http://bugzilla.spamassassin.org/show_bug.cgi?id=4412
------- Additional Comments From [EMAIL PROTECTED] 2005-08-17 09:47 -------
The patch passes make test on my system, but I don't understand what is going on
with this part of the patch that removes two lines of code
- $self->{dest} = sockaddr_in($self->{res}->{port},
- inet_aton($self->{res}->{nameservers}[0]));
I see that $self->{dest} is used later in sub bgsend in the following snippet:
my $data = $pkt->data;
my $dest = $self->{dest};
$self->connect_sock() if !$self->{sock};
if (!$self->{sock}->send ($pkt->data, 0, $self->{dest})) {
warn "dns: sendto() failed: $@";
return;
}
And then $data and $dest are never used, which is a separate issue.
The code in Net::DNS::Resolver::Base.pm version 0.53 in sub send_udp that
computes the equivalent of $self->{dest} here distinguishes between IPv6 and
IPv4 doing something like the following to compute what it names $dst_sockaddr:
In the IPv4 case it looks like
$dst_sockaddr = sockaddr_in($dstport, inet_aton($ns_address));
In the IPv6 case it looks like
# we can use getaddrinfo
no strict 'subs'; # Because of the eval statement in the BEGIN
# AI_NUMERICHOST is not available at compile time.
# The AI_NUMERICHOST surpresses lookups.
my $old_wflag = $^W; #circumvent perl -w warnings about 'udp'
$^W = 0;
my @res = getaddrinfo($ns_address, $dstport, AF_UNSPEC, SOCK_DGRAM,
0, AI_NUMERICHOST);
$^W=$old_wflag ;
use strict 'subs';
my ($sockfamily, $socktype_tmp,
$proto_tmp, $dst_sockaddr, $canonname_tmp) = @res;
So do we need equivalent code to compute $self->{dest}? Why were those lines
just removed in the patch if $self->{dest} is later used in sub bgsend?
I don't have time right now to figure out a test case to demonstrate bgsend
breaking with this patch or to figure out why it doesn't break, but this doesn't
look to me like it could be correct.
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
