Seemingly unpredictable results when calling an overriden parent method on an instance of a child class casted to the parent class

Question

I have a parent class with a sub-vi Override.vi, and a child which overrides this sub-vi.  I create an instance of this child.  I cast this child to it's parent class and store it in an array.  Later, if I invoke the parent's 'Override.vi' on this child (casted to parent) then Labview 2013 seems to randomly choose whether to run the parent or the child override.vi.  In Labview 2011 SP1 it would always call the childs version of override.vi (which while surprising to me was very useful).  This has totally broken an application I have been developing, any insight as to how to control which override.vi is run would be helpful (re-casting to the child class isn't really an option, as there are in fact many child classes each with their own version of override.vi).

Answer 1

The actual data type of the wire is irrelevant in deciding which VI to run. The only thing that is relevant is the class of the object which is actually on the wire, so casting to the parent should not be relevant. *IF* the object really is a child, then LV should always call the child's VI, just like you say it works in 2011.
I suspect that what's happening in your case is that somewhere you're generating a parent and that's what's actually on the wire (e.g. maybe you have an error somewhere and a function outputs the default value, which is a parent). The fact that it didn't happen in 2011 doesn't mean it's a bug in 2013. It could be that something else has changed.
In any case, it's impossible to tell whether this is a misunderstanding, a bug in your code or a bug in LV without actual code. If you can post actual code which shows this, people can help. Otherwise (if it only happens in code you don't want to publish), you should try contacting NI directly so that you can at least show them the code.
Try to take over the world!