[BUG]: Y-axis autorange stops working when updating the xaxes range for subplots with shared x-axes and y-axes

Description

Making a subplot figure with shared_xaxes = True, shared_yaxes=True and updating the figure to with an explicit xaxis_range resets the y-axes ranges to (-1, 4).

• This only happens of both the x-axes and y-axes are shared, and only when updating the x-axes range.
• It only happens if you actually have more than one row, and more than one column in the figure.
• Tried with numpy inputs and plain python lists - same result.

Steps to reproduce

import numpy as np
import plotly.graph_objects as go
from plotly.subplots import make_subplots

fig = make_subplots(rows=2, cols=2, shared_xaxes=True, shared_yaxes=True)
x = np.linspace(0, 100, 200)
y = 10 * x
fig.add_scatter(x=x, y=y, row=1, col=1)
fig.add_scatter(x=x, y=y, row=1, col=2)
fig.add_scatter(x=x, y=y, row=2, col=1)
fig.add_scatter(x=x, y=y, row=2, col=2)
fig.update_xaxes(range=(0, 100))
fig.show()

Start with the provided make_subplots reproduction using shared_xaxes and shared_yaxes, then trace the update_xaxes(range=(0, 100)) entry point in plotly.js. Compare autorange behavior for the 2×2 subplot case with cases that do not share both axes. Done means the explicit x-axis range no longer resets the y-axis ranges and the regression is covered by a test.