Dear All,
An item for the wishlist, could the access to headers by "$header_" and
"def" be expanded? Given that the header names are terminated by colons,
this Wishlist item might not be entirely straight forward.
The reason I'd like to do this is so that I can do things such as, in
the ACLs:
# Always add the score header
warn message = X-${message_id}-Spam-Score: $spam_score
spam = exim:true
Then in the system message filter:
if ("${if def:header_X-${message_id}-Spam-Score: {there}}" is there)
then
if ("${if def:header_X-Spam-Score: {there}}" is there) then
headers remove X-Spam-Score
endif
headers add "X-Spam-Score: $h_X-${message_id}-Spam-Score:"
headers remove X-${message_id}-Spam-Score
Endif
I'm starting to see problems with my own Spam flagging / processing when
messages are received that already contain spam headers. I thought that
I could add *my* headers in containing the message ID and then in the
system filter make sure my version of the header is the only one in the
message.
I guess I could use a "unique" static string in my configuration, but I
though the message ID would provide an excellent unique string that no
one would be able to spoof.
Alternatively, I wouldn't need to do this if it were possible to delete
headers in the DATA ACL (other wishlist item).
Anyone else have this problem? A different solution?
Cheers,
Matt.
