1 <!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook V3.1//EN" [
2 <!entity % dummy "IGNORE">
3 <!entity supported SYSTEM "supported.sgml">
4 <!entity newfeatures SYSTEM "newfeatures.sgml">
5 <!entity p-intro SYSTEM "privoxy.sgml">
6 <!entity seealso SYSTEM "seealso.sgml">
7 <!entity buildsource SYSTEM "buildsource.sgml">
8 <!entity contacting SYSTEM "contacting.sgml">
9 <!entity history SYSTEM "history.sgml">
10 <!entity copyright SYSTEM "copyright.sgml">
11 <!entity license SYSTEM "license.sgml">
12 <!entity GPLv2 SYSTEM "../../LICENSE">
13 <!entity p-authors SYSTEM "p-authors.sgml">
14 <!entity config SYSTEM "p-config.sgml">
15 <!entity changelog SYSTEM "changelog.sgml">
16 <!entity p-version "3.0.25">
17 <!entity p-status "UNRELEASED">
18 <!entity % p-authors-formal "INCLUDE"> <!-- include additional text, etc -->
19 <!entity % p-not-stable "INCLUDE">
20 <!entity % p-stable "IGNORE">
21 <!entity % p-text "IGNORE"> <!-- define we are not a text only doc -->
22 <!entity % p-doc "INCLUDE"> <!-- and we are a formal doc -->
23 <!entity % p-readme "IGNORE">
24 <!entity % user-man "IGNORE">
25 <!entity % config-file "IGNORE">
26 <!entity % p-supp-userman "IGNORE"> <!-- Omit some from supported.sgml -->
27 <!entity my-copy "©"> <!-- kludge for docbook2man -->
28 <!entity % draft "IGNORE"> <!-- WIP stuff -->
29 <!entity % seealso-extra "INCLUDE"> <!-- extra stuff from seealso.sgml -->
30 <!entity my-app "<application>Privoxy</application>">
33 File : $Source: /cvsroot/ijbswa/current/doc/source/user-manual.sgml,v $
36 This file belongs into
37 ijbswa.sourceforge.net:/home/groups/i/ij/ijbswa/htdocs/
39 $Id: user-manual.sgml,v 2.206 2016/03/17 10:43:07 fabiankeil Exp $
41 Copyright (C) 2001-2014 Privoxy Developers http://www.privoxy.org/
44 ========================================================================
45 NOTE: Please read developer-manual/documentation.html before touching
46 anything in this, or other Privoxy documentation.
47 ========================================================================
54 <title>Privoxy &p-version; User Manual</title>
58 <!-- Completely the wrong markup, but very little is allowed -->
59 <!-- in this part of an article. FIXME -->
60 <link linkend="copyright">Copyright</link> &my-copy; 2001-2014 by
61 <ulink url="http://www.privoxy.org/">Privoxy Developers</ulink>
65 <pubdate>$Id: user-manual.sgml,v 2.206 2016/03/17 10:43:07 fabiankeil Exp $</pubdate>
69 Note: the following should generate a separate page, and a live link to it,
70 all nicely done. But it doesn't for some mysterious reason. Please leave
71 commented unless it can be fixed proper. For the time being, the
72 copyright/license declarations will be in their own sgml.
85 This is here to keep vim syntax file from breaking :/
86 If I knew enough to fix it, I would.
87 PLEASE DO NOT REMOVE! HB: hal@foobox.net
93 The <citetitle>Privoxy User Manual</citetitle> gives users information on how to
94 install, configure and use <ulink
95 url="http://www.privoxy.org/">Privoxy</ulink>.
98 <!-- Include privoxy.sgml boilerplate: -->
100 <!-- end privoxy.sgml -->
103 You can find the latest version of the <citetitle>Privoxy User Manual</citetitle> at <ulink
104 url="http://www.privoxy.org/user-manual/">http://www.privoxy.org/user-manual/</ulink>.
105 Please see the <link linkend="contact">Contact section</link> on how to
106 contact the developers.
110 <!-- Feel free to send a note to the developers at <email>ijbswa-developers@lists.sourceforge.net</email>. -->
116 <!-- ~~~~~ New section ~~~~~ -->
117 <sect1 label="1" id="introduction"><title>Introduction</title>
119 This documentation is included with the current &p-status; version of
120 <application>Privoxy</application>, &p-version;<![%p-not-stable;[,
121 and is mostly complete at this point. The most up to date reference for the
122 time being is still the comments in the source files and in the individual
123 configuration files. Development of a new version is currently nearing
124 completion, and includes significant changes and enhancements over
128 <!-- include only in non-stable versions -->
131 Since this is a &p-status; version, not all new features are well tested. This
132 documentation may be slightly out of sync as a result (especially with
133 CVS sources). And there <emphasis>may be</emphasis> bugs, though hopefully
138 <!-- ~~~~~ New section ~~~~~ -->
139 <sect2 id="features"><title>Features</title>
141 In addition to the core
142 features of ad blocking and
143 <ulink url="http://en.wikipedia.org/wiki/Browser_cookie">cookie</ulink> management,
144 <application>Privoxy</application> provides many supplemental
145 features<![%p-not-stable;[, some of them currently under development]]>,
146 that give the end-user more control, more privacy and more freedom:
148 <!-- Include newfeatures.sgml boilerplate here: -->
150 <!-- end boilerplate -->
155 <!-- ~ End section ~ -->
158 <!-- ~~~~~ New section ~~~~~ -->
159 <sect1 id="installation"><title>Installation</title>
162 <application>Privoxy</application> is available both in convenient pre-compiled
163 packages for a wide range of operating systems, and as raw source code.
164 For most users, we recommend using the packages, which can be downloaded from our
165 <ulink url="http://sourceforge.net/projects/ijbswa/">Privoxy Project
171 On some platforms, the installer may remove previously installed versions, if
172 found. (See below for your platform). In any case <emphasis>be sure to backup
173 your old configuration if it is valuable to you.</emphasis> See the <link
174 linkend="upgradersnote">note to upgraders</link> section below.
177 <!-- ~~~~~ New section ~~~~~ -->
178 <sect2 id="installation-packages"><title>Binary Packages</title>
180 How to install the binary packages depends on your operating system:
183 <!-- XXX: The installation sections should be sorted -->
185 <!-- ~~~~~ New section ~~~~~ -->
186 <sect3 id="installation-deb"><title>Debian and Ubuntu</title>
188 DEBs can be installed with <literal>apt-get install privoxy</literal>,
189 and will use <filename>/etc/privoxy</filename> for the location of
194 <!-- ~~~~~ New section ~~~~~ -->
195 <sect3 id="installation-pack-win"><title>Windows</title>
198 Just double-click the installer, which will guide you through
199 the installation process. You will find the configuration files
200 in the same directory as you installed <application>Privoxy</application> in.
203 Version 3.0.5 beta introduced full <application>Windows</application> service
204 functionality. On Windows only, the <application>Privoxy</application>
205 program has two new command line arguments to install and uninstall
206 <application>Privoxy</application> as a <emphasis>service</emphasis>.
210 <term>Arguments:</term>
213 <replaceable class="parameter">--install</replaceable>[:<replaceable class="parameter">service_name</replaceable>]
216 <replaceable class="parameter">--uninstall</replaceable>[:<replaceable class="parameter">service_name</replaceable>]
222 After invoking <application>Privoxy</application> with
223 <command>--install</command>, you will need to bring up the
224 <application>Windows</application> service console to assign the user you
225 want <application>Privoxy</application> to run under, and whether or not you
226 want it to run whenever the system starts. You can start the
227 <application>Windows</application> services console with the following
228 command: <command>services.msc</command>. If you do not take the manual step
229 of modifying <application>Privoxy's</application> service settings, it will
230 not start. Note too that you will need to give Privoxy a user account that
231 actually exists, or it will not be permitted to
232 write to its log and configuration files.
237 <!-- ~~~~~ New section ~~~~~ -->
238 <sect3 id="installation-os2"><title>OS/2</title>
241 First, make sure that no previous installations of
242 <application>Junkbuster</application> and / or
243 <application>Privoxy</application> are left on your
244 system. Check that no <application>Junkbuster</application>
245 or <application>Privoxy</application> objects are in
251 Then, just double-click the WarpIN self-installing archive, which will
252 guide you through the installation process. A shadow of the
253 <application>Privoxy</application> executable will be placed in your
254 startup folder so it will start automatically whenever OS/2 starts.
258 The directory you choose to install <application>Privoxy</application>
259 into will contain all of the configuration files.
263 <!-- ~~~~~ New section ~~~~~ -->
264 <sect3 id="installation-mac"><title>Mac OS X</title>
266 Installation instructions for the OS X platform depend upon whether
267 you downloaded a ready-built installation package (.pkg or .mpkg) or have
268 downloaded the source code.
271 <sect3 renderas="sect4" id="OS-X-install-from-package">
272 <title>Installation from ready-built package</title>
274 The downloaded file will either be a .pkg (for OS X 10.5 upwards) or a bzipped
275 .mpkg file (for OS X 10.4). The former can be double-clicked as is and the
276 installation will start; double-clicking the latter will unzip the .mpkg file
277 which can then be double-clicked to commence the installation.
280 The privoxy service will automatically start after a successful installation
281 (and thereafter every time your computer starts up) however you will need to
282 configure your web browser(s) to use it. To do so, configure them to use a
283 proxy for HTTP and HTTPS at the address 127.0.0.1:8118.
286 To prevent the privoxy service from automatically starting when your computer
287 starts up, remove or rename the file <literal>/Library/LaunchDaemons/org.ijbswa.privoxy.plist</literal>
288 (on OS X 10.5 and higher) or the folder named
289 <literal>/Library/StartupItems/Privoxy</literal> (on OS X 10.4 'Tiger').
292 To manually start or stop the privoxy service, use the scripts startPrivoxy.sh
293 and stopPrivoxy.sh supplied in /Applications/Privoxy. They must be run from an
294 administrator account, using sudo.
297 To uninstall, run /Applications/Privoxy/uninstall.command as sudo from an
298 administrator account.
301 <sect3 renderas="sect4" id="OS-X-install-from-source">
302 <title>Installation from source</title>
304 To build and install the Privoxy source code on OS X you will need to obtain
305 the macsetup module from the Privoxy Sourceforge CVS repository (refer to
306 Sourceforge help for details of how to set up a CVS client to have read-only
307 access to the repository). This module contains scripts that leverage the usual
308 open-source tools (available as part of Apple's free of charge Xcode
309 distribution or via the usual open-source software package managers for OS X
310 (MacPorts, Homebrew, Fink etc.) to build and then install the privoxy binary
311 and associated files. The macsetup module's README file contains complete
312 instructions for its use.
315 The privoxy service will automatically start after a successful installation
316 (and thereafter every time your computer starts up) however you will need to
317 configure your web browser(s) to use it. To do so, configure them to use a
318 proxy for HTTP and HTTPS at the address 127.0.0.1:8118.
321 To prevent the privoxy service from automatically starting when your computer
322 starts up, remove or rename the file <literal>/Library/LaunchDaemons/org.ijbswa.privoxy.plist</literal>
323 (on OS X 10.5 and higher) or the folder named
324 <literal>/Library/StartupItems/Privoxy</literal> (on OS X 10.4 'Tiger').
327 To manually start or stop the privoxy service, use the Privoxy Utility
328 for Mac OS X (also part of the macsetup module). This application can start
329 and stop the privoxy service and display its log and configuration files.
332 To uninstall, run the macsetup module's uninstall.sh as sudo from an
333 administrator account.
337 <!-- ~~~~~ New section ~~~~~ -->
338 <sect3 id="installation-freebsd"><title>FreeBSD</title>
341 Privoxy is part of FreeBSD's Ports Collection, you can build and install
342 it with <literal>cd /usr/ports/www/privoxy; make install clean</literal>.
348 <!-- ~~~~~ New section ~~~~~ -->
349 <sect2 id="installation-source"><title>Building from Source</title>
352 The most convenient way to obtain the <application>Privoxy</application> sources
353 is to download the source tarball from our
354 <ulink url="http://sourceforge.net/project/showfiles.php?group_id=11118&package_id=10571">project download
359 If you like to live on the bleeding edge and are not afraid of using
360 possibly unstable development versions, you can check out the up-to-the-minute
361 version directly from <ulink url="http://sourceforge.net/cvs/?group_id=11118">the
362 CVS repository</ulink>.
364 deprecated...out of business.
365 or simply download <ulink
366 url="http://cvs.sourceforge.net/cvstarballs/ijbswa-cvsroot.tar.bz2">the nightly CVS
371 <!-- include buildsource.sgml boilerplate: -->
373 <!-- end boilerplate -->
376 <!-- ~~~~~ New section ~~~~~ -->
377 <sect2 id="installation-keepupdated"><title>Keeping your Installation Up-to-Date</title>
380 If you wish to receive an email notification whenever we release updates of
381 <application>Privoxy</application> or the actions file, <ulink
382 url="http://lists.sourceforge.net/lists/listinfo/ijbswa-announce/">subscribe
383 to our announce mailing list</ulink>, ijbswa-announce@lists.sourceforge.net.
387 In order not to lose your personal changes and adjustments when updating
388 to the latest <literal>default.action</literal> file we <emphasis>strongly
389 recommend</emphasis> that you use <literal>user.action</literal> and
390 <literal>user.filter</literal> for your local
391 customizations of <application>Privoxy</application>. See the <link
392 linkend="actions-file">Chapter on actions files</link> for details.
400 <!-- ~ End section ~ -->
402 <!-- ~~~~~ New section ~~~~~ -->
403 <sect1 id="whatsnew">
404 <title>What's New in this Release</title>
408 <!-- ~~~~~ New section ~~~~~ -->
410 <sect2 id="upgradersnote">
411 <title>Note to Upgraders</title>
414 A quick list of things to be aware of before upgrading from earlier
415 versions of <application>Privoxy</application>:
423 The recommended way to upgrade &my-app; is to backup your old
424 configuration files, install the new ones, verify that &my-app;
425 is working correctly and finally merge back your changes using
426 <application>diff</application> and maybe <application>patch</application>.
429 There are a number of new features in each &my-app; release and
430 most of them have to be explicitly enabled in the configuration
431 files. Old configuration files obviously don't do that and due
432 to syntax changes using old configuration files with a new
433 &my-app; isn't always possible anyway.
438 Note that some installers remove earlier versions completely,
439 including configuration files, therefore you should really save
440 any important configuration files!
445 On the other hand, other installers don't overwrite existing configuration
446 files, thinking you will want to do that yourself.
451 In the default configuration only fatal errors are logged now.
452 You can change that in the <link linkend="DEBUG">debug section</link>
453 of the configuration file. You may also want to enable more verbose
454 logging until you verified that the new &my-app; version is working
461 Three other config file settings are now off by default:
462 <link linkend="enable-remote-toggle">enable-remote-toggle</link>,
463 <link linkend="enable-remote-http-toggle">enable-remote-http-toggle</link>,
464 and <link linkend="enable-edit-actions">enable-edit-actions</link>.
465 If you use or want these, you will need to explicitly enable them, and
466 be aware of the security issues involved.
473 What constitutes a <quote>default</quote> configuration has changed,
474 and you may want to review which actions are <quote>on</quote> by
475 default. This is primarily a matter of emphasis, but some features
476 you may have been used to, may now be <quote>off</quote> by default.
477 There are also a number of new actions and filters you may want to
478 consider, most of which are not fully incorporated into the default
479 settings as yet (see above).
486 The default actions setting is now <literal>Cautious</literal>. Previous
487 releases had a default setting of <literal>Medium</literal>. Experienced
488 users may want to adjust this, as it is fairly conservative by &my-app;
489 standards and past practices. See <ulink
490 url="http://config.privoxy.org/edit-actions-list?f=default">
491 http://config.privoxy.org/edit-actions-list?f=default</ulink>. New users
492 should try the default settings for a while before turning up the volume.
498 The default setting has filtering turned <emphasis>off</emphasis>, which
499 subsequently means that compression is <emphasis>on</emphasis>. Remember
500 that filtering does not work on compressed pages, so if you use, or want to
501 use, filtering, you will need to force compression off. Example:
505 { +<link linkend="filter">filter</link>{google} +<link linkend="prevent-compression">prevent-compression</link> }
509 Or if you use a number of filters, or filter many sites, you may just want
510 to turn off compression for all sites in
511 <filename>default.action</filename> (or
512 <filename>user.action</filename>).
519 Also, <link linkend="SESSION-COOKIES-ONLY">session-cookies-only</link> is
520 off by default now. If you've liked this feature in the past, you may want
521 to turn it back on in <filename>user.action</filename> now.
528 Some installers may not automatically start
529 <application>Privoxy</application> after installation.
540 <!-- ~~~~~ New section ~~~~~ -->
541 <sect1 id="quickstart"><title>Quickstart to Using Privoxy</title>
547 Install <application>Privoxy</application>. See the <link
548 linkend="installation">Installation Section</link> below for platform specific
555 Advanced users and those who want to offer <application>Privoxy</application>
556 service to more than just their local machine should check the <link
557 linkend="config">main config file</link>, especially the <link
558 linkend="access-control">security-relevant</link> options. These are
565 Start <application>Privoxy</application>, if the installation program has
566 not done this already (may vary according to platform). See the section
567 <link linkend="startup">Starting <application>Privoxy</application></link>.
573 Set your browser to use <application>Privoxy</application> as HTTP and
574 HTTPS (SSL) <ulink url="http://en.wikipedia.org/wiki/Proxy_server">proxy</ulink>
575 by setting the proxy configuration for address of
576 <literal>127.0.0.1</literal> and port <literal>8118</literal>.
577 <emphasis>DO NOT</emphasis> activate proxying for <literal>FTP</literal> or
578 any protocols besides HTTP and HTTPS (SSL) unless you intend to prevent your
579 browser from using these protocols.
585 Flush your browser's disk and memory caches, to remove any cached ad images.
586 If using <application>Privoxy</application> to manage
587 <ulink url="http://en.wikipedia.org/wiki/Browser_cookie">cookies</ulink>,
588 you should remove any currently stored cookies too.
594 A default installation should provide a reasonable starting point for
595 most. There will undoubtedly be occasions where you will want to adjust the
596 configuration, but that can be dealt with as the need arises. Little
597 to no initial configuration is required in most cases, you may want
599 <ulink url="config.html#ENABLE-EDIT-ACTIONS">web-based action editor</ulink> though.
600 Be sure to read the warnings first.
603 See the <link linkend="configuration">Configuration section</link> for more
604 configuration options, and how to customize your installation.
605 You might also want to look at the <link
606 linkend="quickstart-ad-blocking">next section</link> for a quick
607 introduction to how <application>Privoxy</application> blocks ads and
614 If you experience ads that slip through, innocent images that are
615 blocked, or otherwise feel the need to fine-tune
616 <application>Privoxy's</application> behavior, take a look at the <link
617 linkend="actions-file">actions files</link>. As a quick start, you might
618 find the <link linkend="act-examples">richly commented examples</link>
619 helpful. You can also view and edit the actions files through the <ulink
620 url="http://config.privoxy.org">web-based user interface</ulink>. The
621 Appendix <quote><link linkend="actionsanat">Troubleshooting: Anatomy of an
622 Action</link></quote> has hints on how to understand and debug actions that
623 <quote>misbehave</quote>.
629 Please see the section <link linkend="contact">Contacting the
630 Developers</link> on how to report bugs, problems with websites or to get
637 Now enjoy surfing with enhanced control, comfort and privacy!
645 <!-- ~~~~~ New section ~~~~~ -->
647 <sect2 id="quickstart-ad-blocking">
648 <title>Quickstart to Ad Blocking</title>
650 NOTE: This section is deliberately redundant for those that don't
651 want to read the whole thing (which is getting lengthy).
654 Ad blocking is but one of <application>Privoxy's</application>
655 array of features. Many of these features are for the technically minded advanced
656 user. But, ad and banner blocking is surely common ground for everybody.
659 This section will provide a quick summary of ad blocking so
660 you can get up to speed quickly without having to read the more extensive
661 information provided below, though this is highly recommended.
664 First a bit of a warning ... blocking ads is much like blocking SPAM: the
665 more aggressive you are about it, the more likely you are to block
666 things that were not intended. And the more likely that some things
667 may not work as intended. So there is a trade off here. If you want
668 extreme ad free browsing, be prepared to deal with more
669 <quote>problem</quote> sites, and to spend more time adjusting the
670 configuration to solve these unintended consequences. In short, there is
671 not an easy way to eliminate <emphasis>all</emphasis> ads. Either take
672 the easy way and settle for <emphasis>most</emphasis> ads blocked with the
673 default configuration, or jump in and tweak it for your personal surfing
674 habits and preferences.
677 Secondly, a brief explanation of <application>Privoxy's </application>
678 <quote>actions</quote>. <quote>Actions</quote> in this context, are
679 the directives we use to tell <application>Privoxy</application> to perform
680 some task relating to HTTP transactions (i.e. web browsing). We tell
681 <application>Privoxy</application> to take some <quote>action</quote>. Each
682 action has a unique name and function. While there are many potential
683 <application>actions</application> in <application>Privoxy's</application>
684 arsenal, only a few are used for ad blocking. <link
685 linkend="actions">Actions</link>, and <link linkend="actions-file">action
686 configuration files</link>, are explained in depth below.
689 Actions are specified in <application>Privoxy's</application> configuration,
690 followed by one or more URLs to which the action should apply. URLs
691 can actually be URL type <link linkend="af-patterns">patterns</link> that use
692 wildcards so they can apply potentially to a range of similar URLs. The
693 actions, together with the URL patterns are called a section.
696 When you connect to a website, the full URL will either match one or more
697 of the sections as defined in <application>Privoxy's</application> configuration,
698 or not. If so, then <application>Privoxy</application> will perform the
699 respective actions. If not, then nothing special happens. Furthermore, web
700 pages may contain embedded, secondary URLs that your web browser will
701 use to load additional components of the page, as it parses the
702 original page's HTML content. An ad image for instance, is just an URL
703 embedded in the page somewhere. The image itself may be on the same server,
704 or a server somewhere else on the Internet. Complex web pages will have many
705 such embedded URLs. &my-app; can deal with each URL individually, so, for
706 instance, the main page text is not touched, but images from such-and-such
711 The most important actions for basic ad blocking are: <literal><link
712 linkend="block">block</link></literal>, <literal><link
713 linkend="handle-as-image">handle-as-image</link></literal>,
715 linkend="handle-as-empty-document">handle-as-empty-document</link></literal>,and
716 <literal><link linkend="set-image-blocker">set-image-blocker</link></literal>:
724 <literal><link linkend="block">block</link></literal> - this is perhaps
725 the single most used action, and is particularly important for ad blocking.
726 This action stops any contact between your browser and any URL patterns
727 that match this action's configuration. It can be used for blocking ads,
728 but also anything that is determined to be unwanted. By itself, it simply
729 stops any communication with the remote server and sends
730 <application>Privoxy</application>'s own built-in BLOCKED page instead to
731 let you now what has happened (with some exceptions, see below).
737 <literal><link linkend="handle-as-image">handle-as-image</link></literal> -
738 tells <application>Privoxy</application> to treat this URL as an image.
739 <application>Privoxy</application>'s default configuration already does this
740 for all common image types (e.g. GIF), but there are many situations where this
741 is not so easy to determine. So we'll force it in these cases. This is particularly
742 important for ad blocking, since only if we know that it's an image of
743 some kind, can we replace it with an image of our choosing, instead of the
744 <application>Privoxy</application> BLOCKED page (which would only result in
745 a <quote>broken image</quote> icon). There are some limitations to this
746 though. For instance, you can't just brute-force an image substitution for
747 an entire HTML page in most situations.
753 <literal><link linkend="handle-as-empty-document">handle-as-empty-document</link></literal> -
754 sends an empty document instead of <application>Privoxy's</application>
755 normal BLOCKED HTML page. This is useful for file types that are neither
756 HTML nor images, such as blocking JavaScript files.
763 linkend="set-image-blocker">set-image-blocker</link></literal> - tells
764 <application>Privoxy</application> what to display in place of an ad image that
765 has hit a block rule. For this to come into play, the URL must match a
766 <literal><link linkend="block">block</link></literal> action somewhere in the
767 configuration, <emphasis>and</emphasis>, it must also match an
768 <literal><link linkend="handle-as-image">handle-as-image</link></literal> action.
771 The configuration options on what to display instead of the ad are:
775 <emphasis>pattern</emphasis> - a checkerboard pattern, so that an ad
776 replacement is obvious. This is the default.
781 <emphasis>blank</emphasis> - A very small empty GIF image is displayed.
782 This is the so-called <quote>invisible</quote> configuration option.
787 <emphasis>http://<URL></emphasis> - A redirect to any image anywhere
788 of the user's choosing (advanced usage).
797 Advanced users will eventually want to explore &my-app;
798 <literal><link linkend="filter">filters</link></literal> as well. Filters
799 are very different from <literal><link
800 linkend="block">blocks</link></literal>.
801 A <quote>block</quote> blocks a site, page, or unwanted contented. Filters
802 are a way of filtering or modifying what is actually on the page. An example
803 filter usage: a text replacement of <quote>no-no</quote> for
804 <quote>nasty-word</quote>. That is a very simple example. This process can be
805 used for ad blocking, but it is more in the realm of advanced usage and has
806 some pitfalls to be wary off.
810 The quickest way to adjust any of these settings is with your browser through
811 the special <application>Privoxy</application> editor at <ulink
812 url="http://config.privoxy.org/show-status">http://config.privoxy.org/show-status</ulink>
813 (shortcut: <ulink url="http://p.p/">http://p.p/show-status</ulink>). This
814 is an internal page, and does not require Internet access.
818 Note that as of <application>Privoxy</application> 3.0.7 beta the
819 action editor is disabled by default. Check the
820 <ulink url="config.html#ENABLE-EDIT-ACTIONS">enable-edit-actions
821 section in the configuration file</ulink> to learn why and in which
822 cases it's safe to enable again.
826 If you decided to enable the action editor, select the appropriate
827 <quote>actions</quote> file, and click
828 <quote><guibutton>Edit</guibutton></quote>. It is best to put personal or
829 local preferences in <filename>user.action</filename> since this is not
830 meant to be overwritten during upgrades, and will over-ride the settings in
831 other files. Here you can insert new <quote>actions</quote>, and URLs for ad
832 blocking or other purposes, and make other adjustments to the configuration.
833 <application>Privoxy</application> will detect these changes automatically.
837 A quick and simple step by step example:
845 Right click on the ad image to be blocked, then select
846 <quote><guimenuitem>Copy Link Location</guimenuitem></quote> from the
854 url="http://config.privoxy.org/show-status">http://config.privoxy.org/show-status</ulink>
859 Find <filename>user.action</filename> in the top section, and click
860 on <quote><guibutton>Edit</guibutton></quote>:
863 <!-- image of editor and actions files selections -->
865 <figure pgwide="0" float="0"><title>Actions Files in Use</title>
868 <imagedata fileref="files-in-use.jpg" format="jpg">
871 <phrase>[ Screenshot of Actions Files in Use ]</phrase>
880 You should have a section with only
881 <literal><link linkend="block">block</link></literal> listed under
882 <quote>Actions:</quote>.
883 If not, click a <quote><guibutton>Insert new section below</guibutton></quote>
884 button, and in the new section that just appeared, click the
885 <guibutton>Edit</guibutton> button right under the word <quote>Actions:</quote>.
886 This will bring up a list of all actions. Find
887 <literal><link linkend="block">block</link></literal> near the top, and click
888 in the <quote>Enabled</quote> column, then <quote><guibutton>Submit</guibutton></quote>
894 Now, in the <literal><link linkend="block">block</link></literal> actions section,
895 click the <quote><guibutton>Add</guibutton></quote> button, and paste the URL the
896 browser got from <quote><guimenuitem>Copy Link Location</guimenuitem></quote>.
897 Remove the <literal>http://</literal> at the beginning of the URL. Then, click
898 <quote><guibutton>Submit</guibutton></quote> (or
899 <quote><guibutton>OK</guibutton></quote> if in a pop-up window).
904 Now go back to the original page, and press <keycap>SHIFT-Reload</keycap>
905 (or flush all browser caches). The image should be gone now.
913 This is a very crude and simple example. There might be good reasons to use a
914 wildcard pattern match to include potentially similar images from the same
915 site. For a more extensive explanation of <quote>patterns</quote>, and
916 the entire actions concept, see <link linkend="actions-file">the Actions
921 For advanced users who want to hand edit their config files, you might want
922 to now go to the <link linkend="act-examples">Actions Files Tutorial</link>.
923 The ideas explained therein also apply to the web-based editor.
926 There are also various
927 <link linkend="filter">filters</link> that can be used for ad blocking
928 (filters are a special subset of actions). These
929 fall into the <quote>advanced</quote> usage category, and are explained in
930 depth in later sections.
