BUG: loc-indexing with a slice on a CategoricalIndex

[topper-123]

• followup to BUG: loc-indexing with a CategoricalIndex with non-string categories #29922
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

#29922 didn't handle loc-indexing using a slice on non-string CategoricalIndexes. This fixes that. E.g. we had:

>>> cat_idx = pd.CategoricalIndex([1,2,3], ordered=True)
>>> df = pd.DataFrame({"A": ["foo", "bar", "baz"]}, index=cat_idx)
>>> df.loc[1:2, "A"]
TypeError: cannot do slice indexing on...

The above works now.

[jreback]

thanks @topper-123 very nice