[Ukfreebsd] Notes from yesterdays meeting

Vanderdenduur rvrnt at icloud.com
Thu Oct 27 16:57:33 BST 2016


Hello,

On 26/10/2016 à 18:10, Sevan Janiyan wrote:

> Using privoxy to filter trackers

In connexion with Privoxy, I wanted to be a little more accurate. 
Please, visit their web page <https://www.privoxy.org/> and where I 
humbly contribute to 
<https://www.prxbx.com/forums/forumdisplay.php?fid=49>. PRXBX forum 
seems to be a good resource to start with Privoxy. I’m known as 
/Faxopita/ in the field Privoxy/ProxHTTPSProxy.

Privoxy, a local web-filtering proxy, is a superb tool to filter 
everything–including ads and trackers–modify web pages on the fly before 
they actually reach your web browser and block untrusted domain names. 
Natively, Privoxy can filter URLs based on either the host 
name–including HTTPS–or the path-side of the URL–excluding HTTPS.

    The pattern matching syntax is different for the host and path parts
    of the URL
    <http://www.privoxy.org/user-manual/actions-file.html#AF-PATTERNS>.
    The host part uses a simple globbing type matching technique, while
    the path part uses more flexible “Regular Expressions” (POSIX 1003.2).

However, if you want extended REGEX for hosts, you can compile with the 
option <https://sourceforge.net/p/ijbswa/support-requests/1706/#6376> 
|--enable-extended-host-patterns|, which is a “switch” to enable PCRE 
filtering for hostnames. Visit this blacklist page 
<http://185.82.200.39/blocklist.txt> as well.

The beauty is that blocking unwanted domains with Privoxy is seriously 
more flexible than resorting to a /hosts/ file. With metacharacters, you 
can really compress your block list versus hosts files. Also, If you 
want to block |www.ugly_domain.tld| and |ugly_domain.tld|, just input 
|.ugly_domain.tld| in your Privoxy-friendly block list. So, no need to 
double your entries because of the additional |www.|

Since Privoxy cannot filter the path side of encrypted requests–Privoxy 
is neither a MITM (nor caching web) proxy–you must enhance it with an 
“add-on”, such as ProxHTTPSProxy 
<https://www.prxbx.com/forums/forumdisplay.php?fid=48> (written in 
Python) to fix this undesirable outcome.

Drawback: security-wise, one must be careful and “trust” our 
machine/network… Indeed, using that kind of add-on will enable clear 
text queries made locally between your browser and ProxHTTPSProxy! You 
can exclude certain domains from being viewed by the add-on, such as 
banks, e-commercers, etc. if you prefer.

ProxHTTPSProxy is great for personal use. You can triple check the 
source code as well if you’re unsure. After 15 months of use, I must say 
it works great and I had no security issues so far.

Visit the PRXBX forum regularly for regular updates. RSS feed #1 
<https://www.prxbx.com/forums/syndication.php?fid=49&limit=15>, RSS feed 
#2 <https://www.prxbx.com/forums/syndication.php?fid=48&limit=15>.

Have a great Friday!

Raf

​


More information about the Ukfreebsd mailing list