937 <!-- ~ End section ~ -->
940 <!-- ~~~~~ New section ~~~~~ -->
942 <title>Starting Privoxy</title>
944 Before launching <application>Privoxy</application> for the first time, you
945 will want to configure your browser(s) to use
946 <application>Privoxy</application> as a HTTP and HTTPS (SSL)
947 <ulink url="http://en.wikipedia.org/wiki/Proxy_server">proxy</ulink>. The default is
948 127.0.0.1 (or localhost) for the proxy address, and port 8118 (earlier versions
949 used port 8000). This is the one configuration step <emphasis>that must be done
953 Please note that <application>Privoxy</application> can only proxy HTTP and
954 HTTPS traffic. It will not work with FTP or other protocols.
957 <!-- image of Mozilla Proxy configuration -->
959 <figure pgwide="0" float="0"><title>Proxy Configuration Showing
960 Mozilla/Netscape HTTP and HTTPS (SSL) Settings</title>
963 <imagedata fileref="proxy_setup.jpg" format="jpg">
966 <phrase>[ Screenshot of Mozilla Proxy Configuration ]</phrase>
974 With <application>Firefox</application>, this is typically set under:
978 <guibutton>Tools</guibutton> -> <guibutton>Options</guibutton> -> <guibutton>Advanced</guibutton> -> <guibutton>Network</guibutton> -><guibutton>Connection</guibutton> -> <guibutton>Settings</guibutton>
983 Or optionally on some platforms:
987 <guibutton>Edit</guibutton> -> <guibutton>Preferences</guibutton> -> <guibutton>General</guibutton> -> <guibutton>Connection Settings</guibutton> -> <guibutton>Manual Proxy Configuration</guibutton>
993 With <application>Netscape</application> (and
994 <application>Mozilla</application>), this can be set under:
999 <!-- Mix ascii and gui art, something for everybody -->
1000 <!-- spacing on this is tricky -->
1001 <guibutton>Edit</guibutton> -> <guibutton>Preferences</guibutton> -> <guibutton>Advanced</guibutton> -> <guibutton>Proxies</guibutton> -> <guibutton>HTTP Proxy</guibutton>
1006 For <application>Internet Explorer v.5-7</application>:
1010 <guibutton>Tools</guibutton> -> <guibutton>Internet Options</guibutton> -> <guibutton>Connections</guibutton> -> <guibutton>LAN Settings</guibutton>
1014 Then, check <quote>Use Proxy</quote> and fill in the appropriate info
1015 (Address: 127.0.0.1, Port: 8118). Include HTTPS (SSL), if you want HTTPS
1016 proxy support too (sometimes labeled <quote>Secure</quote>). Make sure any
1017 checkboxes like <quote>Use the same proxy server for all protocols</quote> is
1018 <emphasis>UNCHECKED</emphasis>. You want only HTTP and HTTPS (SSL)!
1021 <!-- image of IE Proxy configuration -->
1023 <figure pgwide="0" float="0"><title>Proxy Configuration Showing
1024 Internet Explorer HTTP and HTTPS (Secure) Settings</title>
1027 <imagedata fileref="proxy2.jpg" format="jpg">
1030 <phrase>[ Screenshot of IE Proxy Configuration ]</phrase>
1038 After doing this, flush your browser's disk and memory caches to force a
1039 re-reading of all pages and to get rid of any ads that may be cached. Remove
1040 any <ulink url="http://en.wikipedia.org/wiki/Browser_cookie">cookies</ulink>,
1041 if you want <application>Privoxy</application> to manage that. You are now
1042 ready to start enjoying the benefits of using
1043 <application>Privoxy</application>!
1047 <application>Privoxy</application> itself is typically started by specifying the
1048 main configuration file to be used on the command line. If no configuration
1049 file is specified on the command line, <application>Privoxy</application>
1050 will look for a file named <filename>config</filename> in the current
1051 directory. Except on Win32 where it will try <filename>config.txt</filename>.
1054 <sect2 id="start-debian">
1055 <title>Debian</title>
1057 We use a script. Note that Debian typically starts &my-app; upon booting per
1058 default. It will use the file
1059 <filename>/etc/privoxy/config</filename> as its main configuration
1064 # /etc/init.d/privoxy start
1069 <sect2 id="start-freebsd">
1070 <title>FreeBSD and ElectroBSD</title>
1072 To start <application>Privoxy</application> upon booting, add
1073 "privoxy_enable='YES'" to <filename>/etc/rc.conf</filename>.
1074 <application>Privoxy</application> will use
1075 <filename>/usr/local/etc/privoxy/config</filename> as its main
1079 If you installed <application>Privoxy</application> into a jail, the
1080 paths above are relative to the jail root.
1083 To start <application>Privoxy</application> manually, run:
1087 # service privoxy onestart
1092 <sect2 id="start-windows">
1093 <title>Windows</title>
1095 Click on the &my-app; Icon to start <application>Privoxy</application>. If no configuration file is
1096 specified on the command line, <application>Privoxy</application> will look
1097 for a file named <filename>config.txt</filename>. Note that Windows will
1098 automatically start &my-app; when the system starts if you chose that option
1102 <application>Privoxy</application> can run with full Windows service functionality.
1103 On Windows only, the &my-app; program has two new command line arguments
1104 to install and uninstall &my-app; as a service. See the
1105 <link linkend="installation-pack-win">Windows Installation
1106 instructions</link> for details.
1110 <sect2 id="start-unices">
1111 <title>Generic instructions for Unix derivates (Solaris, NetBSD, HP-UX etc.)</title>
1113 Example Unix startup command:
1117 # /usr/sbin/privoxy --user privoxy /etc/privoxy/config
1121 Note that if you installed <application>Privoxy</application> through
1122 a package manager, the package will probably contain a platform-specific
1123 script or configuration file to start <application>Privoxy</application>
1128 <sect2 id="start-os2">
1131 During installation, <application>Privoxy</application> is configured to
1132 start automatically when the system restarts. You can start it manually by
1133 double-clicking on the <application>Privoxy</application> icon in the
1134 <application>Privoxy</application> folder.
1138 <sect2 id="start-macosx">
1139 <title>Mac OS X</title>
1141 The privoxy service will automatically start after a successful installation
1142 (and thereafter every time your computer starts up) however you will need to
1143 configure your web browser(s) to use it. To do so, configure them to use a
1144 proxy for HTTP and HTTPS at the address 127.0.0.1:8118.
1147 To prevent the privoxy service from automatically starting when your computer
1148 starts up, remove or rename the file <literal>/Library/LaunchDaemons/org.ijbswa.privoxy.plist</literal>
1149 (on OS X 10.5 and higher) or the folder named
1150 <literal>/Library/StartupItems/Privoxy</literal> (on OS X 10.4 'Tiger').
1153 To manually start or stop the privoxy service, use the scripts startPrivoxy.sh
1154 and stopPrivoxy.sh supplied in /Applications/Privoxy. They must be run from an
1155 administrator account, using sudo.
1163 See the section <link linkend="cmdoptions">Command line options</link> for
1167 must find a better place for this paragraph
1170 The included default configuration files should give a reasonable starting
1171 point. Most of the per site configuration is done in the
1172 <ulink url="actions-file.html"><quote>actions</quote></ulink> files. These are
1173 where various cookie actions are defined, ad and banner blocking, and other
1174 aspects of <application>Privoxy</application> configuration. There are several
1175 such files included, with varying levels of aggressiveness.
1179 You will probably want to keep an eye out for sites for which you may prefer
1180 persistent cookies, and add these to your actions configuration as needed. By
1181 default, most of these will be accepted only during the current browser
1182 session (aka <quote>session cookies</quote>), unless you add them to the
1183 configuration. If you want the browser to handle this instead, you will need
1184 to edit <filename>user.action</filename> (or through the web based interface)
1185 and disable this feature. If you use more than one browser, it would make
1186 more sense to let <application>Privoxy</application> handle this. In which
1187 case, the browser(s) should be set to accept all cookies.
1191 Another feature where you will probably want to define exceptions for trusted
1192 sites is the popup-killing (through <ulink
1193 url="actions-file.html#FILTER-POPUPS"><quote>+filter{popups}</quote></ulink>),
1194 because your favorite shopping, banking, or leisure site may need
1195 popups (explained below).
1199 <application>Privoxy</application> does not support all of the optional HTTP/1.1
1200 features yet. In the unlikely event that you experience inexplicable problems
1201 with browsers that use HTTP/1.1 per default
1202 (like <application>Mozilla</application> or recent versions of I.E.), you might
1203 try to force HTTP/1.0 compatibility. For Mozilla, look under <literal>Edit ->
1204 Preferences -> Debug -> Networking</literal>.
1205 Alternatively, set the <quote>+downgrade-http-version</quote> config option in
1206 <filename>default.action</filename> which will downgrade your browser's HTTP
1207 requests from HTTP/1.1 to HTTP/1.0 before processing them.
1211 After running <application>Privoxy</application> for a while, you can
1212 start to fine tune the configuration to suit your personal, or site,
1213 preferences and requirements. There are many, many aspects that can
1214 be customized. <quote>Actions</quote>
1215 can be adjusted by pointing your browser to
1216 <ulink url="http://config.privoxy.org/">http://config.privoxy.org/</ulink>
1217 (shortcut: <ulink url="http://p.p/">http://p.p/</ulink>),
1218 and then follow the link to <quote>View & Change the Current Configuration</quote>.
1219 (This is an internal page and does not require Internet access.)
1223 In fact, various aspects of <application>Privoxy</application>
1224 configuration can be viewed from this page, including
1225 current configuration parameters, source code version numbers,
1226 the browser's request headers, and <quote>actions</quote> that apply
1227 to a given URL. In addition to the actions file
1228 editor mentioned above, <application>Privoxy</application> can also
1229 be turned <quote>on</quote> and <quote>off</quote> (toggled) from this page.
1233 If you encounter problems, try loading the page without
1234 <application>Privoxy</application>. If that helps, enter the URL where
1235 you have the problems into <ulink url="http://p.p/show-url-info">the browser
1236 based rule tracing utility</ulink>. See which rules apply and why, and
1237 then try turning them off for that site one after the other, until the problem
1238 is gone. When you have found the culprit, you might want to turn the rest on
1243 If the above paragraph sounds gibberish to you, you might want to <link
1244 linkend="actions-file">read more about the actions concept</link>
1245 or even dive deep into the <link linkend="actionsanat">Appendix
1250 If you can't get rid of the problem at all, think you've found a bug in
1251 Privoxy, want to propose a new feature or smarter rules, please see the
1252 section <link linkend="contact"><quote>Contacting the
1253 Developers</quote></link> below.
1258 <!-- ~~~~~ New section ~~~~~ -->
1259 <sect2 id="cmdoptions">
1260 <title>Command Line Options</title>
1262 <application>Privoxy</application> may be invoked with the following
1263 command-line options:
1271 <emphasis>--config-test</emphasis>
1274 Exit after loading the configuration files before binding to
1275 the listen address. The exit code signals whether or not the
1276 configuration files have been successfully loaded.
1279 If the exit code is 1, at least one of the configuration files
1280 is invalid, if it is 0, all the configuration files have been
1281 successfully loaded (but may still contain errors that can
1282 currently only be detected at run time).
1285 This option doesn't affect the log setting, combination with
1286 <emphasis>--no-daemon</emphasis> is recommended if a configured
1287 log file shouldn't be used.
1292 <emphasis>--version</emphasis>
1295 Print version info and exit. Unix only.
1300 <emphasis>--help</emphasis>
1303 Print short usage info and exit. Unix only.
1308 <emphasis>--no-daemon</emphasis>
1311 Don't become a daemon, i.e. don't fork and become process group
1312 leader, and don't detach from controlling tty. Unix only.
1317 <emphasis>--pidfile FILE</emphasis>
1320 On startup, write the process ID to <emphasis>FILE</emphasis>. Delete the
1321 <emphasis>FILE</emphasis> on exit. Failure to create or delete the
1322 <emphasis>FILE</emphasis> is non-fatal. If no <emphasis>FILE</emphasis>
1323 option is given, no PID file will be used. Unix only.
1328 <emphasis>--user USER[.GROUP]</emphasis>
1331 After (optionally) writing the PID file, assume the user ID of
1332 <emphasis>USER</emphasis>, and if included the GID of GROUP. Exit if the
1333 privileges are not sufficient to do so. Unix only.
1338 <emphasis>--chroot</emphasis>
1341 Before changing to the user ID given in the <emphasis>--user</emphasis> option,
1342 chroot to that user's home directory, i.e. make the kernel pretend to the &my-app;
1343 process that the directory tree starts there. If set up carefully, this can limit
1344 the impact of possible vulnerabilities in &my-app; to the files contained in that hierarchy.
1350 <emphasis>--pre-chroot-nslookup hostname</emphasis>
1353 Specifies a hostname (for example www.privoxy.org) to look up before doing a chroot.
1354 On some systems, initializing the resolver library involves reading config files from
1355 /etc and/or loading additional shared libraries from /lib.
1356 On these systems, doing a hostname lookup before the chroot reduces
1357 the number of files that must be copied into the chroot tree.
1360 For fastest startup speed, a good value is a hostname that is not in /etc/hosts but that
1361 your local name server (listed in /etc/resolv.conf) can resolve without recursion
1362 (that is, without having to ask any other name servers). The hostname need not exist,
1363 but if it doesn't, an error message (which can be ignored) will be output.
1369 <emphasis>configfile</emphasis>
1372 If no <emphasis>configfile</emphasis> is included on the command line,
1373 <application>Privoxy</application> will look for a file named
1374 <quote>config</quote> in the current directory (except on Win32
1375 where it will look for <quote>config.txt</quote> instead). Specify
1376 full path to avoid confusion. If no config file is found,
1377 <application>Privoxy</application> will fail to start.
1385 On <application>MS Windows</application> only there are two additional
1386 command-line options to allow <application>Privoxy</application> to install and
1387 run as a <emphasis>service</emphasis>. See the
1388 <link linkend="installation-pack-win">Window Installation section</link>
1396 <!-- ~ End section ~ -->
1399 <!-- ~~~~~ New section ~~~~~ -->
1400 <sect1 id="configuration"><title>Privoxy Configuration</title>
1402 All <application>Privoxy</application> configuration is stored
1403 in text files. These files can be edited with a text editor.
1404 Many important aspects of <application>Privoxy</application> can
1405 also be controlled easily with a web browser.
1409 <!-- ~~~~~ New section ~~~~~ -->
1411 <sect2 id="control-with-webbrowser">
1412 <title>Controlling Privoxy with Your Web Browser</title>
1414 <application>Privoxy</application>'s user interface can be reached through the special
1415 URL <ulink url="http://config.privoxy.org/">http://config.privoxy.org/</ulink>
1416 (shortcut: <ulink url="http://p.p/">http://p.p/</ulink>),
1417 which is a built-in page and works without Internet access.
1418 You will see the following section:
1422 <!-- Needs to be put in a table and colorized -->
1425 <bridgehead renderas="sect2"> Privoxy Menu</bridgehead>
1429 ▪ <ulink url="http://config.privoxy.org/show-status">View & change the current configuration</ulink>
1432 ▪ <ulink url="http://config.privoxy.org/show-version">View the source code version numbers</ulink>
1435 ▪ <ulink url="http://config.privoxy.org/show-request">View the request headers.</ulink>
1438 ▪ <ulink url="http://config.privoxy.org/show-url-info">Look up which actions apply to a URL and why</ulink>
1441 ▪ <ulink url="http://config.privoxy.org/toggle">Toggle Privoxy on or off</ulink>
1444 ▪ <ulink
1445 url="http://www.privoxy.org/&p-version;/user-manual/">Documentation</ulink>
1453 This should be self-explanatory. Note the first item leads to an editor for the
1454 <link linkend="actions-file">actions files</link>, which is where the ad, banner,
1455 cookie, and URL blocking magic is configured as well as other advanced features of
1456 <application>Privoxy</application>. This is an easy way to adjust various
1457 aspects of <application>Privoxy</application> configuration. The actions
1458 file, and other configuration files, are explained in detail below.
1462 <quote>Toggle Privoxy On or Off</quote> is handy for sites that might
1463 have problems with your current actions and filters. You can in fact use
1464 it as a test to see whether it is <application>Privoxy</application>
1465 causing the problem or not. <application>Privoxy</application> continues
1466 to run as a proxy in this case, but all manipulation is disabled, i.e.
1467 <application>Privoxy</application> acts like a normal forwarding proxy.
1471 Note that several of the features described above are disabled by default
1472 in <application>Privoxy</application> 3.0.7 beta and later.
1474 <ulink url="config.html">configuration file</ulink> to learn why
1475 and in which cases it's safe to enable them again.
1480 <!-- ~ End section ~ -->
1485 <!-- ~~~~~ New section ~~~~~ -->
1487 <sect2 id="confoverview">
1488 <title>Configuration Files Overview</title>
1490 For Unix, *BSD and Linux, all configuration files are located in
1491 <filename>/etc/privoxy/</filename> by default. For MS Windows, OS/2, and
1492 AmigaOS these are all in the same directory as the
1493 <application>Privoxy</application> executable. <![%p-not-stable;[ The name
1494 and number of configuration files has changed from previous versions, and is
1495 subject to change as development progresses.]]>
1499 The installed defaults provide a reasonable starting point, though
1500 some settings may be aggressive by some standards. For the time being, the
1501 principle configuration files are:
1509 The <link linkend="config">main configuration file</link> is named <filename>config</filename>
1510 on Linux, Unix, BSD, OS/2, and AmigaOS and <filename>config.txt</filename>
1511 on Windows. This is a required file.
1517 <filename>match-all.action</filename> is used to define which <quote>actions</quote>
1518 relating to banner-blocking, images, pop-ups, content modification, cookie handling
1519 etc should be applied by default. It should be the first actions file loaded.
1522 <filename>default.action</filename> defines many exceptions (both positive and negative)
1523 from the default set of actions that's configured in <filename>match-all.action</filename>.
1524 It should be the second actions file loaded and shouldn't be edited by the user.
1527 Multiple actions files may be defined in <filename>config</filename>. These
1528 are processed in the order they are defined. Local customizations and locally
1529 preferred exceptions to the default policies as defined in
1530 <filename>match-all.action</filename> (which you will most probably want
1531 to define sooner or later) are best applied in <filename>user.action</filename>,
1532 where you can preserve them across upgrades. The file isn't installed by all
1533 installers, but you can easily create it yourself with a text editor.
1536 There is also a web based editor that can be accessed from
1538 url="http://config.privoxy.org/show-status">http://config.privoxy.org/show-status</ulink>
1540 url="http://p.p/show-status">http://p.p/show-status</ulink>) for the
1541 various actions files.
1547 <quote>Filter files</quote> (the <link linkend="filter-file">filter
1548 file</link>) can be used to re-write the raw page content, including
1549 viewable text as well as embedded HTML and JavaScript, and whatever else
1550 lurks on any given web page. The filtering jobs are only pre-defined here;
1551 whether to apply them or not is up to the actions files.
1552 <filename>default.filter</filename> includes various filters made
1553 available for use by the developers. Some are much more intrusive than
1554 others, and all should be used with caution. You may define additional
1555 filter files in <filename>config</filename> as you can with
1556 actions files. We suggest <filename>user.filter</filename> for any
1557 locally defined filters or customizations.
1565 The syntax of the configuration and filter files may change between different
1566 Privoxy versions, unfortunately some enhancements cost backwards compatibility.
1567 <!-- Add link to documentation-->
1571 All files use the <quote><literal>#</literal></quote> character to denote a
1572 comment (the rest of the line will be ignored) and understand line continuation
1573 through placing a backslash ("<literal>\</literal>") as the very last character
1574 in a line. If the <literal>#</literal> is preceded by a backslash, it looses
1575 its special function. Placing a <literal>#</literal> in front of an otherwise
1576 valid configuration line to prevent it from being interpreted is called "commenting
1577 out" that line. Blank lines are ignored.
1581 The actions files and filter files
1582 can use Perl style <link linkend="regex">regular expressions</link> for
1583 maximum flexibility.
1587 After making any changes, there is no need to restart
1588 <application>Privoxy</application> in order for the changes to take
1589 effect. <application>Privoxy</application> detects such changes
1590 automatically. Note, however, that it may take one or two additional
1591 requests for the change to take effect. When changing the listening address
1592 of <application>Privoxy</application>, these <quote>wake up</quote> requests
1593 must obviously be sent to the <emphasis>old</emphasis> listening address.
1598 While under development, the configuration content is subject to change.
1599 The below documentation may not be accurate by the time you read this.
1600 Also, what constitutes a <quote>default</quote> setting, may change, so
1601 please check all your configuration files on important issues.
1607 <!-- ~ End section ~ -->
1610 <!-- ~~~~~~~~ New section Header ~~~~~~~~~ -->
1612 <!-- **************************************************** -->
1613 <!-- Include config.sgml here -->
1614 <!-- This is where the entire config file is detailed. -->
1616 <!-- end include -->
1619 <!-- ~ End section ~ -->
1623 <!-- ~~~~~~~~ New section Header ~~~~~~~~~ -->
1625 <sect1 id="actions-file"><title>Actions Files</title>
1629 XXX: similar descriptions are in the Configuration Files sections.
1630 We should only describe them at one place.
1633 The actions files are used to define what <emphasis>actions</emphasis>
1634 <application>Privoxy</application> takes for which URLs, and thus determines
1635 how ad images, cookies and various other aspects of HTTP content and
1636 transactions are handled, and on which sites (or even parts thereof).
1637 There are a number of such actions, with a wide range of functionality.
1638 Each action does something a little different.
1639 These actions give us a veritable arsenal of tools with which to exert
1640 our control, preferences and independence. Actions can be combined so that
1641 their effects are aggregated when applied against a given set of URLs.
1645 are three action files included with <application>Privoxy</application> with
1652 <filename>match-all.action</filename> - is used to define which
1653 <quote>actions</quote> relating to banner-blocking, images, pop-ups,
1654 content modification, cookie handling etc should be applied by default.
1655 It should be the first actions file loaded
1660 <filename>default.action</filename> - defines many exceptions (both
1661 positive and negative) from the default set of actions that's configured
1662 in <filename>match-all.action</filename>. It is a set of rules that should
1663 work reasonably well as-is for most users. This file is only supposed to
1664 be edited by the developers. It should be the second actions file loaded.
1669 <filename>user.action</filename> - is intended to be for local site
1670 preferences and exceptions. As an example, if your ISP or your bank
1671 has specific requirements, and need special handling, this kind of
1672 thing should go here. This file will not be upgraded.
1677 <guibutton>Edit</guibutton> <guibutton>Set to Cautious</guibutton> <guibutton>Set to Medium</guibutton> <guibutton>Set to Advanced</guibutton>
1680 These have increasing levels of aggressiveness <emphasis>and have no
1681 influence on your browsing unless you select them explicitly in the
1682 editor</emphasis>. A default installation should be pre-set to
1683 <literal>Cautious</literal>. New users should try this for a while before
1684 adjusting the settings to more aggressive levels. The more aggressive
1685 the settings, then the more likelihood there is of problems such as sites
1686 not working as they should.
1689 The <guibutton>Edit</guibutton> button allows you to turn each
1690 action on/off individually for fine-tuning. The <guibutton>Cautious</guibutton>
1691 button changes the actions list to low/safe settings which will activate
1692 ad blocking and a minimal set of &my-app;'s features, and subsequently
1693 there will be less of a chance for accidental problems. The
1694 <guibutton>Medium</guibutton> button sets the list to a medium level of
1695 other features and a low level set of privacy features. The
1696 <guibutton>Advanced</guibutton> button sets the list to a high level of
1697 ad blocking and medium level of privacy. See the chart below. The latter
1698 three buttons over-ride any changes via with the
1699 <guibutton>Edit</guibutton> button. More fine-tuning can be done in the
1700 lower sections of this internal page.
1703 While the actions file editor allows to enable these settings in all
1704 actions files, they are only supposed to be enabled in the first one
1705 to make sure you don't unintentionally overrule earlier rules.
1708 The default profiles, and their associated actions, as pre-defined in
1709 <filename>default.action</filename> are:
1712 <table frame=all><title>Default Configurations</title>
1713 <tgroup cols=4 align=left colsep=1 rowsep=1>
1714 <colspec colname=c1>
1715 <colspec colname=c2>
1716 <colspec colname=c3>
1717 <colspec colname=c4>
1720 <entry>Feature</entry>
1721 <entry>Cautious</entry>
1722 <entry>Medium</entry>
1723 <entry>Advanced</entry>
1728 <!-- <entry>f1</entry> -->
1729 <!-- <entry>f2</entry> -->
1730 <!-- <entry>f3</entry> -->
1731 <!-- <entry>f4</entry> -->
1737 <entry>Ad-blocking Aggressiveness</entry>
1738 <entry>medium</entry>
1744 <entry>Ad-filtering by size</entry>
1751 <entry>Ad-filtering by link</entry>
1757 <entry>Pop-up killing</entry>
1758 <entry>blocks only</entry>
1759 <entry>blocks only</entry>
1760 <entry>blocks only</entry>
1764 <entry>Privacy Features</entry>
1766 <entry>medium</entry>
1767 <entry>medium/high</entry>
1771 <entry>Cookie handling</entry>
1773 <entry>session-only</entry>
1778 <entry>Referer forging</entry>
1785 <entry>GIF de-animation</entry>
1792 <entry>Fast redirects</entry>
1799 <entry>HTML taming</entry>
1806 <entry>JavaScript taming</entry>
1813 <entry>Web-bug killing</entry>
1820 <entry>Image tag reordering</entry>
1836 The list of actions files to be used are defined in the main configuration
1837 file, and are processed in the order they are defined (e.g.
1838 <filename>default.action</filename> is typically processed before
1839 <filename>user.action</filename>). The content of these can all be viewed and
1841 url="http://config.privoxy.org/show-status">http://config.privoxy.org/show-status</ulink>.
1842 The over-riding principle when applying actions, is that the last action that
1843 matches a given URL wins. The broadest, most general rules go first
1844 (defined in <filename>default.action</filename>),
1845 followed by any exceptions (typically also in
1846 <filename>default.action</filename>), which are then followed lastly by any
1847 local preferences (typically in <emphasis>user</emphasis><filename>.action</filename>).
1848 Generally, <filename>user.action</filename> has the last word.
1852 An actions file typically has multiple sections. If you want to use
1853 <quote>aliases</quote> in an actions file, you have to place the (optional)
1854 <link linkend="aliases">alias section</link> at the top of that file.
1855 Then comes the default set of rules which will apply universally to all
1856 sites and pages (be <emphasis>very careful</emphasis> with using such a
1857 universal set in <filename>user.action</filename> or any other actions file after
1858 <filename>default.action</filename>, because it will override the result
1859 from consulting any previous file). And then below that,
1860 exceptions to the defined universal policies. You can regard
1861 <filename>user.action</filename> as an appendix to <filename>default.action</filename>,
1862 with the advantage that it is a separate file, which makes preserving your
1863 personal settings across <application>Privoxy</application> upgrades easier.
1867 Actions can be used to block anything you want, including ads, banners, or
1868 just some obnoxious URL whose content you would rather not see. Cookies can be accepted
1869 or rejected, or accepted only during the current browser session (i.e. not
1870 written to disk), content can be modified, some JavaScripts tamed, user-tracking
1871 fooled, and much more. See below for a <link linkend="actions">complete list
1875 <!-- ~~~~~ New section ~~~~~ -->
1876 <sect2 id="right-mix">
1877 <title>Finding the Right Mix</title>
1879 Note that some <link linkend="actions">actions</link>, like cookie suppression
1880 or script disabling, may render some sites unusable that rely on these
1881 techniques to work properly. Finding the right mix of actions is not always easy and
1882 certainly a matter of personal taste. And, things can always change, requiring
1883 refinements in the configuration. In general, it can be said that the more
1884 <quote>aggressive</quote> your default settings (in the top section of the
1885 actions file) are, the more exceptions for <quote>trusted</quote> sites you
1886 will have to make later. If, for example, you want to crunch all cookies per
1887 default, you'll have to make exceptions from that rule for sites that you
1888 regularly use and that require cookies for actually useful purposes, like maybe
1889 your bank, favorite shop, or newspaper.
1893 We have tried to provide you with reasonable rules to start from in the
1894 distribution actions files. But there is no general rule of thumb on these
1895 things. There just are too many variables, and sites are constantly changing.
1896 Sooner or later you will want to change the rules (and read this chapter again :).
1900 <!-- ~~~~~ New section ~~~~~ -->
1901 <sect2 id="how-to-edit">
1902 <title>How to Edit</title>
1904 The easiest way to edit the actions files is with a browser by
1905 using our browser-based editor, which can be reached from <ulink
1906 url="http://config.privoxy.org/show-status">http://config.privoxy.org/show-status</ulink>.
1907 Note: the config file option <link
1908 linkend="enable-edit-actions">enable-edit-actions</link> must be enabled for
1909 this to work. The editor allows both fine-grained control over every single
1910 feature on a per-URL basis, and easy choosing from wholesale sets of defaults
1911 like <quote>Cautious</quote>, <quote>Medium</quote> or
1912 <quote>Advanced</quote>. Warning: the <quote>Advanced</quote> setting is more
1913 aggressive, and will be more likely to cause problems for some sites.
1914 Experienced users only!
1918 If you prefer plain text editing to GUIs, you can of course also directly edit the
1919 the actions files with your favorite text editor. Look at
1920 <filename>default.action</filename> which is richly commented with many
1926 <sect2 id="actions-apply">
1927 <title>How Actions are Applied to Requests</title>
1929 Actions files are divided into sections. There are special sections,
1930 like the <quote><link linkend="aliases">alias</link></quote> sections which will
1931 be discussed later. For now let's concentrate on regular sections: They have a
1932 heading line (often split up to multiple lines for readability) which consist
1933 of a list of actions, separated by whitespace and enclosed in curly braces.
1934 Below that, there is a list of URL and tag patterns, each on a separate line.
1938 To determine which actions apply to a request, the URL of the request is
1939 compared to all URL patterns in each <quote>action file</quote>.
1940 Every time it matches, the list of applicable actions for the request is
1941 incrementally updated, using the heading of the section in which the
1942 pattern is located. The same is done again for tags and tag patterns later on.
1946 If multiple applying sections set the same action differently,
1947 the last match wins. If not, the effects are aggregated.
