username_1: I'm a little surprised that the source maps are so large, but you're right that we definitely need to optimize them. I don't know what the most efficient approach is for this in Node -- would the best bet be to create an intermediate file and have it read from there? username_0: You could just use a temp file as a cache instead of creating a separate one, then you wouldn’t need to worry about clearing it. I think that’s what Babel does anyway but I haven’t checked recently