This was a different bug than the one in issue 70:
When translating a delete effect of an operator that does not have an add effect
on that variable but a precondition on a different value, we do not want to
represent the delete effect. At this point there was a bug in the code that did
not properly remove the delete effect from the effect list, but only set the
effect precondition list to an empty list. This lead to a side effect that
removed the precondition on the variable.
|