work around reversed(range()) bug by disabling optimised looping for non-trivial step values (ticket #763)

Add "nogil" to the stdc++ pxd files

NumPy is starting to seriously deprecating access to the member fields
in an ndarray (it was always frowned upon, but now it is starting to
become enforced). To support the large body of Cython code out there
accessing these fields (arr.shape[0] and so on), we special-case
PyArrayObject in Cython, with special knowledge of the NumPy API.

Ideally, we may introduce features in Cython in the future that allows
specifying this kind of magic with syntax in pxd files, and then we can
move away from special-casing NumPy.

DOC - partial rewrite of fused types doc

Ignore byte compiled elisp modules in the repository.

* When cython.git is included as a submodule in an
  Emacs configuration to load cython-mode.el and is
  byte-compiled, it generates an .elc file which causes
  the submodule to appear dirty unnecessarily in the parent
  git repository. Adding this rule will avoid dirty submodules.
Signed-off-by: Yesudeep Mangalapilly <yesudeep@google.com>

--HG--
extra : transplant_source : %28%90%91h.%EE%94%1AE%F5%A9%DA33J%D0%60h%F5t

After conversation with Mark F today. Change from pointers to scalars
for clarity.  Add quickstart section.  Explain case where the same fused
type in an argument list must be specialized to same type.

--HG--
extra : transplant_source : %9F%0B%FF%0F%16%0C%25%B9%FA7%D8%25%9EDpJSE%5E%BB

--HG--
extra : transplant_source : %AE%FF9%0ALO%CE%5E%C7%93%FA%B1d%E73%E8%CA%BE%DC%C8

--HG--
extra : transplant_source : %B2Q%A0%A7c%3B%EE%40%93%08%E1%C9-%83%88%BA%ACU%7D%8D

--HG--
extra : transplant_source : %D5V%9A%D2%DE%A7%A8%20%21%05%FFp%DEk%C4f%3Ct%0DT

Surprisingly, a global "nogil" did not work

libcpp.string supports s[0]= and s.at(0)=