Re: Training dspam

[Ted Zlatanov <tzz@lifelogs.com>]

On Sat, 29 Mar 2008 16:38:03 -0400 James Cloos <cloos@jhcloos.com> wrote: 

JC> The manual suggests a couple of (defun)s to report erroneously collated
JC> spam and innocent mail.  The synchronous nature of the recommended
JC> functions has proven painfully slow to use on my aging laptop, so I came
JC> up with these modified (defun)s.

JC> Perhaps they will be useful to someone else.

JC> (defun gnus-summary-dspam-report-spam ()
JC>   "Submit spam to dspam, then mark it as expirable."
JC>   (interactive)
JC>   (gnus-summary-show-raw-article)
JC>   (gnus-eval-in-buffer-window gnus-article-buffer
JC>     (save-restriction
JC>       (widen)
JC>       (call-process-region (point-min) (point-max) "/usr/bin/dspam"
JC>         nil 0 nil "--mode=teft" "--user" (getenv "USER")
JC>         "--class=spam" "--source=error")))
JC>   (gnus-summary-mark-as-expirable 1))

JC> (defun gnus-summary-dspam-report-innocent ()
JC>   "Submit false-positive to dspam."
JC>   (interactive)
JC>   (gnus-summary-show-raw-article)
JC>   (gnus-eval-in-buffer-window gnus-article-buffer
JC>     (save-restriction
JC>       (widen)
JC>       (call-process-region (point-min) (point-max) "/usr/bin/dspam"
JC>         nil 0 nil "--mode=teft" "--user"  (getenv "USER")
JC>         "--class=innocent" "--source=error"))))

Thanks for the code.  It looks useful, can it go into spam.el (do you
have the papers)?  I may use the same approach for all the reporting
functions if it's faster than what's there already.  I don't use dspam
personally and need to review to make sure the code is consistent, but
the papers come first...

Ted