1948 E.g. a URL might match a regular section with a heading line of <literal>{
1949 +<link linkend="handle-as-image">handle-as-image</link> }</literal>,
1950 then later another one with just <literal>{
1951 +<link linkend="block">block</link> }</literal>, resulting
1952 in <emphasis>both</emphasis> actions to apply. And there may well be
1953 cases where you will want to combine actions together. Such a section then
1959 { +<literal>handle-as-image</literal> +<literal>block{Banner ads.}</literal> }
1960 # Block these as if they were images. Send no block page.
1962 media.example.com/.*banners
1963 .example.com/images/ads/</screen>
1967 You can trace this process for URL patterns and any given URL by visiting <ulink
1968 url="http://config.privoxy.org/show-url-info">http://config.privoxy.org/show-url-info</ulink>.
1972 Examples and more detail on this is provided in the Appendix, <link linkend="ACTIONSANAT">
1973 Troubleshooting: Anatomy of an Action</link> section.
1977 <!-- ~~~~~ New section ~~~~~ -->
1978 <sect2 id="af-patterns">
1979 <title>Patterns</title>
1981 As mentioned, <application>Privoxy</application> uses <quote>patterns</quote>
1982 to determine what <emphasis>actions</emphasis> might apply to which sites and
1983 pages your browser attempts to access. These <quote>patterns</quote> use wild
1984 card type <emphasis>pattern</emphasis> matching to achieve a high degree of
1985 flexibility. This allows one expression to be expanded and potentially match
1986 against many similar patterns.
1990 Generally, an URL pattern has the form
1991 <literal><host><port>/<path></literal>, where the
1992 <literal><host></literal>, the <literal><port></literal>
1993 and the <literal><path></literal> are optional. (This is why the special
1994 <literal>/</literal> pattern matches all URLs). Note that the protocol
1995 portion of the URL pattern (e.g. <literal>http://</literal>) should
1996 <emphasis>not</emphasis> be included in the pattern. This is assumed already!
1999 The pattern matching syntax is different for the host and path parts of
2000 the URL. The host part uses a simple globbing type matching technique,
2001 while the path part uses more flexible
2002 <ulink url="http://en.wikipedia.org/wiki/Regular_expressions"><quote>Regular
2003 Expressions</quote></ulink> (POSIX 1003.2).
2006 The port part of a pattern is a decimal port number preceded by a colon
2007 (<literal>:</literal>). If the host part contains a numerical IPv6 address,
2008 it has to be put into angle brackets
2009 (<literal><</literal>, <literal>></literal>).
2014 <term><literal>www.example.com/</literal></term>
2017 is a host-only pattern and will match any request to <literal>www.example.com</literal>,
2018 regardless of which document on that server is requested. So ALL pages in
2019 this domain would be covered by the scope of this action. Note that a
2020 simple <literal>example.com</literal> is different and would NOT match.
2025 <term><literal>www.example.com</literal></term>
2028 means exactly the same. For host-only patterns, the trailing <literal>/</literal> may
2034 <term><literal>www.example.com/index.html</literal></term>
2037 matches all the documents on <literal>www.example.com</literal>
2038 whose name starts with <literal>/index.html</literal>.
2043 <term><literal>www.example.com/index.html$</literal></term>
2046 matches only the single document <literal>/index.html</literal>
2047 on <literal>www.example.com</literal>.
2052 <term><literal>/index.html$</literal></term>
2055 matches the document <literal>/index.html</literal>, regardless of the domain,
2056 i.e. on <emphasis>any</emphasis> web server anywhere.
2061 <term><literal>/</literal></term>
2064 Matches any URL because there's no requirement for either the
2065 domain or the path to match anything.
2070 <term><literal>:8000/</literal></term>
2073 Matches any URL pointing to TCP port 8000.
2078 <term><literal>10.0.0.1/</literal></term>
2081 Matches any URL with the host address <literal>10.0.0.1</literal>.
2082 (Note that the real URL uses plain brackets, not angle brackets.)
2087 <term><literal><2001:db8::1>/</literal></term>
2090 Matches any URL with the host address <literal>2001:db8::1</literal>.
2091 (Note that the real URL uses plain brackets, not angle brackets.)
2096 <term><literal>index.html</literal></term>
2099 matches nothing, since it would be interpreted as a domain name and
2100 there is no top-level domain called <literal>.html</literal>. So its
2108 <!-- ~~~~~ New section ~~~~~ -->
2109 <sect3 id="host-pattern"><title>The Host Pattern</title>
2112 The matching of the host part offers some flexible options: if the
2113 host pattern starts or ends with a dot, it becomes unanchored at that end.
2114 The host pattern is often referred to as domain pattern as it is usually
2115 used to match domain names and not IP addresses.
2121 <term><literal>.example.com</literal></term>
2124 matches any domain with first-level domain <literal>com</literal>
2125 and second-level domain <literal>example</literal>.
2126 For example <literal>www.example.com</literal>,
2127 <literal>example.com</literal> and <literal>foo.bar.baz.example.com</literal>.
2128 Note that it wouldn't match if the second-level domain was <literal>another-example</literal>.
2133 <term><literal>www.</literal></term>
2136 matches any domain that <emphasis>STARTS</emphasis> with
2137 <literal>www.</literal> (It also matches the domain
2138 <literal>www</literal> but most of the time that doesn't matter.)
2143 <term><literal>.example.</literal></term>
2146 matches any domain that <emphasis>CONTAINS</emphasis> <literal>.example.</literal>.
2147 And, by the way, also included would be any files or documents that exist
2148 within that domain since no path limitations are specified. (Correctly
2149 speaking: It matches any FQDN that contains <literal>example</literal> as
2150 a domain.) This might be <literal>www.example.com</literal>,
2151 <literal>news.example.de</literal>, or
2152 <literal>www.example.net/cgi/testing.pl</literal> for instance. All these
2160 Additionally, there are wild-cards that you can use in the domain names
2161 themselves. These work similarly to shell globbing type wild-cards:
2162 <quote>*</quote> represents zero or more arbitrary characters (this is
2164 <ulink url="http://en.wikipedia.org/wiki/Regular_expressions"><quote>Regular
2165 Expression</quote></ulink> based syntax of <quote>.*</quote>),
2166 <quote>?</quote> represents any single character (this is equivalent to the
2167 regular expression syntax of a simple <quote>.</quote>), and you can define
2168 <quote>character classes</quote> in square brackets which is similar to
2169 the same regular expression technique. All of this can be freely mixed:
2174 <term><literal>ad*.example.com</literal></term>
2177 matches <quote>adserver.example.com</quote>,
2178 <quote>ads.example.com</quote>, etc but not <quote>sfads.example.com</quote>
2183 <term><literal>*ad*.example.com</literal></term>
2186 matches all of the above, and then some.
2191 <term><literal>.?pix.com</literal></term>
2194 matches <literal>www.ipix.com</literal>,
2195 <literal>pictures.epix.com</literal>, <literal>a.b.c.d.e.upix.com</literal> etc.
2200 <term><literal>www[1-9a-ez].example.c*</literal></term>
2203 matches <literal>www1.example.com</literal>,
2204 <literal>www4.example.cc</literal>, <literal>wwwd.example.cy</literal>,
2205 <literal>wwwz.example.com</literal> etc., but <emphasis>not</emphasis>
2206 <literal>wwww.example.com</literal>.
2213 While flexible, this is not the sophistication of full regular expression based syntax.
2218 <!-- ~ End section ~ -->
2221 <!-- ~~~~~ New section ~~~~~ -->
2222 <sect3 id="path-pattern"><title>The Path Pattern</title>
2225 <application>Privoxy</application> uses <quote>modern</quote> POSIX 1003.2
2226 <ulink url="http://en.wikipedia.org/wiki/Regular_expressions"><quote>Regular
2227 Expressions</quote></ulink> for matching the path portion (after the slash),
2228 and is thus more flexible.
2232 There is an <link linkend="regex">Appendix</link> with a brief quick-start into regular
2233 expressions, you also might want to have a look at your operating system's documentation
2234 on regular expressions (try <literal>man re_format</literal>).
2238 Note that the path pattern is automatically left-anchored at the <quote>/</quote>,
2239 i.e. it matches as if it would start with a <quote>^</quote> (regular expression speak
2240 for the beginning of a line).
2244 Please also note that matching in the path is <emphasis>CASE INSENSITIVE</emphasis>
2245 by default, but you can switch to case sensitive at any point in the pattern by using the
2246 <quote>(?-i)</quote> switch: <literal>www.example.com/(?-i)PaTtErN.*</literal> will match
2247 only documents whose path starts with <literal>PaTtErN</literal> in
2248 <emphasis>exactly</emphasis> this capitalization.
2253 <term><literal>.example.com/.*</literal></term>
2256 Is equivalent to just <quote>.example.com</quote>, since any documents
2257 within that domain are matched with or without the <quote>.*</quote>
2258 regular expression. This is redundant
2263 <term><literal>.example.com/.*/index.html$</literal></term>
2266 Will match any page in the domain of <quote>example.com</quote> that is
2267 named <quote>index.html</quote>, and that is part of some path. For
2268 example, it matches <quote>www.example.com/testing/index.html</quote> but
2269 NOT <quote>www.example.com/index.html</quote> because the regular
2270 expression called for at least two <quote>/'s</quote>, thus the path
2271 requirement. It also would match
2272 <quote>www.example.com/testing/index_html</quote>, because of the
2273 special meta-character <quote>.</quote>.
2278 <term><literal>.example.com/(.*/)?index\.html$</literal></term>
2281 This regular expression is conditional so it will match any page
2282 named <quote>index.html</quote> regardless of path which in this case can
2283 have one or more <quote>/'s</quote>. And this one must contain exactly
2284 <quote>.html</quote> (but does not have to end with that!).
2289 <term><literal>.example.com/(.*/)(ads|banners?|junk)</literal></term>
2292 This regular expression will match any path of <quote>example.com</quote>
2293 that contains any of the words <quote>ads</quote>, <quote>banner</quote>,
2294 <quote>banners</quote> (because of the <quote>?</quote>) or <quote>junk</quote>.
2295 The path does not have to end in these words, just contain them.
2300 <term><literal>.example.com/(.*/)(ads|banners?|junk)/.*\.(jpe?g|gif|png)$</literal></term>
2303 This is very much the same as above, except now it must end in either
2304 <quote>.jpg</quote>, <quote>.jpeg</quote>, <quote>.gif</quote> or <quote>.png</quote>. So this
2305 one is limited to common image formats.
2312 There are many, many good examples to be found in <filename>default.action</filename>,
2313 and more tutorials below in <link linkend="regex">Appendix on regular expressions</link>.
2318 <!-- ~ End section ~ -->
2321 <!-- ~~~~~ New section ~~~~~ -->
2322 <sect3 id="tag-pattern"><title>The Request Tag Pattern</title>
2325 Request tag patterns are used to change the applying actions based on the
2326 request's tags. Tags can be created based on HTTP headers with either
2327 the <link linkend="CLIENT-HEADER-TAGGER">client-header-tagger</link>
2328 or the <link linkend="SERVER-HEADER-TAGGER">server-header-tagger</link> action.
2332 Request tag patterns have to start with <quote>TAG:</quote>, so &my-app;
2333 can tell them apart from other patterns. Everything after the colon
2334 including white space, is interpreted as a regular expression with
2335 path pattern syntax, except that tag patterns aren't left-anchored
2336 automatically (&my-app; doesn't silently add a <quote>^</quote>,
2337 you have to do it yourself if you need it).
2341 To match all requests that are tagged with <quote>foo</quote>
2342 your pattern line should be <quote>TAG:^foo$</quote>,
2343 <quote>TAG:foo</quote> would work as well, but it would also
2344 match requests whose tags contain <quote>foo</quote> somewhere.
2345 <quote>TAG: foo</quote> wouldn't work as it requires white space.
2349 Sections can contain URL and request tag patterns at the same time,
2350 but request tag patterns are checked after the URL patterns and thus
2351 always overrule them, even if they are located before the URL patterns.
2355 Once a new request tag is added, Privoxy checks right away if it's matched by one
2356 of the request tag patterns and updates the action settings accordingly. As a result
2357 request tags can be used to activate other tagger actions, as long as these other
2358 taggers look for headers that haven't already be parsed.
2362 For example you could tag client requests which use the
2363 <literal>POST</literal> method,
2364 then use this tag to activate another tagger that adds a tag if cookies
2365 are sent, and then use a block action based on the cookie tag. This allows
2366 the outcome of one action, to be input into a subsequent action. However if
2367 you'd reverse the position of the described taggers, and activated the
2368 method tagger based on the cookie tagger, no method tags would be created.
2369 The method tagger would look for the request line, but at the time
2370 the cookie tag is created, the request line has already been parsed.
2374 While this is a limitation you should be aware of, this kind of
2375 indirection is seldom needed anyway and even the example doesn't
2376 make too much sense.
2381 <!-- ~~~~~ New section ~~~~~ -->
2382 <sect3 id="negative-tag-patterns"><title>The Negative Request Tag Patterns</title>
2385 To match requests that do not have a certain request tag, specify a negative tag pattern
2386 by prefixing the tag pattern line with either <quote>NO-REQUEST-TAG:</quote>
2387 or <quote>NO-RESPONSE-TAG:</quote> instead of <quote>TAG:</quote>.
2391 Negative request tag patterns created with <quote>NO-REQUEST-TAG:</quote> are checked
2392 after all client headers are scanned, the ones created with <quote>NO-RESPONSE-TAG:</quote>
2393 are checked after all server headers are scanned. In both cases all the created
2394 tags are considered.
2397 <!-- ~~~~~ New section ~~~~~ -->
2398 <sect3 id="client-tag-pattern"><title>The Client Tag Pattern</title>
2400 <!-- XXX: This section contains duplicates content from the
2401 client-specific-tag documentation. -->
2405 This is an experimental feature. The syntax is likely to change in future versions.
2410 Client tag patterns are not set based on HTTP headers but based on
2411 the client's IP address. Users can enable them themselves, but the
2412 Privoxy admin controls which tags are available and what their effect
2417 After a client-specific tag has been defined with the
2418 <link linkend="client-specific-tag">client-specific-tag</link>,
2419 directive, action sections can be activated based on the tag by using a
2420 CLIENT-TAG pattern. The CLIENT-TAG pattern is evaluated at the same priority
2421 as URL patterns, as a result the last matching pattern wins. Tags that
2422 are created based on client or server headers are evaluated later on
2423 and can overrule CLIENT-TAG and URL patterns!
2426 The tag is set for all requests that come from clients that requested
2427 it to be set. Note that "clients" are differentiated by IP address,
2428 if the IP address changes the tag has to be requested again.
2431 Clients can request tags to be set by using the CGI interface <ulink
2432 url="http://config.privoxy.org/show-client-tags">http://config.privoxy.org/show-client-tags</ulink>.
2441 # If the admin defined the client-specific-tag circumvent-blocks,
2442 # and the request comes from a client that previously requested
2443 # the tag to be set, overrule all previous +block actions that
2444 # are enabled based on URL to CLIENT-TAG patterns.
2446 CLIENT-TAG:^circumvent-blocks$
2448 # This section is not overruled because it's located after
2450 {+block{Nobody is supposed to request this.}}
2451 example.org/blocked-example-page</screen>
2456 <!-- ~ End section ~ -->
2459 <!-- ~~~~~ New section ~~~~~ -->
2461 <sect2 id="actions">
2462 <title>Actions</title>
2464 All actions are disabled by default, until they are explicitly enabled
2465 somewhere in an actions file. Actions are turned on if preceded with a
2466 <quote>+</quote>, and turned off if preceded with a <quote>-</quote>. So a
2467 <literal>+action</literal> means <quote>do that action</quote>, e.g.
2468 <literal>+block</literal> means <quote>please block URLs that match the
2469 following patterns</quote>, and <literal>-block</literal> means <quote>don't
2470 block URLs that match the following patterns, even if <literal>+block</literal>
2471 previously applied.</quote>
2476 Again, actions are invoked by placing them on a line, enclosed in curly braces and
2477 separated by whitespace, like in
2478 <literal>{+some-action -some-other-action{some-parameter}}</literal>,
2479 followed by a list of URL patterns, one per line, to which they apply.
2480 Together, the actions line and the following pattern lines make up a section
2481 of the actions file.
2485 Actions fall into three categories:
2492 Boolean, i.e the action can only be <quote>enabled</quote> or
2493 <quote>disabled</quote>. Syntax:
2497 +<replaceable class="function">name</replaceable> # enable action <replaceable class="parameter">name</replaceable>
2498 -<replaceable class="function">name</replaceable> # disable action <replaceable class="parameter">name</replaceable></screen>
2501 Example: <literal>+handle-as-image</literal>
2508 Parameterized, where some value is required in order to enable this type of action.
2513 +<replaceable class="function">name</replaceable>{<replaceable class="parameter">param</replaceable>} # enable action and set parameter to <replaceable class="parameter">param</replaceable>,
2514 # overwriting parameter from previous match if necessary
2515 -<replaceable class="function">name</replaceable> # disable action. The parameter can be omitted</screen>
2518 Note that if the URL matches multiple positive forms of a parameterized action,
2519 the last match wins, i.e. the params from earlier matches are simply ignored.
2522 Example: <literal>+hide-user-agent{Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.8.1.4) Gecko/20070602 Firefox/2.0.0.4}</literal>
2528 Multi-value. These look exactly like parameterized actions,
2529 but they behave differently: If the action applies multiple times to the
2530 same URL, but with different parameters, <emphasis>all</emphasis> the parameters
2531 from <emphasis>all</emphasis> matches are remembered. This is used for actions
2532 that can be executed for the same request repeatedly, like adding multiple
2533 headers, or filtering through multiple filters. Syntax:
2537 +<replaceable class="function">name</replaceable>{<replaceable class="parameter">param</replaceable>} # enable action and add <replaceable class="parameter">param</replaceable> to the list of parameters
2538 -<replaceable class="function">name</replaceable>{<replaceable class="parameter">param</replaceable>} # remove the parameter <replaceable class="parameter">param</replaceable> from the list of parameters
2539 # If it was the last one left, disable the action.
2540 <replaceable class="parameter">-name</replaceable> # disable this action completely and remove all parameters from the list</screen>
2543 Examples: <literal>+add-header{X-Fun-Header: Some text}</literal> and
2544 <literal>+filter{html-annoyances}</literal>
2552 If nothing is specified in any actions file, no <quote>actions</quote> are
2553 taken. So in this case <application>Privoxy</application> would just be a
2554 normal, non-blocking, non-filtering proxy. You must specifically enable the
2555 privacy and blocking features you need (although the provided default actions
2556 files will give a good starting point).
2560 Later defined action sections always over-ride earlier ones of the same type.
2561 So exceptions to any rules you make, should come in the latter part of the file (or
2562 in a file that is processed later when using multiple actions files such
2563 as <filename>user.action</filename>). For multi-valued actions, the actions
2564 are applied in the order they are specified. Actions files are processed in
2565 the order they are defined in <filename>config</filename> (the default
2566 installation has three actions files). It also quite possible for any given
2567 URL to match more than one <quote>pattern</quote> (because of wildcards and
2568 regular expressions), and thus to trigger more than one set of actions! Last
2572 <!-- start actions listing -->
2574 The list of valid <application>Privoxy</application> actions are:
2578 <!-- ********************************************************** -->
2579 <!-- Please note the below defined actions use id's that are -->
2580 <!-- probably linked from other places, so please don't change. -->
2582 <!-- ********************************************************** -->
2585 <!-- ~~~~~ New section ~~~~~ -->
2587 <sect3 renderas="sect4" id="add-header">
2588 <title>add-header</title>
2592 <term>Typical use:</term>
2594 <para>Confuse log analysis, custom applications</para>
2599 <term>Effect:</term>
2602 Sends a user defined HTTP header to the web server.
2609 <!-- boolean, parameterized, Multi-value -->
2611 <para>Multi-value.</para>
2616 <term>Parameter:</term>
2619 Any string value is possible. Validity of the defined HTTP headers is not checked.
2620 It is recommended that you use the <quote><literal>X-</literal></quote> prefix
2630 This action may be specified multiple times, in order to define multiple
2631 headers. This is rarely needed for the typical user. If you don't know what
2632 <quote>HTTP headers</quote> are, you definitely don't need to worry about this
2636 Headers added by this action are not modified by other actions.
2642 <term>Example usage:</term>
2645 <screen># Add a DNT ("Do not track") header to all requests,
2646 # event to those that already have one.
2648 # This is just an example, not a recommendation.
2650 # There is no reason to believe that user-tracking websites care
2651 # about the DNT header and depending on the User-Agent, adding the
2652 # header may make user-tracking easier.
2653 {+add-header{DNT: 1}}
2662 <!-- ~~~~~ New section ~~~~~ -->
2663 <sect3 renderas="sect4" id="block">
2664 <title>block</title>
2668 <term>Typical use:</term>
2670 <para>Block ads or other unwanted content</para>
2675 <term>Effect:</term>
2678 Requests for URLs to which this action applies are blocked, i.e. the
2679 requests are trapped by &my-app; and the requested URL is never retrieved,
2680 but is answered locally with a substitute page or image, as determined by
2682 linkend="handle-as-image">handle-as-image</link></literal>,
2684 linkend="set-image-blocker">set-image-blocker</link></literal>, and
2686 linkend="handle-as-empty-document">handle-as-empty-document</link></literal> actions.
2694 <!-- boolean, parameterized, Multi-value -->
2696 <para>Parameterized.</para>
2701 <term>Parameter:</term>
2703 <para>A block reason that should be given to the user.</para>
2711 <application>Privoxy</application> sends a special <quote>BLOCKED</quote> page
2712 for requests to blocked pages. This page contains the block reason given as
2713 parameter, a link to find out why the block action applies, and a click-through
2714 to the blocked content (the latter only if the force feature is available and
2718 A very important exception occurs if <emphasis>both</emphasis>
2719 <literal>block</literal> and <literal><link linkend="handle-as-image">handle-as-image</link></literal>,
2720 apply to the same request: it will then be replaced by an image. If
2721 <literal><link linkend="set-image-blocker">set-image-blocker</link></literal>
2722 (see below) also applies, the type of image will be determined by its parameter,
2723 if not, the standard checkerboard pattern is sent.
2726 It is important to understand this process, in order
2727 to understand how <application>Privoxy</application> deals with
2728 ads and other unwanted content. Blocking is a core feature, and one
2729 upon which various other features depend.
2732 The <literal><link linkend="filter">filter</link></literal>
2733 action can perform a very similar task, by <quote>blocking</quote>
2734 banner images and other content through rewriting the relevant URLs in the
2735 document's HTML source, so they don't get requested in the first place.
2736 Note that this is a totally different technique, and it's easy to confuse the two.
2742 <term>Example usage (section):</term>
2745 <screen>{+block{No nasty stuff for you.}}
2746 # Block and replace with "blocked" page
2747 .nasty-stuff.example.com
2749 {+block{Doubleclick banners.} +handle-as-image}
2750 # Block and replace with image
2754 {+block{Layered ads.} +handle-as-empty-document}
2755 # Block and then ignore
2756 adserver.example.net/.*\.js$</screen>
2766 <!-- ~~~~~ New section ~~~~~ -->
2767 <sect3 renderas="sect4" id="change-x-forwarded-for">
2768 <title>change-x-forwarded-for</title>
2772 <term>Typical use:</term>
2774 <para>Improve privacy by not forwarding the source of the request in the HTTP headers.</para>
2779 <term>Effect:</term>
2782 Deletes the <quote>X-Forwarded-For:</quote> HTTP header from the client request,
2790 <!-- Boolean, Parameterized, Multi-value -->
2792 <para>Parameterized.</para>
2797 <term>Parameter:</term>
2801 <para><quote>block</quote> to delete the header.</para>
2805 <quote>add</quote> to create the header (or append
2806 the client's IP address to an already existing one).
2817 It is safe and recommended to use <literal>block</literal>.
2820 Forwarding the source address of the request may make
2821 sense in some multi-user setups but is also a privacy risk.
2826 <term>Example usage:</term>
2829 <screen>+change-x-forwarded-for{block}</screen>
2836 <!-- ~~~~~ New section ~~~~~ -->
2837 <sect3 renderas="sect4" id="client-header-filter">
2838 <title>client-header-filter</title>
2842 <term>Typical use:</term>
2845 Rewrite or remove single client headers.
2851 <term>Effect:</term>
2854 All client headers to which this action applies are filtered on-the-fly through
2855 the specified regular expression based substitutions.
2862 <!-- boolean, parameterized, Multi-value -->
2864 <para>Multi-value.</para>
2869 <term>Parameter:</term>
2872 The name of a client-header filter, as defined in one of the
2873 <link linkend="filter-file">filter files</link>.
2882 Client-header filters are applied to each header on its own, not to
2883 all at once. This makes it easier to diagnose problems, but on the downside
2884 you can't write filters that only change header x if header y's value is z.
2885 You can do that by using tags though.
2888 Client-header filters are executed after the other header actions have finished
2889 and use their output as input.
2892 If the request URI gets changed, &my-app; will detect that and use the new
2893 one. This can be used to rewrite the request destination behind the client's
2894 back, for example to specify a Tor exit relay for certain requests.
2897 Please refer to the <link linkend="filter-file">filter file chapter</link>
2898 to learn which client-header filters are available by default, and how to
2906 <term>Example usage (section):</term>
2910 # Hide Tor exit notation in Host and Referer Headers
2911 {+client-header-filter{hide-tor-exit-notation}}
2922 <!-- ~~~~~ New section ~~~~~ -->
2923 <sect3 renderas="sect4" id="client-header-tagger">
2924 <title>client-header-tagger</title>
2928 <term>Typical use:</term>
2931 Block requests based on their headers.
2937 <term>Effect:</term>
2940 Client headers to which this action applies are filtered on-the-fly through
2941 the specified regular expression based substitutions, the result is used as
2949 <!-- boolean, parameterized, Multi-value -->
2951 <para>Multi-value.</para>
2956 <term>Parameter:</term>
2959 The name of a client-header tagger, as defined in one of the
2960 <link linkend="filter-file">filter files</link>.
2969 Client-header taggers are applied to each header on its own,
2970 and as the header isn't modified, each tagger <quote>sees</quote>
2974 Client-header taggers are the first actions that are executed
2975 and their tags can be used to control every other action.
2981 <term>Example usage (section):</term>
2985 # Tag every request with the User-Agent header
2986 {+client-header-tagger{user-agent}}
2989 # Tagging itself doesn't change the action
2990 # settings, sections with TAG patterns do:
2992 # If it's a download agent, use a different forwarding proxy,
2993 # show the real User-Agent and make sure resume works.
2994 {+forward-override{forward-socks5 10.0.0.2:2222 .} \
2995 -hide-if-modified-since \
2996 -overwrite-last-modified \
3001 TAG:^User-Agent: NetBSD-ftp/
3002 TAG:^User-Agent: Novell ZYPP Installer
3003 TAG:^User-Agent: RPM APT-HTTP/
3004 TAG:^User-Agent: fetch libfetch/
3005 TAG:^User-Agent: Ubuntu APT-HTTP/
3006 TAG:^User-Agent: MPlayer/
3011 # Tag all requests with the Range header set
3012 {+client-header-tagger{range-requests}}
3015 # Disable filtering for the tagged requests.
3017 # With filtering enabled Privoxy would remove the Range headers
3018 # to be able to filter the whole response. The downside is that
3019 # it prevents clients from resuming downloads or skipping over
3020 # parts of multimedia files.
3021 {-filter -deanimate-gifs}
3032 <!-- ~~~~~ New section ~~~~~ -->
3033 <sect3 renderas="sect4" id="content-type-overwrite">
3034 <title>content-type-overwrite</title>
3038 <term>Typical use:</term>
3040 <para>Stop useless download menus from popping up, or change the browser's rendering mode</para>
3045 <term>Effect:</term>
3048 Replaces the <quote>Content-Type:</quote> HTTP server header.
3055 <!-- Boolean, Parameterized, Multi-value -->
3057 <para>Parameterized.</para>
3062 <term>Parameter:</term>
3074 The <quote>Content-Type:</quote> HTTP server header is used by the
3075 browser to decide what to do with the document. The value of this
3076 header can cause the browser to open a download menu instead of
3077 displaying the document by itself, even if the document's format is
3078 supported by the browser.
3081 The declared content type can also affect which rendering mode
3082 the browser chooses. If XHTML is delivered as <quote>text/html</quote>,
3083 many browsers treat it as yet another broken HTML document.
3084 If it is send as <quote>application/xml</quote>, browsers with
3085 XHTML support will only display it, if the syntax is correct.
3088 If you see a web site that proudly uses XHTML buttons, but sets
3089 <quote>Content-Type: text/html</quote>, you can use &my-app;
3090 to overwrite it with <quote>application/xml</quote> and validate
3091 the web master's claim inside your XHTML-supporting browser.
3092 If the syntax is incorrect, the browser will complain loudly.
3095 You can also go the opposite direction: if your browser prints
3096 error messages instead of rendering a document falsely declared
3097 as XHTML, you can overwrite the content type with
3098 <quote>text/html</quote> and have it rendered as broken HTML document.
3101 By default <literal>content-type-overwrite</literal> only replaces
3102 <quote>Content-Type:</quote> headers that look like some kind of text.
3103 If you want to overwrite it unconditionally, you have to combine it with
3104 <literal><link linkend="force-text-mode">force-text-mode</link></literal>.
3105 This limitation exists for a reason, think twice before circumventing it.
3108 Most of the time it's easier to replace this action with a custom
3109 <literal><link linkend="server-header-filter">server-header filter</link></literal>.
3110 It allows you to activate it for every document of a certain site and it will still
3111 only replace the content types you aimed at.
