In my opinion, the Java 1.8 implementation has another issue with "Europe/Vienna". Yes, the time zone "Europe/Vienna" is special at 1980-04-06 in that the clock change happens at the first hour, so this day effectively starts at 1 o'clock local time, not at 0 o'clock as, to the best of my knowledge, anywhere else in the world even at the days of spring clock changes. But that's how Austria wanted it to happen back then, and that's how "Europe/Vienna" has been defined in the TZ / Olson database. This is what Java correctly shows, given the time zone has been set to "Europe/Vienna".

The real issue comes into play if you issue an Calendar.add(Calendar.DAY_OF_MONTH, 1), of course after having adjusted the Calendar instance to "Europe/Vienna". I'd expect the first hour of the next day to be 0 in local, daylight saving time, but Java computes 1, which is wrong.