Sven wrote:

> Hi Mark,

> thanks for your reply,

> On Fri, Aug 12, 2005 at 12:16:53PM +0200, Mark Martinec told us:
>> As long as the effects of spam checking for such recipients are not visible
>> (e.g. in headers or in paying the cost of additional service), they wouldn't
>> know the difference nor care. This is why I'm saying that bypass* could
>> in principle be derived from *lovers. The bypass* is just an optimization
>> mechanism.
>>
>> If there is no effect, then it is irrelevant whether a mail was checked or 
>> not
>> (except for wasted time).

> that is what I meant: see this from the point of the person(s) running
> a busy mail server. With respect to spam scanning there are three
> kind of people:

> 1. People who want spam scanning and want the spam to be discarded/
>    quarantined/whatever.

bypass lover
  0      0     useful, check and block malware
  
> 2. People who want spam scanning but who want to receive all mails to
>    act based on the spam scanning headers, e.g. in their MUA (those
>    would be the ones for spam_lovers). Those people probably also
>    want to define a threshold for marking mail as spam so they can
>    filter on the "X-Spam-Status: Yes" header.

bypass lover
  0      1     useful, check but deliver tagged
  
> 3. People who want no spam scanning and don't want to pay for it. Those
>    people can be put in bypass_spam_checks, no work will be wasted
>    in spam scanning their mail, thus more resources on your mail
>    server for scanning other people's mails.

bypass lover
  1      1     useful, no checks if possible, and no effects

You have not mentioned a use for the fourth scenario which is the one
that could no longer be configured if the change were made.

bypass lover
  1      0     not too useful, free riding

> Sven

> > If I knew how much confusion the semantics of bypass_*  vs.  *_lovers
> > is causing, I would have provide only _lovers, and derive bypass_
> > automatically from it. Perhaps some day...

Mark, are you saying you want to automatically assign all *lovers to
bypass*? Then how would you accomplish this?:

bypass lover
  0      1     useful, check but deliver tagged

I don't find it confusing. I'm would favor flexibility. At first, I
did not understand what *lovers did exactly. Bypass seemed obvious. I
think that a note that including a recipient in a @*_lovers_maps is
functionally equivalent to setting $final_*_destiny = D_PASS; for
that recipient helps explain it (at least is does for me). I think it
would be more confusing if you told people, "If you want to bypass
spam checks for a recipient, put then in a spam_lovers map."

I think that you can only do something like:
@spam_lovers_maps = @bypass_spam_checks_maps;
(or)
@bypass_spam_checks_maps = @spam_lovers_maps;
under certain circumstances, not all. I think a note explaining under
which circumstances these could be used would be appropriate.

Gary V



-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
AMaViS-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/amavis-user
AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3
AMaViS-HowTos:http://www.amavis.org/howto/

Reply via email to