I thought about using a transformer (9VAC output) as Russel suggested. In fact, I still using it for testing purposes.

My intention using the resistor based voltage divider is to reduce size. If everything works ok (IE, the project works) I intend to assemble it in a wall-wart like package, so small devices like resistors would be a plus.

I found some RMS to DC converters from Analog Devices, but I would have to do some voltage division to feed the input. So I prefer do the math in my program, and use resistors.

My schematics show two grounds, one in the transformer primary and the other on the circuit. They are not the same and aren't connected. The ground on the primary is due to a limitation o EWB, that I used to a little simulation. So since the grounds aren't connected I don't see a hazard if the live/neutral are reversed as michael notted(quite common here in Brazil, where mains outlets aren't "polarized".

Steve, have you seen something like that in some device? I don't have to worry about strict safety standards, it's not a commercial product, so I'm thinking to insulate the AC input with epoxy.

I don't see the mains input "migrating" to the circuit, but high voltage is not my expertise, and everytime that I work with something above 5V I got cautious