Make inscription type more flexible

[casey]

We need to be quite lenient in what inscriptions we accept, to maintain global consensus over symbols.

Ideally, this means performing a minimum of validity checks, since the more validity checks we have and the more complex they are, the more likely we'll accidentally change them and change the set of valid inscriptions, and the more likely that different implementations will disagree about the set of valid inscriptions.

This PR converts the inscription enum to a struct, and removes all validity checks associated with content and content type. This isn't strictly necessary at the moment, but it'll be more code churn the longer we wait to do it.