On 2008-07-11 at 23:13 +0200, Alexandre Busquets Triola wrote:
> I have problems with emails with sender spamassassin headers
> X-Spam-Score: -0.9 (/)
Before these ACL checks, have:
warn set acl_m_is_spam = 0
> warn message = X-Spam-Flag:YES
> !authenticated = *
> condition = ${if <{$message_size}{1200k}{1}{0}}
> spam = spamd:true
> condition = ${if >{$spam_score_int}{50}{1}{0}}
Add to this: set acl_m_is_spam = 1
Hereafter, $acl_m_is_spam is either 1 or 0 and is unaffected by
anything other than your own spam score. You've separated out the
security-sensitive data to a different namespace which can't be directly
added to by the remote untrusted data source.
> virtual_delivery:
> driver = appendfile
> mode = 0600
> maildir_format = true
> delivery_date_add
> envelope_to_add
> return_path_add
> create_directory = true
> headers_remove = Subject : X-Spam-Flag
> headers_add = Subject: ${if eq{$h_X-Spam-Flag:}{YES} {*****SPAM*****
> $h_Subject:}{$h_Subject:}}
${if >{$acl_m_is_spam}{0} {*****SPAM***** $rh_Subject:}{$rh_Subject:}}
Note that this is preserving any needed MIME mangling in the original
"Subject:" header by using the raw form. There's a slight risk that the
first line will end up overlong, but it's less wrong than putting
decoded data in.
-Phil
