+<sect3 renderas="sect4" id="keep-alive-timeout"><title>keep-alive-timeout</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ Number of seconds after which an open connection will no longer be reused.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>Time in seconds.</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>None</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if unset:</term>
+ <listitem>
+ <para>
+ Connections are not kept alive.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ This option allows clients to keep the connection to &my-app;
+ alive. If the server supports it, &my-app; will keep
+ the connection to the server alive as well. Under certain
+ circumstances this may result in speed-ups.
+ </para>
+ <para>
+ By default, &my-app; will close the connection to the server if
+ the client connection gets closed, or if the specified timeout
+ has been reached without a new request coming in. This behaviour
+ can be changed with the <ulink
+ url="#CONNECTION-SHARING">connection-sharing</ulink> option.
+ </para>
+ <para>
+ This option has no effect if <application>Privoxy</application>
+ has been compiled without keep-alive support.
+ </para>
+ <para>
+ Note that a timeout of five seconds as used in the default
+ configuration file significantly decreases the number of
+ connections that will be reused. The value is used because
+ some browsers limit the number of connections they open to
+ a single host and apply the same limit to proxies. This can
+ result in a single website <quote>grabbing</quote> all the
+ connections the browser allows, which means connections to
+ other websites can't be opened until the connections currently
+ in use time out.
+ </para>
+ <para>
+ Several users have reported this as a Privoxy bug, so the
+ default value has been reduced. Consider increasing it to
+ 300 seconds or even more if you think your browser can handle
+ it. If your browser appears to be hanging, it probably can't.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ keep-alive-timeout 300
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@keep-alive-timeout 5</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="tolerate-pipelining"><title>tolerate-pipelining</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ Whether or not pipelined requests should be served.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>0 or 1.</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>None</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if unset:</term>
+ <listitem>
+ <para>
+ If Privoxy receives more than one request at once, it terminates the
+ client connection after serving the first one.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ &my-app; currently doesn't pipeline outgoing requests,
+ thus allowing pipelining on the client connection is not
+ guaranteed to improve the performance.
+ </para>
+ <para>
+ By default &my-app; tries to discourage clients from pipelining
+ by discarding aggressively pipelined requests, which forces the
+ client to resend them through a new connection.
+ </para>
+ <para>
+ This option lets &my-app; tolerate pipelining. Whether or not
+ that improves performance mainly depends on the client configuration.
+ </para>
+ <para>
+ If you are seeing problems with pages not properly loading,
+ disabling this option could work around the problem.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ tolerate-pipelining 1
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@tolerate-pipelining 1</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="default-server-timeout"><title>default-server-timeout</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ Assumed server-side keep-alive timeout if not specified by the server.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>Time in seconds.</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>None</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if unset:</term>
+ <listitem>
+ <para>
+ Connections for which the server didn't specify the keep-alive
+ timeout are not reused.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ Enabling this option significantly increases the number of connections
+ that are reused, provided the <ulink
+ url="#KEEP-ALIVE-TIMEOUT">keep-alive-timeout</ulink> option
+ is also enabled.
+ </para>
+ <para>
+ While it also increases the number of connections problems
+ when &my-app; tries to reuse a connection that already has
+ been closed on the server side, or is closed while &my-app;
+ is trying to reuse it, this should only be a problem if it
+ happens for the first request sent by the client. If it happens
+ for requests on reused client connections, &my-app; will simply
+ close the connection and the client is supposed to retry the
+ request without bothering the user.
+ </para>
+ <para>
+ Enabling this option is therefore only recommended if the
+ <ulink
+ url="#CONNECTION-SHARING">connection-sharing</ulink> option
+ is disabled.
+ </para>
+ <para>
+ It is an error to specify a value larger than the <ulink
+ url="#KEEP-ALIVE-TIMEOUT">keep-alive-timeout</ulink> value.
+ </para>
+ <para>
+ This option has no effect if <application>Privoxy</application>
+ has been compiled without keep-alive support.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ default-server-timeout 60
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@#default-server-timeout 60</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="connection-sharing"><title>connection-sharing</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ Whether or not outgoing connections that have been kept alive
+ should be shared between different incoming connections.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>0 or 1</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>None</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if unset:</term>
+ <listitem>
+ <para>
+ Connections are not shared.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ This option has no effect if <application>Privoxy</application>
+ has been compiled without keep-alive support, or if it's disabled.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ Note that reusing connections doesn't necessary cause speedups.
+ There are also a few privacy implications you should be aware of.
