Import the ctypes module.

Small bjit improvements + fixes the build

Add tip on how to debug infinite recursive bugs from tp_ slots.

It wouldn't rerun cmake if cmakelists.txt got updated.

have the llvm jit start emitting some capi-style calls

The test was passing before by using the fallback behavior, but
importing ctypes successfully makes it try to use it and fails.

- Temporarily disable code for ctypes that requires class descriptors.
- Skip the attribute lookup for __new__ for the base type class and
  directly call a special function assigned to tp_new.
- Add a #define from CPython's ./configure
- Add test.

Eventually we'll want to have a more sophisticated way of
determining when to switch to capi exceptions; they do add
noticeable overhead when no exception gets thrown.

None are enabled in this commit, but add some of the
helpers for handling exceptions if they were to be thrown.

If both a metaclass and an instance of it are defined in C extensions,
the metaclass doesn't have attrs_offset != 0 or tp_dictoffset != 0
which causes the call to PyType_Ready on the instance to fail.

Refactors and cleanup in objmodel.cpp

Since it's one class and it's big enough, might as well have it's own
file.

It's not too big of a move, hopefully this won't cause merge conflicts
for anyone. Just part of an attempt to reduce the size of objmodel.cpp

All it does is a field access.

Force instantiate these templates as well

Add some PerThreadSet features

- some extra debugging for the threading_local issue
- remove the data for threads that don't come along with a fork()

Templatize runtimeCall for different exception styles

bjit: fix problems when setting bjit thresholds to 1

- we hit an assert when we ran out of scratch space
- 'classobjSetattr' returned void but all runtime funcs must return a Box* object.
- sneak in a change which removes a unused arg from '_setupCall'

Which also involves templatizing most of the things that it
calls as well.

CompiledFunction's now have an additional parameter "exception_style",
and we will try to call a version of the function with the requested
style.  All of our runtime functions are written to the CXX style,
and there's not a whole lot of benefit until we templatize the runtime
functions individually.