genhol defect in 1931

[christophsax]

This seems wrong:

genhol(easter, start = -46, end = 0, frequency = 12)

You obtain:

           Jan        Feb        Mar        Apr        May        
1931 0.00000000 0.23404255 0.00000000 0.10638298 0.00000000

and the expected value for Mar is: 0.659574468

It seems that it only happens with Mar in 1931 (maybe because it is the first year of the series, according to vector Easter).

[christophsax]

This probably should be rewritten, rather than fixed.

The defect is only affecting the first year, and only if you choose a very long easter effect. So it is a quite exotic setting. Low prio.