[Spambayes] unbounded "warning: unhandled exception"

[Tony Meyer]

> It seems likely that the error is generated more than once
> for each attempt to check mail.
> Eudora is configured to check mail every 5 minutes.

Can anyone else that has experienced this problem verify this?  If I unplug
my network cable (to simulate a loss of connection), then I get one error
message ("Can't connect to pop.ihug.co.nz:110: (7, 'getaddrinfo failed')")
every time Eudora (6.0) tries to check for mail (manually, or via the
timer).  I tried pulling the plug in the midst of downloading messages, and
that didn't change the result.

There seem to be two separate cases of this.  One has the "getaddrinfo
failed" error that I can duplicate, and the other has the "unhandled
exception" error.  I don't know what it is that causes that - can anyone
generate that error repeatedly?  Is it just by removing the network
connection?

Are you using SpamBayes with all defaults in a normal sort of setup, without
other things chained in the way, and so forth?

> I cleared out 40 gigabytes of log, and had a 4 gigabyte log
> inside of 4 hours (it may have been even faster.)

That's writing this error nearly 10,000 times per second
(10**9/60/60/len("warning: unhandled exception\n")).  Whatever is happening,
it's happening pretty quickly.

> I notice that when I don't have network connectivity, it 
> takes a long time for Eudora to fail when checking email.

With and without SpamBayes?

> My guess is that something about the current interaction is
> not making the problem clear to Eudora, and therefore resulting
> in a LOT of interactions, each of which produce an error line.

Errors simply pass through SpamBayes, so it'll get whatever the mail server
is sending.

As above, if anyone can give me steps to generate an "unhandled exception"
error, then I'll happily look into this.  For the moment, I can't think of
anything that SpamBayes would do that would cause 10,000 unhandled
exceptions per second.

=Tony.Meyer

-- 
Please always include the list (spambayes at python.org) in your replies
(reply-all), and please don't send me personal mail about SpamBayes.
http://www.massey.ac.nz/~tameyer/writing/reply_all.html explains this.