It only parses as little data as possible to get just file names, on some files like ZIPs it'll be just the header but on others like tar it'll have to walk through and seek to the start of each file. It should be pretty fast even on big files though