MISRA Discussion Forums

There shall be no more than one definition of each virtual function on each path through the inheritance hierarchy.

If I'm interpreting this rule correctly, once you define a virtual function for a derived class with the virtual base class as an ancestor, no more derivations in the chain can override the first definition.

This would seem to cripple a key advantage of using classes and an object-oriented design.

Isn't it pretty common to override a virtual member function in a child class even if the parent class has provided a different definition? Please explain.

This is intentional and, as stated in the text for the rule, is done to increase clarity for review and maintenance. It would be acceptable to deviate if a more complex hierarchy is required.

We will review this when working on the next version of MISRA C++.

Under the impression of MISRA Compliance i came up with two deviation permits so far:

• Reliability (Fault tolerance): Implementing default/fallback behaviour/strategy for polymorphic types
  
• Maintainability (Reusability): Extending base class functionality for polymorphic types

Thanks