Making a post request with CppCMS

We are using CppCMS server and its performance is very low for mail requests when message messages reach a certain size of some kB. This is already happening with the minimal primer server examples slightly modified:

#include <cppcms/application.h>  
#include <cppcms/applications_pool.h>  
#include <cppcms/service.h>  
#include <cppcms/http_request.h>  
#include <cppcms/http_response.h>  
#include <cppcms/mount_point.h>  
#include <booster/intrusive_ptr.h>

#include <utility>
#include <iostream>  
#include <boost/timer.hpp>

class hello 
  : public cppcms::application 
{  
  public:  
    hello(cppcms::service &srv) :  
      cppcms::application(srv), counter(0) {}  
    virtual void main(std::string url);  
    boost::timer my_timer;
    int counter;
};

void hello::main(std::string url)  
{  
    response().out() << "Hello from asynchron CppCMS after " << my_timer.elapsed() << "s\n";
    std::cout << "[" << url.size() << "] = " << url.substr(0, 50) << '\n';
} 

int main(int argc,char ** argv)  
{  
    try {  
        cppcms::service srv(argc,argv); 
    booster::intrusive_ptr<hello> p= new hello(srv);
    srv.applications_pool().mount( p, cppcms::mount_point("(.*)") ); 
    srv.run();  
    }  
    catch(std::exception const &e) {  
    std::cerr << e.what() << std::endl;  
    } 
    return 0;
}    

      

        
        
        
      

    

It's weird that the server only spends a few milliseconds on the main function, but one or two seconds between two calls.

• What is CppCMS doing during this time?
• And more importantly, how can this time be shortened?
• Is this a thread pool problem?
• Is there something expensive about messages? (I read something about compression) Can it be turned off or accelerated?

9 mail requests with XML files between 8 and 10kB take 9 seconds without even processing the requests. The same requests take less than half a second with a node.js server. I've tried different configurations and read a lot of dox and posts that couldn't find anything that explains these long downtime. Any help is appreciated.

Thanks in advance, Peter