Apps should not pend forever on software bus

[jphickey]

The SAMPLE_APP, which is supposed to exemplify the current best practices, shouldn't be pending forever on a software bus message as it does here:

sample_app/fsw/src/sample_app.c

Lines 85 to 87 in b956292

	status = CFE_SB_RcvMsg(&SAMPLE_AppData.MsgPtr,
	SAMPLE_AppData.CommandPipe,
	CFE_SB_PEND_FOREVER);

The problem with pending forever is that the app also needs to perform an orderly exit if a restart/reload/delete command is sent to ES, which is checked by CFE_ES_RunLoop(). If an app is pending forever for software bus messages but none are sent, then the shutdown request will remain pending indefinitely.

[jphickey]

Update - I did test this and confirm that the ES will actually forcibly delete the task via an OS_TaskDelete if it doesn't exit itself within the delay period. So the app actually will still get restarted, but forcibly deleting an app via that method is not good because it will leak any resources that the app was using at the time it was deleted.

As a best-practice example, it should use a time-limited pend on the software bus, so it can also poll the CFE_ES_RunLoop() function.

[jphickey]

Calling this an "enhancement" rather than a "bug" because it is also saved by the fact that SCH_LAB by default sends a message every few seconds, so while less than ideal, its not fully broken.

[skliper]

Added milestone, feel free to fix if you want