Use ID-based equality check for TemplateFormat consistently

[izeye]

There are three ways to check equality for TemplateFormats constants:

• getId().equals()
• equals()
• ==

As they are constants, == will be sufficient. So this PR changes to use == for equality check consistently.

[wilkinsona]

Thanks, @izeye. While there are constants for Asciidoctor and Markdown, there's nothing to stop someone creating their own TemplateFormat implementation with the same ID as one of those constants. If someone does that, their TemplateFormat should be considered equal to the built-in one. In other words, getId().equals() is the right comparison to use in main code. It's less clear-cut in the tests as then we know we'll only be dealing with the constants.

[izeye]

@wilkinsona Thanks for the feedback! If so, how about implementing TemplateFormat.equals() with TemplateFormat.getId() and using TemplateFormat.equals() consistently?

[wilkinsona]

With TemplateFormat being an interface, it's difficult to enforce an implementation of equals. I think I'd prefer to keep the existing id-based equality check in the main code. I don't really feel strongly about the tests. Given that the constants exist and they're expected to be used, a same instance check is not unreasonable. On the other hand, lots of the tests shouldn't really know or care about the constants and a id-based equality check would keep things more loosely coupled.

[izeye]

@wilkinsona Sorry, I missed the fact that TemplateFormat is an interface. I'm not sure I understood your last comment correctly. Do you prefer an ID-based equality check or just to leave things as they are?

[wilkinsona]

Sorry, I wasn't very clear. I think the test's current use of equalTo to compare TemplateFormat instances could be improved. It should either use sameInstance (thereby verifying that the constant is being used), or it should compare the formats' IDs. I think I prefer the latter as it minimises any coupling to the constants.

[izeye]

@wilkinsona Thanks for the clarification! I updated this as you suggested.

[izeye]

I missed two places and polished them a bit in 4e75e00.

[wilkinsona]

Thanks again, @izeye. These changes are now in 1.2.x and master.