This is a pretty big simplification: the user should first use pip to
install everything in `requirements.txt`, and we no longer try to
cater to cases where Cython is missing, etc.
simplejson was there for python <= 2.6 compatibility.
It's not needed now, and installing it with pip can lead to segfaults
because pip is dumb.
(see e.g. https://github.com/simplejson/simplejson/issues/114)
We don't actually use nilmdb.utils.threadsafety.verify_proxy in the
main NilmDB code, but it's useful for finding errors.
It found an issue with __getattr__ in SerializerProxy which
(1) can't be avoided?
(2) is now commented in the code
(3) shouldn't matter in real use
__getattr__ is only called for attributes that don't exist; callers
should never purposely see the private __* symbols, so we don't need
to check for them.
Previous commits went back and forth a bit on whether the various APIs
should use bytes or strings, but bytes appears to be a better answer,
because actual data in streams will always be 7-bit ASCII or raw
binary. There's no reason to apply the performance penalty of
constantly converting between bytes and strings.
One drawback now is that lots of code now has to have "b" prefixes on
strings, especially in tests, which inflates this commit quite a bit.