I still come back to, how does this solve the subscription confirmation
process? That will only work by sending a mail to the person that sent the
request initially.



Matt Deatherage said:
> On 5/7/07 at 12:14 PM, a.h.s. boy (lists) <[EMAIL PROTECTED]>
> wrote:
>
>> >2) RE: Spam engine. All of those messages will end up in a reply to
>> >the person that sent the mail (or that Letterrip thinks sent the
>> >mail). You are not redistributing SPAM. Letterrip is rejecting it.
>> >This is exactly the same behavior that any other mail server would do.
>>
>> Well, like it or not, LR _is_ redistributing spam. Spam is coming from
>> Machine 1 (spoofed as Machine 2), to LR, and LR is _sending the spam_
>> to Machine 2. That's redistribution. I TOTALLY agree with you that's
>> it's doing so for rather legitimate reasons, but the inadvertent side
>> effect is that it winds up redistributing spam.
>
> I'd have to agree, reluctantly - if person A sends spam to LRP that's
> spoofed as from person B and contains an invalid command, and if LRP
> sends a reply to person B containing the original message, then there's
> no way to get around the conclusion "LRP is distributing spam."
>
> At first I thought the answer to this might be a quick processor -
> filter incoming messages for commands, process them, and don't send
> errors back to the sender, just eat the message.  Unfortunately, I have
> no idea if this is true - the sample processors in my LRP installation
> are all Classic applications; the most useful-looking of them are saved
> as "run-only" so I can't see the AppleScript, and the ones that are
> readable contain handlers for events that aren't in the LRP 4.07
> dictionary (like <<BassPreP>> and <<BassInit>>), so I have no idea if I
> need to do anything there or not.
>
> If you have EIMS and anti-spam features (like in EIMS 3.3 or DNS-based
> filtering like Denes), you could put LRP behind the spam protection so
> that commands from fake IP addresses don't get through, or (presumably)
> you could write a processor in C that looks up the MX that should go
> with the sending address and eats the message (without error) if it came
> from a different SMTP server.  Or you could disable the "unsubscribe"
> address and put up a Web page with a CGI to do the work instead.
>
> In other words, I think there are things you can do today to alleviate
> the problem.  For all I know, there's an STR# or TEXT resource you can
> change that will keep LRP from substituting the original message into a
> template reply, but I don't know what you'd change to get that effect
> (or if it's true, it just seems possible given LRP's design).  Some are
> more work than others, but I think you can do *something*.
>
> --Matt
>
> --
> Matt Deatherage                              <[EMAIL PROTECTED]>
> GCSF, Incorporated                      <http://www.macjournals.com>
>
> "I was going to be a neo-deconstructionist, but mom wouldn't let me."
>    -- Calvin, to Hobbes
>
>
> --
> This message is from the Letterrip-Talk Mailing list.
> To unsubscribe, send mail to: [EMAIL PROTECTED]
> Archive: http://www.mail-archive.com/letterrip-talk%40lists.letterrip.com/
>
>

--
This message is from the Letterrip-Talk Mailing list.
To unsubscribe, send mail to: [EMAIL PROTECTED]
Archive: http://www.mail-archive.com/letterrip-talk%40lists.letterrip.com/

Reply via email to