Releasing messages from quarantine

[Steve Freegard]

Julian Field wrote:
> Not quite.
> When I remove the dangerous attachment from a message, I send the 
> message on with its original Message-ID: header, which I believe is what 
> I should be doing. Otherwise I'll break threads, among other things. Not 
> every recipient of an attachment in (for example) a mailing list thread 
> is interested in receiving that attachment, and having the thread broken 
> as a result.
> 
> When someone chooses to release a message from the MailWatch quarantine, 
> they don't change the Message-ID: to a new value before sending it. So 
> personally I reckon the ball is in your court. Sites without MailWatch 
> wouldn't want their Message-ID: threads breaking for every message that 
> happened to contain a dodgy attachment the recipient wasn't interested 
> in anyway. Surely it's MailWatch's job to create a new Message-ID: when 
> a message is re-posted with its attachments, now the user has chosen to 
> retrieve them?
> 
> I just see this as a problem for the implementers of quarantine release 
> mechanisms, not for me.
> 
> What do you think?

You're right of course - I'm forgetting that MailScanner can strip 
attachments and deliver the original message.

I can strip the Message-ID on my end and get the MTA to generate a new 
one automatically; but this is going to break threading as well.  I 
guess there is no clean way around this.

Regards,
Steve.