Skip encoding defaults - #94

This adds an option to skip encoding fields if their value matches the default value. This is done on a best-effort basis - supporting generic == tests has a measurable performance cost. Rather, we check for common and efficient cases, opting for an efficient 90% solution rather than an inefficient 100% solution. In python code, the check would be something like:

def is_default(x, default):
    if x is default:
        return True
    if type(x) != type(default):
        return False
    if type(x) in (list, set, dict) and (len(x) == len(default) == 0):
        return True
    return False

If the value matches the default value, it's not encoded in the message body. For some messages with lots of optional fields, this can have a measurable impact on encoding an decoding performance.

Todo:

• Maybe pick a better kwarg name
• mypy support
• Docs
• Tests
• Benchmark changes to ensure negligible performance slowdowns in happy path