John Rudd wrote:
> As long as people are thinking about new implementations of
> spamc/spamd, I'd like to make a couple suggestions:
>
> 1) make room for passing a reference to a message instead of the message
> itself, to reduce network latency when it is not necessary (for
> example, if the spamd chosen just happens to be on localhost (because
> it's part of a list of servers that are being balanced), then you'll
> waste a lot of time transfering the message through the socket,
> instead of just transfering the path to the file that contains the
> message).
Good idea. clamd works this way for example -
SCAN <filename>
Of course this only works for sockets and TCP connections to localhost.
I could see this being really useful for mbox files, for example.
> 2) make room for using other tools in addition to spam assassin (for
> example, running an external virus scanner), making spamc/spamd a
> possible nice general purpose means of distributing the load for email
> scanning (not just for spam marking type scanning) and even email
> delivery.
MIMEDefang does a lot of what you're suggesting. Check out
message_contains_virus_clamd in mimedefang.pl, for example... the inspiration
for this whole thing (on my part) was the desire for a message_is_spam_spamd.
[EMAIL PROTECTED] 805.964.4554 x902
Hispanic Business Inc./HireDiversity.com Software Engineer
perl -e"map{y/a-z/l-za-k/;print}shift" "Jjhi pcdiwtg Ptga wprztg,"
