Filtering out HTML/IMG/vB code in Guestbook Entries

ChefFrank · 03-12-2006, 12:14 PM

Lately, I've been plagued with a number of spurious entries in my guestbook, advertising everything from "low, low, mortgage rates", to "increase length, girth and volume". At least they give me props for "a well designed website" and promise to "come on your site often". However, they always leave a link to some website, none of which are food related, which is what my site is about.

I'm running the 'standard' Perl guestbook (from Matt's Free Scripts), and the engine itself is actually built-in to the server - one of the benefits of my domain host is that they actually have all the work done for me - just copy the HTML code into my page, and it works! I would like to build a pre-processor that would either eliminate any links in the text section, or, just simply tell the user that the entry is not allowed.

I had thought about building a table of IP addresses or emails from the Guestlog that are originating these messages, but there isn't any pattern to where they're coming from.

Anyone got a routine that will scan a block of text, and either sanitize it and return the clean text, or return a value of 'good or bad' that I can then use to control the correct action (accept or reject text)?

Frank

sde · 03-12-2006, 01:19 PM

you get those entries from bots. people want to place links to their site for search engine optomization. you could look around for a guestbook with an image verification. most bots can't read an image, so it would probably prevent a lot of those entries.

my advice would be to just get rid of the guestbook all together. you have forums and chat that people can interact with. guestbooks are a bit dated.

ChefFrank · 03-12-2006, 01:44 PM

Yeah - that thought has crossed my mind, too . . . .

I did some checking, and found that there's version of the Perl Script that has this capability built in - there's a variable that won't allow HTML in the text. However, it may take some time before my host decides to install it. I may just install the Perl part in my cgi-bin directory and run from there instead of the version on my host's cgi-sys.

If anyone has a script though, I'd still be interested in looking at it just to see how it's done - I can do it in a couple of other languages, but nothing that would run on my host and has a CGI interface.

Frank

teknomage1 · 03-18-2006, 12:43 AM

If you post image generating portion of the perl code I can try to help you understand it / convert it to php.

Also, I agree with sde, guest books are very 90's. Blog style comment sections are all the rage these days. You could do cool stuff by having comments at the bottom of each recipe page.

DJMaze · 03-18-2006, 07:17 AM

http://opensourcecms.com/

ChefFrank · 03-22-2006, 10:17 AM

Thanks - I pretty much killed the guestbook. The Perl part of the code is actually under the control of my host, and while I could have put a version in my own cgi-bin directory, why bother?

I'm looking at the response to each page approch though. I have enough models about to try to figure it out myself. If I get stuck, I'll be back!!!

Thanks!

Frank