anomalous subjects

[Tom Anderson]

On Wed, 2004-11-24 at 23:16, David Relson wrote:
> I'm assuming 0.93.1, yes?  How are you running bogofilter?  What
> options?  What MTA?

Sendmail 8.12.11 -> Procmail 3.22 -> Bogofilter 0.92.8 -uep

> No attachments :-<  Best method is to create a .tgz or a.zip file and
> attach it.

Yeah, I suggested allowing multipart/digest through... that's how
Evolution forwards as attachment.

Well, let me shelve this for a moment, as now I don't think it's a
bogofilter issue with those emails.  I looked through my procmail logs
and it shows this:

>From Kennygui at emailgaul.com  Wed Nov 24 21:53:07 2004
 Subject: [SPAM] Work and Live in the USA 
  Folder: /var/spool/mail/tanderso					   1639

In my inbox, this was one of the spams missing the "[SPAM]", but it
shows it in the log file, which means that bogofilter output it
correctly, but it was stripped somewhere else.  I'll have investigate
further to find out what went on.

However, I don't know if this is coincidence or not, but just four
messages later, I had this error:

bogofilter: (db) db_get_dbvalue( 'aperiodic' ), err: -30982,
DB_RUNRECOVERY: Fatal error, run database recovery
bogofilter: (db) db_close err: -30982, DB_RUNRECOVERY: Fatal error, run
database recovery
procmail: Program failure (3) of "bogofilter"
procmail: Rescue of unfiltered data succeeded

Then two more successful emails, and then:

bogofilter: (db) db_get_dbvalue( 'ape' ), err: -30982, DB_RUNRECOVERY:
Fatal error, run database recovery
bogofilter: (db) db_close err: -30982, DB_RUNRECOVERY: Fatal error, run
database recovery
procmail: Program failure (3) of "bogofilter"
procmail: Rescue of unfiltered data succeeded

And then I ran a bfproxy correction run, and in the results, it
correctly processed 20 emails before bailing on the last two with an
exit code of 3.

I just verified this wordlist earlier today, so this just happened. 
I've rebuilt my wordlist again and verified it.  Before I upgraded from
17.x, I never had these kinds of problems.

I'll keep monitoring closely for any further anomalies.

Tom

P.S. In the FAQ, it might be wise to mention that db_verify may in fact
be db4.1_verify, which is otherwise difficult to locate.