Infinite Loop: LogLocator Colorbar & update_ticks

[ocehugo]

• [ 1.5.1,3.5.1,OSX] Matplotlib version, Python version and Platform (Windows, OSX, Linux ...)
• [pip ] How did you install Matplotlib and Python (pip, anaconda, from source ...)

import matplotlib.pyplot as plt
import matplotlib as mpl
from matplotlib.colors import LogNorm
import numpy as np
from matplotlib.mlab import bivariate_normal


N = 100
X, Y = np.mgrid[-3:3:complex(0, N), -2:2:complex(0, N)]

# A low hump with a spike coming out of the top right.
# Needs to have z/colour axis on a log scale so we see both hump and spike.
# linear scale only shows the spike.
Z1 = bivariate_normal(X, Y, 0.1, 0.2, 1.0, 1.0) + 0.1 * bivariate_normal(X, Y, 1.0, 1.0, 0.0, 0.0)

plt.subplot(2, 1, 1)
plt.pcolor(X, Y, Z1, norm=LogNorm(vmin=Z1.min(), vmax=Z1.max()), cmap='PuBu_r')
cbar = plt.colorbar()
cbar.locator = mpl.ticker.LogLocator(numticks=1)
cbar.update_ticks()

The code above follow from the pcolor_log.py example.

PS: I would be great if mpl include a LogMaxNLocator just like the linear one.

[tacaswell]

The simple 'fix' is that to get what you want with numticks=1 I think you really want to pass numticks=2

[ocehugo]

I know, but it's easy to forgot that the numticks argument is in fact numticks-1.

No one likes to restart their notebooks because of a colorbar hehe .

[tacaswell]

Well, the solution I am looking at will just raise instead of infinite loop which is slightly better.

[yongzhez]

So I was wondering what would be a preferred solution. I would be guessing that there's a better form of validation for input that would filter out ALL possible unwanted ticker values? I apologize for the pull request since I didn't see that you already added a fix similar to mine.

[tacaswell]

I am hesitant to invest too much in validation of this sort of thing right now. On of the mid-term (by summer) goals is to get the entire library traitlet-ified which will provide much greater more flexible validation for one giant upfront cost.