Johannes Schindelin <Johannes.Schindelin@xxxxxx> writes: > The code would also be incorrect, as the `minutes` variable can be > negative, which the `unsigned_add_overflows()` macro cannot handle (it > would report very, very false positives, and it would hurt you more than > me because I live East of Greenwich). Yes and no. If we were to care about integer wraparound to do this check, we'd still have to worry about negative offset applied to a time very close to the epoch that turns into a timestamp far far in the future (unsigned_sub_underflows, that is).