Automate checking for broken links

[arthurattwell]

As discussed IRL with @ChristinaRookie, it would be great to have an easy way to find broken links in a book. In books with large TOCs, we need to check for broken links in navigation, for instance, which might be only broken IDs. (EpubCheck does a good job of spotting these, but often we aren't doing epub output till late in a project, if at all.) Spotting broken external links would also be useful for quality assurance.

Some options:

1. We could use an HTML validator, though that might give us far more feedback than we really want.
2. We could use a crawler, like simplecrawler. We might run this with gulp (e.g. like this) on output documents we've opened with Phantom (so that we're checking fully rendered HTML), much like we do for refreshing search indexes.
3. We could use a dedicated link-checking package line broken-link-checker.

Right now I like (2).