Some diffs are suboptimal

[iver56]

This diff could be better

Might be related to #379

[stianjensen]

How would the optimal diff look in this case?

[cristeahub]

This reminds me of how git diffs look !

[stianjensen]

Yup, git has different diff algorithms.

One is the fast one (which is default) and produces similar results to this.

Another is the 'patience' diff, which I believe is a little slower, but might handle situations like this better.

[iver56]

@stianjensen A better diff could look like this:

[stianjensen]

I think this might probably be the same issue. What's actually happened here is that the 'by' has been removed.