Having the input type `char` makes the output of `tsrb_get_one()`
incomparable to the input of `tsrb_add_one()`. By changing it to
`uint8_t` we not only definitively fix it to an octet, but also ensure
that the input and output are the same.
Modified file doc header so that important info about usage and buffer
size show up on main API doc page instead of file doc page;
added comment in a couple of places to remind reader that buffer size
must be powewr of 2.