+ </para>
+ <para>
+ If this option is effective, outgoing connections are shared between
+ clients (if there are more than one) and closing the browser that initiated
+ the outgoing connection does no longer affect the connection between &my-app;
+ and the server unless the client's request hasn't been completed yet.
+ </para>
+ <para>
+ If the outgoing connection is idle, it will not be closed until either
+ <application>Privoxy's</application> or the server's timeout is reached.
+ While it's open, the server knows that the system running &my-app; is still
+ there.
+ </para>
+ <para>
+ If there are more than one client (maybe even belonging to multiple users),
+ they will be able to reuse each others connections. This is potentially
+ dangerous in case of authentication schemes like NTLM where only the
+ connection is authenticated, instead of requiring authentication for
+ each request.
+ </para>
+ <para>
+ If there is only a single client, and if said client can keep connections
+ alive on its own, enabling this option has next to no effect. If the client
+ doesn't support connection keep-alive, enabling this option may make sense
+ as it allows &my-app; to keep outgoing connections alive even if the client
+ itself doesn't support it.
+ </para>
+ <para>
+ You should also be aware that enabling this option increases the likelihood
+ of getting the "No server or forwarder data" error message, especially if you
+ are using a slow connection to the Internet.
+ </para>
+ <para>
+ This option should only be used by experienced users who
+ understand the risks and can weight them against the benefits.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ connection-sharing 1
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@#connection-sharing 1</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="socket-timeout"><title>socket-timeout</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ Number of seconds after which a socket times out if
+ no data is received.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>Time in seconds.</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>None</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if unset:</term>
+ <listitem>
+ <para>
+ A default value of 300 seconds is used.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ The default is quite high and you probably want to reduce it.
+ If you aren't using an occasionally slow proxy like Tor, reducing
+ it to a few seconds should be fine.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ socket-timeout 300
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@socket-timeout 300</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="max-client-connections"><title>max-client-connections</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ Maximum number of client connections that will be served.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>Positive number.</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>128</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if unset:</term>
+ <listitem>
+ <para>
+ Connections are served until a resource limit is reached.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ &my-app; creates one thread (or process) for every incoming client
+ connection that isn't rejected based on the access control settings.
+ </para>
+ <para>
+ If the system is powerful enough, &my-app; can theoretically deal with
+ several hundred (or thousand) connections at the same time, but some
+ operating systems enforce resource limits by shutting down offending
+ processes and their default limits may be below the ones &my-app; would
+ require under heavy load.
+ </para>
+ <para>
+ Configuring &my-app; to enforce a connection limit below the thread
+ or process limit used by the operating system makes sure this doesn't
+ happen. Simply increasing the operating system's limit would work too,
+ but if &my-app; isn't the only application running on the system,
+ you may actually want to limit the resources used by &my-app;.
+ </para>
+ <para>
+ If &my-app; is only used by a single trusted user, limiting the
+ number of client connections is probably unnecessary. If there
+ are multiple possibly untrusted users you probably still want to
+ additionally use a packet filter to limit the maximal number of
+ incoming connections per client. Otherwise a malicious user could
+ intentionally create a high number of connections to prevent other
+ users from using &my-app;.
+ </para>
+ <para>
+ Obviously using this option only makes sense if you choose a limit
+ below the one enforced by the operating system.
+ </para>
+ <para>
+ One most POSIX-compliant systems &my-app; can't properly deal with
+ more than FD_SETSIZE file descriptors at the same time and has to reject
+ connections if the limit is reached. This will likely change in a
+ future version, but currently this limit can't be increased without
+ recompiling &my-app; with a different FD_SETSIZE limit.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ max-client-connections 256
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@#max-client-connections 256</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="listen-backlog"><title>listen-backlog</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ Connection queue length requested from the operating system.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>Number.</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>128</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if unset:</term>
+ <listitem>
+ <para>
+ A connection queue length of 128 is requested from the operating system.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ Under high load incoming connection may queue up before Privoxy
+ gets around to serve them. The queue length is limitted by the
+ operating system. Once the queue is full, additional connections
+ are dropped before Privoxy can accept and serve them.
+ </para>
+ <para>
+ Increasing the queue length allows Privoxy to accept more
+ incomming connections that arrive roughly at the same time.
+ </para>
+ <para>
+ Note that Privoxy can only request a certain queue length,
+ whether or not the requested length is actually used depends
+ on the operating system which may use a different length instead.
