Debian GNU/Linux 6.0.10 (squeeze) and OSX 10.11.4, Python 3.5.1, matplotlib 1.5.1 (anaconda 4.0.5)
Setting the prop_cycle for an axis breaks the usage of keyword aliases in the plot function. In the following snippets the keyword lw=0.1 is ignored, but linewidth=0.1 produces the wanted result.
Maybe it is worth mentioning that color is still accessible via its alias c after setting a prop_cycle.
If it seems weird to set the prop_cycle to just ignore it: I stumbled across this problem while playing around with stylesheets to set a default cycle through different linestyles. But sometimes you want to explicitly change this behaviour.
fig, ax = plt.subplots()
ax.set_prop_cycle('linewidth', [2, 3, 4])
for c in range(5):
ax.plot(np.arange(10), c * np.arange(10), lw=0.1)

fig, ax = plt.subplots()
ax.set_prop_cycle('linewidth', [2, 3, 4])
for c in range(5):
ax.plot(np.arange(10), c * np.arange(10), linewidth=0.1)

Debian GNU/Linux 6.0.10 (squeeze) and OSX 10.11.4, Python 3.5.1, matplotlib 1.5.1 (anaconda 4.0.5)
Setting the prop_cycle for an axis breaks the usage of keyword aliases in the plot function. In the following snippets the keyword
lw=0.1is ignored, butlinewidth=0.1produces the wanted result.Maybe it is worth mentioning that
coloris still accessible via its aliascafter setting a prop_cycle.If it seems weird to set the prop_cycle to just ignore it: I stumbled across this problem while playing around with stylesheets to set a default cycle through different linestyles. But sometimes you want to explicitly change this behaviour.