DB won't train, 'Unsure' about everything

Since my previous thread I've narrowed down the problem.  I can now
describe the specific issue in the hope someone can explain it.

I created a wordlist.db from a text dump, and it doesn't work.

$ bogofilter -s -v -B <any file>
# 0 words, 0 messages

$ bogofilter -t -v -B <any file>
<any file> U 0.520000

db_verify says it's OK.

The source text dump looks OK, though it has a lot of non-ASCII such as
AU<C2><F2> 0 1 20230305   (as displayed by less)

I renamed this DB away and ran
$ bogofilter -n -v -B ./INBOX/63957
# 208 words, 1 message
which created a new small DB.  After training with a few ham and spams.
it gives sensible spam scores.

-but I need my old database and its years of training!

Are there any lines in the text dump that might be causing this, that I
can remove?

Should I chuck Berkeley DB and install the Sqlite bogofilter?

And why should a text dump that loads without error result in a DB that
doesn't work??!







_______________________________________________
bogofilter mailing list
[email protected]
https://www.bogofilter.org/mailman/listinfo/bogofilter