Revision history for Set-IntSpan-Fast-XS
0.01    2008-06-30
        Initial release.
0.02    2008-06-30
        Adaptive algorithm selection should mean we're always
        faster than Set::IntSpan::Fast.
        Implemented union in terms of _merge. Much faster. Also
        speeds up, e.g. intersection and anything else that
        calls union directly or indirectly.
0.03    2008-07-01
        Magic @ISA bashing allows us to be loaded automatically by
        Set::IntSpan::Fast.
0.04    2008-07-01
        New version of S::I::F makes @ISA bashing redundant.
0.05    2009-03-08
        Don't mask _list_to_ranges with out own incompatible version.