[Fwd: Re: Ubuntu irssi 0.8.12-4ubuntu2]

[Paul O'Malley - gnu's not unix -]

Aaron Toponce wrote:
> On Sun, Oct 12, 2008 at 02:10:27PM +0100, Paul O'Malley - gnu's not unix - wrote:
>> the moving the port is specific to the service run on Freenode so hard 
>> luck - no can help
> 
> Port 8001 is specific to Freenode. Not that hard to change.
Depends on person running server, skills, time, interest - other 
implications can therefore be understood as port changing.

This does not solve any problems, it routes around them.

>> speak to your irc server people
> 
> Wrong. Speak to your router hardware vendor, or just change your router.

[begin lolcats speak] Wrong use of wrong! [end lolcats speak]

Incorrect assumption, this is not a router issue alone.
In fact it is a router and damage by exploit issue.

The problem is not alone for the person disconnected.
A single scenario we might consider may be:
In a busy channel let us assume the exploit is triggered.
You can loose some percentage of your users, disrupting many 
conversations, some of which may be at a critical moment.
For these individuals this is unfortunate, however the channel 
management now has several seconds of disruption as people leave and 
join, with the consequent attempts to explain what is going on.

This situation is now rather well managed in #ubuntu and its offshoots.

The real problem which my terse answer did not address this, it is not 
where does the problem come from, it is a how to manage situation.

More appropriate questions, and not very complete or very verbose:

What is the real problem to be solved?

Consequences of bad firmware in some routers.

Can it be solved in a generic way?

No.
This follows from the simple fact that you can't make people get new 
routers or be better informed, as much as you might like to.
(See bug 1 for a parallel proof, only time can improve the situation.)

Can we advance the general level of awareness of the problem, and the 
port movement method for avoiding it?

Yes, advise about upgrading firmware, which in some cases does not fix 
the problem, advocating the use of some third party port to avoid it.

However, this leaves us with a real problem, if we actually document all 
the bad stuff you get some silly programmer who copies the bad code to a 
whole lot of new ports, now the use of 8001 and friends is of no use.

So I guess I stand by my original, implied, if not very verbose answer, 
each of the servers will handle it in their own way, if we try to do a 
global fix we will have programmers of bad firmware incorporating our 
fix into their code and thus breaking the hack we use to fix it.

Fun for all!

Regards,

P.