Drop pandas 1.X support?

[fjetter]

pandas 1.X has been released about 10 months ago and pandas 3.0 is on our doorstep. We're keeping a lot of compatibility code around, both in tests and in actual code. Additionally, we soon want to switch the default backend of dask.dataframe to dask-expr which requires >=2.X.

Therefore, I want to propose to drop support for 1.X with the next release.

Thoughts or concerns?

[crusaderky]

Now dask[dataframe] has a hard dependency on dask-expr, which in turn depends on pandas>=2.0.

CC @jrbourbeau a conscious choice on this point should be a blocker fo the next release.