Embedded Python in C++ socket programming -

can not send , receive udp packets in single python embedded c++ program in send , receive scripts run in different threads? unhandled exception when exe file run. when 1 of send or receive command commented out i.e either pyrun_simplestring(sendpy) or pyrun_simplestring(recpy) program works fine. mistake here?

the code given below:

dword winapi sendpack(lpvoid ivalue) {   while(1){     const char* sendpy = "udpsocksend.sendto('10707',('10.107.35.167',21567))";    pyrun_simplestring(sendpy);      } return 0; }   dword winapi receive(lpvoid ivalue){   while(1){      py_initialize();     recpy = "data,addr = udpsockrcv.recvfrom(99000)";               pyrun_simplestring(recpy);  }  return 0;  }  int threads()     {      handle sendpackthread, receivethread;      dword dwgenericthread;           char lszthreadparam[4];       receivethread =      createthread(null,0,receive,&lszthreadparam,0,&dwgenericthread);               if(receivethread == null){                    dword dwerror = getlasterror();                    return 0;          }        sendpackthread =          createthread(null,0,sendpack,&lszthreadparam,0,&dwgenericthread);              if(sendpackthread == null){                    dword dwerror = getlasterror();                    std::cout<<"scm:error in creating send sample thread"<<dwerror<<"\n" ;                return 0;              }      return 1; }   int main(int argc, char* argv[]) {      using namespace std;     py_initialize();      const char * initpy = "import socket;      udpsocksend = socket.socket(socket.af_inet,socket.sock_dgram);      udpsockrcv = socket.socket(socket.af_inet,socket.sock_dgram);     listen_addr = ('',2000);udpsockrcv.bind(listen_addr)";     pyrun_simplestring(initpy);       int thd = threads();      system("pause");     return 0;  } 

thanks in advance

i think problem here is, python not thread safe. can't access interpreter 2 threads , expect works. see http://docs.python.org/2/c-api/init.html#non-python-created-threads more details. have acquire , release gil (the global interpreter lock).

this thing kind of mutex ensures, 1 thread accesses python objects @ time. gil reason why python multithread performance bad.