DEPR: Deprecate xlrd and pyxlsb excel engines

[phofl]

We added calamine support for 2.2, so we can theoretically deprecate these 2 engines in favour of calamine to reduce our dependencies and just code that isn't needed anymore

[rhshadrach]

+1

In general I'm okay with supporting any "good enough" engine even if it means we have multiple engines for certain formats, but neither of these fit that criteria in my opinion. I would wait a bit for users to give feedback on calamine though. Also might think about deprecating the default to become calamine first, then deprecating these altogether?

[asishm]

I like the idea of making calamine the default first and then deprecating (similar transition had taken place when the default engine was changed for xlsx files)

[jbrockmendel]

For my edification, is calamine better in some way than the others? IIRC xlrd is not maintained; is that accurate?

[rhshadrach]

xlrd is not maintained; is that accurate?

Yes - and for pyxlsb, it doesn't offer some features that I would consider basic, like reading datetimes.