Document and test options to createPatch

[oBusk]

The documentation states that createPatch only supports context, when in actuality it passes the options on to the diffLines implementation.

This PR updates the documentation to clarify this and also adds tests to verify this behavior.

Since these methods are supposed to output a patch, it's arguable that the options doesn't make sense, but I'd argue that the ignoreWhitespace at least makes sense since outputting a patch is a pretty normalized way to display a diff and it makes sense to be able to use the existing options for it.

newLineIsToken is a bit weirder when creating "patches". I added docs and tests for that in separate commit so that it can be reverted if it's something that is not wanted.

[kpdecker]

LGTM, thank you for the contribution.

[ExplodingCabbage]

newLineIsToken is a bit weirder when creating "patches". I added docs and tests for that in separate commit so that it can be reverted if it's something that is not wanted.

Yeah, IMO allowing newLineIsToken to be passed to patch functions at all is a misfeature. When there are actual changes to newline structure, the "patches" we get - like the one shown in the newlineIsToken: true test case added here - aren't valid patches and can't be applied to the original file by jsdiff or any other tool. I find it difficult to imagine this ever being useful to anyone but easy to imagine people getting caught out by it when they generate a "patch" with this option and then discover it can't be applied.

[ExplodingCabbage]

However, let me also thank you for the tests - without them, I would never have realised this feature existed or that I was unwittingly breaking it in my third commit at #535! I think I want to the feature, but now I can at least be more deliberate about doing so and document it in the release notes...