This has come up for discussion several times on the lists, so it might 
be worth looking at the archives for additional info on what was said, 
etc.  I run into this problem myself occasionally.  It would probably be 
a useful addition.  Rather than 'auto-whitelisting' Tom, it might be 
easier to just extend the to* filter sources to optionally read To: and 
Cc: headers for addresses to match rather than just the envelope 
recipient address as is currently the case.  Then, one could add a 
'to-file' line to the incoming filter pointing at their whitelist, and 
since Charles is in that, a match will occur when Tom replies to Charles 
with you included.

See http://tmda.net/unstable/filter-sources.html

That would solve another often contended point about why to* doesn't 
match the To: header in incoming messages.  The idea of envelope 
recipient is confusing to most users and I don't think they expect the 
current behavior, though I do think that should remain the default.

Bernard Johnson wrote:
> Here is a scenario that I tend to see quite a bit:
> 
> 1) I routinely converse with Charles.  Charles sends me an email, with a
> CC: to Tom.  Charles is on my whitelist so his message is accepted.  Tom
> I barely know or did not previously have an email address for.
> 
> 2) Tom does a "reply all" joining the conversation.  When I receive his
> message it is challenged, even though both Tom and I know Charles who
> initiated the conversation.  I might also know Tom but don't have him in
> the whitelist.
> 
> Has anyone thought about "social whitelisting"?  For example, in the
> above scenario, if I had auto-whitelisted Tom when I got Charles' email
> that contained Tom in the CC:, the conversation would continue normally
> without interruption by TMDA.
> 
> Rationale:
> 1) Charles does not correspond with spammers.
> 2) This trades convenience of my correspondents vs. the possibility I
> may have to blacklist an address now and then.
> 3) If Tom knows Charles well enough to be on a CC: email along with me,
> then he probably deserves to be whitelisted.
> 4) In order for a spammer to exploit this method, he would have to know
> one of my whitelist senders and forge a message from one of them to add
> a sender to my whitelist (I use SPF so this is increasingly hard to
> accomplish)
> 5) Mass virus emails from users' contact list will be dealt with before
> reaching TMDA.
> 
> I don't think this can currently be done with TMDA, but I believe it
> would be a pretty easy thing to add.
> 
> It seems simple enough... Am I missing something?  Are there drawbacks
> that I'm missing?
_________________________________________________
tmda-workers mailing list ([email protected])
http://tmda.net/lists/listinfo/tmda-workers

Reply via email to