preferred mail setup to let users decide on using spamassassin or not

[Benedict Verheyen <benedict.verheyen@gmail.com>]

Hello,

Following a thread here about fetchmail/getmail, is started playing
with getmail.

My current setup is as follows: fetchmail gets the mail for multiple
users and delivers to exim4. Exim then passes the message on to clamav
and spamassassin and the message then goes to exim again.

Last, maildrop gets the message from exim and the messages are
distributed to the Maildir of the user.

Now when i tested getmail, i triggered it via a cronjob (crontab -e) and
specified this in the users ~/.getmail directory:
...
[filter-1]
type = Filter_classifier
path = /usr/bin/clamdscan
arguments = ('--stdout', '--no-summary', '-')
exitcodes_drop = (1,)

[filter-2]
type = Filter_external
path = /usr/bin/spamc

[destination]
type = MDA_external
path = /usr/bin/maildrop
unixfrom = true

Now i can see a number of scenario's and i was wondering what scenario
is best if you the users to be able to customize their spamassassin
settings and maybe set for themselves if they want to check for spam or
not (i would always want to scan for viruses).
maildrop is used in this setup to put the messages in various
directories inside Maildir.

1. fetchmail/getmail -> spamassassin -> clamav -> maildrop

2. fetchmail/getmail -> exim 4 -> spamassassin -> clamav -> exim 4 ->
maildrop

3. fetchmail/getmail -> maildrop -> spamassassin -> clamav

4. fetchmail/getmail -> exim4 -> maildrop -> spamassassin -> clamav

There are of course lot's more options.
Option 2 is easy in a way that spamassassin and clamav are called from 1
place.
Option 1 and 3 look like the best options if you want users to be able
to customize their settings.
1 and 3 boil down to either having getmail call spamassassin and clamav
or having maildrop do it.

Are options 1 & 3 indeed better if you want the user to have a say in
what happens?

Another question i had was when the user runs 'crontab -e', are the
settings stored somewhere?

Thanks,
Benedict