[Openchrome-users] Input for MOTU Meeting on Clamav

[Jelle de Jong]

Scott Kitterman wrote:
> On Tuesday 19 June 2007 11:22, you wrote:
>> Scott Kitterman wrote:
>>> I've decided to try and pick this thread back up.
>>>
>>> From the last MOTU meeting I was at, there was some discussion about
>>> looking into the API differences between clamav 0.8x (libclamav1) and
>>> 0.9x (libclamav2).  I'm hopelessly unqualified for that job.  Who
>>> volunteers?
>>>
>>> The more I think about this, I think the alternative source package
>>> (clamav-alt) proposal isn't the right answer.  I think this is going to
>>> (in the end) need an alternative repository for clamav and it's rdepends.
>>>
>>> I'd like to discuss this idea.
>>>
>>> We'd create a project (similar to the regular backports projects) for
>>> clamav backports.  It would hold the latest clamav (with necessary source
>>> package modifications to go back to all supported releases) and whatever
>>> rdepends volunteers are willing to contribute.  I'd commit to maintaining
>>> the clamav package, clamsmtp, and probably klamav.
>>>
>>> Would this approach work?
>>>
>>> What approvals would we need to get this established?
>>>
>>> Any better ideas?
>>>
>>> Scott K
>> Hi Scott,
>>
>> What is the problem exactly? Do you need testers, I have a few test
>> servers that I can run with the new clamav setup. I can even make
>> howto's how to use it with postfix.
>>
>> Kind regards,
>>
>> Jelle
> 
> The problem is that clamav 0.9x does not have the same external interfaces as 
> the 0.8x versions in Dapper/Edgy, so we can't just backport it in the normal 
> way.  OTOH, the existing Dapper/Edgy versions have outstanding security 
> issues that we don't have patches for, so what to do.
> 
> We will need testers, but not yet.  I'll keep your offer in mind when we get 
> to that.

Give me a notification when you need testers and I will make it a priority.

> How are you integrating Postfix with Clamav (I use Clamsmtp myself)?

I tested clamav with amavisd via postfix filters, but I don't like this
way, there are to many local send and return paths. I am trying out
better ways to get a more direct and more controllable approche like
with separated smtp daemons

> Scott K

Kind regards,

Jelle