Anyone wanting to test: use
this one instead.
I've lengthed the timeouts and retries on some of the AT routines, as well as adding short pauses when initialising the mouse. As a result, all 3 rodents I had to test with initialise correctly.
The zip also contains a potentially useful AT "terminal", with which you can send commands to the connected device. Curiously, the keyboard I've been using sends F5 rather than FA as an ack... maybe why the PC it used to be on would compain that there wasn't a keyboard attached
As for mixing mouse and cursor keys; the mouse updates the (16-bit) values of Mouse._x and Mouse._y by an offset every Mouse._update, so you can manually change the values if you so wish. It's very easy to mix direct input and the mouse routines, especially as none of the calls in Emerson are blocking
