Unneeded and it just uses unnecessary resources.

Support universal wheels

Changed os.mkdir to os.makedirs to make it support nested directory creation

Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>

Added extra newline to test pr testing by travis

fixes issue GH-259.
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>

A MissingSectionHeader error now also reports the cached file

Reason: it doesn't only apply to URLs downloaded to the extends-cache, but
also to URLs downloaded as tempfiles.

Fixes #245 mostly

See #245

On windows, two files are generated (a something.exe and a
something-script.py). The if/else I used is used elsewhere in other tests,
too.

Enable downloading behind a proxy by using urllib2

As per @reinout comments in PR #253, let's make it clear that we're
assigning request._urlopener to a patched URL opener.

Check the 'use-dependency-links' option earlier.

This can give a small speed increase.

Before this, for each dist we would first check if it had a dependency
links option, which is almost always the case, but it is usually
empty: the package *will* have a dependency_links.txt in its EGG-INFO,
but there is only an empty line in it.

When you have a lot of packages and several buildout parts that use
them, the dependency links can be checked a lot of times.

On an extended Plone buildout with 'use-dependency-links=false', where
get_dist was called over 2500 times, the difference was 0.1 seconds,
so not much.  When I/O is slow, the difference could be more
noticeable.

Note that actually setting 'use-dependency-links' to false, already
helps much more.  For this buildout it shaved off 1.5 seconds.

Note that the 0.1 seconds win is because we do not have to call
'os.exists' for all those files, and the 1.5 seconds win is because we
do not have to actually read those files.

Also, for completeness sake, note that in this buildout get_dist was
called lots of times, but the list of dists that we checked for
dependency links in this method was always just a list of one.  So we
could have simply switched the order of the three parts of the
condition around with the same effect (and a smaller diff).  Still,
the way I did it now seems better.

And I am probably explaining myself much much more than is needed. :-)

Fixed bootstrap help text: --buildout-version.

Behind a proxy, `urllib` does not work with Python2 due to [Python issue
24599](https://bugs.python.org/issue24599).

Instead of using `urllib`, let's implement `urlretrieve` with `urllib2`, which
instead works and leave the Python3 implementation untouched (since that
one works too).

This fixes buildout issue #32.

There's no need to change the `try: except:IOError` block where `urlretrieve` is
used (line 195 in download.py) since the exceptions raised by `urllib2`
are subclasses of `IOError` and will then be caught.