include instructions for users of packaging systems (e.g. MacPorts).

…t for everything

… majority are grayscale http://www.research.ibm.com/dx/proceedings/pravda/truevis.htm

…verging colormaps for bipolar data

…faces

…ff the edge of the image.

…ctly dealt with.

hexbin_demo.py crashes with gtk backend

…ff" toward the end of the filename so files are more appropriately grouped when viewed in file managers and thumbnail tools (eg. gthumb)

….6. Continue to use histogram for earlier Numpy versions.

…ache the baseline images, since the test results are never the same twice given different dictionary orderings in the file.

… removes titles and ticks from a figure before comparison. Use this throughout to make more of the tests less dependent on text differences.

Changed wx.NamedColor to wx.NamedColour for compatibility with wxpython 2.9.

Update examples/user_interfaces/wxcursor_demo.py

Update examples/user_interfaces/mathtext_wx.py

Update examples/user_interfaces/embedding_in_wx4.py

Update examples/user_interfaces/embedding_in_wx2.py

Move Python 3 import of reload() to the module that uses it

GUI icon in Tkinter

Fixed logformatting for non integer bases.

.

Made tests faster

Test framework cleanups

Filled contours now use more of the available colormap range than
previously.

This change should have no effect on any plot, but it avoids
warnings that can arise when an inf occurs under a mask.

Previously, although vmin, vmax could be specified manually
in a norm used to initialize a ContourSet, the values would
be used only if both were specified.  Now one can set either
or both of vmin, vmax in the norm, and they will take precedence
over the default.

This changeset also simplifies and clarifies the vmin, vmax
explanation and handling for pcolor and pcolormesh.

…ntour

Formerly, colorbar would not place ticks at contour levels outside the
vmin-vmax range, and would not add line contour marks at such levels.

Contour norm scaling

…ib into pelson-multiaxes_colorbar

Conflicts:
	lib/matplotlib/colorbar.py
	lib/matplotlib/tests/test_colorbar.py

I also had to change some result_images because of the change in
contourf colormapping, and fix the default orientation code to
restore the "vertical" default.