Re: [Squid3-ipv6] on SNMP

From: Amos Jeffries <squid3@dont-contact.us>
Date: Wed, 26 Sep 2007 10:45:05 +1200 (NZST)

> Sorry for the rest, I lost private contact with Amos...
>
> Amos, your server is filtering ours, and I can read, but no send you any
> message... Could you check the list your servers are asking ?
>

I found a few of these died this week. Should not be used anymore from now.

>
> <squid3@treenet.co.nz>: host rio.treenetnz.com[60.234.233.111] said: 554
> 5.7.1
> Service unavailable; Client host [147.96.1.101] blocked using
> bogons.dnsiplists.completewhois.com (in reply to RCPT TO command)
>
> I had problems to send you it since my smtp server must be filtered on
> yours... I would never commit anything before you know it...
>
> Answer me, as always, to rmartine@fdi.ucm.es
>
> ----- Mensaje original -----
> De: "Rafael Martinez (Squid development)" <rmartine@fdi.ucm.es>
> Fecha: Lunes, Septiembre 24, 2007 9:46 pm
> Asunto: About SNMP IPv6
> A: Amos Jeffries <squid3@treenet.co.nz>
>
>> Hi Amos.
>>
>> As you can imagine, most I can do is try to understand your great
>> effort. How is your pinger6 going ?
>>

Structure #3 being dreamed up presently, 1,2 were horrible.

>> I read your NOTES-IPV6 file, and I took an eye into SNMP subject...
>> After inspecting the code, I saw there are no IP dependencies on
>> sourcesunder snmplib/, but there are at least under
>> src/snmp_agent.cc,snmp_core.cc
>>
>> Regarding snmp_agent.cc, I modified the code to deal with IPv6
>> MIB values.
>> I read somewhat about that and I did it according to standards...
>>
>> Don't stop watching the changes I may commit... It would be sad
>> to disarrange the great
>> you made... I think my change is OK.

Works, just me being pedantic pointing out some slightly faster method.
When you thinks its going I'll update the website marking it as Experimental.

>>
>> In a medium-term I will set-up an Squid Agent and a MRTG
>> client-manager... But I'm not sure on how to setup the basic
>> communitieson SQUID...I think they are populated automatically
>> on snmpInit().
>>
>> By the way, you see in the code I'm forced to call the
>> deprecated methods
>> GetInAddr(struct in_addr) and GetInAddr(struct
>> in6_addr). What about to refactoring

Hmm, 'deprecated' is probably the wrong word for those ones. I just called
them deprecated to prevent the itch to use them when not needed.
This is one case when they should be used.

>> the class IPaddress.h in such a way as Java does?
>>
>> I mean: Make IPaddress a virtual class
>> with a virtual method (bool IPAddr::GetRawAddr(void *addr,int *len))
>> and let the derived ones to implement the method ? By the way,
>> the virtual IPAddr would not have
>> any content, only IPv4Addr and IPv6Addr...
>>

IPAddress is just a sockaddr_in* with a 'library' of extra methods.
Intended to be a POD class same as the original sockaddr* are.

Amos
Received on Tue Sep 25 2007 - 16:45:25 MDT

This archive was generated by hypermail pre-2.1.9 : Mon Oct 01 2007 - 12:00:05 MDT