fix #759 - removes GetLastSenderId()

[CDKnightNASA]

Fixes #759
Fixes #608

Removes CFE_SB_GetLastSenderId(), which is supremely broken.

Testing performed
sb_UT passes

Expected behavior changes
Removes the CFE_SB_GetLastSenderId() API.

System(s) tested on
Debian 10.3

Contributor Info - All information REQUIRED for consideration of pull request
Christopher.D.Knight@nasa.gov

[skliper]

Likely needs a requirements update.

[CDKnightNASA]

Likely needs a requirements update.

Done. I can squash prior to CCB if that's preferred.

Looking at https://github.com/nasa/cFE/blob/master/docs/cfe%20requirements.docx I have some architectural questions about the SB in general in a multi-processor environment. Perhaps we can discuss at a CCB meeting or have a splinter.

[skliper]

@dmknutsen Heads up this change (assuming stakeholders accept it) will need to be reflected in JIRA, with impacts to test cases.

[jphickey]

The requirement necessitates a different discussion than the API. Reading that requirement text/rationale, it actually seems reasonable to me, but the problem is the implementation didn't match, so I'm not sure it should be removed.

The CFE_SB_GetLastSenderId() didn't actually implement the function described or fulfill this requirement. It was tracking/returning the wrong thing and was not adequate for the rationale described in the requirement text. That's why we need to remove/deprecate this existing API - it doesn't match.

Is it OK to keep a requirement but note that it is not met in the current version(s)? This wouldn't really be a change - as CFE_SB_GetLastSenderId() has never met this requirement - so in theory every CFE version to date has not actually met this requirement.

There are three things we need to do:

1. Remove/deprecate the existing API as it doesn't correctly implement/fulfill a requirement
2. Revalidate/discuss the requirement itself and determine if it is still applicable or needs an update
3. Implement whatever was the outcome of (2) above.

This PR should only do item 1, nothing else at this time. There is certainly a requirements aspect to this, but I wouldn't jump to the conclusion of removing the existing requirement.

[jphickey]

Also, if my understanding is correct, the requirements CSV file is an export from a Jira repository maintained internally by GSFC - so we should not change/modify the CSV file directly anyway.

I'm good with commit 2a17ed2 alone.

[astrogeco]

@CDKnightNASA want to discuss today?

[astrogeco]

CCB 2020-07-01: Will have a design discussion later. Needs some more conversations about Omit_deprecated, certification, etc.

Add a short documentation change explaining that the function does not work as expected.

[CDKnightNASA]

replaced by #784