+ </para>
+ <para>
+ On many operating systems a limit of -1 can be specified to
+ instruct the operating system to use the maximum queue length
+ allowed. Check the listen man page to see if your platform allows this.
+ </para>
+ <para>
+ On some platforms you can use "netstat -Lan -p tcp" to see the effective
+ queue length.
+ </para>
+ <para>
+ Effectively using a value above 128 usually requires changing
+ the system configuration as well. On FreeBSD-based system the
+ limit is controlled by the kern.ipc.soacceptqueue sysctl.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ listen-backlog 4096
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@#listen-backlog -1</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="enable-accept-filter"><title>enable-accept-filter</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ Whether or not Privoxy should use an accept filter
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>0 or 1</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>0</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if unset:</term>
+ <listitem>
+ <para>
+ No accept filter is enabled.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ Accept filters reduce the number of context switches by not
+ passing sockets for new connections to Privoxy until a complete
+ HTTP request is available.
+ </para>
+ <para>
+ As a result, Privoxy can process the whole request right away
+ without having to wait for additional data first.
+ </para>
+ <para>
+ For this option to work, Privoxy has to be compiled with
+ FEATURE_ACCEPT_FILTER and the operating system has to support
+ it (which may require loading a kernel module).
+ </para>
+ <para>
+ Currently accept filters are only supported on FreeBSD-based
+ systems. Check the
+ <ulink url="https://www.freebsd.org/cgi/man.cgi?query=accf_http">accf_http(9)
+ man page</ulink>
+ to learn how to enable the support in the operating system.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ enable-accept-filter 1
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@#enable-accept-filter 1</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="handle-as-empty-doc-returns-ok"><title>handle-as-empty-doc-returns-ok</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ The status code Privoxy returns for pages blocked with
+ <!-- URL will only end up in the user manual so the relative link should work. -->
+ <literal><ulink url="actions-file.html#HANDLE-AS-EMPTY-DOCUMENT">+handle-as-empty-document</ulink></literal>.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>0 or 1</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>0</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if unset:</term>
+ <listitem>
+ <para>
+ Privoxy returns a status 403(forbidden) for all blocked pages.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if set:</term>
+ <listitem>
+ <para>
+ Privoxy returns a status 200(OK) for pages blocked with +handle-as-empty-document
+ and a status 403(Forbidden) for all other blocked pages.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ This directive was added as a work-around for Firefox bug 492459:
+ <quote>Websites are no longer rendered if SSL requests for JavaScripts are blocked by a proxy.</quote>
+ (<ulink url="https://bugzilla.mozilla.org/show_bug.cgi?id=492459"
+ >https://bugzilla.mozilla.org/show_bug.cgi?id=492459</ulink>),
+ the bug has been fixed for quite some time, but this directive is also useful
+ to make it harder for websites to detect whether or not resources are being
+ blocked.
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@#handle-as-empty-doc-returns-ok 1</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="enable-compression"><title>enable-compression</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ Whether or not buffered content is compressed before delivery.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>0 or 1</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>0</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if unset:</term>
+ <listitem>
+ <para>
+ Privoxy does not compress buffered content.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Effect if set:</term>
+ <listitem>
+ <para>
+ Privoxy compresses buffered content before delivering it to the client,
+ provided the client supports it.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ This directive is only supported if Privoxy has been compiled with
+ FEATURE_COMPRESSION, which should not to be confused with FEATURE_ZLIB.
+ </para>
+ <para>
+ Compressing buffered content is mainly useful if Privoxy and the
+ client are running on different systems. If they are running on the
+ same system, enabling compression is likely to slow things down.
+ If you didn't measure otherwise, you should assume that it does
+ and keep this option disabled.
+ </para>
+ <para>
+ Privoxy will not compress buffered content below a certain length.
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@#enable-compression 1</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="compression-level"><title>compression-level</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ The compression level that is passed to the zlib library when compressing buffered content.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>Positive number ranging from 0 to 9.</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>1</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ Compressing the data more takes usually longer than compressing
+ it less or not compressing it at all. Which level is best depends
+ on the connection between Privoxy and the client. If you can't
+ be bothered to benchmark it for yourself, you should stick with
+ the default and keep compression disabled.
+ </para>
+ <para>
+ If compression is disabled, the compression level is irrelevant.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ <screen>
+ # Best speed (compared to the other levels)
+ compression-level 1
+ # Best compression
+ compression-level 9
+ # No compression. Only useful for testing as the added header
+ # slightly increases the amount of data that has to be sent.
+ # If your benchmark shows that using this compression level
+ # is superior to using no compression at all, the benchmark
+ # is likely to be flawed.
