I think what you meant was both transistors "ON,' Jan, and I agree, in that one should KNOW what multiple keypresses will do.

However, I've never seen an MCU that has so low an output impedance in the sourcing circuit, and also in the sinking circuit that it will damage the part to have a short-term connection through these transistors, whether CMOS or NMOS/HMOS. There's generally a specification for how long ONE pin can drive a short circuit to GND or to Vcc, too. That shouldn't be exceeded in any case.

Whether N-channel or P-channel, these mosfets behave like resistors when "ON." As they warm up, they become more resistive, not that this characteristic is, in any relevant way, important. If one turns the connection between row and column on only long enough to sample the state of the column signal, as shown in the subject schematic, and then turns the row off again, it won't be conducting significant current long enough to damage anything. If you leave it turned on all the time, you get what you deserve!

RE