Mailman 3 February 2008 - spambayes-dev

imap filter traceback
by skip＠pobox.com Feb. 13, 2008

Feb. 13, 2008

I've been trying to get sb_imapfilter working with Gmail. It works, sort of, though I'm unconvinced it's really doing a lot. Running with the latest version from CVS (including Dave Abrahams' patch), I get lots of output like this when running in verbose mode: *[imapfilter] can't find saved message after 100 iterations: 25984 .[imapfilter] classified as None: 25985 Persisting messageinfo.fs state in database Persisting messageinfo.fs state in database *[imapfilter] can't find saved message after 100 iterations: 25985 .[imapfilter] classified as None: 25986 Persisting messageinfo.fs state in database Persisting messageinfo.fs state in database It would appear that it's failing to pull many messages and thus failing to classify them. In last night's run I saw lots of "classified as None" messages and every once in awhile an "already classified" message. Eventually it croaks with the following traceback: Traceback (most recent call last): File "/Users/skip/local/bin/sb_imapfilter.py", line 1328, in <module> run() File "/Users/skip/local/bin/sb_imapfilter.py", line 1310, in run imap_filter.Filter() File "/Users/skip/local/bin/sb_imapfilter.py", line 1116, in Filter self.unsure_folder, self.ham_folder) File "/Users/skip/local/bin/sb_imapfilter.py", line 962, in Filter for msg in self: File "/Users/skip/local/bin/sb_imapfilter.py", line 818, in __iter__ yield self[key] File "/Users/skip/local/bin/sb_imapfilter.py", line 847, in __getitem__ response = self.imap_server.uid("FETCH", key, "RFC822.HEADER") File "/Users/skip/local/lib/python2.5/imaplib.py", line 752, in uid typ, dat = self._simple_command(name, command, *args) File "/Users/skip/local/lib/python2.5/imaplib.py", line 1055, in _simple_command return self._command_complete(name, self._command(name, *args)) File "/Users/skip/local/lib/python2.5/imaplib.py", line 890, in _command_complete self._check_bye() File "/Users/skip/local/lib/python2.5/imaplib.py", line 807, in _check_bye raise self.abort(bye[-1]) imaplib.abort: System Error I'll open a bug report when I'm at work and back on the net. I know nothing about IMAP though, so I can't tell if this is caused by a bug in sb_imapfilter.py or if it reveals some sort of shortcoming with Gmail's IMAP implementation. Skip

1 0

Initial thoughts - SpamBayes + Gmail w/ IMAP *and* POP3
by skip＠pobox.com Feb. 11, 2008

Feb. 11, 2008

Okay, so tonight I got the SpamBayes IMAP filter working with Gmail after applying a patch that Dave Abrahams submitted about a year ago. (I've yet to check that in, but will do it tomorrow.) My current setup is Gmail w/ POP3 for download and SpamBayes filtering on my laptop. To this I added the IMAP filter. I haven't done too much with it yet, but it works. After a fashion. I'm not particularly keen on the way the IMAP filter operates, but that's probably just my inexperience with IMAP and the way it forces you to work. I wound up creating two labels, "sbham" and "sbspam". Those appear as folders to the IMAP filter. Any message I wanted it to train on get one of those labels. I have it classify messages in both my INBOX and Spam folders. Unfortunately, I tend to dump my spam periodically via the Gmail interface. Most messages with the sbspam label wind up in my Spam folder. I'll have to get in the habit of not deleting spam tagged with that label so I wind up with something to (re)train on later. Since Gmail has its own spam filter (which you can't disable) I can't help but feeling that the two systems are going to duke it out a bit. We'll see how that works. While I think about it, I also have a Yahoo! Mail address. That tool supports IMAP as well, but you have to upgrade to a "premium" mail service. I imagine it's not all that expensive, but I'll continue to experiment with free mail options for now. Skip

2 2

Major change to the website front page - feedback please
by skip＠pobox.com Feb. 8, 2008

Feb. 8, 2008

With all the questions today about what SpamBayes will work on, I think it would be a good idea to make it easier for people to decide what to download for different operating system/mail program combinations. I've attached a tentative replacement. I would like feedback on the table right at the top of the page. For combinations which haven't been tested (e.g., IncrediMail, Yahoo! Mail, etc), should I offer any download options? Maybe just a link to the 1.1a4 source? At any rate, feedback please. I'd like to install a replacement tomorrow if possible to try and head off many of the types of questions we received today. Thanks, Skip

2 2

Re: [spambayes-dev] [Spambayes] Major change to the website front page - feedback please
by David Reichard Feb. 8, 2008

Feb. 8, 2008

Skip: Looks great. Most open-source websites expect novice users to somehow guess which code to download. Same with manufacturer's websites for printer drivers. Wish more of them would have a table like the one you propose. One suggestion: Add a note (or footnote) "source code only" where relevant. I've never mastered installing form source code and so only download compiled code. It's nice to know which links lead only to source code so I can avoid them :( As for untested apps, yes, please list, with a note. - David -----Original Message----- From: spambayes-bounces(a)python.org [mailto:spambayes-bounces@python.org] On Behalf Of skip(a)pobox.com Sent: Thursday, February 07, 2008 8:31 PM To: spambayes-dev(a)python.org; spambayes(a)python.org Subject: [Spambayes] Major change to the website front page - feedback please With all the questions today about what SpamBayes will work on, I think it would be a good idea to make it easier for people to decide what to download for different operating system/mail program combinations. I've attached a tentative replacement. I would like feedback on the table right at the top of the page. For combinations which haven't been tested (e.g., IncrediMail, Yahoo! Mail, etc), should I offer any download options? Maybe just a link to the 1.1a4 source? At any rate, feedback please. I'd like to install a replacement tomorrow if possible to try and head off many of the types of questions we received today. Thanks, Skip

1 0