Moved __call__ functionality in subclasses to parent class
Slight PEP8 cleanup

Fixes Py2k compatibility

Add six import for consistency

The scale passed to the `pprint_val` method needs to also be
scaled by the log so that the range used to format the tick labels
matches the actual range of the tick labels (not the underlying values).

reported via http://stackoverflow.com/questions/33975758/matplotlib-logformatterexponent-e-in-the-exponent-labels-of-cbar

Makes sure that there are never trailing 'e' with no exponent.

Code used to generate the test data to establish bench mark.

First print out is human-readable, second is for testing

import itertools
import matplotlib.ticker as mticker
import numpy as np

domains = sorted([1e-3, 1.5e-2, 1e6, 100, 5, .5])
float_values = sorted([np.pi * (10**i) for i in range(-5, 6)])
int_values = sorted([1 * (10 ** i) for i in range(-5, 6)])
fmt = mticker.LogFormatter()
print()
for d in domains:
    print('Domain ', d)
    for f in float_values:
        print('     {: >10.7f}: {: <10}'.format(f, fmt.pprint_val(f, d)))
    for f in int_values:
        print('     {: >10g}: {: <10}'.format(f, fmt.pprint_val(f, d)))

print()
print()

print(',\n'.join(['        ({v:.10g}, {d!r}, {res!r})'.format(
    v=f, d=d, res=fmt.pprint_val(f, d))
                  for d, f in itertools.product(
                          domains, float_values + int_values)]))

Replace the old label, which included the legthy repr of the axes
object, to a nicer label which emphasizes the title or axes labels
(depending on what is present).  Also drops the use of the label of the
Axes object itself, for which I haven't found any other reference or use
in general.  (Another option would be to use solely `Axes.get_label()`
when defined, as this doesn't seem to have any other use.)

See #5468.

FIX: formatting in LogFormatterExponent

qt: raise each new window

Fix #5670. No double endpoints in Path.to_polygon

Nicer axes names in selector for figure options.

DOC: Update banner logo on main website

Fix errorbar extension arrows

Remove setup.py tests and adapt docs to use tests.py

DOC: Add documentation for mpl_toolkits.axes_grid1.inset_locator

Fix #5573: Use SVG in docs

This patch allows the user to interactively set an image's colormap from the
axes and lines (and now images) parameters editor (in the Qt backend).

Colormaps are listed from those registered with matplotlib.cm.  Colorbars seem
to be handled properly.

Image options

DOC: Improved documentation for FuncFormatter formatter class

if Patch.linestyle  is 'none', 'None', ' ', '' or None then no edge
rendered (as in Line2D).