3114 Of course you can apply <literal>content-type-overwrite</literal>
3115 to a whole site and then make URL based exceptions, but it's a lot
3116 more work to get the same precision.
3122 <term>Example usage (sections):</term>
3125 <screen># Check if www.example.net/ really uses valid XHTML
3126 { +content-type-overwrite{application/xml} }
3129 # but leave the content type unmodified if the URL looks like a style sheet
3130 {-content-type-overwrite}
3131 www.example.net/.*\.css$
3132 www.example.net/.*style
3141 <!-- ~~~~~ New section ~~~~~ -->
3142 <sect3 renderas="sect4" id="crunch-client-header">
3146 <title>crunch-client-header</title>
3150 <term>Typical use:</term>
3152 <para>Remove a client header <application>Privoxy</application> has no dedicated action for.</para>
3157 <term>Effect:</term>
3160 Deletes every header sent by the client that contains the string the user supplied as parameter.
3167 <!-- Boolean, Parameterized, Multi-value -->
3169 <para>Parameterized.</para>
3174 <term>Parameter:</term>
3186 This action allows you to block client headers for which no dedicated
3187 <application>Privoxy</application> action exists.
3188 <application>Privoxy</application> will remove every client header that
3189 contains the string you supplied as parameter.
3192 Regular expressions are <emphasis>not supported</emphasis> and you can't
3193 use this action to block different headers in the same request, unless
3194 they contain the same string.
3197 <literal>crunch-client-header</literal> is only meant for quick tests.
3198 If you have to block several different headers, or only want to modify
3199 parts of them, you should use a
3200 <literal><link linkend="client-header-filter">client-header filter</link></literal>.
3204 Don't block any header without understanding the consequences.
3211 <term>Example usage (section):</term>
3214 <screen># Block the non-existent "Privacy-Violation:" client header
3215 { +crunch-client-header{Privacy-Violation:} }
3225 <!-- ~~~~~ New section ~~~~~ -->
3226 <sect3 renderas="sect4" id="crunch-if-none-match">
3227 <title>crunch-if-none-match</title>
3233 <term>Typical use:</term>
3235 <para>Prevent yet another way to track the user's steps between sessions.</para>
3240 <term>Effect:</term>
3243 Deletes the <quote>If-None-Match:</quote> HTTP client header.
3250 <!-- Boolean, Parameterized, Multi-value -->
3252 <para>Boolean.</para>
3257 <term>Parameter:</term>
3269 Removing the <quote>If-None-Match:</quote> HTTP client header
3270 is useful for filter testing, where you want to force a real
3271 reload instead of getting status code <quote>304</quote> which
3272 would cause the browser to use a cached copy of the page.
3275 It is also useful to make sure the header isn't used as a cookie
3276 replacement (unlikely but possible).
3279 Blocking the <quote>If-None-Match:</quote> header shouldn't cause any
3280 caching problems, as long as the <quote>If-Modified-Since:</quote> header
3281 isn't blocked or missing as well.
3284 It is recommended to use this action together with
3285 <literal><link linkend="hide-if-modified-since">hide-if-modified-since</link></literal>
3287 <literal><link linkend="overwrite-last-modified">overwrite-last-modified</link></literal>.
3293 <term>Example usage (section):</term>
3296 <screen># Let the browser revalidate cached documents but don't
3297 # allow the server to use the revalidation headers for user tracking.
3298 {+hide-if-modified-since{-60} \
3299 +overwrite-last-modified{randomize} \
3300 +crunch-if-none-match}
3309 <!-- ~~~~~ New section ~~~~~ -->
3310 <sect3 renderas="sect4" id="crunch-incoming-cookies">
3311 <title>crunch-incoming-cookies</title>
3315 <term>Typical use:</term>
3318 Prevent the web server from setting HTTP cookies on your system
3324 <term>Effect:</term>
3327 Deletes any <quote>Set-Cookie:</quote> HTTP headers from server replies.
3334 <!-- Boolean, Parameterized, Multi-value -->
3336 <para>Boolean.</para>
3341 <term>Parameter:</term>
3353 This action is only concerned with <emphasis>incoming</emphasis> HTTP cookies. For
3354 <emphasis>outgoing</emphasis> HTTP cookies, use
3355 <literal><link linkend="crunch-outgoing-cookies">crunch-outgoing-cookies</link></literal>.
3356 Use <emphasis>both</emphasis> to disable HTTP cookies completely.
3359 It makes <emphasis>no sense at all</emphasis> to use this action in conjunction
3360 with the <literal><link linkend="session-cookies-only">session-cookies-only</link></literal> action,
3361 since it would prevent the session cookies from being set. See also
3362 <literal><link linkend="filter-content-cookies">filter-content-cookies</link></literal>.
3368 <term>Example usage:</term>
3371 <screen>+crunch-incoming-cookies</screen>
3379 <!-- ~~~~~ New section ~~~~~ -->
3380 <sect3 renderas="sect4" id="crunch-server-header">
3381 <title>crunch-server-header</title>
3387 <term>Typical use:</term>
3389 <para>Remove a server header <application>Privoxy</application> has no dedicated action for.</para>
3394 <term>Effect:</term>
3397 Deletes every header sent by the server that contains the string the user supplied as parameter.
3404 <!-- Boolean, Parameterized, Multi-value -->
3406 <para>Parameterized.</para>
3411 <term>Parameter:</term>
3423 This action allows you to block server headers for which no dedicated
3424 <application>Privoxy</application> action exists. <application>Privoxy</application>
3425 will remove every server header that contains the string you supplied as parameter.
3428 Regular expressions are <emphasis>not supported</emphasis> and you can't
3429 use this action to block different headers in the same request, unless
3430 they contain the same string.
3433 <literal>crunch-server-header</literal> is only meant for quick tests.
3434 If you have to block several different headers, or only want to modify
3435 parts of them, you should use a custom
3436 <literal><link linkend="server-header-filter">server-header filter</link></literal>.
3440 Don't block any header without understanding the consequences.
3447 <term>Example usage (section):</term>
3450 <screen># Crunch server headers that try to prevent caching
3451 { +crunch-server-header{no-cache} }
3460 <!-- ~~~~~ New section ~~~~~ -->
3461 <sect3 renderas="sect4" id="crunch-outgoing-cookies">
3462 <title>crunch-outgoing-cookies</title>
3466 <term>Typical use:</term>
3469 Prevent the web server from reading any HTTP cookies from your system
3475 <term>Effect:</term>
3478 Deletes any <quote>Cookie:</quote> HTTP headers from client requests.
3485 <!-- Boolean, Parameterized, Multi-value -->
3487 <para>Boolean.</para>
3492 <term>Parameter:</term>
3504 This action is only concerned with <emphasis>outgoing</emphasis> HTTP cookies. For
3505 <emphasis>incoming</emphasis> HTTP cookies, use
3506 <literal><link linkend="crunch-incoming-cookies">crunch-incoming-cookies</link></literal>.
3507 Use <emphasis>both</emphasis> to disable HTTP cookies completely.
3510 It makes <emphasis>no sense at all</emphasis> to use this action in conjunction
3511 with the <literal><link linkend="session-cookies-only">session-cookies-only</link></literal> action,
3512 since it would prevent the session cookies from being read.
3518 <term>Example usage:</term>
3521 <screen>+crunch-outgoing-cookies</screen>
3530 <!-- ~~~~~ New section ~~~~~ -->
3531 <sect3 renderas="sect4" id="deanimate-gifs">
3532 <title>deanimate-gifs</title>
3536 <term>Typical use:</term>
3538 <para>Stop those annoying, distracting animated GIF images.</para>
3543 <term>Effect:</term>
3546 De-animate GIF animations, i.e. reduce them to their first or last image.
3553 <!-- boolean, parameterized, Multi-value -->
3555 <para>Parameterized.</para>
3560 <term>Parameter:</term>
3563 <quote>last</quote> or <quote>first</quote>
3572 This will also shrink the images considerably (in bytes, not pixels!). If
3573 the option <quote>first</quote> is given, the first frame of the animation
3574 is used as the replacement. If <quote>last</quote> is given, the last
3575 frame of the animation is used instead, which probably makes more sense for
3576 most banner animations, but also has the risk of not showing the entire
3577 last frame (if it is only a delta to an earlier frame).
3580 You can safely use this action with patterns that will also match non-GIF
3581 objects, because no attempt will be made at anything that doesn't look like
3588 <term>Example usage:</term>
3591 <screen>+deanimate-gifs{last}</screen>
3598 <!-- ~~~~~ New section ~~~~~ -->
3599 <sect3 renderas="sect4" id="downgrade-http-version">
3600 <title>downgrade-http-version</title>
3604 <term>Typical use:</term>
3606 <para>Work around (very rare) problems with HTTP/1.1</para>
3611 <term>Effect:</term>
3614 Downgrades HTTP/1.1 client requests and server replies to HTTP/1.0.
3621 <!-- boolean, parameterized, Multi-value -->
3623 <para>Boolean.</para>
3628 <term>Parameter:</term>
3640 This is a left-over from the time when <application>Privoxy</application>
3641 didn't support important HTTP/1.1 features well. It is left here for the
3642 unlikely case that you experience HTTP/1.1-related problems with some server
3646 Note that enabling this action is only a workaround. It should not
3647 be enabled for sites that work without it. While it shouldn't break
3648 any pages, it has an (usually negative) performance impact.
3651 If you come across a site where enabling this action helps, please report it,
3652 so the cause of the problem can be analyzed. If the problem turns out to be
3653 caused by a bug in <application>Privoxy</application> it should be
3654 fixed so the following release works without the work around.
3660 <term>Example usage (section):</term>
3663 <screen>{+downgrade-http-version}
3664 problem-host.example.com</screen>
3672 <!-- ~~~~~ New section ~~~~~ -->
3673 <sect3 renderas="sect4" id="external-filter">
3674 <title>external-filter</title>
3678 <term>Typical use:</term>
3680 <para>Modify content using a programming language of your choice.</para>
3685 <term>Effect:</term>
3688 All instances of text-based type, most notably HTML and JavaScript, to which
3689 this action applies, can be filtered on-the-fly through the specified external
3691 By default plain text documents are exempted from filtering, because web
3692 servers often use the <literal>text/plain</literal> MIME type for all files
3693 whose type they don't know.)
3700 <!-- boolean, parameterized, Multi-value -->
3702 <para>Multi-value.</para>
3707 <term>Parameter:</term>
3710 The name of an external content filter, as defined in the
3711 <link linkend="filter-file">filter file</link>.
3712 External filters can be defined in one or more files as defined by the
3713 <literal><link linkend="filterfile">filterfile</link></literal>
3714 option in the <link linkend="config">config file</link>.
3717 When used in its negative form,
3718 and without parameters, <emphasis>all</emphasis> filtering with external
3719 filters is completely disabled.
3728 External filters are scripts or programs that can modify the content in
3729 case common <literal><link linkend="filter">filters</link></literal>
3730 aren't powerful enough. With the exception that this action doesn't
3731 use pcrs-based filters, the notes in the
3732 <literal><link linkend="filter">filter</link></literal> section apply.
3736 Currently external filters are executed with &my-app;'s privileges.
3737 Only use external filters you understand and trust.
3741 This feature is experimental, the <literal><link
3742 linkend="external-filter-syntax">syntax</link></literal>
3743 may change in the future.
3750 <term>Example usage:</term>
3753 <screen>+external-filter{fancy-filter}</screen>
3760 <!-- ~~~~~ New section ~~~~~ -->
3761 <sect3 renderas="sect4" id="fast-redirects">
3762 <title>fast-redirects</title>
3766 <term>Typical use:</term>
3768 <para>Fool some click-tracking scripts and speed up indirect links.</para>
3773 <term>Effect:</term>
3776 Detects redirection URLs and redirects the browser without contacting
3777 the redirection server first.
3784 <!-- boolean, parameterized, Multi-value -->
3786 <para>Parameterized.</para>
3791 <term>Parameter:</term>
3796 <quote>simple-check</quote> to just search for the string <quote>http://</quote>
3797 to detect redirection URLs.
3802 <quote>check-decoded-url</quote> to decode URLs (if necessary) before searching
3803 for redirection URLs.
3814 Many sites, like yahoo.com, don't just link to other sites. Instead, they
3815 will link to some script on their own servers, giving the destination as a
3816 parameter, which will then redirect you to the final target. URLs
3817 resulting from this scheme typically look like:
3818 <quote>http://www.example.org/click-tracker.cgi?target=http%3a//www.example.net/</quote>.
3821 Sometimes, there are even multiple consecutive redirects encoded in the
3822 URL. These redirections via scripts make your web browsing more traceable,
3823 since the server from which you follow such a link can see where you go
3824 to. Apart from that, valuable bandwidth and time is wasted, while your
3825 browser asks the server for one redirect after the other. Plus, it feeds
3829 This feature is currently not very smart and is scheduled for improvement.
3830 If it is enabled by default, you will have to create some exceptions to
3831 this action. It can lead to failures in several ways:
3834 Not every URLs with other URLs as parameters is evil.
3835 Some sites offer a real service that requires this information to work.
3836 For example a validation service needs to know, which document to validate.
3837 <literal>fast-redirects</literal> assumes that every URL parameter that
3838 looks like another URL is a redirection target, and will always redirect to
3839 the last one. Most of the time the assumption is correct, but if it isn't,
3840 the user gets redirected anyway.
3843 Another failure occurs if the URL contains other parameters after the URL parameter.
3845 <quote>http://www.example.org/?redirect=http%3a//www.example.net/&foo=bar</quote>.
3846 contains the redirection URL <quote>http://www.example.net/</quote>,
3847 followed by another parameter. <literal>fast-redirects</literal> doesn't know that
3848 and will cause a redirect to <quote>http://www.example.net/&foo=bar</quote>.
3849 Depending on the target server configuration, the parameter will be silently ignored
3850 or lead to a <quote>page not found</quote> error. You can prevent this problem by
3851 first using the <literal><link linkend="redirect">redirect</link></literal> action
3852 to remove the last part of the URL, but it requires a little effort.
3855 To detect a redirection URL, <literal>fast-redirects</literal> only
3856 looks for the string <quote>http://</quote>, either in plain text
3857 (invalid but often used) or encoded as <quote>http%3a//</quote>.
3858 Some sites use their own URL encoding scheme, encrypt the address
3859 of the target server or replace it with a database id. In theses cases
3860 <literal>fast-redirects</literal> is fooled and the request reaches the
3861 redirection server where it probably gets logged.
3867 <term>Example usage:</term>
3871 { +fast-redirects{simple-check} }
3874 { +fast-redirects{check-decoded-url} }
3875 another.example.com/testing</screen>
3884 <!-- ~~~~~ New section ~~~~~ -->
3885 <sect3 renderas="sect4" id="filter">
3886 <title>filter</title>
3890 <term>Typical use:</term>
3892 <para>Get rid of HTML and JavaScript annoyances, banner advertisements (by size),
3893 do fun text replacements, add personalized effects, etc.</para>
3898 <term>Effect:</term>
3901 All instances of text-based type, most notably HTML and JavaScript, to which
3902 this action applies, can be filtered on-the-fly through the specified regular
3903 expression based substitutions. (Note: as of version 3.0.3 plain text documents
3904 are exempted from filtering, because web servers often use the
3905 <literal>text/plain</literal> MIME type for all files whose type they don't know.)
3912 <!-- boolean, parameterized, Multi-value -->
3914 <para>Multi-value.</para>
3919 <term>Parameter:</term>
3922 The name of a content filter, as defined in the <link linkend="filter-file">filter file</link>.
3923 Filters can be defined in one or more files as defined by the
3924 <literal><link linkend="filterfile">filterfile</link></literal>
3925 option in the <link linkend="config">config file</link>.
3926 <filename>default.filter</filename> is the collection of filters
3927 supplied by the developers. Locally defined filters should go
3928 in their own file, such as <filename>user.filter</filename>.
3931 When used in its negative form,
3932 and without parameters, <emphasis>all</emphasis> filtering is completely disabled.
3941 For your convenience, there are a number of pre-defined filters available
3942 in the distribution filter file that you can use. See the examples below for
3946 Filtering requires buffering the page content, which may appear to
3947 slow down page rendering since nothing is displayed until all content has
3948 passed the filters. (The total time until the page is completely rendered
3949 doesn't change much, but it may be perceived as slower since the page is
3950 not incrementally displayed.)
3951 This effect will be more noticeable on slower connections.
3954 <quote>Rolling your own</quote>
3955 filters requires a knowledge of
3956 <ulink url="http://en.wikipedia.org/wiki/Regular_expressions"><quote>Regular
3957 Expressions</quote></ulink> and
3958 <ulink url="http://en.wikipedia.org/wiki/Html"><quote>HTML</quote></ulink>.
3959 This is very powerful feature, and potentially very intrusive.
3960 Filters should be used with caution, and where an equivalent
3961 <quote>action</quote> is not available.
3964 The amount of data that can be filtered is limited to the
3965 <literal><link linkend="buffer-limit">buffer-limit</link></literal>
3966 option in the main <link linkend="config">config file</link>. The
3967 default is 4096 KB (4 Megs). Once this limit is exceeded, the buffered
3968 data, and all pending data, is passed through unfiltered.
3971 Inappropriate MIME types, such as zipped files, are not filtered at all.
3972 (Again, only text-based types except plain text). Encrypted SSL data
3973 (from HTTPS servers) cannot be filtered either, since this would violate
3974 the integrity of the secure transaction. In some situations it might
3975 be necessary to protect certain text, like source code, from filtering
3976 by defining appropriate <literal>-filter</literal> exceptions.
3979 Compressed content can't be filtered either, but if &my-app;
3980 is compiled with zlib support and a supported compression algorithm
3981 is used (gzip or deflate), &my-app; can first decompress the content
3985 If you use a &my-app; version without zlib support, but want filtering to work on
3986 as much documents as possible, even those that would normally be sent compressed,
3987 you must use the <literal><link linkend="prevent-compression">prevent-compression</link></literal>
3988 action in conjunction with <literal>filter</literal>.
3991 Content filtering can achieve some of the same effects as the
3992 <literal><link linkend="block">block</link></literal>
3993 action, i.e. it can be used to block ads and banners. But the mechanism
3994 works quite differently. One effective use, is to block ad banners
3995 based on their size (see below), since many of these seem to be somewhat
3999 <link linkend="contact">Feedback</link> with suggestions for new or
4000 improved filters is particularly welcome!
4003 The below list has only the names and a one-line description of each
4004 predefined filter. There are <link linkend="predefined-filters">more
4005 verbose explanations</link> of what these filters do in the <link
4006 linkend="filter-file">filter file chapter</link>.
4012 <term>Example usage (with filters from the distribution <filename>default.filter</filename> file).
4013 See <link linkend="PREDEFINED-FILTERS">the Predefined Filters section</link> for
4014 more explanation on each:</term>
4017 <anchor id="filter-js-annoyances">
4018 <screen>+filter{js-annoyances} # Get rid of particularly annoying JavaScript abuse.</screen>
4021 <anchor id="filter-js-events">
4022 <screen>+filter{js-events} # Kill JavaScript event bindings and timers (Radically destructive! Only for extra nasty sites).</screen>
4025 <anchor id="filter-html-annoyances">
4026 <screen>+filter{html-annoyances} # Get rid of particularly annoying HTML abuse.</screen>
4029 <anchor id="filter-content-cookies">
4030 <screen>+filter{content-cookies} # Kill cookies that come in the HTML or JS content.</screen>
4033 <anchor id="filter-refresh-tags">
4034 <screen>+filter{refresh-tags} # Kill automatic refresh tags if refresh time is larger than 9 seconds.</screen>
4037 <anchor id="filter-unsolicited-popups">
4038 <screen>+filter{unsolicited-popups} # Disable only unsolicited pop-up windows.</screen>
4041 <anchor id="filter-all-popups">
4042 <screen>+filter{all-popups} # Kill all popups in JavaScript and HTML.</screen>
4045 <anchor id="filter-img-reorder">
4046 <screen>+filter{img-reorder} # Reorder attributes in <img> tags to make the banners-by-* filters more effective.</screen>
4049 <anchor id="filter-banners-by-size">
4050 <screen>+filter{banners-by-size} # Kill banners by size.</screen>
4053 <anchor id="filter-banners-by-link">
4054 <screen>+filter{banners-by-link} # Kill banners by their links to known clicktrackers.</screen>
4057 <anchor id="filter-webbugs">
4058 <screen>+filter{webbugs} # Squish WebBugs (1x1 invisible GIFs used for user tracking).</screen>
4061 <anchor id="filter-tiny-textforms">
4062 <screen>+filter{tiny-textforms} # Extend those tiny textareas up to 40x80 and kill the hard wrap.</screen>
4065 <anchor id="filter-jumping-windows">
4066 <screen>+filter{jumping-windows} # Prevent windows from resizing and moving themselves.</screen>
4069 <anchor id="filter-frameset-borders">
4070 <screen>+filter{frameset-borders} # Give frames a border and make them resizable.</screen>
4073 <anchor id="filter-iframes">
4074 <screen>+filter{iframes} # Removes all detected iframes. Should only be enabled for individual sites.</screen>
4077 <anchor id="filter-demoronizer">
4078 <screen>+filter{demoronizer} # Fix MS's non-standard use of standard charsets.</screen>
4081 <anchor id="filter-shockwave-flash">
4082 <screen>+filter{shockwave-flash} # Kill embedded Shockwave Flash objects.</screen>
4085 <anchor id="filter-quicktime-kioskmode">
4086 <screen>+filter{quicktime-kioskmode} # Make Quicktime movies saveable.</screen>
4089 <anchor id="filter-fun">
4090 <screen>+filter{fun} # Text replacements for subversive browsing fun!</screen>
4093 <anchor id="filter-crude-parental">
4094 <screen>+filter{crude-parental} # Crude parental filtering. Note that this filter doesn't work reliably.</screen>
4097 <anchor id="filter-ie-exploits">
4098 <screen>+filter{ie-exploits} # Disable some known Internet Explorer bug exploits.</screen>
4101 <anchor id="filter-site-specifics">
4102 <screen>+filter{site-specifics} # Cure for site-specific problems. Don't apply generally!</screen>
4105 <anchor id="filter-no-ping">
4106 <screen>+filter{no-ping} # Removes non-standard ping attributes in <a> and <area> tags.</screen>
4109 <anchor id="filter-google">
4110 <screen>+filter{google} # CSS-based block for Google text ads. Also removes a width limitation and the toolbar advertisement.</screen>
4113 <anchor id="filter-yahoo">
4114 <screen>+filter{yahoo} # CSS-based block for Yahoo text ads. Also removes a width limitation.</screen>
4117 <anchor id="filter-msn">
4118 <screen>+filter{msn} # CSS-based block for MSN text ads. Also removes tracking URLs and a width limitation.</screen>
4121 <anchor id="filter-blogspot">
4122 <screen>+filter{blogspot} # Cleans up some Blogspot blogs. Read the fine print before using this.</screen>
4130 <!-- ~~~~~ New section ~~~~~ -->
4131 <sect3 renderas="sect4" id="force-text-mode">
4132 <title>force-text-mode</title>
4138 <term>Typical use:</term>
4140 <para>Force <application>Privoxy</application> to treat a document as if it was in some kind of <emphasis>text</emphasis> format. </para>
4145 <term>Effect:</term>
4148 Declares a document as text, even if the <quote>Content-Type:</quote> isn't detected as such.
4155 <!-- Boolean, Parameterized, Multi-value -->
4157 <para>Boolean.</para>
4162 <term>Parameter:</term>
4174 As explained <literal><link linkend="filter">above</link></literal>,
4175 <application>Privoxy</application> tries to only filter files that are
4176 in some kind of text format. The same restrictions apply to
4177 <literal><link linkend="content-type-overwrite">content-type-overwrite</link></literal>.
4178 <literal>force-text-mode</literal> declares a document as text,
4179 without looking at the <quote>Content-Type:</quote> first.
4183 Think twice before activating this action. Filtering binary data
4184 with regular expressions can cause file damage.
4191 <term>Example usage:</term>
4204 <!-- ~~~~~ New section ~~~~~ -->
4205 <sect3 renderas="sect4" id="forward-override">
4206 <title>forward-override</title>
4212 <term>Typical use:</term>
4214 <para>Change the forwarding settings based on User-Agent or request origin</para>
4219 <term>Effect:</term>
4222 Overrules the forward directives in the configuration file.
4229 <!-- Boolean, Parameterized, Multi-value -->
4231 <para>Parameterized.</para>
4236 <term>Parameter:</term>
4240 <para><quote>forward .</quote> to use a direct connection without any additional proxies.</para>
4244 <quote>forward 127.0.0.1:8123</quote> to use the HTTP proxy listening at 127.0.0.1 port 8123.
4249 <quote>forward-socks4a 127.0.0.1:9050 .</quote> to use the socks4a proxy listening at
4250 127.0.0.1 port 9050. Replace <quote>forward-socks4a</quote> with <quote>forward-socks4</quote>
4251 to use a socks4 connection (with local DNS resolution) instead, use <quote>forward-socks5</quote>
4252 for socks5 connections (with remote DNS resolution).
4257 <quote>forward-socks4a 127.0.0.1:9050 proxy.example.org:8000</quote> to use the socks4a proxy
4258 listening at 127.0.0.1 port 9050 to reach the HTTP proxy listening at proxy.example.org port 8000.
4259 Replace <quote>forward-socks4a</quote> with <quote>forward-socks4</quote> to use a socks4 connection
4260 (with local DNS resolution) instead, use <quote>forward-socks5</quote>
4261 for socks5 connections (with remote DNS resolution).
4266 <quote>forward-webserver 127.0.0.1:80</quote> to use the HTTP
4267 server listening at 127.0.0.1 port 80 without adjusting the
4271 This makes it more convenient to use Privoxy to make
4272 existing websites available as onion services as well.
4275 Many websites serve content with hardcoded URLs and
4276 can't be easily adjusted to change the domain based
4277 on the one used by the client.
4280 Putting Privoxy between Tor and the webserver (or an stunnel
4281 that forwards to the webserver) allows to rewrite headers and
4282 content to make client and server happy at the same time.
4285 Using Privoxy for webservers that are only reachable through
4286 onion addresses and whose location is supposed to be secret
4287 is not recommended and should not be necessary anyway.
4298 This action takes parameters similar to the
4299 <link linkend="forwarding">forward</link> directives in the configuration
4300 file, but without the URL pattern. It can be used as replacement, but normally it's only
4301 used in cases where matching based on the request URL isn't sufficient.
4305 Please read the description for the <link linkend="forwarding">forward</link> directives before
4306 using this action. Forwarding to the wrong people will reduce your privacy and increase the
4307 chances of man-in-the-middle attacks.
4310 If the ports are missing or invalid, default values will be used. This might change
4311 in the future and you shouldn't rely on it. Otherwise incorrect syntax causes Privoxy
4312 to exit. Due to design limitations, invalid parameter syntax isn't detected until the
4313 action is used the first time.
4316 Use the <ulink url="http://config.privoxy.org/show-url-info">show-url-info CGI page</ulink>
4317 to verify that your forward settings do what you thought the do.
4324 <term>Example usage:</term>
4328 # Use an ssh tunnel for requests previously tagged as
4329 # <quote>User-Agent: fetch libfetch/2.0</quote> and make sure
4330 # resuming downloads continues to work.
4332 # This way you can continue to use Tor for your normal browsing,
4333 # without overloading the Tor network with your FreeBSD ports updates
4334 # or downloads of bigger files like ISOs.
4336 # Note that HTTP headers are easy to fake and therefore their
4337 # values are as (un)trustworthy as your clients and users.
4338 {+forward-override{forward-socks5 10.0.0.2:2222 .} \
4339 -hide-if-modified-since \
4340 -overwrite-last-modified \
4342 TAG:^User-Agent: fetch libfetch/2\.0$
4351 <!-- ~~~~~ New section ~~~~~ -->
4352 <sect3 renderas="sect4" id="handle-as-empty-document">
4353 <title>handle-as-empty-document</title>
4359 <term>Typical use:</term>
4361 <para>Mark URLs that should be replaced by empty documents <emphasis>if they get blocked</emphasis></para>
4366 <term>Effect:</term>
4369 This action alone doesn't do anything noticeable. It just marks URLs.
4370 If the <literal><link linkend="block">block</link></literal> action <emphasis>also applies</emphasis>,
4371 the presence or absence of this mark decides whether an HTML <quote>BLOCKED</quote>
4372 page, or an empty document will be sent to the client as a substitute for the blocked content.
4373 The <emphasis>empty</emphasis> document isn't literally empty, but actually contains a single space.
4380 <!-- Boolean, Parameterized, Multi-value -->
4382 <para>Boolean.</para>
4387 <term>Parameter:</term>
4399 Some browsers complain about syntax errors if JavaScript documents
4400 are blocked with <application>Privoxy's</application>
4401 default HTML page; this option can be used to silence them.
4402 And of course this action can also be used to eliminate the &my-app;
4403 BLOCKED message in frames.
4406 The content type for the empty document can be specified with
4407 <literal><link linkend="content-type-overwrite">content-type-overwrite{}</link></literal>,
4408 but usually this isn't necessary.
4414 <term>Example usage:</term>
4417 <screen># Block all documents on example.org that end with ".js",
4418 # but send an empty document instead of the usual HTML message.
