[Mimedefang] Re: changes to md_log subroutine

[John Kirkland]

Alan,

I wouldn't introduce a new comma in the mdlog line.  The graphdefang
regular expression for mdlog divides out text based on commas, and you've
introduced one in the recipients field that was previously not there.

This is your choice, but using graphdefang, I wouldn't use the new mdlog
option.  You lose user-level granularity every time you receive spam
addressed to more than one person at your site.  I would argue that a
single spam sent to 5 distinct people at my site is 5 spam(s?), not 1 spam
(even though mimedefang only scans the message once).

Regards,
John

> Hi,
>
>  just thought i'd share this with the list.
>
> after applying the patch to mimedefang.pl to implement to new features of
> me_log_enable() and md_log() i was starting to notice something *I* thought
> was strange in my graphdefang graphs.
>
> user 'rcpt=2' was getting a decent amount of mail.  (although I had no idea
> who that actually was).
>
> I've modified one line in mimedefang.pl to make it a little bit more "user
> friendly" in my opinion.
>
> i'm not an Uber programmer, so i'm not sure how to submit a diff patch, but
> here's the line in the md_log() function of mimedefang.pl:
>
>
> I changed:
>         my $lcrecipient = "rcpts=" . scalar(@Recipients);
>
> to:
>         my $lcrecipient = "$Recipients[0], rcpts=" . scalar(@Recipients);
>
>
> I think this is more appropriate as then graphdefang will show the 1st (of
> $#Recipients) as the recipient and then append a field with the number of
> recipients. (which will probably be less confusing in the long run)
>
> I'm not entirely sure how this will effect graphdefang, if at all, but I
> figured i'd share it.  If it's something that you feel would be a worthy
> addition to the release code (or possibly a patch), i'd rather have it
> included with the distribution than to have a custom-hacked mimedefang.pl
> that's gonna get clobbered when I eventually upgrade.
>
> any thoughts/comments/suggestions welcome
>
>
> thanks
>
> alan