Recently, Somebody Somewhere wrote these words
> On Fri, Nov 18, 2005 at 10:23:05AM +0000, Declan Moriarty wrote:
> >
> > Third, If I applied things like reject_non_fqdn_sender,
> > reject_unknown_sender_domain or reject_unverified_sender I would
> > bounce several list members with half cocked mail setups, I suspect.
>
> Let me try to clarify some those checks. Assume you had postfix on
> your box receiving mail. That is, lfs list mail was sent from belg to
> genius. No ISP, no fetchmail. Many of the checks would pass regardless
> of what the author used for an email address because belg is sending
> you the mail, not the author.
>
> >From [EMAIL PROTECTED] Fri Nov 18 07:30:30
> >2005
>
> This is the envelope header of your email which I'm replying to now.
> This is who the email is from as far as postfix is concerned.
>
> >From: Declan Moriarty <[EMAIL PROTECTED]>
>
> This is who the email claims to be from. Just like a real letter, you
> can claim to be one person on the envelope of a letter, and yet use a
> different address on the letter inside the envelope. The postal worker
> does not know that.
If I had smtp without an isp (No ISP, no fetchmail), I would be loving
these smtp checks. So with the smtp checks, then, and spam from the
list, envelope checks would always pass. Sender, it depends, right?
But you guys can't check senders, or you'll bounce any half set up
mailer someone struggling with LFS has just set up. That was where all
this began. These ones
<[EMAIL PROTECTED]>
<[EMAIL PROTECTED]>
in the last 10 spam to get through from the list don't resolve, which
seems to prove you don't check senders.
>
> > Fourthly, I would also note that I receive no mail until my isp has
> > terminated his transaction with the spammer, and then scanned it
> > fairly thoroughly(some seconds), and then held it for collection
> > (Some minutes or hours). That leaves me unclear what percentage of
> > bounces will actually reach a spammer, or more likely waste
> > bandwidth on legit servers until it double bounces somewhere and
> > gets ditched.
>
> It happens like this: 1) Either the server admin sets a policy that
> email is processed while the sending smtp server sits on the line
> (that allows for a rejection code) or 2) It tells the sending server
> that it received it okay, disconnects, and then processes the mail. I
> would imagine #2 is more common because it increases throughput and
> prevents extra failures due to smtp timeouts. However, now you can't
> send an error code so you are left with another policy decision:
> bounce, discard or send it on to the intended recipient. My ISP, for
> example, just marks the subject line and lets the customer decide
> since it is their email and false positives are always possible.
>
Esat/iol have stringent processing while sending. Believe me, I had
quite a job ever getting out, and only managed because their error
messages were informative enough. I also have to scrape past their
spamassassin installation.
After lengthy correspondence with Spamcop, Spamhaus, and others, ntl
implemented smtp rejection. The fact that they lost 2000+ customers
together here gave them a few free servers to play with. Before that,
they were regularly featured as viruses like mydoom spammed everyone
worldwide when they were current, and kids were wondering why their
connections were so slow. They supplied m$ie4 and outlook, and only
support windows installations :-(. They never made money as an isp, but
they made a fortune on technical support on a premium rate number!
This has been a very educational thread. One last thing. When a bounce
is done at smtp level, does it go to envelope sender or 'original'
sender? If the latter, I probably couldn't bounce in this email 'cage'
anyhow.
--
With best Regards,
Declan Moriarty.
--
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
