To have overlap detection work properly and to return reasonable error codes
to the user, CanHandleAreas() now first checks the alignment of all areas and
performs overlap detection afterwards.
I don't quite see what the advantage is in having an extra 'for' loop.
The error messages are still the same, aren't they?
Shouldn't
for (int i = 0; i < NumAreas; i++) {
if (Areas[i].x1 > Areas[i].x2 || Areas[i].y1 > Areas[i].y2 || Areas[i].x1 < 0 || Areas[i].y1 < 0)
return oeWrongAlignment;
for (int j = i + 1; j < NumAreas; j++) {
if (Areas[i].Intersects(Areas[j]))
return oeAreasOverlap;
}
}
do just the same?
Well, the difference is that your version might return early with
oeAreasOverlap without having checked Area[j] for correct alignment first.