+ compression-level 0
+ </screen>
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@#compression-level 1</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="client-header-order"><title>client-header-order</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ The order in which client headers are sorted before forwarding them.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>Client header names delimited by spaces or tabs</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>None</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ By default &my-app; leaves the client headers in the order they
+ were sent by the client. Headers are modified in-place, new headers
+ are added at the end of the already existing headers.
+ </para>
+ <para>
+ The header order can be used to fingerprint client requests
+ independently of other headers like the User-Agent.
+ </para>
+ <para>
+ This directive allows to sort the headers differently to better
+ mimic a different User-Agent. Client headers will be emitted
+ in the order given, headers whose name isn't explicitly specified
+ are added at the end.
+ </para>
+ <para>
+ Note that sorting headers in an uncommon way will make fingerprinting
+ actually easier. Encrypted headers are not affected by this directive.
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+<![%config-file;[<literallayout>@@#client-header-order Host \
+ User-Agent \
+ Accept \
+ Accept-Language \
+ Accept-Encoding \
+ Proxy-Connection \
+ Referer \
+ Cookie \
+ DNT \
+ If-Modified-Since \
+ Cache-Control \
+ Content-Length \
+ Content-Type
+</literallayout>]]>
+</sect3>
+
+
+<sect3 renderas="sect4" id="client-specific-tag"><title>client-specific-tag</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ The name of a tag that will always be set for clients that
+ requested it through the webinterface.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>Tag name followed by a description that will be shown in the webinterface</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>None</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <warning>
+ <para>
+ This is an experimental feature. The syntax is likely to change
+ in future versions.
+ </para>
+ </warning>
+ <para>
+ Client-specific tags allow Privoxy admins to create different
+ profiles and let the users chose which one they want without
+ impacting other users.
+ </para>
+ <para>
+ One use case is allowing users to circumvent certain blocks
+ without having to allow them to circumvent all blocks.
+ This is not possible with the
+ <link linkend="enable-remote-toggle">enable-remote-toggle feature</link>
+ because it would bluntly disable all blocks for all users and also affect
+ other actions like filters.
+ It also is set globally which renders it useless in most multi-user setups.
+ </para>
+ <para>
+ After a client-specific tag has been defined with the client-specific-tag
+ directive, action sections can be activated based on the tag by using a
+ <ulink url="actions-file.html#CLIENT-TAG-PATTERN">CLIENT-TAG</ulink> pattern.
+ The CLIENT-TAG pattern is evaluated at the same priority
+ as URL patterns, as a result the last matching pattern wins.
+ Tags that are created based on client or server headers are evaluated
+ later on and can overrule CLIENT-TAG and URL patterns!
+ </para>
+ <para>
+ The tag is set for all requests that come from clients that requested
+ it to be set.
+ Note that "clients" are differentiated by IP address,
+ if the IP address changes the tag has to be requested again.
+ </para>
+ <para>
+ Clients can request tags to be set by using the CGI interface <ulink
+ url="http://config.privoxy.org/client-tags">http://config.privoxy.org/client-tags</ulink>.
+ The specific tag description is only used on the web page and should
+ be phrased in away that the user understand the effect of the tag.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ <screen>
+ # Define a couple of tags, the described effect requires action sections
+ # that are enabled based on CLIENT-TAG patterns.
+ client-specific-tag circumvent-blocks Overrule blocks but do not affect other actions
+ disable-content-filters Disable content-filters but do not affect other actions
+ </screen>
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+</sect3>
+
+<!-- ~ End section ~ -->
+
+<sect3 renderas="sect4" id="client-tag-lifetime"><title>client-tag-lifetime</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ How long a temporarily enabled tag remains enabled.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>Time in seconds.</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>60</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <warning>
+ <para>
+ This is an experimental feature. The syntax is likely to change
+ in future versions.
+ </para>
+ </warning>
+ <para>
+ In case of some tags users may not want to enable them permanently,
+ but only for a short amount of time, for example to circumvent a block
+ that is the result of an overly-broad URL pattern.
+ </para>
+ <para>
+ The CGI interface <ulink
+ url="http://config.privoxy.org/client-tags">http://config.privoxy.org/client-tags</ulink>
+ therefore provides a "enable this tag temporarily" option.
