Merge pull request matplotlib#6146 from maqifrnswa/master

API: ticker.LinearLocator view_limits algorithm changes

closes matplotlib#6142

Author: tacaswell

doc/api/api_changes/2015-03-14-SSH.rst

@@ -0,0 +1,16 @@
+``matplotlib.ticker.LinearLocator`` algorithm update
+```````````````````````````
+
+The ``matplotlib.ticker.LinearLocator`` is used to define the range and location
+of tickmarks of a plot when the user wants a exact number of ticks.
+``LinearLocator`` thus differs from the default locator ``MaxNLocator``, which
+does not necessarily return the exact user requested number of ticks.
+
+The view range algorithm in ``matplotlib.ticker.LinearLocator`` has been
+changed so that more convenient tick location are chosen. The new algorithm
+returns a plot view range that is a multiple of the user requested number of
+ticks. This ensures tick marks to be located at whole integers more
+constistently. For example, when both y-axis of a``twinx`` plot use
+``matplotlib.ticker.LinearLocator`` with the same number of ticks, the grids of
+both axis will be properly aligned at convenient tick locations.
+

lib/matplotlib/axes/_base.py

@@ -3762,7 +3762,8 @@ def twinx(self):
         create a twin of Axes for generating a plot with a sharex
         x-axis but independent y axis.  The y-axis of self will have
         ticks on left and the returned axes will have ticks on the
-        right.
+        right. To ensure tick marks of both axis align, see
+        :class:`~matplotlib.ticker.LinearLocator`
 
         .. note::
             For those who are 'picking' artists while using twinx, pick

lib/matplotlib/ticker.py

@@ -1348,10 +1348,11 @@ def view_limits(self, vmin, vmax):
             vmax += 1
 
         if rcParams['axes.autolimit_mode'] == 'round_numbers':
-            exponent, remainder = _divmod(math.log10(vmax - vmin), 1)
+            exponent, remainder = _divmod(math.log10(vmax - vmin),
+                                         math.log10(max([self.numticks-1, 1])))
             if remainder < 0.5:
                 exponent -= 1
-            scale = 10 ** (-exponent)
+            scale = max([self.numticks-1, 1]) ** (-exponent)
             vmin = math.floor(scale * vmin) / scale
             vmax = math.ceil(scale * vmax) / scale