On Tue, July 5, 2005 2:55 pm, Rene Brehmer said: > > Documented research indicate that on Mon, 4 Jul 2005 15:22:51 +0100, Gaby > vanhegan wrote: > >> On 4 Jul 2005, at 15:09, Miles Thompson wrote: >> >>> There is a lot of JUNK showing up on this list, and for me this one >>> was the last straw. From "Post Office" at [EMAIL PROTECTED] it had >>> an attachment named >> >> Likewise on the php-install list as well. Of the 20 emails I've had >> over the last few days, 3 have been legitimate posts... >> >> Gaby > > PHP DB list is pretty flooded as well ... I set up my filters to simply > dump all messages that are remotely like those to the trash ... might > loose > some valid in between, but the trash is starting to take over to the point > where it's getting too much work to dump it manually ...
Recently, I decided to Take Steps. Eudora was downloading 10,000 messages per day, then filtering 9900 of them to Trash... But my poor cable-modem and Mac 9100/100 just couldn't keep up. I spent more time waiting for Eudora to process email than I did working! Squirrel Mail is okay, so long as you log in frequently... Because its filters only run when you log in! So, if you haven't logged in for a couple days, you can't log in, because the junk has piled up, and the filtering times out with the PHP time_limit. Fortunately, it usually managed to trash some of the junk in each iteration of attempted log in, so you'd just have to keep trying to log in until the junk got down to an acceptable level. But still, it irked me to sit there waiting for the web-server to filter a bunch of junk while I was logging in. What numb-nuts thought that was a good idea? Run the filters all day, every day, and take out the trash every minute. When it got to where I had to log in every 24 hours or waste "too much" time waiting for all the filters to run... It started with a PHP script to pick through the stuff Spam Assassin hadn't caught, and get rid of even more junk: http://l-i-e.com/imap/index.phps I was winning for awhile... Over time, I built up my subject/body keywords of junk. Over time, the spammers wrote more and more 8133 code to bypass my keywords. Over time, the spammers cranked up more gear, and send more and more junk. I have conceded the arms race. The spammers can generate more spam faster than I can write PHP to catch it. And that's saying something. :-^ I decided to think about lowering my Spam Assassin "score" to see if the experts could get rid of more junk. I was worried about losing real email. I had NO IDEA how much of my real email was scoring how high. NO IDEA where to set the cutoff. Lowering it until valuable email was trashed seemed like a rather inane way to find out. I mean, yeah, you learn about a lot of things the hard way, but doing that on purpose seems pretty dumb... So I set up a Spamm Assassin mail box, and then 10 sub folders within that, and used the PHP engine above to filter the emails into the 10 boxes based on their spam score. Anything scoring 9 in Spam Assassin goes in the 9 box. Then scores of 8 in the 8 box. And so on. It's about a week or two now. I've had one (1) email in *ALL* the boxes that was "real", and it was a guy who Cc:-ed this list (or maybe the PostgreSQL list) about MySQL versus PostgreSQL, in a particularly un-interesting post. EVERYTHING ELSE that had any Spam Score at all was junk. My Inbox is back to managable levels, at least for now. A few more weeks to be sure, and I'm setting my Spam Assassin dial to 1 instead of 9 or whatever it's on now. Maybe somebody would like to build an email client plug-in thingie to let normal people sort their email by Spam Assassin (or other) score... It was certainly instructive for me! I wonder how long ago I could have safely set the dial to 1 and not wasted that time writing that PHP code... Well, I guess I learned how to do the IMAP thing, and that's not bad, eh? > But there's quite a few, atleast on the email version of the lists, that > have their vCard attached to their messages (they may not even know it), > so > we'd loose all those (valid) messages as well if the list has a flat rule > about no attachments. I don't really care, I just have to regularly empty > my attachments folder because of all those vCards.... Ah, yes. Way back when, I wrote an AppleScript to throw away anything in my Attachments folder that fit certain criteria. *.vcf was definitely one of the first to go. I'd give it to you but, A) that hard drive crashed, and B) you should do in PHP shell script with OS X anyway :-) You can also configure Eudora to not download attachments until you ask it to, which turned out to be best in the long run. Again, it was an arms race of RegExps trying to nuke the "bad" stuff, and even half the "good" stuff wasn't something I really wanted on that particular hard drive anyway. -- Like Music? http://l-i-e.com/artists.htm -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
