Currently when we output the optimized bytecode, any jumps are referencing jumpdests from the unoptimized bytecode. We'll have to keep track of all the removed bytes and their locations so that any jumps which point to instructions after these values are adjusted accordingly