4419 {+block{Blocked JavaScript} +handle-as-empty-document}
4429 <!-- ~~~~~ New section ~~~~~ -->
4430 <sect3 renderas="sect4" id="handle-as-image">
4431 <title>handle-as-image</title>
4435 <term>Typical use:</term>
4437 <para>Mark URLs as belonging to images (so they'll be replaced by images <emphasis>if they do get blocked</emphasis>, rather than HTML pages)</para>
4442 <term>Effect:</term>
4445 This action alone doesn't do anything noticeable. It just marks URLs as images.
4446 If the <literal><link linkend="block">block</link></literal> action <emphasis>also applies</emphasis>,
4447 the presence or absence of this mark decides whether an HTML <quote>blocked</quote>
4448 page, or a replacement image (as determined by the <literal><link
4449 linkend="set-image-blocker">set-image-blocker</link></literal> action) will be sent to the
4450 client as a substitute for the blocked content.
4457 <!-- Boolean, Parameterized, Multi-value -->
4459 <para>Boolean.</para>
4464 <term>Parameter:</term>
4476 The below generic example section is actually part of <filename>default.action</filename>.
4477 It marks all URLs with well-known image file name extensions as images and should
4481 Users will probably only want to use the handle-as-image action in conjunction with
4482 <literal><link linkend="block">block</link></literal>, to block sources of banners, whose URLs don't
4483 reflect the file type, like in the second example section.
4486 Note that you cannot treat HTML pages as images in most cases. For instance, (in-line) ad
4487 frames require an HTML page to be sent, or they won't display properly.
4488 Forcing <literal>handle-as-image</literal> in this situation will not replace the
4489 ad frame with an image, but lead to error messages.
4495 <term>Example usage (sections):</term>
4498 <screen># Generic image extensions:
4501 /.*\.(gif|jpg|jpeg|png|bmp|ico)$
4503 # These don't look like images, but they're banners and should be
4504 # blocked as images:
4506 {+block{Nasty banners.} +handle-as-image}
4507 nasty-banner-server.example.com/junk.cgi\?output=trash
4516 <!-- ~~~~~ New section ~~~~~ -->
4517 <sect3 renderas="sect4" id="hide-accept-language">
4518 <title>hide-accept-language</title>
4524 <term>Typical use:</term>
4526 <para>Pretend to use different language settings.</para>
4531 <term>Effect:</term>
4534 Deletes or replaces the <quote>Accept-Language:</quote> HTTP header in client requests.
4541 <!-- Boolean, Parameterized, Multi-value -->
4543 <para>Parameterized.</para>
4548 <term>Parameter:</term>
4551 Keyword: <quote>block</quote>, or any user defined value.
4560 Faking the browser's language settings can be useful to make a
4561 foreign User-Agent set with
4562 <literal><link linkend="hide-user-agent">hide-user-agent</link></literal>
4566 However some sites with content in different languages check the
4567 <quote>Accept-Language:</quote> to decide which one to take by default.
4568 Sometimes it isn't possible to later switch to another language without
4569 changing the <quote>Accept-Language:</quote> header first.
4572 Therefore it's a good idea to either only change the
4573 <quote>Accept-Language:</quote> header to languages you understand,
4574 or to languages that aren't wide spread.
4577 Before setting the <quote>Accept-Language:</quote> header
4578 to a rare language, you should consider that it helps to
4579 make your requests unique and thus easier to trace.
4580 If you don't plan to change this header frequently,
4581 you should stick to a common language.
4587 <term>Example usage (section):</term>
4590 <screen># Pretend to use Canadian language settings.
4591 {+hide-accept-language{en-ca} \
4592 +hide-user-agent{Mozilla/5.0 (X11; U; OpenBSD i386; en-CA; rv:1.8.0.4) Gecko/20060628 Firefox/1.5.0.4} \
4602 <!-- ~~~~~ New section ~~~~~ -->
4603 <sect3 renderas="sect4" id="hide-content-disposition">
4604 <title>hide-content-disposition</title>
4610 <term>Typical use:</term>
4612 <para>Prevent download menus for content you prefer to view inside the browser.</para>
4617 <term>Effect:</term>
4620 Deletes or replaces the <quote>Content-Disposition:</quote> HTTP header set by some servers.
4627 <!-- Boolean, Parameterized, Multi-value -->
4629 <para>Parameterized.</para>
4634 <term>Parameter:</term>
4637 Keyword: <quote>block</quote>, or any user defined value.
4646 Some servers set the <quote>Content-Disposition:</quote> HTTP header for
4647 documents they assume you want to save locally before viewing them.
4648 The <quote>Content-Disposition:</quote> header contains the file name
4649 the browser is supposed to use by default.
4652 In most browsers that understand this header, it makes it impossible to
4653 <emphasis>just view</emphasis> the document, without downloading it first,
4654 even if it's just a simple text file or an image.
4657 Removing the <quote>Content-Disposition:</quote> header helps
4658 to prevent this annoyance, but some browsers additionally check the
4659 <quote>Content-Type:</quote> header, before they decide if they can
4660 display a document without saving it first. In these cases, you have
4661 to change this header as well, before the browser stops displaying
4665 It is also possible to change the server's file name suggestion
4666 to another one, but in most cases it isn't worth the time to set
4670 This action will probably be removed in the future,
4671 use server-header filters instead.
4677 <term>Example usage:</term>
4680 <screen># Disarm the download link in Sourceforge's patch tracker
4682 +content-type-overwrite{text/plain}\
4683 +hide-content-disposition{block} }
4684 .sourceforge.net/tracker/download\.php</screen>
4692 <!-- ~~~~~ New section ~~~~~ -->
4693 <sect3 renderas="sect4" id="hide-if-modified-since">
4694 <title>hide-if-modified-since</title>
4700 <term>Typical use:</term>
4702 <para>Prevent yet another way to track the user's steps between sessions.</para>
4707 <term>Effect:</term>
4710 Deletes the <quote>If-Modified-Since:</quote> HTTP client header or modifies its value.
4717 <!-- Boolean, Parameterized, Multi-value -->
4719 <para>Parameterized.</para>
4724 <term>Parameter:</term>
4727 Keyword: <quote>block</quote>, or a user defined value that specifies a range of hours.
4736 Removing this header is useful for filter testing, where you want to force a real
4737 reload instead of getting status code <quote>304</quote>, which would cause the
4738 browser to use a cached copy of the page.
4741 Instead of removing the header, <literal>hide-if-modified-since</literal> can
4742 also add or subtract a random amount of time to/from the header's value.
4743 You specify a range of minutes where the random factor should be chosen from and
4744 <application>Privoxy</application> does the rest. A negative value means
4745 subtracting, a positive value adding.
4748 Randomizing the value of the <quote>If-Modified-Since:</quote> makes
4749 it less likely that the server can use the time as a cookie replacement,
4750 but you will run into caching problems if the random range is too high.
4753 It is a good idea to only use a small negative value and let
4754 <literal><link linkend="overwrite-last-modified">overwrite-last-modified</link></literal>
4755 handle the greater changes.
4758 It is also recommended to use this action together with
4759 <literal><link linkend="crunch-if-none-match">crunch-if-none-match</link></literal>,
4760 otherwise it's more or less pointless.
4766 <term>Example usage (section):</term>
4769 <screen># Let the browser revalidate but make tracking based on the time less likely.
4770 {+hide-if-modified-since{-60} \
4771 +overwrite-last-modified{randomize} \
4772 +crunch-if-none-match}
4781 <!-- ~~~~~ New section ~~~~~ -->
4782 <sect3 renderas="sect4" id="hide-from-header">
4783 <title>hide-from-header</title>
4787 <term>Typical use:</term>
4789 <para>Keep your (old and ill) browser from telling web servers your email address</para>
4794 <term>Effect:</term>
4797 Deletes any existing <quote>From:</quote> HTTP header, or replaces it with the
4805 <!-- Boolean, Parameterized, Multi-value -->
4807 <para>Parameterized.</para>
4812 <term>Parameter:</term>
4815 Keyword: <quote>block</quote>, or any user defined value.
4824 The keyword <quote>block</quote> will completely remove the header
4825 (not to be confused with the <literal><link linkend="block">block</link></literal>
4829 Alternately, you can specify any value you prefer to be sent to the web
4830 server. If you do, it is a matter of fairness not to use any address that
4831 is actually used by a real person.
4834 This action is rarely needed, as modern web browsers don't send
4835 <quote>From:</quote> headers anymore.
4841 <term>Example usage:</term>
4844 <screen>+hide-from-header{block}</screen> or
4845 <screen>+hide-from-header{spam-me-senseless@sittingduck.example.com}</screen>
4853 <!-- ~~~~~ New section ~~~~~ -->
4854 <sect3 renderas="sect4" id="hide-referrer">
4855 <title>hide-referrer</title>
4856 <anchor id="hide-referer">
4859 <term>Typical use:</term>
4861 <para>Conceal which link you followed to get to a particular site</para>
4866 <term>Effect:</term>
4869 Deletes the <quote>Referer:</quote> (sic) HTTP header from the client request,
4870 or replaces it with a forged one.
4877 <!-- Boolean, Parameterized, Multi-value -->
4879 <para>Parameterized.</para>
4884 <term>Parameter:</term>
4888 <para><quote>conditional-block</quote> to delete the header completely if the host has changed.</para>
4891 <para><quote>conditional-forge</quote> to forge the header if the host has changed.</para>
4894 <para><quote>block</quote> to delete the header unconditionally.</para>
4897 <para><quote>forge</quote> to pretend to be coming from the homepage of the server we are talking to.</para>
4900 <para>Any other string to set a user defined referrer.</para>
4910 <literal>conditional-block</literal> is the only parameter,
4911 that isn't easily detected in the server's log file. If it blocks the
4912 referrer, the request will look like the visitor used a bookmark or
4913 typed in the address directly.
4916 Leaving the referrer unmodified for requests on the same host
4917 allows the server owner to see the visitor's <quote>click path</quote>,
4918 but in most cases she could also get that information by comparing
4919 other parts of the log file: for example the User-Agent if it isn't
4920 a very common one, or the user's IP address if it doesn't change between
4924 Always blocking the referrer, or using a custom one, can lead to
4925 failures on servers that check the referrer before they answer any
4926 requests, in an attempt to prevent their content from being
4927 embedded or linked to elsewhere.
4930 Both <literal>conditional-block</literal> and <literal>forge</literal>
4931 will work with referrer checks, as long as content and valid referring page
4932 are on the same host. Most of the time that's the case.
4935 <literal>hide-referer</literal> is an alternate spelling of
4936 <literal>hide-referrer</literal> and the two can be can be freely
4937 substituted with each other. (<quote>referrer</quote> is the
4938 correct English spelling, however the HTTP specification has a bug - it
4939 requires it to be spelled as <quote>referer</quote>.)
4945 <term>Example usage:</term>
4948 <screen>+hide-referrer{forge}</screen> or
4949 <screen>+hide-referrer{http://www.yahoo.com/}</screen>
4957 <!-- ~~~~~ New section ~~~~~ -->
4958 <sect3 renderas="sect4" id="hide-user-agent">
4959 <title>hide-user-agent</title>
4963 <term>Typical use:</term>
4965 <para>Try to conceal your type of browser and client operating system</para>
4970 <term>Effect:</term>
4973 Replaces the value of the <quote>User-Agent:</quote> HTTP header
4974 in client requests with the specified value.
4981 <!-- Boolean, Parameterized, Multi-value -->
4983 <para>Parameterized.</para>
4988 <term>Parameter:</term>
4991 Any user-defined string.
5001 This can lead to problems on web sites that depend on looking at this header in
5002 order to customize their content for different browsers (which, by the
5003 way, is <emphasis>NOT</emphasis> the right thing to do: good web sites
5004 work browser-independently).
5008 Using this action in multi-user setups or wherever different types of
5009 browsers will access the same <application>Privoxy</application> is
5010 <emphasis>not recommended</emphasis>. In single-user, single-browser
5011 setups, you might use it to delete your OS version information from
5012 the headers, because it is an invitation to exploit known bugs for your
5013 OS. It is also occasionally useful to forge this in order to access
5014 sites that won't let you in otherwise (though there may be a good
5015 reason in some cases).
5018 More information on known user-agent strings can be found at
5019 <ulink url="http://www.user-agents.org/">http://www.user-agents.org/</ulink>
5021 <ulink url="http://en.wikipedia.org/wiki/User_agent">http://en.wikipedia.org/wiki/User_agent</ulink>.
5027 <term>Example usage:</term>
5030 <screen>+hide-user-agent{Netscape 6.1 (X11; I; Linux 2.4.18 i686)}</screen>
5038 <!-- ~~~~~ New section ~~~~~ -->
5039 <sect3 renderas="sect4" id="limit-connect">
5040 <title>limit-connect</title>
5044 <term>Typical use:</term>
5046 <para>Prevent abuse of <application>Privoxy</application> as a TCP proxy relay or disable SSL for untrusted sites</para>
5051 <term>Effect:</term>
5054 Specifies to which ports HTTP CONNECT requests are allowable.
5061 <!-- Boolean, Parameterized, Multi-value -->
5063 <para>Parameterized.</para>
5068 <term>Parameter:</term>
5071 A comma-separated list of ports or port ranges (the latter using dashes, with the minimum
5072 defaulting to 0 and the maximum to 65K).
5081 By default, i.e. if no <literal>limit-connect</literal> action applies,
5082 <application>Privoxy</application> allows HTTP CONNECT requests to all
5083 ports. Use <literal>limit-connect</literal> if fine-grained control
5084 is desired for some or all destinations.
5087 The CONNECT methods exists in HTTP to allow access to secure websites
5088 (<quote>https://</quote> URLs) through proxies. It works very simply:
5089 the proxy connects to the server on the specified port, and then
5090 short-circuits its connections to the client and to the remote server.
5091 This means CONNECT-enabled proxies can be used as TCP relays very easily.
5094 <application>Privoxy</application> relays HTTPS traffic without seeing
5095 the decoded content. Websites can leverage this limitation to circumvent &my-app;'s
5096 filters. By specifying an invalid port range you can disable HTTPS entirely.
5102 <term>Example usages:</term>
5104 <!-- I had trouble getting the spacing to look right in my browser -->
5105 <!-- I probably have the wrong font setup, bollocks. -->
5106 <!-- Apparently the emphasis tag uses a proportional font no matter what -->
5108 <screen>+limit-connect{443} # Port 443 is OK.
5109 +limit-connect{80,443} # Ports 80 and 443 are OK.
5110 +limit-connect{-3, 7, 20-100, 500-} # Ports less than 3, 7, 20 to 100 and above 500 are OK.
5111 +limit-connect{-} # All ports are OK
5112 +limit-connect{,} # No HTTPS/SSL traffic is allowed</screen>
5120 <!-- ~~~~~ New section ~~~~~ -->
5121 <sect3 renderas="sect4" id="limit-cookie-lifetime">
5122 <title>limit-cookie-lifetime</title>
5126 <term>Typical use:</term>
5128 <para>Limit the lifetime of HTTP cookies to a couple of minutes or hours.</para>
5133 <term>Effect:</term>
5136 Overwrites the expires field in Set-Cookie server headers if it's above the specified limit.
5143 <!-- Boolean, Parameterized, Multi-value -->
5145 <para>Parameterized.</para>
5150 <term>Parameter:</term>
5153 The lifetime limit in minutes, or 0.
5162 This action reduces the lifetime of HTTP cookies coming from the
5163 server to the specified number of minutes, starting from the time
5164 the cookie passes Privoxy.
5167 Cookies with a lifetime below the limit are not modified.
5168 The lifetime of session cookies is set to the specified limit.
5171 The effect of this action depends on the server.
5174 In case of servers which refresh their cookies with each response
5175 (or at least frequently), the lifetime limit set by this action
5177 Thus, a session associated with the cookie continues to work with
5178 this action enabled, as long as a new request is made before the
5179 last limit set is reached.
5182 However, some servers send their cookies once, with a lifetime of several
5183 years (the year 2037 is a popular choice), and do not refresh them
5184 until a certain event in the future, for example the user logging out.
5185 In this case this action may limit the absolute lifetime of the session,
5186 even if requests are made frequently.
5189 If the parameter is <quote>0</quote>, this action behaves like
5190 <literal><link linkend="session-cookies-only">session-cookies-only</link></literal>.
5196 <term>Example usages:</term>
5199 <screen>+limit-cookie-lifetime{60}
5207 <!-- ~~~~~ New section ~~~~~ -->
5208 <sect3 renderas="sect4" id="prevent-compression">
5209 <title>prevent-compression</title>
5213 <term>Typical use:</term>
5216 Ensure that servers send the content uncompressed, so it can be
5217 passed through <literal><link linkend="filter">filter</link></literal>s.
5223 <term>Effect:</term>
5226 Removes the Accept-Encoding header which can be used to ask for compressed transfer.
5233 <!-- Boolean, Parameterized, Multi-value -->
5235 <para>Boolean.</para>
5240 <term>Parameter:</term>
5252 More and more websites send their content compressed by default, which
5253 is generally a good idea and saves bandwidth. But the <literal><link
5254 linkend="filter">filter</link></literal> and
5255 <literal><link linkend="deanimate-gifs">deanimate-gifs</link></literal>
5256 actions need access to the uncompressed data.
5259 When compiled with zlib support (available since &my-app; 3.0.7), content that should be
5260 filtered is decompressed on-the-fly and you don't have to worry about this action.
5261 If you are using an older &my-app; version, or one that hasn't been compiled with zlib
5262 support, this action can be used to convince the server to send the content uncompressed.
5265 Most text-based instances compress very well, the size is seldom decreased by less than 50%,
5266 for markup-heavy instances like news feeds saving more than 90% of the original size isn't
5270 Not using compression will therefore slow down the transfer, and you should only
5271 enable this action if you really need it. As of &my-app; 3.0.7 it's disabled in all
5272 predefined action settings.
5275 Note that some (rare) ill-configured sites don't handle requests for uncompressed
5276 documents correctly. Broken PHP applications tend to send an empty document body,
5277 some IIS versions only send the beginning of the content. If you enable
5278 <literal>prevent-compression</literal> per default, you might want to add
5279 exceptions for those sites. See the example for how to do that.
5285 <term>Example usage (sections):</term>
5289 # Selectively turn off compression, and enable a filter
5291 { +filter{tiny-textforms} +prevent-compression }
5292 # Match only these sites
5297 # Or instead, we could set a universal default:
5299 { +prevent-compression }
5302 # Then maybe make exceptions for broken sites:
5304 { -prevent-compression }
5305 .compusa.com/</screen>
5314 <!-- ~~~~~ New section ~~~~~ -->
5315 <sect3 renderas="sect4" id="overwrite-last-modified">
5316 <title>overwrite-last-modified</title>
5322 <term>Typical use:</term>
5324 <para>Prevent yet another way to track the user's steps between sessions.</para>
5329 <term>Effect:</term>
5332 Deletes the <quote>Last-Modified:</quote> HTTP server header or modifies its value.
5339 <!-- Boolean, Parameterized, Multi-value -->
5341 <para>Parameterized.</para>
5346 <term>Parameter:</term>
5349 One of the keywords: <quote>block</quote>, <quote>reset-to-request-time</quote>
5350 and <quote>randomize</quote>
5359 Removing the <quote>Last-Modified:</quote> header is useful for filter
5360 testing, where you want to force a real reload instead of getting status
5361 code <quote>304</quote>, which would cause the browser to reuse the old
5362 version of the page.
5365 The <quote>randomize</quote> option overwrites the value of the
5366 <quote>Last-Modified:</quote> header with a randomly chosen time
5367 between the original value and the current time. In theory the server
5368 could send each document with a different <quote>Last-Modified:</quote>
5369 header to track visits without using cookies. <quote>Randomize</quote>
5370 makes it impossible and the browser can still revalidate cached documents.
5373 <quote>reset-to-request-time</quote> overwrites the value of the
5374 <quote>Last-Modified:</quote> header with the current time. You could use
5375 this option together with
5376 <literal><link linkend="hide-if-modified-since">hide-if-modified-since</link></literal>
5377 to further customize your random range.
5380 The preferred parameter here is <quote>randomize</quote>. It is safe
5381 to use, as long as the time settings are more or less correct.
5382 If the server sets the <quote>Last-Modified:</quote> header to the time
5383 of the request, the random range becomes zero and the value stays the same.
5384 Therefore you should later randomize it a second time with
5385 <literal><link linkend="hide-if-modified-since">hided-if-modified-since</link></literal>,
5389 It is also recommended to use this action together with
5390 <literal><link linkend="crunch-if-none-match">crunch-if-none-match</link></literal>.
5396 <term>Example usage:</term>
5399 <screen># Let the browser revalidate without being tracked across sessions
5400 { +hide-if-modified-since{-60} \
5401 +overwrite-last-modified{randomize} \
5402 +crunch-if-none-match}
5411 <!-- ~~~~~ New section ~~~~~ -->
5412 <sect3 renderas="sect4" id="redirect">
5413 <title>redirect</title>
5419 <term>Typical use:</term>
5422 Redirect requests to other sites.
5428 <term>Effect:</term>
5431 Convinces the browser that the requested document has been moved
5432 to another location and the browser should get it from there.
5439 <!-- Boolean, Parameterized, Multi-value -->
5441 <para>Parameterized</para>
5446 <term>Parameter:</term>
5449 An absolute URL or a single pcrs command.
5458 Requests to which this action applies are answered with a
5459 HTTP redirect to URLs of your choosing. The new URL is
5460 either provided as parameter, or derived by applying a
5461 single pcrs command to the original URL.
5464 The syntax for pcrs commands is documented in the
5465 <link linkend="filter-file">filter file</link> section.
5468 Requests can't be blocked and redirected at the same time,
5469 applying this action together with
5470 <literal><link linkend="block">block</link></literal>
5471 is a configuration error. Currently the request is blocked
5472 and an error message logged, the behavior may change in the
5473 future and result in Privoxy rejecting the action file.
5476 This action can be combined with
5477 <literal><link linkend="fast-redirects">fast-redirects{check-decoded-url}</link></literal>
5478 to redirect to a decoded version of a rewritten URL.
5481 Use this action carefully, make sure not to create redirection loops
5482 and be aware that using your own redirects might make it
5483 possible to fingerprint your requests.
5486 In case of problems with your redirects, or simply to watch
5487 them working, enable <link linkend="DEBUG">debug 128</link>.
5493 <term>Example usages:</term>
5496 <screen># Replace example.com's style sheet with another one
5497 { +redirect{http://localhost/css-replacements/example.com.css} }
5498 example.com/stylesheet\.css
5500 # Create a short, easy to remember nickname for a favorite site
5501 # (relies on the browser to accept and forward invalid URLs to &my-app;)
5502 { +redirect{http://www.privoxy.org/user-manual/actions-file.html} }
5505 # Always use the expanded view for Undeadly.org articles
5506 # (Note the $ at the end of the URL pattern to make sure
5507 # the request for the rewritten URL isn't redirected as well)
5508 {+redirect{s@$@&mode=expanded@}}
5509 undeadly.org/cgi\?action=article&sid=\d*$
5511 # Redirect Google search requests to MSN
5512 {+redirect{s@^http://[^/]*/search\?q=([^&]*).*@http://search.msn.com/results.aspx?q=$1@}}
5515 # Redirect MSN search requests to Yahoo
5516 {+redirect{s@^http://[^/]*/results\.aspx\?q=([^&]*).*@http://search.yahoo.com/search?p=$1@}}
5517 search.msn.com//results\.aspx\?q=
5519 # Redirect http://example.com/&bla=fasel&toChange=foo (and any other value but "bar")
5520 # to http://example.com/&bla=fasel&toChange=bar
5522 # The URL pattern makes sure that the following request isn't redirected again.
5523 {+redirect{s@toChange=[^&]+@toChange=bar@}}
5524 example.com/.*toChange=(?!bar)
5526 # Add a shortcut to look up illumos bugs
5527 {+redirect{s@^http://i([0-9]+)/.*@https://www.illumos.org/issues/$1@}}
5528 # Redirected URL = http://i4974/
5529 # Redirect Destination = https://www.illumos.org/issues/4974
5530 i[0-9][0-9][0-9][0-9]*/
5532 # Redirect remote requests for this manual
5533 # to the local version delivered by Privoxy
5534 {+redirect{s@^http://www@http://config@}}
5535 www.privoxy.org/user-manual/</screen>
5544 <!-- ~~~~~ New section ~~~~~ -->
5545 <sect3 renderas="sect4" id="server-header-filter">
5546 <title>server-header-filter</title>
5550 <term>Typical use:</term>
5553 Rewrite or remove single server headers.
5559 <term>Effect:</term>
5562 All server headers to which this action applies are filtered on-the-fly
5563 through the specified regular expression based substitutions.
5570 <!-- boolean, parameterized, Multi-value -->
5572 <para>Multi-value.</para>
5577 <term>Parameter:</term>
5580 The name of a server-header filter, as defined in one of the
5581 <link linkend="filter-file">filter files</link>.
5590 Server-header filters are applied to each header on its own, not to
5591 all at once. This makes it easier to diagnose problems, but on the downside
5592 you can't write filters that only change header x if header y's value is z.
5593 You can do that by using tags though.
5596 Server-header filters are executed after the other header actions have finished
5597 and use their output as input.
5600 Please refer to the <link linkend="filter-file">filter file chapter</link>
5601 to learn which server-header filters are available by default, and how to
5608 <term>Example usage (section):</term>
5612 {+server-header-filter{html-to-xml}}
5613 example.org/xml-instance-that-is-delivered-as-html
5615 {+server-header-filter{xml-to-html}}
5616 example.org/instance-that-is-delivered-as-xml-but-is-not
5626 <!-- ~~~~~ New section ~~~~~ -->
5627 <sect3 renderas="sect4" id="server-header-tagger">
5628 <title>server-header-tagger</title>
5632 <term>Typical use:</term>
5635 Enable or disable filters based on the Content-Type header.
5641 <term>Effect:</term>
5644 Server headers to which this action applies are filtered on-the-fly through
5645 the specified regular expression based substitutions, the result is used as
5653 <!-- boolean, parameterized, Multi-value -->
5655 <para>Multi-value.</para>
5660 <term>Parameter:</term>
5663 The name of a server-header tagger, as defined in one of the
5664 <link linkend="filter-file">filter files</link>.
5673 Server-header taggers are applied to each header on its own,
5674 and as the header isn't modified, each tagger <quote>sees</quote>
5678 Server-header taggers are executed before all other header actions
5679 that modify server headers. Their tags can be used to control
5680 all of the other server-header actions, the content filters
5681 and the crunch actions (<link linkend="redirect">redirect</link>
5682 and <link linkend="block">block</link>).
5685 Obviously crunching based on tags created by server-header taggers
5686 doesn't prevent the request from showing up in the server's log file.
5693 <term>Example usage (section):</term>
5697 # Tag every request with the content type declared by the server
5698 {+server-header-tagger{content-type}}
5701 # If the response has a tag starting with 'image/' enable an external
5702 # filter that only applies to images.
5704 # Note that the filter is not available by default, it's just a
5705 # <literal><link linkend="external-filter-syntax">silly example</link></literal>.
5706 {+external-filter{rotate-image} +force-text-mode}
5717 <!-- ~~~~~ New section ~~~~~ -->
5718 <sect3 renderas="sect4" id="session-cookies-only">
5719 <title>session-cookies-only</title>
5723 <term>Typical use:</term>
5726 Allow only temporary <quote>session</quote> cookies (for the current
5727 browser session <emphasis>only</emphasis>).
5733 <term>Effect:</term>
5736 Deletes the <quote>expires</quote> field from <quote>Set-Cookie:</quote>
5737 server headers. Most browsers will not store such cookies permanently and
5738 forget them in between sessions.
5745 <!-- Boolean, Parameterized, Multi-value -->
5747 <para>Boolean.</para>
5752 <term>Parameter:</term>
5764 This is less strict than <literal><link linkend="crunch-incoming-cookies">crunch-incoming-cookies</link></literal> /
5765 <literal><link linkend="crunch-outgoing-cookies">crunch-outgoing-cookies</link></literal> and allows you to browse
5766 websites that insist or rely on setting cookies, without compromising your privacy too badly.
5769 Most browsers will not permanently store cookies that have been processed by
5770 <literal>session-cookies-only</literal> and will forget about them between sessions.
5771 This makes profiling cookies useless, but won't break sites which require cookies so
5772 that you can log in for transactions. This is generally turned on for all
5773 sites, and is the recommended setting.
5776 It makes <emphasis>no sense at all</emphasis> to use <literal>session-cookies-only</literal>
5777 together with <literal><link linkend="crunch-incoming-cookies">crunch-incoming-cookies</link></literal> or
5778 <literal><link linkend="crunch-outgoing-cookies">crunch-outgoing-cookies</link></literal>. If you do, cookies
5779 will be plainly killed.
5782 Note that it is up to the browser how it handles such cookies without an <quote>expires</quote>
5783 field. If you use an exotic browser, you might want to try it out to be sure.
5786 This setting also has no effect on cookies that may have been stored
5787 previously by the browser before starting <application>Privoxy</application>.
5788 These would have to be removed manually.
5791 <application>Privoxy</application> also uses
5792 the <link linkend="filter-content-cookies">content-cookies filter</link>
5793 to block some types of cookies. Content cookies are not effected by
5794 <literal>session-cookies-only</literal>.
5800 <term>Example usage:</term>
5803 <screen>+session-cookies-only</screen>
5811 <!-- ~~~~~ New section ~~~~~ -->
5812 <sect3 renderas="sect4" id="set-image-blocker">
5813 <title>set-image-blocker</title>
5817 <term>Typical use:</term>
5819 <para>Choose the replacement for blocked images</para>
5824 <term>Effect:</term>
5827 This action alone doesn't do anything noticeable. If <emphasis>both</emphasis>
5828 <literal><link linkend="block">block</link></literal> <emphasis>and</emphasis> <literal><link
5829 linkend="handle-as-image">handle-as-image</link></literal> <emphasis>also</emphasis>
5830 apply, i.e. if the request is to be blocked as an image,
5831 <emphasis>then</emphasis> the parameter of this action decides what will be
5832 sent as a replacement.
