OK, this is my first attempt at using bmf, so don't be afraid to say I'm full of it...
I installed bmf-0.9.2 today. I am having no luck training it using mutt macros to run bmf -S and bmf -N. It seems like both treat the email in question as span. If I grep a word (that is pretty unique to the email in question) in the good/spam lists, it seems like bmf -N and bmf -S have the same effect. (The word counts gets bigger in the spam folder) bmf -n and -s seem to behave better.
Have I completely misunderstood what is going on?
Even worse, I think I fixed the code ....
in case mode_s_to_n I changed:
if( !pblist->mergeclose( pblist, &mlist ) ||
!pglist->unmergeclose( pglist, &mlist ) )
if( !pblist->unmergeclose( pblist, &mlist ) ||
!pglist->mergeclose( pglist, &mlist ) )
Ack, you are absolutely correct. I will fix this in the next release tomorrow.
Thanks, it's working great now. I was pretty confused why all my email became spam so quickly. Things are alot saner now, it's doing a pretty good job at ID'ing my spam after 12 hours. Not perfect, but I'll give it some time.
It felt good to find a bug in C after not looking at code for several years!
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.