move declarations of PyObject and PyTypeObject from ref.pxd to object.pxd and extend PyTypeObject with further struct fields

undo CheckExact broadening for RHS of optimised arithmetic expressions as Python would still evaluate it to get the right subtype as result

when it's clear that the LHS of an optimised arithmetic expression is exactly an int/long/float, apply the optimisations also when the RHS is an int/long/float subtype as it does not influence the result

let's assume PY_LONG_LONG isn't going to hold substantially more bits than 4*PyLong_SHIFT on any well configured system

grow optimisation for PyLong operations with constant C longs to use PY_LONG_LONG if it's larger than C long

PyPy: Improve negative overflow error messages in int type conversion

reimplement ord() to make its (C long) return value and input argument type (bytes/unicode/char/unichar) available to the type analysis and avoid redundant round-trips into Python space

move declaration of __Pyx_sst_abs() macro to TypeConversion.c instead of more high-level ModuleSetupCode.c