5839 <!-- Boolean, Parameterized, Multi-value -->
5841 <para>Parameterized.</para>
5846 <term>Parameter:</term>
5851 <quote>pattern</quote> to send a built-in checkerboard pattern image. The image is visually
5852 decent, scales very well, and makes it obvious where banners were busted.
5857 <quote>blank</quote> to send a built-in transparent image. This makes banners disappear
5858 completely, but makes it hard to detect where <application>Privoxy</application> has blocked
5859 images on a given page and complicates troubleshooting if <application>Privoxy</application>
5860 has blocked innocent images, like navigation icons.
5865 <quote><replaceable class="parameter">target-url</replaceable></quote> to
5866 send a redirect to <replaceable class="parameter">target-url</replaceable>. You can redirect
5867 to any image anywhere, even in your local filesystem via <quote>file:///</quote> URL.
5868 (But note that not all browsers support redirecting to a local file system).
5871 A good application of redirects is to use special <application>Privoxy</application>-built-in
5872 URLs, which send the built-in images, as <replaceable class="parameter">target-url</replaceable>.
5873 This has the same visual effect as specifying <quote>blank</quote> or <quote>pattern</quote> in
5874 the first place, but enables your browser to cache the replacement image, instead of requesting
5875 it over and over again.
5886 The URLs for the built-in images are <quote>http://config.privoxy.org/send-banner?type=<replaceable
5887 class="parameter">type</replaceable></quote>, where <replaceable class="parameter">type</replaceable> is
5888 either <quote>blank</quote> or <quote>pattern</quote>.
5891 There is a third (advanced) type, called <quote>auto</quote>. It is <emphasis>NOT</emphasis> to be
5892 used in <literal>set-image-blocker</literal>, but meant for use from <link linkend="filter-file">filters</link>.
5893 Auto will select the type of image that would have applied to the referring page, had it been an image.
5899 <term>Example usage:</term>
5905 <screen>+set-image-blocker{pattern}</screen>
5908 Redirect to the BSD daemon:
5911 <screen>+set-image-blocker{http://www.freebsd.org/gifs/dae_up3.gif}</screen>
5914 Redirect to the built-in pattern for better caching:
5917 <screen>+set-image-blocker{http://config.privoxy.org/send-banner?type=pattern}</screen>
5925 <!-- ~~~~~ New section ~~~~~ -->
5926 <sect3 id="summary">
5927 <title>Summary</title>
5929 Note that many of these actions have the potential to cause a page to
5930 misbehave, possibly even not to display at all. There are many ways
5931 a site designer may choose to design his site, and what HTTP header
5932 content, and other criteria, he may depend on. There is no way to have hard
5933 and fast rules for all sites. See the <link
5934 linkend="ACTIONSANAT">Appendix</link> for a brief example on troubleshooting
5940 <!-- ~~~~~ New section ~~~~~ -->
5941 <sect2 id="aliases">
5942 <title>Aliases</title>
5944 Custom <quote>actions</quote>, known to <application>Privoxy</application>
5945 as <quote>aliases</quote>, can be defined by combining other actions.
5946 These can in turn be invoked just like the built-in actions.
5947 Currently, an alias name can contain any character except space, tab,
5949 <quote>{</quote> and <quote>}</quote>, but we <emphasis>strongly
5950 recommend</emphasis> that you only use <quote>a</quote> to <quote>z</quote>,
5951 <quote>0</quote> to <quote>9</quote>, <quote>+</quote>, and <quote>-</quote>.
5952 Alias names are not case sensitive, and are not required to start with a
5953 <quote>+</quote> or <quote>-</quote> sign, since they are merely textually
5957 Aliases can be used throughout the actions file, but they <emphasis>must be
5958 defined in a special section at the top of the file!</emphasis>
5959 And there can only be one such section per actions file. Each actions file may
5960 have its own alias section, and the aliases defined in it are only visible
5964 There are two main reasons to use aliases: One is to save typing for frequently
5965 used combinations of actions, the other one is a gain in flexibility: If you
5966 decide once how you want to handle shops by defining an alias called
5967 <quote>shop</quote>, you can later change your policy on shops in
5968 <emphasis>one</emphasis> place, and your changes will take effect everywhere
5969 in the actions file where the <quote>shop</quote> alias is used. Calling aliases
5970 by their purpose also makes your actions files more readable.
5973 Currently, there is one big drawback to using aliases, though:
5974 <application>Privoxy</application>'s built-in web-based action file
5975 editor honors aliases when reading the actions files, but it expands
5976 them before writing. So the effects of your aliases are of course preserved,
5977 but the aliases themselves are lost when you edit sections that use aliases
5982 Now let's define some aliases...
5987 # Useful custom aliases we can use later.
5989 # Note the (required!) section header line and that this section
5990 # must be at the top of the actions file!
5994 # These aliases just save typing later:
5995 # (Note that some already use other aliases!)
5997 +crunch-all-cookies = +<link linkend="CRUNCH-INCOMING-COOKIES">crunch-incoming-cookies</link> +<link linkend="CRUNCH-OUTGOING-COOKIES">crunch-outgoing-cookies</link>
5998 -crunch-all-cookies = -<link linkend="CRUNCH-INCOMING-COOKIES">crunch-incoming-cookies</link> -<link linkend="CRUNCH-OUTGOING-COOKIES">crunch-outgoing-cookies</link>
5999 +block-as-image = +block{Blocked image.} +handle-as-image
6000 allow-all-cookies = -crunch-all-cookies -<link linkend="SESSION-COOKIES-ONLY">session-cookies-only</link> -<link linkend="FILTER-CONTENT-COOKIES">filter{content-cookies}</link>
6002 # These aliases define combinations of actions
6003 # that are useful for certain types of sites:
6005 fragile = -<link linkend="BLOCK">block</link> -<link linkend="FILTER">filter</link> -crunch-all-cookies -<link linkend="FAST-REDIRECTS">fast-redirects</link> -<link linkend="HIDE-REFERER">hide-referrer</link> -<link linkend="PREVENT-COMPRESSION">prevent-compression</link>
6007 shop = -crunch-all-cookies -<link linkend="FILTER-ALL-POPUPS">filter{all-popups}</link>
6009 # Short names for other aliases, for really lazy people ;-)
6011 c0 = +crunch-all-cookies
6012 c1 = -crunch-all-cookies</screen>
6016 ...and put them to use. These sections would appear in the lower part of an
6017 actions file and define exceptions to the default actions (as specified further
6018 up for the <quote>/</quote> pattern):
6023 # These sites are either very complex or very keen on
6024 # user data and require minimal interference to work:
6027 .office.microsoft.com
6028 .windowsupdate.microsoft.com
6029 # Gmail is really mail.google.com, not gmail.com
6033 # Allow cookies (for setting and retrieving your customer data)
6037 .worldpay.com # for quietpc.com
6040 # These shops require pop-ups:
6042 {-filter{all-popups} -filter{unsolicited-popups}}
6044 .overclockers.co.uk</screen>
6048 Aliases like <quote>shop</quote> and <quote>fragile</quote> are typically used for
6049 <quote>problem</quote> sites that require more than one action to be disabled
6050 in order to function properly.
6056 <!-- ~~~~~ New section ~~~~~ -->
6057 <sect2 id="act-examples">
6058 <title>Actions Files Tutorial</title>
6060 The above chapters have shown <link linkend="actions-file">which actions files
6061 there are and how they are organized</link>, how actions are <link
6062 linkend="actions">specified</link> and <link linkend="actions-apply">applied
6063 to URLs</link>, how <link linkend="af-patterns">patterns</link> work, and how to
6064 define and use <link linkend="aliases">aliases</link>. Now, let's look at an
6065 example <filename>match-all.action</filename>, <filename>default.action</filename>
6066 and <filename>user.action</filename> file and see how all these pieces come together:
6069 <sect3 id="match-all">
6070 <title>match-all.action</title>
6072 Remember <emphasis>all actions are disabled when matching starts</emphasis>,
6073 so we have to explicitly enable the ones we want.
6077 While the <filename>match-all.action</filename> file only contains a
6078 single section, it is probably the most important one. It has only one
6079 pattern, <quote><literal>/</literal></quote>, but this pattern
6080 <link linkend="af-patterns">matches all URLs</link>. Therefore, the set of
6081 actions used in this <quote>default</quote> section <emphasis>will
6082 be applied to all requests as a start</emphasis>. It can be partly or
6083 wholly overridden by other actions files like <filename>default.action</filename>
6084 and <filename>user.action</filename>, but it will still be largely responsible
6085 for your overall browsing experience.
6089 Again, at the start of matching, all actions are disabled, so there is
6090 no need to disable any actions here. (Remember: a <quote>+</quote>
6091 preceding the action name enables the action, a <quote>-</quote> disables!).
6092 Also note how this long line has been made more readable by splitting it into
6093 multiple lines with line continuation.
6099 +<link linkend="CHANGE-X-FORWARDED-FOR">change-x-forwarded-for{block}</link> \
6100 +<link linkend="HIDE-FROM-HEADER">hide-from-header{block}</link> \
6101 +<link linkend="SET-IMAGE-BLOCKER">set-image-blocker{pattern}</link> \
6108 The default behavior is now set.
6112 <sect3 id="default-action">
6113 <title>default.action</title>
6116 If you aren't a developer, there's no need for you to edit the
6117 <filename>default.action</filename> file. It is maintained by
6118 the &my-app; developers and if you disagree with some of the
6119 sections, you should overrule them in your <filename>user.action</filename>.
6123 Understanding the <filename>default.action</filename> file can
6124 help you with your <filename>user.action</filename>, though.
6128 The first section in this file is a special section for internal use
6129 that prevents older &my-app; versions from reading the file:
6134 ##########################################################################
6135 # Settings -- Don't change! For internal Privoxy use ONLY.
6136 ##########################################################################
6138 for-privoxy-version=3.0.11</screen>
6142 After that comes the (optional) alias section. We'll use the example
6143 section from the above <link linkend="aliases">chapter on aliases</link>,
6144 that also explains why and how aliases are used:
6149 ##########################################################################
6151 ##########################################################################
6154 # These aliases just save typing later:
6155 # (Note that some already use other aliases!)
6157 +crunch-all-cookies = +<link linkend="CRUNCH-INCOMING-COOKIES">crunch-incoming-cookies</link> +<link linkend="CRUNCH-OUTGOING-COOKIES">crunch-outgoing-cookies</link>
6158 -crunch-all-cookies = -<link linkend="CRUNCH-INCOMING-COOKIES">crunch-incoming-cookies</link> -<link linkend="CRUNCH-OUTGOING-COOKIES">crunch-outgoing-cookies</link>
6159 +block-as-image = +block{Blocked image.} +handle-as-image
6160 mercy-for-cookies = -crunch-all-cookies -<link linkend="SESSION-COOKIES-ONLY">session-cookies-only</link> -<link linkend="FILTER-CONTENT-COOKIES">filter{content-cookies}</link>
6162 # These aliases define combinations of actions
6163 # that are useful for certain types of sites:
6165 fragile = -<link linkend="BLOCK">block</link> -<link linkend="FILTER">filter</link> -crunch-all-cookies -<link linkend="FAST-REDIRECTS">fast-redirects</link> -<link linkend="HIDE-REFERER">hide-referrer</link>
6166 shop = -crunch-all-cookies -<link linkend="FILTER-ALL-POPUPS">filter{all-popups}</link></screen>
6170 The first of our specialized sections is concerned with <quote>fragile</quote>
6171 sites, i.e. sites that require minimum interference, because they are either
6172 very complex or very keen on tracking you (and have mechanisms in place that
6173 make them unusable for people who avoid being tracked). We will simply use
6174 our pre-defined <literal>fragile</literal> alias instead of stating the list
6175 of actions explicitly:
6180 ##########################################################################
6181 # Exceptions for sites that'll break under the default action set:
6182 ##########################################################################
6184 # "Fragile" Use a minimum set of actions for these sites (see alias above):
6187 .office.microsoft.com # surprise, surprise!
6188 .windowsupdate.microsoft.com
6189 mail.google.com</screen>
6193 Shopping sites are not as fragile, but they typically
6194 require cookies to log in, and pop-up windows for shopping
6195 carts or item details. Again, we'll use a pre-defined alias:
6204 .worldpay.com # for quietpc.com
6206 .scan.co.uk</screen>
6210 The <literal><link linkend="FAST-REDIRECTS">fast-redirects</link></literal>
6211 action, which may have been enabled in <filename>match-all.action</filename>,
6212 breaks some sites. So disable it for popular sites where we know it misbehaves:
6217 { -<link linkend="FAST-REDIRECTS">fast-redirects</link> }
6221 .altavista.com/.*(like|url|link):http
6222 .altavista.com/trans.*urltext=http
6223 .nytimes.com</screen>
6227 It is important that <application>Privoxy</application> knows which
6228 URLs belong to images, so that <emphasis>if</emphasis> they are to
6229 be blocked, a substitute image can be sent, rather than an HTML page.
6230 Contacting the remote site to find out is not an option, since it
6231 would destroy the loading time advantage of banner blocking, and it
6232 would feed the advertisers information about you. We can mark any
6233 URL as an image with the <literal><link
6234 linkend="handle-as-image">handle-as-image</link></literal> action,
6235 and marking all URLs that end in a known image file extension is a
6241 ##########################################################################
6243 ##########################################################################
6245 # Define which file types will be treated as images, in case they get
6246 # blocked further down this file:
6248 { +<link linkend="HANDLE-AS-IMAGE">handle-as-image</link> }
6249 /.*\.(gif|jpe?g|png|bmp|ico)$</screen>
6253 And then there are known banner sources. They often use scripts to
6254 generate the banners, so it won't be visible from the URL that the
6255 request is for an image. Hence we block them <emphasis>and</emphasis>
6256 mark them as images in one go, with the help of our
6257 <literal>+block-as-image</literal> alias defined above. (We could of
6258 course just as well use <literal>+<link linkend="block">block</link>
6259 +<link linkend="handle-as-image">handle-as-image</link></literal> here.)
6260 Remember that the type of the replacement image is chosen by the
6261 <literal><link linkend="set-image-blocker">set-image-blocker</link></literal>
6262 action. Since all URLs have matched the default section with its
6263 <literal>+<link linkend="set-image-blocker">set-image-blocker</link>{pattern}</literal>
6264 action before, it still applies and needn't be repeated:
6269 # Known ad generators:
6274 .ad.*.doubleclick.net
6275 .a.yimg.com/(?:(?!/i/).)*$
6276 .a[0-9].yimg.com/(?:(?!/i/).)*$
6282 One of the most important jobs of <application>Privoxy</application>
6283 is to block banners. Many of these can be <quote>blocked</quote>
6284 by the <literal><link linkend="filter">filter</link>{banners-by-size}</literal>
6285 action, which we enabled above, and which deletes the references to banner
6286 images from the pages while they are loaded, so the browser doesn't request
6287 them anymore, and hence they don't need to be blocked here. But this naturally
6288 doesn't catch all banners, and some people choose not to use filters, so we
6289 need a comprehensive list of patterns for banner URLs here, and apply the
6290 <literal><link linkend="block">block</link></literal> action to them.
6293 First comes many generic patterns, which do most of the work, by
6294 matching typical domain and path name components of banners. Then comes
6295 a list of individual patterns for specific sites, which is omitted here
6296 to keep the example short:
6301 ##########################################################################
6302 # Block these fine banners:
6303 ##########################################################################
6304 { <link linkend="BLOCK">+block{Banner ads.}</link> }
6312 /.*count(er)?\.(pl|cgi|exe|dll|asp|php[34]?)
6313 /(?:.*/)?(publicite|werbung|rekla(ma|me|am)|annonse|maino(kset|nta|s)?)/
6315 # Site-specific patterns (abbreviated):
6317 .hitbox.com</screen>
6321 It's quite remarkable how many advertisers actually call their banner
6322 servers ads.<replaceable>company</replaceable>.com, or call the directory
6323 in which the banners are stored simply <quote>banners</quote>. So the above
6324 generic patterns are surprisingly effective.
6327 But being very generic, they necessarily also catch URLs that we don't want
6328 to block. The pattern <literal>.*ads.</literal> e.g. catches
6329 <quote>nasty-<emphasis>ads</emphasis>.nasty-corp.com</quote> as intended,
6330 but also <quote>downlo<emphasis>ads</emphasis>.sourcefroge.net</quote> or
6331 <quote><emphasis>ads</emphasis>l.some-provider.net.</quote> So here come some
6332 well-known exceptions to the <literal>+<link linkend="BLOCK">block</link></literal>
6336 Note that these are exceptions to exceptions from the default! Consider the URL
6337 <quote>downloads.sourcefroge.net</quote>: Initially, all actions are deactivated,
6338 so it wouldn't get blocked. Then comes the defaults section, which matches the
6339 URL, but just deactivates the <literal><link linkend="BLOCK">block</link></literal>
6340 action once again. Then it matches <literal>.*ads.</literal>, an exception to the
6341 general non-blocking policy, and suddenly
6342 <literal><link linkend="BLOCK">+block</link></literal> applies. And now, it'll match
6343 <literal>.*loads.</literal>, where <literal><link linkend="BLOCK">-block</link></literal>
6344 applies, so (unless it matches <emphasis>again</emphasis> further down) it ends up
6345 with no <literal><link linkend="BLOCK">block</link></literal> action applying.
6350 ##########################################################################
6351 # Save some innocent victims of the above generic block patterns:
6352 ##########################################################################
6356 { -<link linkend="BLOCK">block</link> }
6357 adv[io]*. # (for advogato.org and advice.*)
6358 adsl. # (has nothing to do with ads)
6359 adobe. # (has nothing to do with ads either)
6360 ad[ud]*. # (adult.* and add.*)
6361 .edu # (universities don't host banners (yet!))
6362 .*loads. # (downloads, uploads etc)
6370 www.globalintersec.com/adv # (adv = advanced)
6371 www.ugu.com/sui/ugu/adv</screen>
6375 Filtering source code can have nasty side effects,
6376 so make an exception for our friends at sourceforge.net,
6377 and all paths with <quote>cvs</quote> in them. Note that
6378 <literal>-<link linkend="FILTER">filter</link></literal>
6379 disables <emphasis>all</emphasis> filters in one fell swoop!
6384 # Don't filter code!
6386 { -<link linkend="FILTER">filter</link> }
6391 .sourceforge.net</screen>
6395 The actual <filename>default.action</filename> is of course much more
6396 comprehensive, but we hope this example made clear how it works.
6401 <sect3 id="user-action"><title>user.action</title>
6404 So far we are painting with a broad brush by setting general policies,
6405 which would be a reasonable starting point for many people. Now,
6406 you might want to be more specific and have customized rules that
6407 are more suitable to your personal habits and preferences. These would
6408 be for narrowly defined situations like your ISP or your bank, and should
6409 be placed in <filename>user.action</filename>, which is parsed after all other
6410 actions files and hence has the last word, over-riding any previously
6411 defined actions. <filename>user.action</filename> is also a
6412 <emphasis>safe</emphasis> place for your personal settings, since
6413 <filename>default.action</filename> is actively maintained by the
6414 <application>Privoxy</application> developers and you'll probably want
6415 to install updated versions from time to time.
6419 So let's look at a few examples of things that one might typically do in
6420 <filename>user.action</filename>:
6424 <!-- brief sample user.action here -->
6428 # My user.action file. <fred@example.com></screen>
6432 As <link linkend="aliases">aliases</link> are local to the actions
6433 file that they are defined in, you can't use the ones from
6434 <filename>default.action</filename>, unless you repeat them here:
6439 # Aliases are local to the file they are defined in.
6440 # (Re-)define aliases for this file:
6444 # These aliases just save typing later, and the alias names should
6445 # be self explanatory.
6447 +crunch-all-cookies = +crunch-incoming-cookies +crunch-outgoing-cookies
6448 -crunch-all-cookies = -crunch-incoming-cookies -crunch-outgoing-cookies
6449 allow-all-cookies = -crunch-all-cookies -session-cookies-only
6450 allow-popups = -filter{all-popups}
6451 +block-as-image = +block{Blocked as image.} +handle-as-image
6452 -block-as-image = -block
6454 # These aliases define combinations of actions that are useful for
6455 # certain types of sites:
6457 fragile = -block -crunch-all-cookies -filter -fast-redirects -hide-referrer
6458 shop = -crunch-all-cookies allow-popups
6460 # Allow ads for selected useful free sites:
6462 allow-ads = -block -filter{banners-by-size} -filter{banners-by-link}
6464 # Alias for specific file types that are text, but might have conflicting
6465 # MIME types. We want the browser to force these to be text documents.
6466 handle-as-text = -<link linkend="FILTER">filter</link> +-<link linkend="content-type-overwrite">content-type-overwrite{text/plain}</link> +-<link linkend="FORCE-TEXT-MODE">force-text-mode</link> -<link linkend="HIDE-CONTENT-DISPOSITION">hide-content-disposition</link></screen>
6471 Say you have accounts on some sites that you visit regularly, and
6472 you don't want to have to log in manually each time. So you'd like
6473 to allow persistent cookies for these sites. The
6474 <literal>allow-all-cookies</literal> alias defined above does exactly
6475 that, i.e. it disables crunching of cookies in any direction, and the
6476 processing of cookies to make them only temporary.
6481 { allow-all-cookies }
6485 .redhat.com</screen>
6489 Your bank is allergic to some filter, but you don't know which, so you disable them all:
6494 { -<link linkend="FILTER">filter</link> }
6495 .your-home-banking-site.com</screen>
6499 Some file types you may not want to filter for various reasons:
6504 # Technical documentation is likely to contain strings that might
6505 # erroneously get altered by the JavaScript-oriented filters:
6510 # And this stupid host sends streaming video with a wrong MIME type,
6511 # so that Privoxy thinks it is getting HTML and starts filtering:
6513 stupid-server.example.com/</screen>
6517 Example of a simple <link linkend="BLOCK">block</link> action. Say you've
6518 seen an ad on your favourite page on example.com that you want to get rid of.
6519 You have right-clicked the image, selected <quote>copy image location</quote>
6520 and pasted the URL below while removing the leading http://, into a
6521 <literal>{ +block{} }</literal> section. Note that <literal>{ +handle-as-image
6522 }</literal> need not be specified, since all URLs ending in
6523 <literal>.gif</literal> will be tagged as images by the general rules as set
6524 in default.action anyway:
6529 { +<link linkend="BLOCK">block</link>{Nasty ads.} }
6530 www.example.com/nasty-ads/sponsor\.gif
6531 another.example.net/more/junk/here/</screen>
6535 The URLs of dynamically generated banners, especially from large banner
6536 farms, often don't use the well-known image file name extensions, which
6537 makes it impossible for <application>Privoxy</application> to guess
6538 the file type just by looking at the URL.
6539 You can use the <literal>+block-as-image</literal> alias defined above for
6541 Note that objects which match this rule but then turn out NOT to be an
6542 image are typically rendered as a <quote>broken image</quote> icon by the
6543 browser. Use cautiously.
6552 ar.atwola.com/</screen>
6556 Now you noticed that the default configuration breaks Forbes Magazine,
6557 but you were too lazy to find out which action is the culprit, and you
6558 were again too lazy to give <link linkend="contact">feedback</link>, so
6559 you just used the <literal>fragile</literal> alias on the site, and
6560 -- <emphasis>whoa!</emphasis> -- it worked. The <literal>fragile</literal>
6561 aliases disables those actions that are most likely to break a site. Also,
6562 good for testing purposes to see if it is <application>Privoxy</application>
6563 that is causing the problem or not. We later find other regular sites
6564 that misbehave, and add those to our personalized list of troublemakers:
6572 .mybank.com</screen>
6576 You like the <quote>fun</quote> text replacements in <filename>default.filter</filename>,
6577 but it is disabled in the distributed actions file.
6578 So you'd like to turn it on in your private,
6579 update-safe config, once and for all:
6584 { +<link linkend="filter-fun">filter{fun}</link> }
6585 / # For ALL sites!</screen>
6589 Note that the above is not really a good idea: There are exceptions
6590 to the filters in <filename>default.action</filename> for things that
6591 really shouldn't be filtered, like code on CVS->Web interfaces. Since
6592 <filename>user.action</filename> has the last word, these exceptions
6593 won't be valid for the <quote>fun</quote> filtering specified here.
6597 You might also worry about how your favourite free websites are
6598 funded, and find that they rely on displaying banner advertisements
6599 to survive. So you might want to specifically allow banners for those
6600 sites that you feel provide value to you:
6612 Note that <literal>allow-ads</literal> has been aliased to
6613 <literal>-<link linkend="block">block</link></literal>,
6614 <literal>-<link linkend="filter-banners-by-size">filter{banners-by-size}</link></literal>, and
6615 <literal>-<link linkend="filter-banners-by-link">filter{banners-by-link}</link></literal> above.
6619 Invoke another alias here to force an over-ride of the MIME type <literal>
6620 application/x-sh</literal> which typically would open a download type
6621 dialog. In my case, I want to look at the shell script, and then I can save
6622 it should I choose to.
6632 <filename>user.action</filename> is generally the best place to define
6633 exceptions and additions to the default policies of
6634 <filename>default.action</filename>. Some actions are safe to have their
6635 default policies set here though. So let's set a default policy to have a
6636 <quote>blank</quote> image as opposed to the checkerboard pattern for
6637 <emphasis>ALL</emphasis> sites. <quote>/</quote> of course matches all URL
6643 { +<link linkend="set-image-blocker">set-image-blocker{blank}</link> }
6644 / # ALL sites</screen>
6650 <!-- ~ End section ~ -->
6654 <!-- ~ End section ~ -->
6656 <!-- ~~~~~~~~ New section Header ~~~~~~~~~ -->
6658 <sect1 id="filter-file">
6659 <title>Filter Files</title>
6662 On-the-fly text substitutions need
6663 to be defined in a <quote>filter file</quote>. Once defined, they
6664 can then be invoked as an <quote>action</quote>.
6668 &my-app; supports three different pcrs-based filter actions:
6669 <literal><link linkend="filter">filter</link></literal> to
6670 rewrite the content that is send to the client,
6671 <literal><link linkend="client-header-filter">client-header-filter</link></literal>
6672 to rewrite headers that are send by the client, and
6673 <literal><link linkend="server-header-filter">server-header-filter</link></literal>
6674 to rewrite headers that are send by the server.
6678 &my-app; also supports two tagger actions:
6679 <literal><link linkend="client-header-tagger">client-header-tagger</link></literal>
6681 <literal><link linkend="server-header-tagger">server-header-tagger</link></literal>.
6682 Taggers and filters use the same syntax in the filter files, the difference
6683 is that taggers don't modify the text they are filtering, but use a rewritten
6684 version of the filtered text as tag. The tags can then be used to change the
6685 applying actions through sections with <link linkend="tag-pattern">tag-patterns</link>.
6689 Finally &my-app; supports the
6690 <literal><link linkend="external-filter">external-filter</link></literal> action
6691 to enable <literal><link linkend="external-filter-syntax">external filters</link></literal>
6692 written in proper programming languages.
6697 Multiple filter files can be defined through the <literal> <link
6698 linkend="filterfile">filterfile</link></literal> config directive. The filters
6699 as supplied by the developers are located in
6700 <filename>default.filter</filename>. It is recommended that any locally
6701 defined or modified filters go in a separately defined file such as
6702 <filename>user.filter</filename>.
6706 Common tasks for content filters are to eliminate common annoyances in
6707 HTML and JavaScript, such as pop-up windows,
6708 exit consoles, crippled windows without navigation tools, the
6709 infamous <BLINK> tag etc, to suppress images with certain
6710 width and height attributes (standard banner sizes or web-bugs),
6711 or just to have fun.
6715 Enabled content filters are applied to any content whose
6716 <quote>Content Type</quote> header is recognised as a sign
6717 of text-based content, with the exception of <literal>text/plain</literal>.
6718 Use the <link linkend="FORCE-TEXT-MODE">force-text-mode</link> action
6719 to also filter other content.
6723 Substitutions are made at the source level, so if you want to <quote>roll
6724 your own</quote> filters, you should first be familiar with HTML syntax,
6725 and, of course, regular expressions.
6729 Just like the <link linkend="actions-file">actions files</link>, the
6730 filter file is organized in sections, which are called <emphasis>filters</emphasis>
6731 here. Each filter consists of a heading line, that starts with one of the
6732 <emphasis>keywords</emphasis> <literal>FILTER:</literal>,
6733 <literal>CLIENT-HEADER-FILTER:</literal> or <literal>SERVER-HEADER-FILTER:</literal>
6734 followed by the filter's <emphasis>name</emphasis>, and a short (one line)
6735 <emphasis>description</emphasis> of what it does. Below that line
6736 come the <emphasis>jobs</emphasis>, i.e. lines that define the actual
6737 text substitutions. By convention, the name of a filter
6738 should describe what the filter <emphasis>eliminates</emphasis>. The
6739 comment is used in the <ulink url="http://config.privoxy.org/">web-based
6740 user interface</ulink>.
6744 Once a filter called <replaceable>name</replaceable> has been defined
6745 in the filter file, it can be invoked by using an action of the form
6746 +<literal><link linkend="filter">filter</link>{<replaceable>name</replaceable>}</literal>
6747 in any <link linkend="actions-file">actions file</link>.
6751 Filter definitions start with a header line that contains the filter
6752 type, the filter name and the filter description.
