(no email)
Date: Mon Sep 09 2002 - 23:19:33 EDT
On Mon, 9 Sep 2002, Clifton Royston wrote:
> > Postfix does not have logic in the address list matching code to handle
> > 0/0 portably. This is good IMHO.
>
> It appears to be intended to specifically reject 0/0 in a portable
> way, which would be just fine. I've looked into util/match_ops.c and
> it certainly *appears* that this case is being explicitly tested for,
> and that a "fatal" message should be logged on any attempt to use a
> netmask of /0, ignoring the address.
>
> if ((mask = split_at(saved_pattern, '/')) != 0) {
> if ((*mask_shift = atoi(mask)) <= 0 || *mask_shift > BITS_PER_ADDR
> || (*net_bits = inet_addr(saved_pattern)) == INADDR_NONE) {
> msg_fatal("bad net/mask pattern: %s", pattern);
> }
>
The test is " > BITS_PER_ADDR" not " >= BITS_PER_ADDR" INADDR_NONE is
255.255.255.255 (-1L). There is no code to prevent 00, but as I said
before the 0/0 behaviour is not portable. This is IMHO OK.
-- Viktor. - To unsubscribe, send mail to with content (not subject): unsubscribe postfix-users
|
|
|