More remarkably then that the addresslength of the sockaddr_in struct
(which the application is responsible to pass as argument to most
BSD-style network funtions) is totally ignored in Linux. accept(), for
instance has the length passed to it as a pointer, so it can be modified
if it doesnt think all is needed to confirm a structlength (if I have
understood this right). Passing NULL there seems to be accepted by
Linux, actually.
/Anders
Mircea Ciocan wrote:
>
> Hello all,
>
> Today, when trying to show someone the beauty of sockets programming
> via Stevens Bible II-ed. ;) the young lad observed that the sin_len
> field was left out in Linux implementation. While WRS make clear that is
> not required by Posix.1g standard and is usefull only to kernel
> routines, the young aprentice asked why and what was the reason for
> that.
> Even more, he asked if he can "take the bull by the horns" and put the
> member back in the structure without breaking something.
> Now this is little over my head and I've taken the liberty to ask an
> real kernel network about that issues hoping that someone will shed a
> little light over that question.
>
> Thank you,
>
>
> Mircea C.
> -
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to [EMAIL PROTECTED]
