tagging header lines
David Relson
relson at osagesoftware.com
Wed Feb 12 20:47:46 CET 2003
Greetings,
I've just added the ability to tag header tokens to the cvs version of
bogofilter. It is enabled by "tag_header_lines=yes" in the config file (or
by setting to true the boolean variable tag_header_lines in globals.c).
When enabled, all tokens parsed from designated header lines will be given
a prefix (specified in lexer_head.l). At the moment, tagging only applies
to the message's "^Subject: " line. The tag used is "subj:". As an
example, "Subject: this is a test" will parse as tokens "subj:this" and
"subj:test".
Tags can be readily added for other header lines. That will be done after
it is confirmed that this feature is useful and valuable.
I have also added a script for testing config file options. It is named
parmtest.sh and is in the contrib directory. It can
be used to test whether tag_header_lines improves bogofilter's ability to
classify spam or not.
Later, I'll post the results of my testing this new feature. It would be
good if others would do their own tests and report their results.
Enjoy the new feature!
David
More information about the Bogofilter
mailing list