Better handling of bad requests

[aldendaniels]

Currently, we load the index.html files for all requests, but only load the static files for requests that match the domain root or our document ID regex. This results in us simply loading a broken page on mal-formed URLs.

We should:

1. Allowing an optional trailing slash on correct URLs.
2. Give back a 404 for requests with incorrect URLs including pages that do not match the document ID regex and pages that do but do not correspond to a real document.