This is a very common question, there are a number of solutions detailed
in the Wiki, complete with code you can steal in some cases.
Basically you can't forward, as you have discovered. What you CAN
do is:
1 Forward as an attachment. Users will tend to
forget to make it an attachment, and you have to write some code to unwrap the
original spam out of the attachment. Also, it is real easy to open the
spam by mistake doing this, and thus have the assorted web bugs probably
present in the spam confirm that the user is live.
One trick to encourage users is to explain
that if they select a WHOLE BUNCH
at the same time (2 will do) then they can
just hit Ctrl-F or use menu: Action->Forward (Alt-A
w)
If they select one message or do this from a single open
message they will get a
forward but NOT as an attachment.
I derived (wrote would be too strong a word) some very simple
code yesterday to
peel these apart into a subdirectory
./tmp:
#!/usr/bin/perl -w
use MIME::Parser;
my $parser = new
MIME::Parser; # Create
parser
$parser->output_dir("./tmp");
# Tell it where to output
$parser->extract_nested_messages(0); # Extract
messages whole?
$entity =
$parser->parse(\*STDIN); # Parse an input filehandle
print "Entity: $entity\n\n" if $entity;
Create subdirectory ./tmp (or change name in code); run against
each
multipart message (for %a in (*.msg) do perl split.pl <%a);
delete any
"*.txt" files from the tmp directly --- learn the .msg
files.
2 Set up some IMAP folders on the server and export
these as public folders. The users can then just drag/drop the ham/spam
into these folders. Harvest them with a cron job. There isn't as
much chance of opening the spam this way, at least if they have Preview turned
off, as they really should.
Number 2 is the generally suggested solution, and there are a number of
variants on the Wiki on how to do it.
Number two only works for those whose users have IMAP
-- my Outlook users don't
really like IMAP and many others are using Exchange
server.
BTW, you should probably toss your current Bayes DB and rebuild after
implementing one or the other of these methods. The header corruption
has probably skewed the database enough that it will be difficult to get
straightened out.
Loren
