[Extension release] URL Checker For Posts And Signatures (Page 1 of 2)

Ok, so problem fixed and extension seems to be working. It's cross-checking URLs you try to put in your signature or post.

It's checking only precise match with csv file (so the http://www.buildingquotes.co.uk/ isn't fetched as "spam" but the http://www.buildingquotes.co.uk/boiler-service/ is) (because of the youtube and so on )

Oh, and btw I changed the .csv file a little bit - I erased all white-space before quotes. I need that because I'm checking URL with those quotes (for exact match and not only part of it)

The only remaining problem is how to update the list? I think i'll have to check how works updating of official extensions   

Improvement idea:
Like in stopforumspam extension; insert a table that shows

• what user (+IP) tried to submit an URL

• which URL it was

Advantages:

• possible to then ban users by cross checking with that new table

• it's a good way to see if it's working

What do you think?

Done. Download link 

BTW I found nice bug with this topic You can't search it (I tried "checker" "url checker" but I didn't found anything - guess I'll have to look whether search uses caching), but more importantly, this topic didn't show up in topics list for this forum - check out images)

Page 1
Page 2
Topic's last post

Is splitting topics core part of PunBB or is it an extension? Just thinking whether it could have any connection

@Grez

Just found a new variation of url being entered

This inbetween url tags (the backslash being new)

\http://spamsitename.com/\

so

[ url ] http://thespam/ [/ url]

blocked

but if they use

[ url =\http://thespam/\] spam keyword [/ url]

they get past it ....

If I put the link in my .csv like that it causes  when I try to post a topic with that url
An error was encountered

Note: For detailed error information (necessary for troubleshooting), enable "DEBUG mode". To enable "DEBUG mode", open up the file config.php in a text editor, add a line that looks like "define('FORUM_DEBUG', 1);" (without the quotation marks), and re-upload the file. Once you've solved the problem, it is recommended that "DEBUG mode" be turned off again (just remove the line from the file and re-upload it).

Any ideas?

Intention I have:
Block the url being posted AND show to admin (as it doesn't ban the user)

Some bots immediately tried the backslashes.

So what would have to be changed to get the result:
1. Warning as is now
+ if you publish this URL your account will be blocked.
2. If user tries to publish even then:
You  tried to publish a URL on the spam list, you're account is temporarily blocked. A moderator will contact you shortly, ( for example)

Edit: Example - I've had the same spammer come and post 3-8 posts, all with the same 8 links. If that code in URL Checker blocked the user as soon as he confirmed he's wants to send those links, dozens of users would not get an email with his spam links. Very important in my eyes.

Can you post the exact full post in code tags here
or email it to me  see my avatar....

and the IPs they've been using?

Edit: found some of their urls and keywords:

[url=http://www.overstocksilver.com/sterling-silver-rings.html]Silver Rings[/url]
[url=http://www.overstocksilver.com/[/url]]Sterling Silver Necklaces[/url]
[url=http://www.overstocksilver.com/diamondczrings.html]CZ Rings[/url]
[url=http://www.overstocksilver.com/tiinstsidicz.html]tiffany key pendant[/url]

* added

Now we just need our ace Grez to make this code take out those users posting those links... not just warning

Great stuff.  Just installed a slightly moded (style - I needed table font smaller) version of it here (packed into the stopforumspam extension).

Couple of features I'd find neat:

• Data-Forwarding: Collection of IP, Username, Email-Addy, Post messsge from db - then email that info along with the offending URL to the site admin

• and/or send a copy of the email to me/us (running the url list) so that we can add more URLs to the list - I assume a post will sometimes have new in addition to old URLs)

• Delete-message: maybe a message alerting to where they can have the URL delisted (if it got on the list by mistake)

• also echo the admins email addy on the page with banned reason

PS: The list would have already helped delete 3 bad IPs here today - look forward to the first auto delete Thanks again for the great work!!

I like the idea of importing the list into the MySQL table if its faster and more importantly can be updated via repository!
Duplicates in list: you mean www.badurl.xxx  www.badurl.xxx/xxx or  two times the exact same?

If the list has 10'000 domains (atm 1'700) and certain domains like ezinearticles (accounting for 1% of spam URLs) are blocked, it should get expensive and time intensive for the spammers, lowering their margins. That in turn makes the business model less attractive.

As always, it's not a golden bullet, but a further option.

PS: Which captchas do you use they can't get around?

PS: We just had our first auto-delete here Some chinese IP wanting bad signature.

Slavok deleted 20k users that got past that captcha in the last 12 months (or so). And several users including myself have reported that the CAPTCHA on by defaut has been circumvented.
So I don't know if you've changed any other settings (for the captchas or elsewhere) or if your forum is just one not being targeted...

@Audiofeeline
Thanks, have opened a ticket on trac
http://punbb.informer.com/trac/ticket/349

@Grez

How should we deal with such URLs (one with spaces, the other with backslash) ?

Hello everyoneforum Alex in dat house and I'm a dental tech besides cruising the internet I also like metal detecting
[url=\http://hot guitar.info/\]guitar training video[/url]
[url=\http://someauto. info/\]auto repair videos[/url]
[url=http://hot guitar.info/]guitar videos[/url]
[url=http://someauto. info/]auto review[/url]

Bug report:

if url tag is followed by domain with backslashes
\www.url.xxx/\

it doesn't get caught.