Re-implement binary_search in C

[orbitfold]

Here is how I would do it. It works. However there I don't know of a way to raise a Python exception from a pure C function. I can trying doing the others too and C if some kind of speed-up can be achieved.

Please feel free to reject this pull request this is just a proof of concept.

[wkerzendorf]

@orbitfold great! (I'm CCing @ssim and @mklauser here). I think there are a couple of issues we should find a work around: how do we raise an exception in a C-function? Maybe stackoverflow can help with this.
And how do we check if the code has gotten faster or not.

I think for this we can use a little tool called airspeed velocity asv made by a collaborator of mine. http://spacetelescope.github.io/asv/using.html shows that you can write profiling benchmarks very similar to pytest unit testing frameworks. A final benchmark can be seen here http://mdboom.github.io/astropy-benchmark/. I have not used this tool, but it seems like a good idea. If you have any other suggestions, feel free to try them.

[wkerzendorf]

ah, and Travis fails, before we could merge this, we would need to change the testing. If you have any questions about the pytest framework let me know.

[orbitfold]

http://docs.cython.org/src/userguide/wrapping_CPlusPlus.html

Apparently you can do it in C++ easily, it can catch C++ exceptions and propagate them to Python.

There is even this handy table:

C++                      Python
bad_alloc       MemoryError
bad_cast                TypeError
domain_error        ValueError
invalid_argument ValueError
ios_base::failure  IOError
out_of_range        IndexError
overflow_error     OverflowError
range_error         ArithmeticError
underflow_error   ArithmeticError
(all others)        RuntimeError

http://docs.cython.org/src/userguide/language_basics.html#error-return-values

In C it is possible this way

cdef extern int spam() except -1

So whenever the C function returns -1 it will raise a Python exception, apparently. I cannot test this right now though.

[wkerzendorf]

I don't know what the best way forward is. I think sticking with C for now seems logical, but can be convinced otherwise. That means that we could use the return -1 formalism.

[orbitfold]

Well C++ is (in theory) a superset of C, you can write C code and use C++ features as needed...

[orbitfold]

Hmm..

sudo python setup.py install

compiles and installs fine, while

python setup.py build

fails with

Generation of default configuration item failed! Stdout and stderr are shown below.
Stdout:

Stderr:
/usr/lib/python2.7/dist-packages/pkg_resources.py:1031: UserWarning: /home/vytas/.python-eggs is writable by group/others and vulnerable to attack when used with get_resource_filename. Consider a more secure location (set with .set_extraction_path or the PYTHON_EGG_CACHE environment variable).
  warnings.warn(msg, UserWarning)
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/usr/local/lib/python2.7/dist-packages/astropy-0.3.1-py2.7-linux-x86_64.egg/astropy/config/configuration.py", line 654, in generate_all_config_items
    for cfgitem in six.itervalues(get_config_items(nm)):
  File "/usr/local/lib/python2.7/dist-packages/astropy-0.3.1-py2.7-linux-x86_64.egg/astropy/config/configuration.py", line 540, in get_config_items
    __import__(packageormod)
ImportError: dynamic module does not define init function (initcmontecarlo)

running build_scripts

[orbitfold]

Finally passed the build and tests!

[wkerzendorf]

@orbitfold great work. We should merge this as soon as we have asv running - this way we can see if there was any speedup.

[orbitfold]

Unless there is something I missed I think this can be merged in.

[wkerzendorf]

Great! I'm happy with this, but we have a policy that at least two people need to look at it. @ssim or @mklauser can one of you look at this and merge if it's ready.

[ssim]

@orbitfold @wkerzendorf This looks good to me. I will merge it now (and add a line to the change log).