2008 Leap second bug fix
The most important fix in this version is that it now correctly places a leap second at the end of December 2008. Before, due to bad data scraped from NIST, the leap second occured in March 2009, not even synchronized to the DUT1 sign change!
This is the bad NIST data that I incorporated without sufficient checking:
A synthesized DUT1 offset of +400ms is added from 2009-1-1 through 2009-3-11 (inclusive).
A unittest is added for the DUT1 offsets that signals an error if an implied leap second occurs other than in the last day of a month. This test, as well as better manual inspection of the list of leap seconds embedded in a test, could have caught the problem earlier.
1972 leap second added
Simulated DUT1 data back to June 1, 1972 has also been added, which brings in the June 1972 leap second as well.
`wwvbgen` commandline program
The `wwvbgen` commandline program learned to take time strings in `dateutil.parser.parse` format, such as `wwvbgen "2020-6-14 21:25"`. Such time strings must be shell-quoted if they contain spaces. Negative leap seconds can now be forced from the commandline explicitly.
Internally, `wwvbgen` now uses click instead of deprecated optparse. Full compatibility is not guaranteed.