Repository navigation
[Bug]: Constrained layout UserWarning even when False #23986
Description
Bug summary
When using layout settings such as plt.subplots_adjust or bbox_inches='tight, a UserWarning is produced due to incompatibility with constrained_layout, even if constrained_layout = False. This was not the case in previous versions.
Code for reproduction
import matplotlib.pyplot as plt
import numpy as np
a = np.linspace(0,2*np.pi,100)
b = np.sin(a)
c = np.cos(a)
fig,ax = plt.subplots(1,2,figsize=(8,2),constrained_layout=False)
ax[0].plot(a,b)
ax[1].plot(a,c)
plt.subplots_adjust(wspace=0)Actual outcome
The plot works fine but the warning is generated
/var/folders/ss/pfgdfm2x7_s4cyw2v0b_t7q80000gn/T/ipykernel_76923/4170965423.py:7: UserWarning: This figure was using a layout engine that is incompatible with subplots_adjust and/or tight_layout; not calling subplots_adjust. plt.subplots_adjust(wspace=0)
Expected outcome
no warning
Additional information
Warning disappears when constrained_layout=False is removed
Operating system
OS/X
Matplotlib Version
3.6.0
Matplotlib Backend
No response
Python version
No response
Jupyter version
No response
Installation
conda
Activity
tacaswell commented on Sep 22, 2022
Yup, that is indeed a bug
matplotlib/lib/matplotlib/figure.py
Lines 2428 to 2431 in e98d8d0
PR on the way.
tacaswell commented on Sep 22, 2022
@VanWieren Did you mean to close this? We normally keep bugs open until the PR to fix it is actually merged.
VanWieren commented on Sep 22, 2022
oh oops, I did not know that. Will reopen
tacaswell commented on Sep 22, 2022
No worries! Saying "I'll fix this" is way easier than actually fixing it (for example in this case I thought it would be super simple, but then I found a bunch of complications....).
By keeping the issue open we can track that it is not really fixed, the PR will link to this and we use the github cross-ref tools to automatically
FIX: do not set constrained layout on false-y values
FIX: do not set constrained layout on false-y values