Re: My functions for reporting spam and ham to spamassassin and crm114

Re: My functions for reporting spam and ham to spamassassin and crm114

[Ted Zlatanov]

On Sun, 29 Feb 2004, kaede.news@online.ru wrote:

> I wrote this code because I'm also using crm114 in addition to
> SpamAssassin and I also wanted to increase accuracy by removing
> extra headers. Maybe these functions will also be useful to someone
> besides me.

I would appreciate it if you looked at spam.el, which recently got
CRM114 support, and told us if anything could be improved there.  I
used Adam Sjogren's code, but yours may be better in some places (I
don't know that, I'm just suggesting the code review).

Ted

Re: My functions for reporting spam and ham to spamassassin and crm114

[Ted Zlatanov]

On Mon, 01 Mar 2004, hubert@uhoreg.ca wrote:

>>>>>> "Eugene" == Eugene Morozov <kaede.news@online.ru> writes:
> 
> Eugene> Hello!  I wrote this code because I'm also using crm114 in
> Eugene> addition to SpamAssassin and I also wanted to increase
> Eugene> accuracy by removing extra headers. Maybe these functions
> Eugene> will also be useful to someone besides me.

> But it might be a good idea to strip extra headers (if any) that the
> filters add before spam/ham gets sent to the processor functions.
> Comments, Ted?  

I don't know CRM114 well, but I would suggest requesting this feature
from the developers instead of trying to fix it at the MUA.  If CRM114
supports a command-line switch to exclude headers from analysis, I'll
gladly add that to a CRM114 spam.el backend if such a thing should
materialize.

(by the way: I have not needed any new spam.el backends, but if anyone
needs DSpam or CRM114 or whatever else added, let me know)

> We would probably want to have a set of functions like
> "spam-spamassassin-strip-headers", etc. to do this.  Or a variable
> "spam-spamassassin-headers-regexp" that is a regexp to match
> SpamAssassin headers that should be removed.

The problem is that we can't modify the article, so we have to make a
temporary copy and rewrite that, which can be expensive depending on
the article size.  It would really be much cleaner if CRM114 could be
told to ignore certain headers.

Ted