Add new Svc::ComRetry component

[valdaarhun]

Add new component that can be used by other components to resend a message on failure.

Related Issue(s)	#3994
Has Unit Tests (y/n)	y
Documentation Included (y/n)
Generative AI was used in this contribution (y/n)	n

---

Change Description

Add a new component - ComRetry.

Rationale

Other components such as ComQueue can be connected to ComRetry to resend a message on failure.

Testing/Review Recommendations

Future Work

• Add tests
• Add CMakeLists.txt
• Add docs

AI Usage (see policy)

NA

[valdaarhun]

Hi. Apologies, it took much longer to draft this change. A lot has to be refined and the tests have not been added. I had a few questions and I thought I would get them clarified before continuing.

• I am not sure if this is the right way to store the buffer in dataReturnIn_handler.

this->m_buffer = buffer

I was initially thinking of using DpContainer. Based on what I have understood from the documentation though, it would be an overkill to use it here.

Should Fw::ExternalSerializeBufferWithMemberCopy be used instead to store the buffer (similar to what's done in DpContainer::setBuffer())?

• Once I get more clarity on the question above, I'll add in a check in comStatusIn_handler to test the validity of the buffer. I have marked it as TODO for now.
• I am not sure how the following would work:

    else {
        this->dataReturnOut_out(0, this->m_buffer, this->m_context);
        this->m_retry_count = 0;
        Fw::Success condition = Fw::Success::FAILURE;
        this->comStatusOut_out(0, condition);
    }

Based on this comment:

1. Will dataReturnOut here be connected to dataReturnIn in the ComQueue?
2. Similarly, will comStatusOut be connected to ComQueue's comStatusIn?
3. Given that comStatus is an async port while dataReturn is a sync port, can it be said that dataReturnIn will always be called before comStatusIn?
4. How does this work under the hood? I tried reading through the relevant functions in ComQueueComponentAc.cpp but couldn't really grasp it.
5. If I placed comStatusOut_out before dataReturnOut_out in the snippet above, would that make a difference?

[LeStarch]

Minor changes, but this is what I was expecting.

[LeStarch]

First, you are storing things as I would expect. No need to do more. Second, your validity check isn't necessary either, as this component doesn't alter the validity.

To answer your enumerated questions:

1. yes
2. also yes
3. yes, given that you have comstatus after data return in.
4. You know that dataReturnOut finishes before comstatus is ever sent because dataReturnOut is synchronous and thus finishes all the work before the end of the function call.
5. Yes it would. You have the expected order and should not reorder those calls.

[valdaarhun]

First, you are storing things as I would expect. No need to do more. Second, your validity check isn't necessary either, as this component doesn't alter the validity.

To answer your enumerated questions:

1. yes
2. also yes
3. yes, given that you have comstatus after data return in.
4. You know that dataReturnOut finishes before comstatus is ever sent because dataReturnOut is synchronous and thus finishes all the work before the end of the function call.
5. Yes it would. You have the expected order and should not reorder those calls.

Hi. Thank you for the answers.

Regarding the validity check, I was thinking that would take care of the case when ComRetry receives Fw::Success, but there's no buffer stored in the component.