http://bugzilla.spamassassin.org/show_bug.cgi?id=4412
[EMAIL PROTECTED] changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
------- Additional Comments From [EMAIL PROTECTED] 2005-08-18 12:06 -------
Mark, thanks for the corrections. Can you please submit a CLA (
http://www.apache.org/licenses/#clas ) to make it easier to use your patch?
Also, what failures did you see that you addressed with the change
"- avoid 65000 attempts on other than EADDRINUSE failure;"
And how does your patch avoid that? I only see that it checks for EIADDRINUSE
and if that's not the error, it calls _ipv6_ns_warning which only dies if the
error is a problem with IPv6 nameserver.
I originally did not check for EIADDRINUSE because I thought that if it is
something else and it works on trying again on another port, then why not just
keep on trying? If there is some common failure that we would want to die
immediately on, then we can check for that.
Justin, aside from CLA issues, this is simple enough that you can rewrite Mark's
patch from scratch and may want to just because of style issues, perhaps wanting
something other than die in case of error, etc.
The changes I see in the patch are:
1) Add a Domain argument to the IO::Socket constructor call set to: If PeerAddr
is in IPv6 family and we have IO::Socket::INET6 set it to AF_INET6, else set it
to AF_INET
2) In the loop that looks for an unused port, skip the call to _ipv6_ns_warning
if $! is EIADDRINUSE because that's the error we intend to loop over.
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
