Possibly incorrect implementation of crc8 update
henriksod opened this issue · comments
In most practical implementations, especially if two nodes communicate with each other and needs to determine the data integrity by matching the given crc with the digested one, you cannot have a history dependent crc since there might be packet loss along the way and the two nodes do not share clocks.
I propose that a given initial_start
is provided to the CRC8
class, as done before. However, it should not keep the sum between updates.
Will reopen a new one with a better solution