spam.el + sa-learn

[Tom Rauchenwald <sehnsucht.nach.unendlichkeit@quantentunnel.de>]

[-- Attachment #1: Type: text/plain, Size: 633 bytes --]

Hi,

i recently noticed that my spamassassin bayes database didn't get
updated properly. sa-learn didn't pick up messages I fed it (via the
spam exit processors).

The following patch fixes it for me, but I'm not sure if it is the right
thing to do. For me it looks like sa-learn doesn't accept "From " alone
as message seperator anymore, so i added "nobody" and the current time
string. I'm on Debian Sid FWIW (SpamAssassin version 3.3.1).

Could someone more knowing about the issue comment on this?

Thanks,
-tom

-- 
You should never underestimate the predictability of stupidity. 
                -- Bullet Tooth Tony in Snatch

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: patch-spam.diff --]
[-- Type: text/x-diff, Size: 798 bytes --]

diff --git a/lisp/spam.el b/lisp/spam.el
index 194668e..cf729e3 100644
--- a/lisp/spam.el
+++ b/lisp/spam.el
@@ -2726,9 +2726,7 @@ With a non-nil REMOVE, remove the ADDRESSES."
               (with-current-buffer summary-buffer-name
                 (setq article-string (spam-get-article-as-string article)))
               (when (stringp article-string)
-                (insert "From \n") ; mbox separator (sa-learn only checks the
-                                   ; first five chars, so we can get away with
-                                   ; a bogus line))
+                (insert (concat "From nobody " (current-time-string) "\n")) ; mbox separator
                 (insert article-string)
                 (insert "\n"))))
           ;; call sa-learn on all messages at the same time