Add the `slurp` argument to all `_Program.input*()` methods, and pass it
through, so it's possible to supply it with every method. But most
importantly the `_Program.input()`, which also accepts `text`.

Let Python give us the length of the "bytes" it already knows, instead
of doing an strlen(). This improves performance a bit.

Support parsing RS-separated streams, as per RFC 7464.