This moves us toward keeping Source objects simple, and also moves us toward allowing multiple items to have text content, though for now this would just be a "ContentID" referenced from a Source.

For instance, while a Source might have an HTML content that we actually render to the user, it would be great if we could actually have the source actors created by https://bugzilla.mozilla.org/show_bug.cgi?id=1541629 also have content. That way we can eventually do symbol parsing on a per-source-actor basis and avoid the gross logic we have now where we have to actually parse the HTML file to look for <script> tags. The engine knows all this, so we can get it to tell is and avoid needing to parse HTML clientside.

This would also be nice because we can make the ID a hash of the content + content type and then we'd be able to reuse a bunch of information across refreshes. If you have an original file open and don't change it, we won't have to recalculate symbol data for instance because the store could see that it already has this source text in memory.