[Bug]: [Jupyter][Windows] Axes.bar() causes the Jupyter kernel to terminate with no traceback #32218

Bug summary

Description

I found a reproducible issue where Axes.bar() causes the Jupyter kernel to terminate on Windows.

The same plotting operation works normally in a standalone Python process using the same Conda environment.

Environment

• OS: Windows x64
• Python: 3.12.13
• Matplotlib: 3.11.1
• NumPy: 2.5.2
• IPython: 9.16.1
• ipykernel: 7.3.0
• pyzmq: 27.1.0
• FreeType: 2.14.3
• Installation: Miniforge / conda-forge

Python executable:

E:\Software\miniforge3\envs\optimization\python.exe

Minimal reproduction

In a fresh Jupyter notebook using the Python (optimization) kernel:

import matplotlib
import matplotlib.pyplot as plt

print("Matplotlib:", matplotlib.__version__)

fig, ax = plt.subplots()

print("before bar")

ax.bar(
    [1, 2, 3],
    [1, 2, 3]
)

print("after bar")

plt.close(fig)

### Code for reproduction

```Python
import matplotlib
import matplotlib.pyplot as plt

print("Matplotlib:", matplotlib.__version__)

fig, ax = plt.subplots()

print("before bar")

ax.bar(
    [1, 2, 3],
    [1, 2, 3]
)

print("after bar")

plt.close(fig)

Actual outcome

Matplotlib: 3.11.1
before bar

Expected outcome

Matplotlib: 3.11.1
before bar
after bar

Additional information

The issue occurs when running Matplotlib in a Jupyter kernel on Windows.

In the same Jupyter kernel, the following operations work normally:

• importing matplotlib.pyplot
• creating a figure with plt.figure() / plt.subplots()
• ax.plot()
• ax.scatter()
• plt.close()

However, ax.bar() causes the Jupyter kernel to terminate immediately, without a Python exception or traceback.

I also tested the non-interactive "Agg" backend in Jupyter, and ax.bar() still caused the kernel to terminate.

In contrast, the same bar() operation works normally when executed from a standalone Python process using the same Conda environment.

I have not tested an earlier Matplotlib version yet, so I cannot confirm whether this is a regression.

I do not know the cause or a fix yet.

Operating system

Windows 11 x64

Matplotlib Version

3.11.1

Matplotlib Backend

qtagg

Python version

3.12.13

Jupyter version

4.6.3

Installation

conda