http://bugzilla.spamassassin.org/show_bug.cgi?id=2853





------- Additional Comments From [EMAIL PROTECTED]  2004-05-22 21:48 -------
I agree that C would be best; although it would be better if someone fixed bug
3388 :-) (However I believe that your assertion that "*Any* slow-down is
unacceptable" is false)

Henry: thanks anyways. I remember that there was some talk of implementing
perceptron in perl, but I'd forgotten that it had actually been done and that
there really was such a slow down.

I've changed the majority of the scripts in masses/ (I think i only missed the
stuff in rule-QA) and I've written a README for end users. While re-writing, I
added pod documentation for the "major" scripts -- the ones end users will
actually need. They are:
mass-check
extract-message-from-mbox
hit-frequencies
lint-rules-from-freqs
logs-to-c
rewrite-cf-with-new-scores
fp-fn-statistics

I propose that these scripts as well as perceptron be distributed to the end
users (in .deb's, .rpm's etc). I plan on putting this in a spamassassin-tools
package for Debian. (I hope to upload a 3.0.0 cvs package with that to
experimental by the end of this long weekend.)

The major change is that most of the reading/computing stats stuff, etc is in
lib/Mail/SpamAssassin/Masses.pm rather than duplicating the code in the scripts
or using parse-rules-for-masses. Many of the scripts' options have been changed
slightly for consistancy, and they no longer rely on @ARGV or shell
piping/redirects.

I'll post my changes (in giant patch format) later tonight after I finish one
last big test.



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

Reply via email to