Bug summary

The text.usetex rcParam flag can be used to enable TeX rendering on all Text elements instead of Mathtext. When using the pgf backend, I would expect this flag to have no effect and for the Text elements to be rendered identically since pgf also uses TeX to render.

The problem is that the usetex renderer seems to have a different implementation than the pgf renderer.

Code for reproduction

import matplotlib.pyplot as plt
import matplotlib as mpl

mpl.use('pgf')
plt.rcParams.update({
    'figure.dpi': 400,
    'text.usetex': True, # TODO: comment this out to get different results
    'ytick.labelsize': 1,
    })

# Create the figure and axis objects
fig, ax = plt.subplots(figsize=(2, 1.5))

# Plot the data
x = [1, 2, 3, 4, 5]
y = [2, 4, 6, 8, 10]
ax.plot(x, y)

# Set the title and axis labels
ax.set_title('My Plot')
ax.set_xlabel('X-axis')
ax.set_ylabel('Y-axis')

# Save the plot to a file
plt.savefig('my_plot.png')
plt.savefig('my_plot.pgf')

Actual outcome

When text.usetex is True, the tick labels use display math font to render in Tex.
When text.usetex is False, the tick labels use the regular text font to render in Tex.

Expected outcome

When using the pgf backend, I would expect enabling/disabling text.usetex to have no effect on the output.

Additional information

From what I understand, the text.usetex implementation has an unofficially supported rcParam for the preamble named text.latex.preamble. On the other hand, the pgf implementation has an officially supported rcParam for the preamble named pgf.preamble. From the documentation it's unclear to me which renderer/preamble takes precedence when both pgf and usetex are enabled.

The text.usetex renderer also does not support XeLaTeX/LuaLaTeX according to the documentation.

In conclusion:
Is there any reason why these two rendering engines can not (or should not) be unified? I suspect this is also the root cause of a lot of confusion and consistency errors people have.

Related:
#28084

Operating system

Ubuntu

Matplotlib Version

3.7.5

Matplotlib Backend

pgf

Python version

3.8.10

Jupyter version

No response

Installation

pip