DatetimeIndex doesn't validate dtype argument

[TomAugspurger]

In [2]: pd.DatetimeIndex([1, 2], dtype=object())
Out[2]: DatetimeIndex(['1970-01-01 00:00:00.000000001', '1970-01-01 00:00:00.000000002'], dtype='datetime64[ns]', freq=None)

this should probably raise.

[jbrockmendel]

I’ll work this in to the next PR ( after the current DTA constructor PR goes through). There is a dtype validation step, but it is only checking that conflicting time zones aren’t being passed

[alexcwatt]

I believe this can be closed; @TomAugspurger's example does raise an exception now. The dtype is validated with _validate_dt64_dtype to ensure it is either datetime64[ns] or DatetimeTZDtype.

[mroeschke]

Probably worthwhile to add a regression test.