Use std::bitset instead of plain ints for bitfield flags

In many, many places in the code, we currently use uint32_t as places to hold collections of flags. C++ has a class, std::bitset<N>, that behaves like a uint32_t in the ways in which is matters, but has some nice features.

- It can be any length. set at compile time via template arg.
- It can do standard biwise operators <<, &=
- It can also do array-like access, so checking the status of a flag can be done with fields[MY_FLAG]
- It has several readable methods for manipulating things, such as .flip(), .set(), clear(), any(), none(), reset()

If these are used, then it is clear that a particular variable is intended to be used as a collection of flags.

Class overview can be viewed at: http://www.sgi.com/tech/stl/bitset.html