Uninitialized variable possible in EVS_SendViaPorts

[markhermeling]

https://github.com/nasa/cFE/blob/e3b67674630cd098dbdefdc4e727c04a2b23dc31/modules/evs/fsw/src/cfe_evs_utils.c#LL545C3-L545C3

There is a possibility of an uninitialized variable in EVS_SendViaPorts in cue_evs_utils.c. I do not have a running example, problem found using static analysis with GrammaTech CodeSonar, screenshot attached.

Line 545 calls CFE_MSG_GetMsgTime(CFE_MSG_PTR(EVS_PktPtr->TelemetryHeader), &PktTime);
this function is in cue_msg_sechdr_time.c
however, if either the first, or the second parameter is NULL, then that function will return CFE_MSG_BAD_ARGUMENT on line 81 of that file. Second parameter is unlikely to be NULL as it is a pointer to a local variable from EVS_SendViaPorts().

Which means it does not initialize PktTime, which is subsequently used in the CFE_TIME_Print call.

This may not be a problem if the compiler initializes local variables to 0, but there is no guarantee on that.

PktTime is used in a time calculation to convert epoch to calendar time, so worst case this could take a long time, and/or print garbage.

Probably would be better to check the return of the call to CFE_MSG_GetMsgTime and set PktTime to 0 if CFE_MSG_BAD_ARGUMENT is returned.