they have different prototypes.

Instead of using a private property like #value in Option, how about using something like _value or Symbol('value')? I think those would signal effectively that the property isn't meant to be used outside the class.

[alxndrsn]

I've tried to address this in #1297