AbstractRealInterval constructors about imglib2 HOT 3 CLOSED

Thanks for pointing these out, Barry! Both are clearly bugs. I just pushed the fix.

I fixed the copy constructor and removed the AbstractRealInterval(double[] dimensions) constructor.
The way we defined it, min and max are inclusive. So for (integer) Interval with min=0, dimensions=max+1.
Because every Interval is also a RealInterval, we decided that RealInterval has no dimensions() method (exactly because the dimensions=max+1 makes no sense here). Therefore the (double[] dimensions) constructor also makes no sense.

The alternative would have been to define min as inclusive and max as exclusive. Maybe we should have done that, now it would be a lot of work to change it (but not impossible)

from imglib2.

Thanks for the fixes Tobias. Shall we close this issue? Or did you want feedback on the exclusive max idea?

from imglib2.

Sorry, Barry. I just forgot. I'll close the issue now. Feedback on exclusive vs inclusive max is welcome though :)

from imglib2.