//	mtmagicserver.cpp//	Multi-threaded magic squares server//	Runs as CGI (not multi-threaded) or multi-threaded standalone//	Computation of the magic square has been deliberately slowed down to//	demonstrate work load issues (see SLEEP constant)// Run from the command line with arguments IP (which must be the// IP of the current machine you are using) and PORT to run this as a// multi-threaded stand-alone server on a port. For example:// > mtmagicserver.cgi 18081// To let 'magic' client talk to this service, change the URL in code magic.cpp// into "localhost:18081"// This example illustrates two alternative server implementations with threads.// The first implementation recycles gSOAP resources but is bounded to a maximum// number of threads. Each thread needs to be joined, so runaway processes will// halt the server at some point.// The second implementation has no thread limitation. Runaway threads are not// controlled.#include "soapH.h"#include "magic.nsmap"#include <unistd.h>	// import sleep()#include <pthread.h>#define BACKLOG (100)	// Max. request backlog#define MAX_THR (8)	// Max. threads to serve requests#define SLEEP	(0)	// use this to make each thread sleep to mimic work load latency////////////////////////////////////////////////////////////////////////////////////	Multi-Threaded Magic Squares Server//////////////////////////////////////////////////////////////////////////////////void *process_request(void*);int main(int argc, char **argv){ struct soap soap;  soap_init(&soap);  // soap.accept_timeout = 60; // die if no requests are made within 1 minute  if (argc < 2)		// no args: assume this is a CGI application  { soap_serve(&soap);	// serve request    soap_destroy(&soap);// cleanup class instances    soap_end(&soap);	// cleanup  }  else  { pthread_t tid;    int port;    int m, s, i;    port = atoi(argv[1]);    m = soap_bind(&soap, NULL, port, BACKLOG);    if (m < 0)    { soap_print_fault(&soap, stderr);      exit(1);    }    fprintf(stderr, "Socket connection successful %d\n", m);    for (i = 0; ; i++)    { s = soap_accept(&soap);      if (s < 0)      { if (soap.errnum)          soap_print_fault(&soap, stderr);	else	  fprintf(stderr, "Server timed out\n");        break;      }      fprintf(stderr, "Thread %d accepts socket %d connection from IP %d.%d.%d.%d\n", i, s, (int)(soap.ip>>24)&0xFF, (int)(soap.ip>>16)&0xFF, (int)(soap.ip>>8)&0xFF, (int)soap.ip&0xFF);      pthread_create(&tid, NULL, (void*(*)(void*))process_request, (void*)soap_copy(&soap));    }  }  soap_done(&soap);  return 0;}void *process_request(void *soap){ pthread_detach(pthread_self());  soap_serve((struct soap*)soap);  soap_destroy((struct soap*)soap);  soap_end((struct soap*)soap);  soap_done((struct soap*)soap);  free(soap);  return NULL;}////////////////////////////////////////////////////////////////////////////////////	Magic Square Algorithm//////////////////////////////////////////////////////////////////////////////////int ns1__magic(struct soap *soap, int n, matrix *square){ int i, j, k, l, key = 2;  if (n < 1)    return soap_sender_fault(soap, "Negative or zero size", "<error xmlns=\"http://tempuri.org/\">The input parameter must be positive</error>");  if (n > 100)    return soap_sender_fault(soap, "size > 100", "<error xmlns=\"http://tempuri.org/\">The input parameter must not be too large</error>");  square->resize(n, n);  for (i = 0; i < n; i++)    for (j = 0; j < n; j++)      (*square)[i][j] = 0;  i = 0;  j = (n-1)/2;  (*square)[i][j] = 1;  while (key <= n*n)  { if (i-1 < 0)      k = n-1;    else      k = i-1;    if (j-1 < 0)      l = n-1;    else      l = j-1;    if ((*square)[k][l])      i = (i+1) % n;    else    { i = k;      j = l;    }    (*square)[i][j] = key;    key++;  }  sleep(SLEEP);		// mimic work load latency  return SOAP_OK;}////////////////////////////////////////////////////////////////////////////////////	Class vector Methods//////////////////////////////////////////////////////////////////////////////////vector::vector(){ __ptr = 0;  __size = 0;}vector::vector(int n){ __ptr = (int*)soap_malloc(soap, n*sizeof(int));  __size = n;}vector::~vector(){ soap_unlink(soap, this); // not required, but just to make sure if someone calls delete on this}void vector::resize(int n){ int *p;  if (__size == n)    return;  p = (int*)soap_malloc(soap, n*sizeof(int));  if (__ptr)  { for (int i = 0; i < (n <= __size ? n : __size); i++)      p[i] = __ptr[i];    soap_unlink(soap, __ptr);    free(__ptr);  }  __size = n;  __ptr = p;}int& vector::operator[](int i) const{ if (!__ptr || i < 0 || i >= __size)    fprintf(stderr, "Array index out of bounds\n");  return __ptr[i];}////////////////////////////////////////////////////////////////////////////////////	Class matrix Methods//////////////////////////////////////////////////////////////////////////////////matrix::matrix(){ __ptr = 0;  __size = 0;}matrix::matrix(int rows, int cols){ __ptr = soap_new_vector(soap, rows);  for (int i = 0; i < cols; i++)    __ptr[i].resize(cols);  __size = rows;}matrix::~matrix(){ soap_unlink(soap, this); // not required, but just to make sure if someone calls delete on this}void matrix::resize(int rows, int cols){ int i;  vector *p;  if (__size != rows)  { if (__ptr)    { p = soap_new_vector(soap, rows);      for (i = 0; i < (rows <= __size ? rows : __size); i++)      { if (this[i].__size != cols)          (*this)[i].resize(cols);	(p+i)->__ptr = __ptr[i].__ptr;	(p+i)->__size = cols;      }      for (; i < rows; i++)        __ptr[i].resize(cols);    }    else    { __ptr = soap_new_vector(soap, rows);      for (i = 0; i < rows; i++)        __ptr[i].resize(cols);      __size = rows;    }  }  else    for (i = 0; i < __size; i++)      __ptr[i].resize(cols);}vector& matrix::operator[](int i) const{ if (!__ptr || i < 0 || i >= __size)    fprintf(stderr, "Array index out of bounds\n");  return __ptr[i];}