[Modification Proposal] - Cooperative Dual Task Space cpp-class

[sladalley]

In accordance with the project guidelines, and prior to opening a draft pull request, I am proposing the following changes to the Cooperative Dual Task Space (CDTS) cpp-class:

Modification 1: Virtualization of Methods

Member functions within the Cooperative Dual Task Space class to be declared virtual , enabling them to be overridden in derived classes.

Modification 2: Smart Pointer Integration and Backward Compatibility

In line with the architecture of the controller classes in the library, I propose introducing std::shared_ptr support for the robot instances, along with an appropriate constructor for initialization.
For backward compatibility, also including raw pointer access throughout the class by implementing appropriate accessors and updating the related members accordingly.

Modification 3: get_configuration_space function

Introduce a get_configuration_space method to retrieve the combined configuration space of the CDTS object.

[ffasilva]

Hi @dqrobotics/developers,

For context, @sladalley was implementing the classes for the extended cooperative dual task space and wanted to inherit from the DQ_CooperativeDualTaskSpace class. However, as she mentioned, DQ_CooperativeDualTaskSpace doesn't allow its methods to be overridden, as they are not virtual, and hasn't been updated to use smart pointers yet.

If you're okay with refactoring this class, Seyonne could submit a pull request with the proposed modifications, and I would be happy to review it.

Kind regards,
Frederico

[bvadorno]

Hi, @ffasilva and @sladalley,

In principle, some of these modifications would be okay, but they would need to match the MATLAB implementation. So, this cannot be a modification in the C++ version alone. Also, we'd need to see the class diagram for the current implementation versus the class diagram for the envisaged implementation. Additionally, the Python bindings should also be modified accordingly.

Refactoring to use a nice-to-have, modern implementation, such as smart pointers, should be a separate pull request. Pull requests with substantial modifications that are not associated with a single improvement will be rejected with no further review.

Instructions on how to contribute are here: https://github.com/dqrobotics/matlab/blob/master/CONTRIBUTING.md.

Kind regards,
Bruno