Hi, yet another "smart" question: if I call bind() this way

    struct sockaddr_in sin;
    int fd;

    fd = socket(PF_INET, SOCK_STREAM, IPPROTO_TCP);
    bind(fd, (struct sockaddr*) &sin, sizeof(sin);
...

then do I still have to use sockaddr_storage in an accept() call later 
or will sizeof(sin) from above (or PF_INET?) get saved somewhere 
and thus using sockaddr_in is safe enough?

BTW neither write(2) nor accept(2) mention EINTR even though
that value appears in /usr/src/lib/libpthread/uthread/uthread_*.c
(I used a sendbug to report it this morning, but didn't get a reply).

And 1 more question please: where are the functions _thread_sys_accept, 
_thread_sys_write, etc. defined? I can not find them in /usr/src, 
I only see the prototypes in lib/libpthread/uthread/pthread_private.h  :o|

Regards
Alex

Reply via email to