+ If it is used, the tag will be set until the client-tag-lifetime
+ is over.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ <screen>
+ # Increase the time to life for temporarily enabled tags to 3 minutes
+ client-tag-lifetime 180
+ </screen>
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+</sect3>
+
+<!-- ~ End section ~ -->
+
+<sect3 renderas="sect4" id="trust-x-forwarded-for"><title>trust-x-forwarded-for</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ Whether or not Privoxy should use IP addresses specified with the X-Forwarded-For header
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>0 or one</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>0</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <warning>
+ <para>
+ This is an experimental feature. The syntax is likely to change
+ in future versions.
+ </para>
+ </warning>
+ <para>
+ If clients reach Privoxy through another proxy, for example a load
+ balancer, Privoxy can't tell the client's IP address from the connection.
+ If multiple clients use the same proxy, they will share the same
+ client tag settings which is usually not desired.
+ </para>
+ <para>
+ This option lets Privoxy use the X-Forwarded-For header value as
+ client IP address. If the proxy sets the header, multiple clients
+ using the same proxy do not share the same client tag settings.
+ </para>
+ <para>
+ This option should only be enabled if Privoxy can only be reached
+ through a proxy and if the proxy can be trusted to set the header
+ correctly. It is recommended that ACL are used to make sure only
+ trusted systems can reach Privoxy.
+ </para>
+ <para>
+ If access to Privoxy isn't limited to trusted systems, this option
+ would allow malicious clients to change the client tags for other
+ clients or increase Privoxy's memory requirements by registering
+ lots of client tag settings for clients that don't exist.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ <screen>
+ # Allow systems that can reach Privoxy to provide the client
+ # IP address with a X-Forwarded-For header.
+ trust-x-forwarded-for 1
+ </screen>
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+</sect3>
+
+<!-- ~ End section ~ -->
+
+<!-- ~~~~~ New section ~~~~~ -->
+
+<sect3 renderas="sect4" id="receive-buffer-size"><title>receive-buffer-size</title>
+<variablelist>
+ <varlistentry>
+ <term>Specifies:</term>
+ <listitem>
+ <para>
+ The size of the buffer Privoxy uses to receive data from the server.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Type of value:</term>
+ <listitem>
+ <para>
+ <replaceable>Size in bytes</replaceable>
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Default value:</term>
+ <listitem>
+ <para>5000</para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Notes:</term>
+ <listitem>
+ <para>
+ Increasing the receive-buffer-size increases Privoxy's memory usage but
+ can lower the number of context switches and thereby reduce the
+ cpu usage and potentially increase the throughput.
+ </para>
+ <para>
+ This is mostly relevant for fast network connections and
+ large downloads that don't require filtering.
+ </para>
+ <para>
+ Reducing the buffer size reduces the amount of memory Privoxy
+ needs to handle the request but increases the number of systemcalls
+ and may reduce the throughput.
+ </para>
+ <para>
+ A dtrace command like:
+ <quote>sudo dtrace -n 'syscall::read:return /execname == "privoxy"/ { @[execname] = llquantize(arg0, 10, 0, 5, 20); @m = max(arg0)}'</quote>
+ can be used to properly tune the receive-buffer-size.
+ On systems without dtrace, strace or truss may be used as
+ less convenient alternatives.
+ </para>
+ <para>
+ If the buffer is too large it will increase Privoxy's memory
+ footprint without any benefit. As the memory is (currently)
+ cleared before using it, a buffer that is too large can
+ actually reduce the throughput.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Examples:</term>
+ <listitem>
+ <para>
+ <screen>
+ # Increase the receive buffer size
+ receive-buffer-size 32768
+ </screen>
+ </para>
+ </listitem>
+ </varlistentry>
+</variablelist>
+</sect3>
+
+<!-- ~ End section ~ -->
+
+</sect2>
+
+<!-- ~~~~~ New section ~~~~~ -->
+
+<sect2 id="windows-gui">
+<title>Windows GUI Options</title>
+<para>
+ <application>Privoxy</application> has a number of options specific to the
+ Windows GUI interface:
+</para>
+
+<anchor id="activity-animation">
+<![%config-file;[<para>@@</para>]]> <!-- for spacing -->
+<para>
+ If <quote>activity-animation</quote> is set to 1, the
+ <application>Privoxy</application> icon will animate when
+ <quote>Privoxy</quote> is active. To turn off, set to 0.
+</para>
+
+<![%config-file;[<literallayout>@@#activity-animation 1</literallayout>]]>
+<![%user-man;[
+<para>
+ <literal>
+ <msgtext>
+ <literallayout>
+ <emphasis>activity-animation 1</emphasis>
+ </literallayout>
+ </msgtext>