implement tp_str and tp_repr for more of our builtin types

Have function.func_globals always be dict-like

Internally we store it as either-dict-or-module, but when we expose
it to Python it needs to be given back as a mapping.  I looked briefly
into switching the internal representation, but that seemed like a much
more invasive change.

Also, add a pytest test that exposed this issue.

Add sq_slots & mp_slots to instance (issue #1197)

str: use cpythons [r]partition

it support non str separators
fixes #1297

Bump version numbers

docker: we are now using cython 0.24 + update readme for 0.5.1 release

str does not have `__iter__`

which hits an external server

use r12, r15 and rbx in bjit and inside bjit ICs, remove some nonzero checks

when possible

bjit: use a bitset for register tracking in the rewriter and use r12, r15, rbx, rbp in PPs and the bjit
Keeping the available registers in a bitset makes it more memory efficient and
also easier and more performant to calculate a subset of the registers.
I will soon implement the 'otherThan' functionality using it which would fix the current problem of only allowing to exclude one register

keep track of which variables are known to be a python bool so that we can remove some of the nonzero() calls

at somepoint we really should split all this common code into a function...

Some SciPy fixing.

This is hacky, but delve into scipy souce code to fix the numerous
implicit function declaration is not make sense for now.

Because scipy rely on NumPy, so add scipy test to numpy_test.py. And
rename it to scipy_test.py. numpy_full_test.py will check numpy test
suite. The scipy_test.py only for check scipy test suite. But scipy test
will take a long time. So disabled it by default.

In PyDict_Next, the 3rd argument `pkey` and 4th argument pvalue could be
NULL. lxml will try to call it by PyDict_Next(kwdict, &pos, &key, 0).
And a check like CPython did.

Cython try to store argcount, nlocals, and varnames through PyCode_New.
But Pyston don't do anything with those or support getting them back
out.

- special-case classes that we know have a fixed length (ex str)
- do patchpoints for getIterHelper

(it uses old-style iteration)
And crazily, there is code that relies on this.

I wouldn't usually want to give in to code relying on such an
esoteric feature, but this seems to be a speedup as well (though
that deserves looking into, itself).

Some scipy fixes

I had spelled the test name wrong

The repro is:
- import a.b, and in a.b do
- import a.b as ab

This will cause an attribute error, since a.b (the name) is successfully importable,
but a.b (the attribute) doesn't exist yet.