... very nice Stefan, could we have a test or spec please (as this is easy to break at some point) ?

Added a small test (and reenabled some existing time related regression tests which pass) but i'm not sure if this is the best way to test.

@MSNexploder thanks, its not bad (you can leave it in) but does not actually test for the reported issue,
a new test which would test smt like :

t = Time.local(2000, 1, 1) + (300 * 366 * 24 * 60 * 60)
assert_equal 2300, t.year

... would probably do, what do you think?

@kares That was too easy 😎 Was focusing way to much on the underlying problem.

cool thanks a lot ... @headius could you double-check this is good to go for @MSNexploder, please?

Looks good, thank you!