DEPR: PeriodIndex.astype(dt64)

[jbrockmendel]

• closes #xxxx
• tests added / passed
• Ensure all linting tests pass, see here for how to run them
• whatsnew entry

Match the Series/EA behavior

[simonjayhawkins]

web and docs failing

2021-11-11T21:46:19.7568073Z WARNING: 
2021-11-11T21:46:19.7569269Z >>>-------------------------------------------------------------------------
2021-11-11T21:46:19.7570278Z Warning in /home/runner/work/pandas/pandas/doc/source/user_guide/timeseries.rst at block ending on line 2098
2021-11-11T21:46:19.7571558Z Specify :okwarning: as an option in the ipython:: block to suppress this message
2021-11-11T21:46:19.7573968Z ----------------------------------------------------------------------------
2021-11-11T21:46:19.7576034Z <ipython-input-382-a92cd46ab454>:1: FutureWarning: Converting PeriodIndex to DatetimeIndex with 'astype' is deprecated and will raise in a future version. Use `obj.to_timestamp(how).tz_localize(dtype.tz)` instead.
2021-11-11T21:46:19.7577328Z   pi.astype("datetime64[ns]")
2021-11-11T21:46:19.7578226Z <<<-------------------------------------------------------------------------

[jbrockmendel]

Yah, this was premature. I didn't fix the docbuild bc I was having second thoughts. Trouble is that we do allow dt64.astype(period), and we should either have both or neither. The part I have a strong opinion about is that IF we allow period.astype(dt64), it should be in PeriodArray.astype, not in PeriodIndex.astype

[jreback]

ok can revert for now if like