If I avoid calling an OSCriticalSection in the constructor and just init the struct without locking access, things are OK. Furthermore, once the constructor is done, in other methods, I can successfully call OSCritical sections to protect access. But calling an OSCrit section from within the constructor produces an instant crash. I don't see why. I'd like to call it in order to safely init the gps struct. Am I missing something? I tried adding a short delay between Initing the OSCRIT object before calling the OSCritSection macro, but no luck. Partial code follows...
Thanks
Code: Select all
class GPS_C {
private:
bool bGPS_LockInited;
OS_CRIT GPS_Lock;
gps_t gps // a long struct of stuff needing inits...
public:
GPS_C(); //The constructor
//Methods for loading data into the gps struct object:============================
void StuffDataAfterParsingNMEA_GPGGA( gpgga_t & rGPGGA);
//etc...
//Methods for getting data out of the gps struct object:=======================
void GetNMEA_GPGGA( gpgga_t & rGPGGA);
//etc...
}
//The constructor for the GPS_ Class--------------------------------------
GPS_C::GPS_C() {
// init the OSCrit section only once...redundant, I know.
bGPS_LockInited = FALSE;
if (bGPS_LockInited == FALSE) {
if (OSCriticalSectionObj oscs(GPS_Lock);(&GPS_Lock) == OS_NO_ERR) {
bGPS_LockInited = TRUE ;
iprintf("GPS_Lock Init SUCCESSFUL\r\n" );
} else {
bGPS_LockInited = FALSE ;
iprintf("GPS_Lock Init FAILED\r\n" );
}
}
// now that GS_Lock object (of type OS_CRIT) is inited, lock the
// upcoming section of code in order to safely init the gps struct..
//
// The crash happens here:
OSCriticalSectionObj oscs(GPS_Lock);
// If the OSCriticalSectionObj line above is commented, things are OK, although the
// init below happens without any protection...
init_gps_struct( &gps, knots, degrees);
}