[Bug]: Axes.cla() with sharex may wrongly reset the axes limit.

Nobody has claimed this yet.

### Bug summary

For example, when subplot A shares the x-axis with subplot B (A is created with sharex=B), a call to `B.cla()` will set the xlimit to [0, 1], which may not be expected (in contrast, a call to `A.cla()` will not change the xlimit).

I think the main problem comes from the following function, where it will check whether its `_sharex` attribute is `None`; however, in this case, B is shared to A, but no axis is shared to B, so B's `_sharex` attribute is `None`.
https://github.com/matplotlib/matplotlib/blob/9e18a343fb58a2978a8e27df03190ed21c61c343/lib/matplotlib/axes/_base.py#L1369

### Code for reproduction

```Python
ax = plt.subplot(211)
ax2 = plt.subplot(212, sharex=ax)
ax.plot(range(50))
ax2.plot(range(50))
ax.cla()
plt.show()
```

### Actual outcome


### Expected outcome


### Additional information

_No response_

### Operating system

_No response_

### Matplotlib Version

3.8.3

### Matplotlib Backend

'TkAgg'

### Python version

3.9.13

### Jupyter version

_No response_

### Installation

pip

Read lib/matplotlib/axes/_base.py at the linked Axes.cla() code, then run the supplied subplot reproduction with sharex. Trace how the shared-axis relationship is represented when B is shared to A, and make the behavior match the expected outcome without resetting the x-limit. Confirm the reproduction no longer changes the limit unexpectedly.

Written by the indexing model from the issue text.