RE: DATE feature

> Can anyone confirm that when a DATE for a Leap day e.g 29/0feb has > years added to it the result is garbage, but if the addition of years > is multiples of 4 then everything is OK. We had this problem under 4GL v4.1 SCO Unix. It was not just Feb 29th. Any date that had months or years added that would have given a non existant day resulted in garbage. So adding Months to the 31st of something caused problems too. You could argue this is not a bug(!) since Informix would have to choose which way to correct the date. This is what we did: We had a date calculating function. It had a list of days in the month and could determine which years were leap years. It added days first (which always works) then determined if the resulting year and month have a day of that number. If they did it set the date variable, if not it reduced the day number to the last day of the month in the required month and year.