6753 A content filter header line for a filter called <quote>foo</quote> could look
6758 <screen>FILTER: foo Replace all "foo" with "bar"</screen>
6762 Below that line, and up to the next header line, come the jobs that
6763 define what text replacements the filter executes. They are specified
6764 in a syntax that imitates <ulink url="http://www.perl.org/">Perl</ulink>'s
6765 <literal>s///</literal> operator. If you are familiar with Perl, you
6766 will find this to be quite intuitive, and may want to look at the
6767 PCRS documentation for the subtle differences to Perl behaviour.
6771 Most notably, the non-standard option letter <literal>U</literal> is supported,
6772 which turns the default to ungreedy matching (add <literal>?</literal> to
6773 quantifiers to turn them greedy again).
6777 The non-standard option letter <literal>D</literal> (dynamic) allows
6778 to use the variables $host, $origin (the IP address the request came from),
6779 $path and $url. They will be replaced with the value they refer to before
6780 the filter is executed.
6784 Note that '$' is a bad choice for a delimiter in a dynamic filter as you
6785 might end up with unintended variables if you use a variable name
6786 directly after the delimiter. Variables will be resolved without
6787 escaping anything, therefore you also have to be careful not to chose
6788 delimiters that appear in the replacement text. For example '<' should
6789 be save, while '?' will sooner or later cause conflicts with $url.
6793 The non-standard option letter <literal>T</literal> (trivial) prevents
6794 parsing for backreferences in the substitute. Use it if you want to include
6795 text like '$&' in your substitute without quoting.
6800 <ulink url="http://en.wikipedia.org/wiki/Regular_expressions"><quote>Regular
6801 Expressions</quote></ulink>, you might want to take a look at
6802 the <link linkend="regex">Appendix on regular expressions</link>, and
6803 see the <ulink url="http://perldoc.perl.org/perlre.html">Perl
6805 <ulink url="http://perldoc.perl.org/perlop.html">the
6806 <literal>s///</literal> operator's syntax</ulink> and <ulink
6807 url="http://perldoc.perl.org/perlre.html">Perl-style regular
6808 expressions</ulink> in general.
6809 The below examples might also help to get you started.
6813 <!-- ~~~~~~~~ New section Header ~~~~~~~~~ -->
6815 <sect2 id="filter-file-tut"><title>Filter File Tutorial</title>
6817 Now, let's complete our <quote>foo</quote> content filter. We have already defined
6818 the heading, but the jobs are still missing. Since all it does is to replace
6819 <quote>foo</quote> with <quote>bar</quote>, there is only one (trivial) job
6824 <screen>s/foo/bar/</screen>
6828 But wait! Didn't the comment say that <emphasis>all</emphasis> occurrences
6829 of <quote>foo</quote> should be replaced? Our current job will only take
6830 care of the first <quote>foo</quote> on each page. For global substitution,
6831 we'll need to add the <literal>g</literal> option:
6835 <screen>s/foo/bar/g</screen>
6839 Our complete filter now looks like this:
6842 <screen>FILTER: foo Replace all "foo" with "bar"
6843 s/foo/bar/g</screen>
6847 Let's look at some real filters for more interesting examples. Here you see
6848 a filter that protects against some common annoyances that arise from JavaScript
6849 abuse. Let's look at its jobs one after the other:
6855 FILTER: js-annoyances Get rid of particularly annoying JavaScript abuse
6857 # Get rid of JavaScript referrer tracking. Test page: http://www.randomoddness.com/untitled.htm
6859 s|(<script.*)document\.referrer(.*</script>)|$1"Not Your Business!"$2|Usg</screen>
6863 Following the header line and a comment, you see the job. Note that it uses
6864 <literal>|</literal> as the delimiter instead of <literal>/</literal>, because
6865 the pattern contains a forward slash, which would otherwise have to be escaped
6866 by a backslash (<literal>\</literal>).
6870 Now, let's examine the pattern: it starts with the text <literal><script.*</literal>
6871 enclosed in parentheses. Since the dot matches any character, and <literal>*</literal>
6872 means: <quote>Match an arbitrary number of the element left of myself</quote>, this
6873 matches <quote><script</quote>, followed by <emphasis>any</emphasis> text, i.e.
6874 it matches the whole page, from the start of the first <script> tag.
6878 That's more than we want, but the pattern continues: <literal>document\.referrer</literal>
6879 matches only the exact string <quote>document.referrer</quote>. The dot needed to
6880 be <emphasis>escaped</emphasis>, i.e. preceded by a backslash, to take away its
6881 special meaning as a joker, and make it just a regular dot. So far, the meaning is:
6882 Match from the start of the first <script> tag in a the page, up to, and including,
6883 the text <quote>document.referrer</quote>, if <emphasis>both</emphasis> are present
6884 in the page (and appear in that order).
6888 But there's still more pattern to go. The next element, again enclosed in parentheses,
6889 is <literal>.*</script></literal>. You already know what <literal>.*</literal>
6890 means, so the whole pattern translates to: Match from the start of the first <script>
6891 tag in a page to the end of the last <script> tag, provided that the text
6892 <quote>document.referrer</quote> appears somewhere in between.
6896 This is still not the whole story, since we have ignored the options and the parentheses:
6897 The portions of the page matched by sub-patterns that are enclosed in parentheses, will be
6898 remembered and be available through the variables <literal>$1, $2, ...</literal> in
6899 the substitute. The <literal>U</literal> option switches to ungreedy matching, which means
6900 that the first <literal>.*</literal> in the pattern will only <quote>eat up</quote> all
6901 text in between <quote><script</quote> and the <emphasis>first</emphasis> occurrence
6902 of <quote>document.referrer</quote>, and that the second <literal>.*</literal> will
6903 only span the text up to the <emphasis>first</emphasis> <quote></script></quote>
6904 tag. Furthermore, the <literal>s</literal> option says that the match may span
6905 multiple lines in the page, and the <literal>g</literal> option again means that the
6906 substitution is global.
6910 So, to summarize, the pattern means: Match all scripts that contain the text
6911 <quote>document.referrer</quote>. Remember the parts of the script from
6912 (and including) the start tag up to (and excluding) the string
6913 <quote>document.referrer</quote> as <literal>$1</literal>, and the part following
6914 that string, up to and including the closing tag, as <literal>$2</literal>.
6918 Now the pattern is deciphered, but wasn't this about substituting things? So
6919 lets look at the substitute: <literal>$1"Not Your Business!"$2</literal> is
6920 easy to read: The text remembered as <literal>$1</literal>, followed by
6921 <literal>"Not Your Business!"</literal> (<emphasis>including</emphasis>
6922 the quotation marks!), followed by the text remembered as <literal>$2</literal>.
6923 This produces an exact copy of the original string, with the middle part
6924 (the <quote>document.referrer</quote>) replaced by <literal>"Not Your
6925 Business!"</literal>.
6929 The whole job now reads: Replace <quote>document.referrer</quote> by
6930 <literal>"Not Your Business!"</literal> wherever it appears inside a
6931 <script> tag. Note that this job won't break JavaScript syntax,
6932 since both the original and the replacement are syntactically valid
6933 string objects. The script just won't have access to the referrer
6934 information anymore.
6938 We'll show you two other jobs from the JavaScript taming department, but
6939 this time only point out the constructs of special interest:
6944 # The status bar is for displaying link targets, not pointless blahblah
6946 s/window\.status\s*=\s*(['"]).*?\1/dUmMy=1/ig</screen>
6950 <literal>\s</literal> stands for whitespace characters (space, tab, newline,
6951 carriage return, form feed), so that <literal>\s*</literal> means: <quote>zero
6952 or more whitespace</quote>. The <literal>?</literal> in <literal>.*?</literal>
6953 makes this matching of arbitrary text ungreedy. (Note that the <literal>U</literal>
6954 option is not set). The <literal>['"]</literal> construct means: <quote>a single
6955 <emphasis>or</emphasis> a double quote</quote>. Finally, <literal>\1</literal> is
6956 a back-reference to the first parenthesis just like <literal>$1</literal> above,
6957 with the difference that in the <emphasis>pattern</emphasis>, a backslash indicates
6958 a back-reference, whereas in the <emphasis>substitute</emphasis>, it's the dollar.
6962 So what does this job do? It replaces assignments of single- or double-quoted
6963 strings to the <quote>window.status</quote> object with a dummy assignment
6964 (using a variable name that is hopefully odd enough not to conflict with
6965 real variables in scripts). Thus, it catches many cases where e.g. pointless
6966 descriptions are displayed in the status bar instead of the link target when
6967 you move your mouse over links.
6972 # Kill OnUnload popups. Yummy. Test: http://www.zdnet.com/zdsubs/yahoo/tree/yfs.html
6974 s/(<body [^>]*)onunload(.*>)/$1never$2/iU</screen>
6979 <ulink url="http://www.w3.org/TR/2000/REC-DOM-Level-2-Events-20001113/events.html#Events-eventgroupings-htmlevents">OnUnload
6980 event binding</ulink> in the HTML DOM was a <emphasis>CRIME</emphasis>.
6981 When I close a browser window, I want it to close and die. Basta.
6982 This job replaces the <quote>onunload</quote> attribute in
6983 <quote><body></quote> tags with the dummy word <literal>never</literal>.
6984 Note that the <literal>i</literal> option makes the pattern matching
6985 case-insensitive. Also note that ungreedy matching alone doesn't always guarantee
6986 a minimal match: In the first parenthesis, we had to use <literal>[^>]*</literal>
6987 instead of <literal>.*</literal> to prevent the match from exceeding the
6988 <body> tag if it doesn't contain <quote>OnUnload</quote>, but the page's
6993 The last example is from the fun department:
6998 FILTER: fun Fun text replacements
7000 # Spice the daily news:
7002 s/microsoft(?!\.com)/MicroSuck/ig</screen>
7006 Note the <literal>(?!\.com)</literal> part (a so-called negative lookahead)
7007 in the job's pattern, which means: Don't match, if the string
7008 <quote>.com</quote> appears directly following <quote>microsoft</quote>
7009 in the page. This prevents links to microsoft.com from being trashed, while
7010 still replacing the word everywhere else.
7015 # Buzzword Bingo (example for extended regex syntax)
7017 s* industry[ -]leading \
7019 | customer[ -]focused \
7020 | market[ -]driven \
7021 | award[ -]winning # Comments are OK, too! \
7022 | high[ -]performance \
7023 | solutions[ -]based \
7027 *<font color="red"><b>BINGO!</b></font> \
7032 The <literal>x</literal> option in this job turns on extended syntax, and allows for
7033 e.g. the liberal use of (non-interpreted!) whitespace for nicer formatting.
7041 <!-- ~~~~~~~~ New section Header ~~~~~~~~~ -->
7043 <sect2 id="predefined-filters"><title>The Pre-defined Filters</title>
7047 Note each filter is also listed in the +filter action section above. Please
7048 keep these listings in sync.
7053 The distribution <filename>default.filter</filename> file contains a selection of
7054 pre-defined filters for your convenience:
7059 <term><emphasis>js-annoyances</emphasis></term>
7062 The purpose of this filter is to get rid of particularly annoying JavaScript abuse.
7067 replaces JavaScript references to the browser's referrer information
7068 with the string "Not Your Business!". This compliments the <literal><link
7069 linkend="hide-referrer">hide-referrer</link></literal> action on the content level.
7074 removes the bindings to the DOM's
7075 <ulink url="http://www.w3.org/TR/2000/REC-DOM-Level-2-Events-20001113/events.html#Events-eventgroupings-htmlevents">unload
7076 event</ulink> which we feel has no right to exist and is responsible for most <quote>exit consoles</quote>, i.e.
7077 nasty windows that pop up when you close another one.
7082 removes code that causes new windows to be opened with undesired properties, such as being
7083 full-screen, non-resizeable, without location, status or menu bar etc.
7089 Use with caution. This is an aggressive filter, and can break sites that
7090 rely heavily on JavaScript.
7096 <term><emphasis>js-events</emphasis></term>
7099 This is a very radical measure. It removes virtually all JavaScript event bindings, which
7100 means that scripts can not react to user actions such as mouse movements or clicks, window
7101 resizing etc, anymore. Use with caution!
7104 We <emphasis>strongly discourage</emphasis> using this filter as a default since it breaks
7105 many legitimate scripts. It is meant for use only on extra-nasty sites (should you really
7112 <term><emphasis>html-annoyances</emphasis></term>
7115 This filter will undo many common instances of HTML based abuse.
7118 The <literal>BLINK</literal> and <literal>MARQUEE</literal> tags
7119 are neutralized (yeah baby!), and browser windows will be created as
7120 resizeable (as of course they should be!), and will have location,
7121 scroll and menu bars -- even if specified otherwise.
7127 <term><emphasis>content-cookies</emphasis></term>
7130 Most cookies are set in the HTTP dialog, where they can be intercepted
7132 <literal><link linkend="crunch-incoming-cookies">crunch-incoming-cookies</link></literal>
7133 and <literal><link linkend="crunch-outgoing-cookies">crunch-outgoing-cookies</link></literal>
7134 actions. But web sites increasingly make use of HTML meta tags and JavaScript
7135 to sneak cookies to the browser on the content level.
7138 This filter disables most HTML and JavaScript code that reads or sets
7139 cookies. It cannot detect all clever uses of these types of code, so it
7140 should not be relied on as an absolute fix. Use it wherever you would also
7141 use the cookie crunch actions.
7147 <term><emphasis>refresh-tags</emphasis></term>
7150 Disable any refresh tags if the interval is greater than nine seconds (so
7151 that redirections done via refresh tags are not destroyed). This is useful
7152 for dial-on-demand setups, or for those who find this HTML feature
7159 <term><emphasis>unsolicited-popups</emphasis></term>
7162 This filter attempts to prevent only <quote>unsolicited</quote> pop-up
7163 windows from opening, yet still allow pop-up windows that the user
7164 has explicitly chosen to open. It was added in version 3.0.1,
7165 as an improvement over earlier such filters.
7168 Technical note: The filter works by redefining the window.open JavaScript
7169 function to a dummy function, <literal>PrivoxyWindowOpen()</literal>,
7170 during the loading and rendering phase of each HTML page access, and
7171 restoring the function afterward.
7174 This is recommended only for browsers that cannot perform this function
7175 reliably themselves. And be aware that some sites require such windows
7176 in order to function normally. Use with caution.
7182 <term><emphasis>all-popups</emphasis></term>
7185 Attempt to prevent <emphasis>all</emphasis> pop-up windows from opening.
7186 Note this should be used with even more discretion than the above, since
7187 it is more likely to break some sites that require pop-ups for normal
7188 usage. Use with caution.
7194 <term><emphasis>img-reorder</emphasis></term>
7197 This is a helper filter that has no value if used alone. It makes the
7198 <literal>banners-by-size</literal> and <literal>banners-by-link</literal>
7199 (see below) filters more effective and should be enabled together with them.
7205 <term><emphasis>banners-by-size</emphasis></term>
7208 This filter removes image tags purely based on what size they are. Fortunately
7209 for us, many ads and banner images tend to conform to certain standardized
7210 sizes, which makes this filter quite effective for ad stripping purposes.
7213 Occasionally this filter will cause false positives on images that are not ads,
7214 but just happen to be of one of the standard banner sizes.
7217 Recommended only for those who require extreme ad blocking. The default
7218 block rules should catch 95+% of all ads <emphasis>without</emphasis> this filter enabled.
7224 <term><emphasis>banners-by-link</emphasis></term>
7227 This is an experimental filter that attempts to kill any banners if
7228 their URLs seem to point to known or suspected click trackers. It is currently
7229 not of much value and is not recommended for use by default.
7235 <term><emphasis>webbugs</emphasis></term>
7238 Webbugs are small, invisible images (technically 1X1 GIF images), that
7239 are used to track users across websites, and collect information on them.
7240 As an HTML page is loaded by the browser, an embedded image tag causes the
7241 browser to contact a third-party site, disclosing the tracking information
7242 through the requested URL and/or cookies for that third-party domain, without
7243 the user ever becoming aware of the interaction with the third-party site.
7244 HTML-ized spam also uses a similar technique to verify email addresses.
7247 This filter removes the HTML code that loads such <quote>webbugs</quote>.
7253 <term><emphasis>tiny-textforms</emphasis></term>
7256 A rather special-purpose filter that can be used to enlarge textareas (those
7257 multi-line text boxes in web forms) and turn off hard word wrap in them.
7258 It was written for the sourceforge.net tracker system where such boxes are
7259 a nuisance, but it can be handy on other sites, too.
7262 It is not recommended to use this filter as a default.
7268 <term><emphasis>jumping-windows</emphasis></term>
7271 Many consider windows that move, or resize themselves to be abusive. This filter
7272 neutralizes the related JavaScript code. Note that some sites might not display
7273 or behave as intended when using this filter. Use with caution.
7279 <term><emphasis>frameset-borders</emphasis></term>
7282 Some web designers seem to assume that everyone in the world will view their
7283 web sites using the same browser brand and version, screen resolution etc,
7284 because only that assumption could explain why they'd use static frame sizes,
7285 yet prevent their frames from being resized by the user, should they be too
7286 small to show their whole content.
7289 This filter removes the related HTML code. It should only be applied to sites
7296 <term><emphasis>demoronizer</emphasis></term>
7299 Many Microsoft products that generate HTML use non-standard extensions (read:
7300 violations) of the ISO 8859-1 aka Latin-1 character set. This can cause those
7301 HTML documents to display with errors on standard-compliant platforms.
7304 This filter translates the MS-only characters into Latin-1 equivalents.
7305 It is not necessary when using MS products, and will cause corruption of
7306 all documents that use 8-bit character sets other than Latin-1. It's mostly
7307 worthwhile for Europeans on non-MS platforms, if weird garbage characters
7308 sometimes appear on some pages, or user agents that don't correct for this on
7311 My version of Mozilla (ancient) shows litte square boxes for quote
7312 characters, and apostrophes on moronized pages. So many pages have this, I
7313 can read them fine now. HB 08/27/06
7320 <term><emphasis>shockwave-flash</emphasis></term>
7323 A filter for shockwave haters. As the name suggests, this filter strips code
7324 out of web pages that is used to embed shockwave flash objects.
7332 <term><emphasis>quicktime-kioskmode</emphasis></term>
7335 Change HTML code that embeds Quicktime objects so that kioskmode, which
7336 prevents saving, is disabled.
7342 <term><emphasis>fun</emphasis></term>
7345 Text replacements for subversive browsing fun. Make fun of your favorite
7346 Monopolist or play buzzword bingo.
7352 <term><emphasis>crude-parental</emphasis></term>
7355 A demonstration-only filter that shows how <application>Privoxy</application>
7356 can be used to delete web content on a keyword basis.
7362 <term><emphasis>ie-exploits</emphasis></term>
7365 An experimental collection of text replacements to disable malicious HTML and JavaScript
7366 code that exploits known security holes in Internet Explorer.
7369 Presently, it only protects against Nimda and a cross-site scripting bug, and
7370 would need active maintenance to provide more substantial protection.
7376 <term><emphasis>site-specifics</emphasis></term>
7379 Some web sites have very specific problems, the cure for which doesn't apply
7380 anywhere else, or could even cause damage on other sites.
7383 This is a collection of such site-specific cures which should only be applied
7384 to the sites they were intended for, which is what the supplied
7385 <filename>default.action</filename> file does. Users shouldn't need to change
7386 anything regarding this filter.
7392 <term><emphasis>google</emphasis></term>
7395 A CSS based block for Google text ads. Also removes a width limitation
7396 and the toolbar advertisement.
7402 <term><emphasis>yahoo</emphasis></term>
7405 Another CSS based block, this time for Yahoo text ads. And removes
7406 a width limitation as well.
7412 <term><emphasis>msn</emphasis></term>
7415 Another CSS based block, this time for MSN text ads. And removes
7416 tracking URLs, as well as a width limitation.
7422 <term><emphasis>blogspot</emphasis></term>
7425 Cleans up some Blogspot blogs. Read the fine print before using this one!
7428 This filter also intentionally removes some navigation stuff and sets the
7429 page width to 100%. As a result, some rounded <quote>corners</quote> would
7430 appear to early or not at all and as fixing this would require a browser
7431 that understands background-size (CSS3), they are removed instead.
7437 <term><emphasis>xml-to-html</emphasis></term>
7440 Server-header filter to change the Content-Type from xml to html.
7446 <term><emphasis>html-to-xml</emphasis></term>
7449 Server-header filter to change the Content-Type from html to xml.
7455 <term><emphasis>no-ping</emphasis></term>
7458 Removes the non-standard <literal>ping</literal> attribute from
7459 anchor and area HTML tags.
7465 <term><emphasis>hide-tor-exit-notation</emphasis></term>
7468 Client-header filter to remove the <command>Tor</command> exit node notation
7469 found in Host and Referer headers.
7472 If &my-app; and <command>Tor</command> are chained and &my-app;
7473 is configured to use socks4a, one can use <quote>http://www.example.org.foobar.exit/</quote>
7474 to access the host <quote>www.example.org</quote> through the
7475 <command>Tor</command> exit node <quote>foobar</quote>.
7478 As the HTTP client isn't aware of this notation, it treats the
7479 whole string <quote>www.example.org.foobar.exit</quote> as host and uses it
7480 for the <quote>Host</quote> and <quote>Referer</quote> headers. From the
7481 server's point of view the resulting headers are invalid and can cause problems.
7484 An invalid <quote>Referer</quote> header can trigger <quote>hot-linking</quote>
7485 protections, an invalid <quote>Host</quote> header will make it impossible for
7486 the server to find the right vhost (several domains hosted on the same IP address).
7489 This client-header filter removes the <quote>foo.exit</quote> part in those headers
7490 to prevent the mentioned problems. Note that it only modifies
7491 the HTTP headers, it doesn't make it impossible for the server
7492 to detect your <command>Tor</command> exit node based on the IP address
7493 the request is coming from.
7500 <term><emphasis> </emphasis></term>
7513 <!-- ~~~~~~~~ New section Header ~~~~~~~~~ -->
7514 <sect2 id="external-filter-syntax"><title>External filter syntax</title>
7516 External filters are scripts or programs that can modify the content in
7517 case common <literal><link linkend="filter">filters</link></literal>
7518 aren't powerful enough.
7521 External filters can be written in any language the platform &my-app; runs
7525 They are controlled with the
7526 <literal><link linkend="external-filter">external-filter</link></literal> action
7527 and have to be defined in the <literal><link linkend="filterfile">filterfile</link></literal>
7531 The header looks like any other filter, but instead of pcrs jobs, external
7532 filters contain a single job which can be a program or a shell script (which
7533 may call other scripts or programs).
7536 External filters read the content from STDIN and write the rewritten
7537 content to STDOUT. The environment variables PRIVOXY_URL, PRIVOXY_PATH,
7538 PRIVOXY_HOST, PRIVOXY_ORIGIN can be used to get some details about the
7542 &my-app; will temporary store the content to filter in the
7543 <literal><link linkend="temporary-directory">temporary-directory</link></literal>.
7547 EXTERNAL-FILTER: cat Pointless example filter that doesn't actually modify the content
7550 # Incorrect reimplementation of the filter above in POSIX shell.
7552 # Note that it's a single job that spans multiple lines, the line
7553 # breaks are not passed to the shell, thus the semicolons are required.
7555 # If the script isn't trivial, it is recommended to put it into an external file.
7557 # In general, writing external filters entirely in POSIX shell is not
7558 # considered a good idea.
7559 EXTERNAL-FILTER: cat2 Pointless example filter that despite its name may actually modify the content
7565 EXTERNAL-FILTER: rotate-image Rotate an image by 180 degree. Test filter with limited value.
7566 /usr/local/bin/convert - -rotate 180 -
7568 EXTERNAL-FILTER: citation-needed Adds a "[citation needed]" tag to an image. The coordinates may need adjustment.
7569 /usr/local/bin/convert - -pointsize 16 -fill white -annotate +17+418 "[citation needed]" -
7575 Currently external filters are executed with &my-app;'s privileges!
7576 Only use external filters you understand and trust.
7580 External filters are experimental and the syntax may change in the future.
7586 <!-- ~ End section ~ -->
7590 <!-- ~~~~~ New section ~~~~~ -->
7592 <sect1 id="templates">
7593 <title>Privoxy's Template Files</title>
7595 All <application>Privoxy</application> built-in pages, i.e. error pages such as the
7596 <ulink url="http://show-the-404-error.page"><quote>404 - No Such Domain</quote>
7597 error page</ulink>, the <ulink
7598 url="http://ads.bannerserver.example.com/nasty-ads/sponsor.html"><quote>BLOCKED</quote>
7600 and all pages of its <ulink url="http://config.privoxy.org/">web-based
7601 user interface</ulink>, are generated from <emphasis>templates</emphasis>.
7602 (<application>Privoxy</application> must be running for the above links to work as
7607 These templates are stored in a subdirectory of the <link linkend="confdir">configuration
7608 directory</link> called <filename>templates</filename>. On Unixish platforms,
7610 <ulink url="file:///etc/privoxy/templates/"><filename>/etc/privoxy/templates/</filename></ulink>.
7614 The templates are basically normal HTML files, but with place-holders (called symbols
7615 or exports), which <application>Privoxy</application> fills at run time. It
7616 is possible to edit the templates with a normal text editor, should you want
7617 to customize them. (<emphasis>Not recommended for the casual
7618 user</emphasis>). Should you create your own custom templates, you should use
7619 the <filename>config</filename> setting <link linkend="templdir">templdir</link>
7620 to specify an alternate location, so your templates do not get overwritten
7624 Note that just like in configuration files, lines starting
7625 with <literal>#</literal> are ignored when the templates are filled in.
7629 The place-holders are of the form <literal>@name@</literal>, and you will
7630 find a list of available symbols, which vary from template to template,
7631 in the comments at the start of each file. Note that these comments are not
7632 always accurate, and that it's probably best to look at the existing HTML
7633 code to find out which symbols are supported and what they are filled in with.
7637 A special application of this substitution mechanism is to make whole
7638 blocks of HTML code disappear when a specific symbol is set. We use this
7639 for many purposes, one of them being to include the beta warning in all
7640 our user interface (CGI) pages when <application>Privoxy</application>
7641 is in an alpha or beta development stage:
7646 <!-- @if-unstable-start -->
7648 ... beta warning HTML code goes here ...
7650 <!-- if-unstable-end@ --></screen>
7654 If the "unstable" symbol is set, everything in between and including
7655 <literal>@if-unstable-start</literal> and <literal>if-unstable-end@</literal>
7656 will disappear, leaving nothing but an empty comment:
7660 <screen><!-- --></screen>
7664 There's also an if-then-else construct and an <literal>#include</literal>
7665 mechanism, but you'll sure find out if you are inclined to edit the
7670 All templates refer to a style located at
7671 <ulink url="http://config.privoxy.org/send-stylesheet"><literal>http://config.privoxy.org/send-stylesheet</literal></ulink>.
7672 This is, of course, locally served by <application>Privoxy</application>
7673 and the source for it can be found and edited in the
7674 <filename>cgi-style.css</filename> template.
7679 <!-- ~ End section ~ -->
7683 <!-- ~~~~~ New section ~~~~~ -->
7685 <sect1 id="contact"><title>Contacting the Developers, Bug Reporting and Feature
7688 <!-- Include contacting.sgml boilerplate: -->
7690 <!-- end boilerplate -->
7694 <!-- ~ End section ~ -->
7697 <!-- ~~~~~ New section ~~~~~ -->
7698 <sect1 id="copyright"><title>Privoxy Copyright, License and History</title>
7700 <!-- Include copyright.sgml: -->
7702 <!-- end copyright -->
7705 <application>Privoxy</application> is free software; you can
7706 redistribute it and/or modify it under the terms of the
7707 <citetitle>GNU General Public License</citetitle>, version 2,
7708 as published by the Free Software Foundation and included in
7712 <!-- ~~~~~ New section ~~~~~ -->
7713 <sect2 id="license"><title>License</title>
7715 <screen><![ RCDATA [ &GPLv2; ]]></screen>
7719 <!-- ~ End section ~ -->
7722 <!-- ~~~~~ New section ~~~~~ -->
7724 <sect2 id="history"><title>History</title>
7725 <!-- Include history.sgml: -->
7727 <!-- end history -->
7730 <sect2 id="authors"><title>Authors</title>
7731 <!-- Include p-authors.sgml: -->
7733 <!-- end authors -->
7738 <!-- ~ End section ~ -->
7741 <!-- ~~~~~ New section ~~~~~ -->
7742 <sect1 id="seealso"><title>See Also</title>
7743 <!-- Include seealso.sgml: -->
7745 <!-- end seealso -->
7750 <!-- ~~~~~ New section ~~~~~ -->
7751 <sect1 id="appendix"><title>Appendix</title>
7754 <!-- ~~~~~ New section ~~~~~ -->
7756 <title>Regular Expressions</title>
7758 <application>Privoxy</application> uses Perl-style <quote>regular
7759 expressions</quote> in its <link linkend="actions-file">actions
7760 files</link> and <link linkend="filter-file">filter file</link>,
7761 through the <ulink url="http://www.pcre.org/">PCRE</ulink> and
7764 <ulink url="http://www.oesterhelt.org/pcrs/">PCRS</ulink> libraries.
7766 <application>PCRS</application> libraries.
7770 If you are reading this, you probably don't understand what <quote>regular
7771 expressions</quote> are, or what they can do. So this will be a very brief
7772 introduction only. A full explanation would require a <ulink
7773 url="http://www.oreilly.com/catalog/regex/">book</ulink> ;-)
7777 Regular expressions provide a language to describe patterns that can be
7778 run against strings of characters (letter, numbers, etc), to see if they
7779 match the string or not. The patterns are themselves (sometimes complex)
7780 strings of literal characters, combined with wild-cards, and other special
7781 characters, called meta-characters. The <quote>meta-characters</quote> have
7782 special meanings and are used to build complex patterns to be matched against.
