Re: memory leak

[Trigve Siver <trigves yahoo com>]

>> -
>> and a lot of leaks:
>> ==5048== LEAK SUMMARY:
>> ==5048==    definitely lost: 0 bytes in 0 blocks.
>> ==5048==      possibly lost: 49,424 bytes in 54 blocks.
>> ==5048==    still reachable: 237,710 bytes in 4,803 blocks.
>> ==5048==         suppressed: 0 bytes in 0 blocks.
>>
>
>(waving hand in the air) There are no leaks here. </bad_starwars_joke>
>
>These aren't really leaks perse, but memory that never gets
>deallocated. To prove it to yourself, try something along the
>following:
>
>#include <gtkmm.h>
>
>ITERS=1000
>
>int main(int argc, char *argv[]) {
>  Gtk::Main kit(argc, argv);
>
>  //Print memory usage.
>
>   for( int i = 0 ; i < ITERS ; i++ )
>   {
>      Gtk::Label* lbl = new Gtk::Label()
>      delete lbl ;
>   }
>
>   //Print memory usage
>
>  return 0;
>}
>
>The two memory usage numbers should be the same. I can't believe I
>don't remember how to get the current memory usage off the top of my
>head, but its been awhile since I've done something like this. I
>believe there's a system call like 'usage' or some such. Anyway,
>google will have the answer for that detail.
>
>Paul

I think u mean 

Glib::Error::register_cleanup();
Glib::wrap_register_cleanup(); 


int main(int argc, char *argv[])
{
    try
    {    
        Gtk::Main kit(argc, argv);
        // ...
    }
    catch(std::exception const & e)
    {
        std::cout << e.what() << std::endl;
    }
    // Additional cleanup
    Glib::Error::register_cleanup();
    Glib::wrap_register_cleanup();     

    return 0;

} 


Trigve




       
____________________________________________________________________________________
Boardwalk for $500? In 2007? Ha! Play Monopoly Here and Now (it's updated for today's economy) at Yahoo! Games.
http://get.games.yahoo.com/proddesc?gamekey=monopolyherenow