Forum OpenACS Q&A: ACS Classic 3.4.x Spam module

Posted by C. R. Oldham on

Can anyone recall any problems with the ACS 3.4 classic Spam
module?  We are trying to send out a spam and are finding characters
missing from the spam after it is received in the target user's
email.  Those characters do not appear to be missing in the database
copy of the spam.  It sort of appears like there is some sort of
regexp pass that occurs on the spam before it gets emailed, and that
regexp pass is removing characters.  A cursory glance at the code
revealed nothing obvious, so before I dig in deeper I ask the
community's help.

Posted by C. R. Oldham on
OK, I thought I'd document what I found here.

Before the spam module sends a spam out it encodes the plain text and html parts as quoted-printable.  I'm assuming it does that because it protects the contents of the message from being mangled by some mail readers MIME parsers.  That encoder wasn't very smart--it was causing qp line breaks in the middle of some html tag attributes.  Or maybe the qp MIME parser in Outlook isn't very smart.  Anyway I was loathe to remove the qp encoding altogether since didn't really understand why it was necessary.

My solution was to modify the textarea properties for the textarea where the html gets put so it would hard wrap.  That way we could look through the message and manually fix any "bad" wraps before the message gets mangled by the qp encoder.

Any clues from anyone about the need for QP encoding would be appreciated.  I didn't notice any code that creates QP encoding in OpenACS 4.5.  Is it really necessary?