review size limits

[vinniefalco]

Without the table we could use a larger bit width for off_t. And we should review every function to make sure it respects the size limits. Te refactoring of the top level rfc rules may have invalidated existing checks for limits.

[vinniefalco]

I removed checks for max_size

[alandefreitas]

We are just working with std::size_t now and off_t doesn't exist anymore. What's the state of this issue? Does it still make sense?

[vinniefalco]

I'm pretty sure there are a few loose ends where we don't correctly check for overflow when performing unsigned arithmetic. I tried to get edit_segments and edit_params right, maybe you want to look at them and see if they are correct and sufficiently tested?