7783 Perl Compatible Regular Expressions are an especially convenient
7784 <quote>dialect</quote> of the regular expression language.
7788 To make a simple analogy, we do something similar when we use wild-card
7789 characters when listing files with the <command>dir</command> command in DOS.
7790 <literal>*.*</literal> matches all filenames. The <quote>special</quote>
7791 character here is the asterisk which matches any and all characters. We can be
7792 more specific and use <literal>?</literal> to match just individual
7793 characters. So <quote>dir file?.text</quote> would match
7794 <quote>file1.txt</quote>, <quote>file2.txt</quote>, etc. We are pattern
7795 matching, using a similar technique to <quote>regular expressions</quote>!
7799 Regular expressions do essentially the same thing, but are much, much more
7800 powerful. There are many more <quote>special characters</quote> and ways of
7801 building complex patterns however. Let's look at a few of the common ones,
7802 and then some examples:
7807 <emphasis>.</emphasis> - Matches any single character, e.g. <quote>a</quote>,
7808 <quote>A</quote>, <quote>4</quote>, <quote>:</quote>, or <quote>@</quote>.
7810 </simplelist></para>
7814 <emphasis>?</emphasis> - The preceding character or expression is matched ZERO or ONE
7817 </simplelist></para>
7821 <emphasis>+</emphasis> - The preceding character or expression is matched ONE or MORE
7824 </simplelist></para>
7828 <emphasis>*</emphasis> - The preceding character or expression is matched ZERO or MORE
7831 </simplelist></para>
7835 <emphasis>\</emphasis> - The <quote>escape</quote> character denotes that
7836 the following character should be taken literally. This is used where one of the
7837 special characters (e.g. <quote>.</quote>) needs to be taken literally and
7838 not as a special meta-character. Example: <quote>example\.com</quote>, makes
7839 sure the period is recognized only as a period (and not expanded to its
7840 meta-character meaning of any single character).
7842 </simplelist></para>
7846 <emphasis>[ ]</emphasis> - Characters enclosed in brackets will be matched if
7847 any of the enclosed characters are encountered. For instance, <quote>[0-9]</quote>
7848 matches any numeric digit (zero through nine). As an example, we can combine
7849 this with <quote>+</quote> to match any digit one of more times: <quote>[0-9]+</quote>.
7851 </simplelist></para>
7855 <emphasis>( )</emphasis> - parentheses are used to group a sub-expression,
7856 or multiple sub-expressions.
7858 </simplelist></para>
7862 <emphasis>|</emphasis> - The <quote>bar</quote> character works like an
7863 <quote>or</quote> conditional statement. A match is successful if the
7864 sub-expression on either side of <quote>|</quote> matches. As an example:
7865 <quote>/(this|that) example/</quote> uses grouping and the bar character
7866 and would match either <quote>this example</quote> or <quote>that
7867 example</quote>, and nothing else.
7869 </simplelist></para>
7872 These are just some of the ones you are likely to use when matching URLs with
7873 <application>Privoxy</application>, and is a long way from a definitive
7874 list. This is enough to get us started with a few simple examples which may
7875 be more illuminating:
7879 <emphasis><literal>/.*/banners/.*</literal></emphasis> - A simple example
7880 that uses the common combination of <quote>.</quote> and <quote>*</quote> to
7881 denote any character, zero or more times. In other words, any string at all.
7882 So we start with a literal forward slash, then our regular expression pattern
7883 (<quote>.*</quote>) another literal forward slash, the string
7884 <quote>banners</quote>, another forward slash, and lastly another
7885 <quote>.*</quote>. We are building
7886 a directory path here. This will match any file with the path that has a
7887 directory named <quote>banners</quote> in it. The <quote>.*</quote> matches
7888 any characters, and this could conceivably be more forward slashes, so it
7889 might expand into a much longer looking path. For example, this could match:
7890 <quote>/eye/hate/spammers/banners/annoy_me_please.gif</quote>, or just
7891 <quote>/banners/annoying.html</quote>, or almost an infinite number of other
7892 possible combinations, just so it has <quote>banners</quote> in the path
7897 And now something a little more complex:
7901 <emphasis><literal>/.*/adv((er)?ts?|ertis(ing|ements?))?/</literal></emphasis> -
7902 We have several literal forward slashes again (<quote>/</quote>), so we are
7903 building another expression that is a file path statement. We have another
7904 <quote>.*</quote>, so we are matching against any conceivable sub-path, just so
7905 it matches our expression. The only true literal that <emphasis>must
7906 match</emphasis> our pattern is <application>adv</application>, together with
7907 the forward slashes. What comes after the <quote>adv</quote> string is the
7912 Remember the <quote>?</quote> means the preceding expression (either a
7913 literal character or anything grouped with <quote>(...)</quote> in this case)
7914 can exist or not, since this means either zero or one match. So
7915 <quote>((er)?ts?|ertis(ing|ements?))</quote> is optional, as are the
7916 individual sub-expressions: <quote>(er)</quote>,
7917 <quote>(ing|ements?)</quote>, and the <quote>s</quote>. The <quote>|</quote>
7918 means <quote>or</quote>. We have two of those. For instance,
7919 <quote>(ing|ements?)</quote>, can expand to match either <quote>ing</quote>
7920 <emphasis>OR</emphasis> <quote>ements?</quote>. What is being done here, is an
7921 attempt at matching as many variations of <quote>advertisement</quote>, and
7922 similar, as possible. So this would expand to match just <quote>adv</quote>,
7923 or <quote>advert</quote>, or <quote>adverts</quote>, or
7924 <quote>advertising</quote>, or <quote>advertisement</quote>, or
7925 <quote>advertisements</quote>. You get the idea. But it would not match
7926 <quote>advertizements</quote> (with a <quote>z</quote>). We could fix that by
7927 changing our regular expression to:
7928 <quote>/.*/adv((er)?ts?|erti(s|z)(ing|ements?))?/</quote>, which would then match
7933 <emphasis><literal>/.*/advert[0-9]+\.(gif|jpe?g)</literal></emphasis> - Again
7934 another path statement with forward slashes. Anything in the square brackets
7935 <quote>[ ]</quote> can be matched. This is using <quote>0-9</quote> as a
7936 shorthand expression to mean any digit one through nine. It is the same as
7937 saying <quote>0123456789</quote>. So any digit matches. The <quote>+</quote>
7938 means one or more of the preceding expression must be included. The preceding
7939 expression here is what is in the square brackets -- in this case, any digit
7940 one through nine. Then, at the end, we have a grouping: <quote>(gif|jpe?g)</quote>.
7941 This includes a <quote>|</quote>, so this needs to match the expression on
7942 either side of that bar character also. A simple <quote>gif</quote> on one side, and the other
7943 side will in turn match either <quote>jpeg</quote> or <quote>jpg</quote>,
7944 since the <quote>?</quote> means the letter <quote>e</quote> is optional and
7945 can be matched once or not at all. So we are building an expression here to
7946 match image GIF or JPEG type image file. It must include the literal
7947 string <quote>advert</quote>, then one or more digits, and a <quote>.</quote>
7948 (which is now a literal, and not a special character, since it is escaped
7949 with <quote>\</quote>), and lastly either <quote>gif</quote>, or
7950 <quote>jpeg</quote>, or <quote>jpg</quote>. Some possible matches would
7951 include: <quote>//advert1.jpg</quote>,
7952 <quote>/nasty/ads/advert1234.gif</quote>,
7953 <quote>/banners/from/hell/advert99.jpg</quote>. It would not match
7954 <quote>advert1.gif</quote> (no leading slash), or
7955 <quote>/adverts232.jpg</quote> (the expression does not include an
7956 <quote>s</quote>), or <quote>/advert1.jsp</quote> (<quote>jsp</quote> is not
7957 in the expression anywhere).
7961 We are barely scratching the surface of regular expressions here so that you
7962 can understand the default <application>Privoxy</application>
7963 configuration files, and maybe use this knowledge to customize your own
7964 installation. There is much, much more that can be done with regular
7965 expressions. Now that you know enough to get started, you can learn more on
7970 More reading on Perl Compatible Regular expressions:
7971 <ulink url="http://perldoc.perl.org/perlre.html">http://perldoc.perl.org/perlre.html</ulink>
7975 For information on regular expression based substitutions and their applications
7976 in filters, please see the <link linkend="filter-file">filter file tutorial</link>
7981 <!-- ~ End section ~ -->
7984 <!-- ~~~~~ New section ~~~~~ -->
7985 <sect2 id="internal-pages">
7986 <title>Privoxy's Internal Pages</title>
7989 Since <application>Privoxy</application> proxies each requested
7990 web page, it is easy for <application>Privoxy</application> to
7991 trap certain special URLs. In this way, we can talk directly to
7992 <application>Privoxy</application>, and see how it is
7993 configured, see how our rules are being applied, change these
7994 rules and other configuration options, and even turn
7995 <application>Privoxy's</application> filtering off, all with
8001 The URLs listed below are the special ones that allow direct access
8002 to <application>Privoxy</application>. Of course,
8003 <application>Privoxy</application> must be running to access these. If
8004 not, you will get a friendly error message. Internet access is not
8017 <ulink url="http://config.privoxy.org/">http://config.privoxy.org/</ulink>
8021 There is a shortcut: <ulink url="http://p.p/">http://p.p/</ulink> (But it
8022 doesn't provide a fall-back to a real page, in case the request is not
8023 sent through <application>Privoxy</application>)
8029 Show information about the current configuration, including viewing and
8030 editing of actions files:
8034 <ulink url="http://config.privoxy.org/show-status">http://config.privoxy.org/show-status</ulink>
8041 Show the source code version numbers:
8045 <ulink url="http://config.privoxy.org/show-version">http://config.privoxy.org/show-version</ulink>
8052 Show the browser's request headers:
8056 <ulink url="http://config.privoxy.org/show-request">http://config.privoxy.org/show-request</ulink>
8063 Show which actions apply to a URL and why:
8067 <ulink url="http://config.privoxy.org/show-url-info">http://config.privoxy.org/show-url-info</ulink>
8074 Toggle Privoxy on or off. This feature can be turned off/on in the main
8075 <filename>config</filename> file. When toggled <quote>off</quote>, <quote>Privoxy</quote>
8076 continues to run, but only as a pass-through proxy, with no actions taking
8081 <ulink url="http://config.privoxy.org/toggle">http://config.privoxy.org/toggle</ulink>
8085 Short cuts. Turn off, then on:
8089 <ulink url="http://config.privoxy.org/toggle?set=disable">http://config.privoxy.org/toggle?set=disable</ulink>
8094 <ulink url="http://config.privoxy.org/toggle?set=enable">http://config.privoxy.org/toggle?set=enable</ulink>
8105 <!-- ~~~~~ New section ~~~~~ -->
8107 <title>Chain of Events</title>
8109 Let's take a quick look at how some of <application>Privoxy's</application>
8110 core features are triggered, and the ensuing sequence of events when a web
8111 page is requested by your browser:
8118 First, your web browser requests a web page. The browser knows to send
8119 the request to <application>Privoxy</application>, which will in turn,
8120 relay the request to the remote web server after passing the following
8126 <application>Privoxy</application> traps any request for its own internal CGI
8127 pages (e.g <ulink url="http://p.p/">http://p.p/</ulink>) and sends the CGI page back to the browser.
8132 Next, <application>Privoxy</application> checks to see if the URL
8134 linkend="BLOCK"><quote>+block</quote></link> patterns. If
8135 so, the URL is then blocked, and the remote web server will not be contacted.
8136 <link linkend="HANDLE-AS-IMAGE"><quote>+handle-as-image</quote></link>
8138 <link linkend="HANDLE-AS-EMPTY-DOCUMENT"><quote>+handle-as-empty-document</quote></link>
8139 are then checked, and if there is no match, an
8140 HTML <quote>BLOCKED</quote> page is sent back to the browser. Otherwise, if
8141 it does match, an image is returned for the former, and an empty text
8142 document for the latter. The type of image would depend on the setting of
8143 <link linkend="SET-IMAGE-BLOCKER"><quote>+set-image-blocker</quote></link>
8144 (blank, checkerboard pattern, or an HTTP redirect to an image elsewhere).
8149 Untrusted URLs are blocked. If URLs are being added to the
8150 <filename>trust</filename> file, then that is done.
8155 If the URL pattern matches the <link
8156 linkend="FAST-REDIRECTS"><quote>+fast-redirects</quote></link> action,
8157 it is then processed. Unwanted parts of the requested URL are stripped.
8162 Now the rest of the client browser's request headers are processed. If any
8163 of these match any of the relevant actions (e.g. <link
8164 linkend="HIDE-USER-AGENT"><quote>+hide-user-agent</quote></link>,
8165 etc.), headers are suppressed or forged as determined by these actions and
8171 Now the web server starts sending its response back (i.e. typically a web
8177 First, the server headers are read and processed to determine, among other
8178 things, the MIME type (document type) and encoding. The headers are then
8179 filtered as determined by the
8180 <link linkend="CRUNCH-INCOMING-COOKIES"><quote>+crunch-incoming-cookies</quote></link>,
8181 <link linkend="SESSION-COOKIES-ONLY"><quote>+session-cookies-only</quote></link>,
8182 and <link linkend="DOWNGRADE-HTTP-VERSION"><quote>+downgrade-http-version</quote></link>
8188 If any <link linkend="FILTER"><quote>+filter</quote></link> action
8190 linkend="DEANIMATE-GIFS"><quote>+deanimate-gifs</quote></link>
8191 action applies (and the document type fits the action), the rest of the page is
8192 read into memory (up to a configurable limit). Then the filter rules (from
8193 <filename>default.filter</filename> and any other filter files) are
8194 processed against the buffered content. Filters are applied in the order
8195 they are specified in one of the filter files. Animated GIFs, if present,
8196 are reduced to either the first or last frame, depending on the action
8197 setting.The entire page, which is now filtered, is then sent by
8198 <application>Privoxy</application> back to your browser.
8201 If neither a <link linkend="FILTER"><quote>+filter</quote></link> action
8203 linkend="DEANIMATE-GIFS"><quote>+deanimate-gifs</quote></link>
8204 matches, then <application>Privoxy</application> passes the raw data through
8205 to the client browser as it becomes available.
8210 As the browser receives the now (possibly filtered) page content, it
8211 reads and then requests any URLs that may be embedded within the page
8212 source, e.g. ad images, stylesheets, JavaScript, other HTML documents (e.g.
8213 frames), sounds, etc. For each of these objects, the browser issues a
8214 separate request (this is easily viewable in <application>Privoxy's</application>
8215 logs). And each such request is in turn processed just as above. Note that a
8216 complex web page will have many, many such embedded URLs. If these
8217 secondary requests are to a different server, then quite possibly a very
8218 differing set of actions is triggered.
8225 NOTE: This is somewhat of a simplistic overview of what happens with each URL
8226 request. For the sake of brevity and simplicity, we have focused on
8227 <application>Privoxy's</application> core features only.
8233 <!-- ~~~~~ New section ~~~~~ -->
8234 <sect2 id="actionsanat">
8235 <title>Troubleshooting: Anatomy of an Action</title>
8238 The way <application>Privoxy</application> applies
8239 <link linkend="ACTIONS">actions</link> and <link linkend="FILTER">filters</link>
8240 to any given URL can be complex, and not always so
8241 easy to understand what is happening. And sometimes we need to be able to
8242 <emphasis>see</emphasis> just what <application>Privoxy</application> is
8243 doing. Especially, if something <application>Privoxy</application> is doing
8244 is causing us a problem inadvertently. It can be a little daunting to look at
8245 the actions and filters files themselves, since they tend to be filled with
8246 <link linkend="regex">regular expressions</link> whose consequences are not
8251 One quick test to see if <application>Privoxy</application> is causing a problem
8252 or not, is to disable it temporarily. This should be the first troubleshooting
8253 step (be sure to flush caches afterward!). Looking at the
8254 logs is a good idea too. (Note that both the toggle feature and logging are
8255 enabled via <filename>config</filename> file settings, and may need to be
8256 turned <quote>on</quote>.)
8259 Another easy troubleshooting step to try is if you have done any
8260 customization of your installation, revert back to the installed
8261 defaults and see if that helps. There are times the developers get complaints
8262 about one thing or another, and the problem is more related to a customized
8263 configuration issue.
8267 <application>Privoxy</application> also provides the
8268 <ulink url="http://config.privoxy.org/show-url-info">http://config.privoxy.org/show-url-info</ulink>
8269 page that can show us very specifically how <application>actions</application>
8270 are being applied to any given URL. This is a big help for troubleshooting.
8274 First, enter one URL (or partial URL) at the prompt, and then
8275 <application>Privoxy</application> will tell us
8276 how the current configuration will handle it. This will not
8277 help with filtering effects (i.e. the <link
8278 linkend="FILTER"><quote>+filter</quote></link> action) from
8279 one of the filter files since this is handled very
8280 differently and not so easy to trap! It also will not tell you about any other
8281 URLs that may be embedded within the URL you are testing. For instance, images
8282 such as ads are expressed as URLs within the raw page source of HTML pages. So
8283 you will only get info for the actual URL that is pasted into the prompt area
8284 -- not any sub-URLs. If you want to know about embedded URLs like ads, you
8285 will have to dig those out of the HTML source. Use your browser's <quote>View
8286 Page Source</quote> option for this. Or right click on the ad, and grab the
8291 Let's try an example, <ulink url="http://google.com">google.com</ulink>,
8292 and look at it one section at a time in a sample configuration (your real
8293 configuration may vary):
8298 Matches for http://www.google.com:
8300 In file: default.action <guibutton>[ View ]</guibutton> <guibutton>[ Edit ]</guibutton>
8302 {+change-x-forwarded-for{block}
8303 +deanimate-gifs {last}
8304 +fast-redirects {check-decoded-url}
8305 +filter {refresh-tags}
8306 +filter {img-reorder}
8307 +filter {banners-by-size}
8309 +filter {jumping-windows}
8310 +filter {ie-exploits}
8311 +hide-from-header {block}
8312 +hide-referrer {forge}
8313 +session-cookies-only
8314 +set-image-blocker {pattern}
8317 { -session-cookies-only }
8323 In file: user.action <guibutton>[ View ]</guibutton> <guibutton>[ Edit ]</guibutton>
8324 (no matches in this file)
8329 This is telling us how we have defined our
8330 <link linkend="ACTIONS"><quote>actions</quote></link>, and
8331 which ones match for our test case, <quote>google.com</quote>.
8332 Displayed is all the actions that are available to us. Remember,
8333 the <literal>+</literal> sign denotes <quote>on</quote>. <literal>-</literal>
8334 denotes <quote>off</quote>. So some are <quote>on</quote> here, but many
8335 are <quote>off</quote>. Each example we try may provide a slightly different
8336 end result, depending on our configuration directives.
8340 is for our <filename>default.action</filename> file. The large, multi-line
8341 listing, is how the actions are set to match for all URLs, i.e. our default
8342 settings. If you look at your <quote>actions</quote> file, this would be the
8343 section just below the <quote>aliases</quote> section near the top. This
8344 will apply to all URLs as signified by the single forward slash at the end
8345 of the listing -- <quote> / </quote>.
8349 But we have defined additional actions that would be exceptions to these general
8350 rules, and then we list specific URLs (or patterns) that these exceptions
8351 would apply to. Last match wins. Just below this then are two explicit
8352 matches for <quote>.google.com</quote>. The first is negating our previous
8353 cookie setting, which was for <link
8354 linkend="SESSION-COOKIES-ONLY"><quote>+session-cookies-only</quote></link>
8355 (i.e. not persistent). So we will allow persistent cookies for google, at
8356 least that is how it is in this example. The second turns
8357 <emphasis>off</emphasis> any <link
8358 linkend="FAST-REDIRECTS"><quote>+fast-redirects</quote></link>
8359 action, allowing this to take place unmolested. Note that there is a leading
8360 dot here -- <quote>.google.com</quote>. This will match any hosts and
8361 sub-domains, in the google.com domain also, such as
8362 <quote>www.google.com</quote> or <quote>mail.google.com</quote>. But it would not
8363 match <quote>www.google.de</quote>! So, apparently, we have these two actions
8364 defined as exceptions to the general rules at the top somewhere in the lower
8365 part of our <filename>default.action</filename> file, and
8366 <quote>google.com</quote> is referenced somewhere in these latter sections.
8370 Then, for our <filename>user.action</filename> file, we again have no hits.
8371 So there is nothing google-specific that we might have added to our own, local
8372 configuration. If there was, those actions would over-rule any actions from
8373 previously processed files, such as <filename>default.action</filename>.
8374 <filename>user.action</filename> typically has the last word. This is the
8375 best place to put hard and fast exceptions,
8379 And finally we pull it all together in the bottom section and summarize how
8380 <application>Privoxy</application> is applying all its <quote>actions</quote>
8381 to <quote>google.com</quote>:
8392 +change-x-forwarded-for{block}
8393 -client-header-filter{hide-tor-exit-notation}
8394 -content-type-overwrite
8395 -crunch-client-header
8396 -crunch-if-none-match
8397 -crunch-incoming-cookies
8398 -crunch-outgoing-cookies
8399 -crunch-server-header
8400 +deanimate-gifs {last}
8401 -downgrade-http-version
8404 -filter {content-cookies}
8405 -filter {all-popups}
8406 -filter {banners-by-link}
8407 -filter {tiny-textforms}
8408 -filter {frameset-borders}
8409 -filter {demoronizer}
8410 -filter {shockwave-flash}
8411 -filter {quicktime-kioskmode}
8413 -filter {crude-parental}
8414 -filter {site-specifics}
8415 -filter {js-annoyances}
8416 -filter {html-annoyances}
8417 +filter {refresh-tags}
8418 -filter {unsolicited-popups}
8419 +filter {img-reorder}
8420 +filter {banners-by-size}
8422 +filter {jumping-windows}
8423 +filter {ie-exploits}
8430 -handle-as-empty-document
8432 -hide-accept-language
8433 -hide-content-disposition
8434 +hide-from-header {block}
8435 -hide-if-modified-since
8436 +hide-referrer {forge}
8439 -overwrite-last-modified
8440 -prevent-compression
8442 -server-header-filter{xml-to-html}
8443 -server-header-filter{html-to-xml}
8444 -session-cookies-only
8445 +set-image-blocker {pattern} </screen>
8449 Notice the only difference here to the previous listing, is to
8450 <quote>fast-redirects</quote> and <quote>session-cookies-only</quote>,
8451 which are activated specifically for this site in our configuration,
8452 and thus show in the <quote>Final Results</quote>.
8456 Now another example, <quote>ad.doubleclick.net</quote>:
8462 { +block{Domains starts with "ad"} }
8465 { +block{Domain contains "ad"} }
8468 { +block{Doubleclick banner server} +handle-as-image }
8469 .[a-vx-z]*.doubleclick.net
8474 We'll just show the interesting part here - the explicit matches. It is
8475 matched three different times. Two <quote>+block{}</quote> sections,
8476 and a <quote>+block{} +handle-as-image</quote>,
8477 which is the expanded form of one of our aliases that had been defined as:
8478 <quote>+block-as-image</quote>. (<link
8479 linkend="ALIASES"><quote>Aliases</quote></link> are defined in
8480 the first section of the actions file and typically used to combine more
8485 Any one of these would have done the trick and blocked this as an unwanted
8486 image. This is unnecessarily redundant since the last case effectively
8487 would also cover the first. No point in taking chances with these guys
8488 though ;-) Note that if you want an ad or obnoxious
8489 URL to be invisible, it should be defined as <quote>ad.doubleclick.net</quote>
8490 is done here -- as both a <link
8491 linkend="BLOCK"><quote>+block{}</quote></link>
8492 <emphasis>and</emphasis> an
8493 <link linkend="HANDLE-AS-IMAGE"><quote>+handle-as-image</quote></link>.
8494 The custom alias <quote><literal>+block-as-image</literal></quote> just
8495 simplifies the process and make it more readable.
8499 One last example. Let's try <quote>http://www.example.net/adsl/HOWTO/</quote>.
8500 This one is giving us problems. We are getting a blank page. Hmmm ...
8506 Matches for http://www.example.net/adsl/HOWTO/:
8508 In file: default.action <guibutton>[ View ]</guibutton> <guibutton>[ Edit ]</guibutton>
8512 +change-x-forwarded-for{block}
8513 -client-header-filter{hide-tor-exit-notation}
8514 -content-type-overwrite
8515 -crunch-client-header
8516 -crunch-if-none-match
8517 -crunch-incoming-cookies
8518 -crunch-outgoing-cookies
8519 -crunch-server-header
8521 -downgrade-http-version
8522 +fast-redirects {check-decoded-url}
8524 -filter {content-cookies}
8525 -filter {all-popups}
8526 -filter {banners-by-link}
8527 -filter {tiny-textforms}
8528 -filter {frameset-borders}
8529 -filter {demoronizer}
8530 -filter {shockwave-flash}
8531 -filter {quicktime-kioskmode}
8533 -filter {crude-parental}
8534 -filter {site-specifics}
8535 -filter {js-annoyances}
8536 -filter {html-annoyances}
8537 +filter {refresh-tags}
8538 -filter {unsolicited-popups}
8539 +filter {img-reorder}
8540 +filter {banners-by-size}
8542 +filter {jumping-windows}
8543 +filter {ie-exploits}
8550 -handle-as-empty-document
8552 -hide-accept-language
8553 -hide-content-disposition
8554 +hide-from-header{block}
8555 +hide-referer{forge}
8557 -overwrite-last-modified
8558 +prevent-compression
8560 -server-header-filter{xml-to-html}
8561 -server-header-filter{html-to-xml}
8562 +session-cookies-only
8563 +set-image-blocker{blank} }
8566 { +block{Path contains "ads".} +handle-as-image }
8572 Ooops, the <quote>/adsl/</quote> is matching <quote>/ads</quote> in our
8573 configuration! But we did not want this at all! Now we see why we get the
8574 blank page. It is actually triggering two different actions here, and
8575 the effects are aggregated so that the URL is blocked, and &my-app; is told
8576 to treat the block as if it were an image. But this is, of course, all wrong.
8577 We could now add a new action below this (or better in our own
8578 <filename>user.action</filename> file) that explicitly
8579 <emphasis>un</emphasis> blocks (
8580 <link linkend="BLOCK"><quote>{-block}</quote></link>) paths with
8581 <quote>adsl</quote> in them (remember, last match in the configuration
8582 wins). There are various ways to handle such exceptions. Example:
8594 Now the page displays ;-)
8595 Remember to flush your browser's caches when making these kinds of changes to
8596 your configuration to insure that you get a freshly delivered page! Or, try
8597 using <literal>Shift+Reload</literal>.
8601 But now what about a situation where we get no explicit matches like
8608 { +block{Path starts with "ads".} +handle-as-image }
8614 That actually was very helpful and pointed us quickly to where the problem
8615 was. If you don't get this kind of match, then it means one of the default
8616 rules in the first section of <filename>default.action</filename> is causing
8617 the problem. This would require some guesswork, and maybe a little trial and
8618 error to isolate the offending rule. One likely cause would be one of the
8619 <link linkend="FILTER"><quote>+filter</quote></link> actions.
8620 These tend to be harder to troubleshoot.
8621 Try adding the URL for the site to one of aliases that turn off
8622 <link linkend="FILTER"><quote>+filter</quote></link>:
8630 .worldpay.com # for quietpc.com
8638 <quote><literal>{ shop }</literal></quote> is an <quote>alias</quote> that expands to
8639 <quote><literal>{ -filter -session-cookies-only }</literal></quote>.
8640 Or you could do your own exception to negate filtering:
8648 # Disable ALL filter actions for sites in this section
8656 This would turn off all filtering for these sites. This is best
8657 put in <filename>user.action</filename>, for local site
8658 exceptions. Note that when a simple domain pattern is used by itself (without
8659 the subsequent path portion), all sub-pages within that domain are included
8660 automatically in the scope of the action.
8664 Images that are inexplicably being blocked, may well be hitting the
8665 <link linkend="FILTER-BANNERS-BY-SIZE"><quote>+filter{banners-by-size}</quote></link>
8667 that images of certain sizes are ad banners (works well
8668 <emphasis>most of the time</emphasis> since these tend to be standardized).
8672 <quote><literal>{ fragile }</literal></quote> is an alias that disables most
8673 actions that are the most likely to cause trouble. This can be used as a
8674 last resort for problem sites.
8680 # Handle with care: easy to break
8682 mybank.example.com</screen>
8687 <emphasis>Remember to flush caches!</emphasis> Note that the
8688 <literal>mail.google</literal> reference lacks the TLD portion (e.g.
8689 <quote>.com</quote>). This will effectively match any TLD with
8690 <literal>google</literal> in it, such as <literal>mail.google.de.</literal>,
8694 If this still does not work, you will have to go through the remaining
8695 actions one by one to find which one(s) is causing the problem.
8704 This program is free software; you can redistribute it
8705 and/or modify it under the terms of the GNU General
8706 Public License as published by the Free Software
8707 Foundation; either version 2 of the License, or (at
8708 your option) any later version.
8710 This program is distributed in the hope that it will
8711 be useful, but WITHOUT ANY WARRANTY; without even the
8712 implied warranty of MERCHANTABILITY or FITNESS FOR A
8713 PARTICULAR PURPOSE. See the GNU General Public
8714 License for more details.
8716 The GNU General Public License should be included with
8717 this file. If not, you can view it at
8718 http://www.gnu.org/copyleft/gpl.html
8719 or write to the Free Software Foundation, Inc.,
8720 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301,
projects / privoxy.git / blob