Python
def process(input, part2=False):
if part2:
input = re.sub(r'don\'t\(\).+?do\(\)', '', input) # remove everything between don't() and do()
total = [ int(i[0]) * int(i[1]) for i in re.findall(r'mul\((\d+),(\d+)\)', input) ]
return sum(total)
Given the structure of the input file, we just have to ignore everything between don't() and do(), so remove those from the instructions before processing.
I did wonder if that might be the case, I must have been lucky with my input.