Repository navigation
joinstyle is not respected for FancyArrowPatch (either the path or the arrow heads) #13641
Description
Bug report
Bug summary
joinstyle is not respected for FancyArrowPatch (either the path or the arrow heads)
Code for reproduction
import matplotlib as mpl
import matplotlib.pyplot as plt
# high dpi so you can see the difference
fig, ax = plt.subplots(dpi=600)
# iterate through styles
for i, js in enumerate(("miter", "round", "bevel"), start=1):
arrow_head = mpl.patches.ArrowStyle.CurveFilledAB(head_length=4, head_width=2)
connection = mpl.patches.ConnectionStyle.Bar(fraction=0, angle=-90)
double_arrow = mpl.patches.FancyArrowPatch((0.1 * i, 1 - 0.1), (1 - 0.1, 0.1 * i), linewidth=2,
arrowstyle=arrow_head, connectionstyle=connection, joinstyle=js)
ax.add_patch(double_arrow)Actual outcome
Expected outcome
Different join styles for each line (including arrow heads)
Matplotlib version
- Operating system: Windows 10
- Matplotlib version: 3.0.2
- Matplotlib backend (
print(matplotlib.get_backend())): Qt5Agg - Python version:
Python 3.7.2 (default, Feb 21 2019, 17:35:59) [MSC v.1915 64 bit (AMD64)]
Type 'copyright', 'credits' or 'license' for more information
IPython 7.2.0 -- An enhanced Interactive Python. Type '?' for help.
I installed with the default conda channel.
Activity
ImportanceOfBeingErnest commentedon Mar 10, 2019
The joinstyle and capstyle are fixed for
FancyArrowPatch.matplotlib/lib/matplotlib/patches.py
Lines 4287 to 4289 in 21f1355
I don't know why that is the case. If we remove that line the result would look like this:
Is that what you're expecting? Would someone know why that got fixed in the first place?
ImportanceOfBeingErnest commentedon Mar 10, 2019
Ok, here is the deal (in a comment by @anntzer):
Actually, I don't quite see why we can't keep
"round"as default - in that way no baseline images need to change and it remains backwards compatible - but at the same time allow for user input to change it.ImportanceOfBeingErnest commentedon Mar 10, 2019
A possible fix is at #13647, no changes are made to the defaults and tests pass as they are.