in the current implementation only the signals that are part of a message are set. The unused bits are not set. Which makes sense because there are not used.
But when using this codegen for SAE J1939 CAN messages this doesn't work. Because unused bits should be set to 1 according to this standard.
When looking at cantools for Python an option padding can be used during encoding of a messages. See:
https://cantools.readthedocs.io/en/latest/#cantools.database.can.Database.encode_message
My proposal is that padding will be an optional argument for the new function which is default false.