Autor: W B Hacker Data: A: exim users Assumpte: Re: [exim] strict_rfc821_envelopes
Ian Eiloart wrote: >
> --On 13 May 2010 17:39:42 +0100 Ron White <exim.ml@???> wrote:
>
>> Thank you Ian, that is really useful on more than one level. Has anyone
>> ever told you that you are a guru at this Exim stuff and a very helpful
>> chap too :-) Thanks. Really appreciated.
>>
>
> No problem. I'd not really thought of this approach before. I wonder how
> many false positives a strict approach would lead to!
>
Not *quite* as strict, but we apply three strictness checks on format of
envelope and headers (or their absence) and assign cumulative, weighted point
scores to later compare to a per-user threshold.
Our logs show that *of those tagged at all* some 96% eventually pass for either
normal or sequestered delivery.
Only 4% of those so tagged are eventually rejected - and that may be for
unrelated reasons.
So - JMHO - too much potential for falsing, too little useful meat on that bone.
