Re: A memory release promble about ORBit-0.5.3 at connection.c line 448

[Sebastian Wilhelmi <wilhelmi ira uka de>]

Hi Chuyee,

>     I recently read the ORBit-0.5.3 source code, I wonder if there is a
> memory problem in file connection.c line 448. Since the "g_strdup" malloc'd
> a heap memory, while "inet_ntoa" just return a static memory. And I trace
> the program with gdb, I found the memory alloc'd by g_strdup never freed!
>     Can anyone answer me? thanks :)
> 
> Here is the Code:
>     436   switch(parent->icnxtype) {
>     437   case IIOP_IPV4:
>     438     n = sizeof(struct sockaddr_in);
>     439     if(getpeername(GIOP_CONNECTION_GET_FD(fd_cnx), (struct sockaddr
> *)&fd_cnx-
> >u.ipv4.location, &n))
>     440       {
>     441         fd_cnx->u.ipv4.hostname = g_strdup("");
>     442       }
>     443     else
>     444       {
>     445         hent = gethostbyaddr((const char
> *)&fd_cnx->u.ipv4.location.sin_addr.s
> _addr, 4, AF_INET);
>     446         if(hent)
>     447           {
>     448             fd_cnx->u.ipv4.hostname = g_strdup(hent->h_name);
>     449           }
>     450         else
>     451           {
>     452             fd_cnx->u.ipv4.hostname = inet_ntoa(*((struct in_addr
> *)&fd_cnx->u
> .ipv4.location.sin_addr));
>     453           }
>     454       }
>     455     break;

Yes indeed, there should be g_strdup there. It is fixed in CVS.

Bye,
Sebastian
-- 
Sebastian Wilhelmi
mailto:wilhelmi ira uka de
http://goethe.ira.uka.de/~wilhelmi