Currently, if a message has a valid header and an unsupported block id, we throw it away and continue looking for sync bytes. This has a performance penalty because we now parse out a byte at a time over what was known to be a valid message.
We could read out the remaining bytes in the valid message without parsing, but the proptest failure added in #46 presents a challenge to overcome if we choose to do this.
Currently, if a message has a valid header and an unsupported block id, we throw it away and continue looking for sync bytes. This has a performance penalty because we now parse out a byte at a time over what was known to be a valid message.
We could read out the remaining bytes in the valid message without parsing, but the proptest failure added in #46 presents a challenge to overcome if we choose to do this.