RE: (ROSE) virtual methods

[Aleksey Voitovich]

Ok, I see, thank you.

 

But I hoped maybe some feature could be in C++ add-in, that links any inherited method of a child class with corresponded method of the parent class (or "grand-parent"). This option could help in checking that changed signature will affect on child classes. And this is not a functon of UML, but C++ only.

Is it reasonable?

 

Best Regards,

 

Aleksey

-----Original Message-----
From: Aker, Eric [mailto:[EMAIL PROTECTED]]
Sent: 16 ����� 2001 �. 20:25
To: 'Aleksey Voitovich'; Rose Forum (E-mail)
Subject: RE: (ROSE) virtual methods

Yes you are correct in your description of how Rose works.

The reason is that you do not have to implement the virtual operator in the child class.

You may have a grandchild that you want to be the first implementation of the operation.

 

You are correct that Rose does not have any internal link from an abstract operator.

Rose is a language independent tool and the rules of the language determine the rules

of how abreact is handled.

 

Eric

-----Original Message-----
From: Aleksey Voitovich [mailto:[EMAIL PROTECTED]]
Sent: Friday, March 16, 2001 9:04 AM
To: Rose Forum (E-mail)
Subject: (ROSE) virtual methods

Dear all,

 

I'm newcomer in this forum, so maybe my question had been already discussed.

 

but, if I create a Class A having some virtual (or even abstract) methods, and then Class B, that generalize the class A, I get all public and protected methods of A inherited in the B. That's great. But I also want to generate a C++ code of these classes. And to implement those virtual methods in B, I should create them again in B with the same signatures. This is not very convenient. But worse, if I later change signatures of virtual methods in class A, I could forget to change them in class B.

Is this my mistake or it has a conceptual explanation ?

 

hope someone could help me...

Thank you in advance,

 

Aleksey Voitovich