Open jklmnn opened 3 years ago
IMHO this should be rejected altogether. I guess we'll need a way to associate the parameter with a field of the outer message (or something completely different).
I agree that the given refinement is not suitable and should be rejected. It must be enforced that all parameters of an inner messages are defined using an expression (which could refer to fields of the outer message). For the given example, a valid refinement could be specified as:
for Test::Frame use (Payload => Test::Packet (Length));
Using a parameterized message as follows:
generates code that fails to compile with the following error: