Support diffs

[alexbepple]

Cp. http://tjholowaychuk.tumblr.com/post/18574009869

[rluba]

Thank you for the pull request!

Generally, the diff doesn't seem very useful for containsString because the expected substring and the actual text don't necessarily have much in common:

In addition, the expected shown in the diff isn't really the expected value due to the semantics of containsString.

Can you provide an example where the diff is useful when using containsString? If you just want to see the actual string unescaped, I can provide an option for that and maybe even make it the default.

Regardless, the diff seems very useful for the equalTo matcher, so I've extended your changes to add support when using equalTo and/or is:

Unfortunately, there's currently an open bug in mocha (including the latest 1.x and 2.x releases) that causes the diff to be displayed incorrectly: mochajs/mocha#1241

[alexbepple]

When the expected substring and the actual text do have a lot in common, the diff is useful in determining the difference.

Here is a real-life example. I broke my production code by omitting the single quote after the start date. The diff makes it easier to understand the difference.

I agree, that "expected" does not match the semantics of containsString. However, it’s useful. :)

A more contrived example:

I hope you are convinced that a diff can be useful for containsString. The question now becomes whether it’s okay to make it the default. I think it is.

PS: Don’t mind my stacktraces. ;)

[rluba]

Yes, the diffs in your screenshots really look useful. What version of mocha and which reporter are you using to get this diff? As you can see, my screenshot looked completely different.

[rluba]

Merged into master as 0754c0f and released as part of v2.8.0.

[alexbepple]

"mocha": "~2.0.1",
"mocha-unfunk-reporter": "~0.4.0"

Wow I had no idea that it was the reporter that made the diffs so useful. One more reason to stick with mocha-unfunk-reporter. ;) Originally I started using it because the default reporters were unreadable with my dark color scheme.

[rluba]

Thank you for both the pull request and the pluck of mocha-unfunk-reporter :-)

[rluba]

Seems like unfunk-diff is the reason for the better diffs.

[alexbepple]

I am working on providing more useful diffs for IsObjectWithProperties, including only showing the parts that matter. The API will likely have to change for this, because one cannot find out what matters without having both, actual and expected. Then getExpectedForDiff() and formatActualForDiff(actual) won’t work. So perhaps this effort will result in something like getObjectsForDiff(actual) which returns a hash {actual: …, expected: …}.

In light of this, we should warn on the wiki that this is experimental.

Speaking of which, are you going to change the wiki or shall I give it a shot and then create a PR?

[rluba]

It would be great if you could provide a PR after #8 is merged to document the new API. I wouldn't bother with documenting the experimental API. I've added a warning to the release notes of v2.8.0.