Bulk mode in bogofilter 0.12.1

[michael at optusnet.com.au]

# bogofilter -c /tmp/bogofilter.cf -d d/0 -B -v spam.1/1048889620 spam.1/1048889620
spam.1/1048889620 X-Bogosity: Yes, tests=bogofilter, spamicity=1.000000, version=0.12.1
spam.1/1048889620 X-Bogosity: No, tests=bogofilter, spamicity=0.778405, version=0.12.1

Says it all really. :) Feeding it the same file twice gives
different scores. 

Hmm. I think the problem is in the lexer. It's not finding the mime
type the 2nd time around (this is an html email) so it scores
all the xxx.gif parts as 0.415 and thus gives it a low score.
Terrible.

I think about line 140 of main.c initialize() should be
called directly, rather than doing the mime_reset et al
by hand. 

Something like

diff -u -r bogofilter-0.12.1.old/src/main.c bogofilter-0.12.1/src/main.c
--- bogofilter-0.12.1.old/src/main.c    Fri Apr 25 20:17:36 2003
+++ bogofilter-0.12.1/src/main.c        Sun Apr 27 17:44:59 2003
@@ -137,9 +137,7 @@
  
        passthrough_setup();
        do {
-           init_msg_counts();
-           mime_reset();
-           init_charset_table(charset_default, true);
+           initialize(fpin);
  
            status = bogofilter(&spamicity);
            write_message(out, status);


Yes? No?

Michael.