1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
2 "http://www.w3.org/TR/html4/loose.dtd">
6 <meta name="generator" content=
7 "HTML Tidy for Linux/x86 (vers 7 December 2008), see www.w3.org">
9 <title>What's New in this Release</title>
10 <meta name="GENERATOR" content=
11 "Modular DocBook HTML Stylesheet Version 1.79">
12 <link rel="HOME" title="Privoxy 3.0.19 User Manual" href="index.html">
13 <link rel="PREVIOUS" title="Installation" href="installation.html">
14 <link rel="NEXT" title="Quickstart to Using Privoxy" href=
16 <link rel="STYLESHEET" type="text/css" href="../p_doc.css">
17 <meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
18 <link rel="STYLESHEET" type="text/css" href="p_doc.css">
19 <style type="text/css">
21 background-color: #EEEEEE;
24 :link { color: #0000FF }
25 :visited { color: #840084 }
26 :active { color: #0000FF }
27 hr.c1 {text-align: left}
32 <div class="NAVHEADER">
33 <table summary="Header navigation table" width="100%" border="0"
34 cellpadding="0" cellspacing="0">
36 <th colspan="3" align="center">Privoxy 3.0.19 User Manual</th>
40 <td width="10%" align="left" valign="bottom"><a href=
41 "installation.html" accesskey="P">Prev</a></td>
43 <td width="80%" align="center" valign="bottom"></td>
45 <td width="10%" align="right" valign="bottom"><a href=
46 "quickstart.html" accesskey="N">Next</a></td>
49 <hr class="c1" width="100%">
53 <h1 class="SECT1"><a name="WHATSNEW" id="WHATSNEW">3. What's New in this
56 <p><span class="APPLICATION">Privoxy 3.0.19</span> is a stable release.
57 The changes since 3.0.18 stable are:</p>
65 <p>Prevent a segmentation fault when de-chunking buffered
66 content. It could be triggered by malicious web servers if
67 Privoxy was configured to filter the content and running on a
68 platform where SIZE_T_MAX isn't larger than UINT_MAX, which
69 probably includes most 32-bit systems. On those platforms, all
70 Privoxy versions before 3.0.19 appear to be affected. To be on
71 the safe side, this bug should be presumed to allow code
72 execution as proving that it doesn't seems unrealistic.</p>
76 <p>Do not expect a response from the SOCKS4/4A server until it
77 got something to respond to. This regression was introduced in
78 3.0.18 and prevented the SOCKS4/4A negotiation from working.
79 Reported by qqqqqw in #3459781.</p>
85 <p>General improvements:</p>
89 <p>Fix an off-by-one in an error message about connect
94 <p>Use a GNUMakefile variable for the webserver root directory
95 and update the path. Sourceforge changed it which broke various
96 web-related targets.</p>
100 <p>Update the CODE_STATUS description.</p>
106 <p>The following changes were made between 3.0.17 and 3.0.18:</p>
114 <p>If a generated redirect URL contains characters RFC 3986
115 doesn't permit, they are (re)encoded. Not doing this makes
116 Privoxy versions from 3.0.5 to 3.0.17 susceptible to HTTP
117 response splitting (CWE-113) attacks if the
118 +fast-redirects{check-decoded-url} action is used.</p>
122 <p>Fix a logic bug that could cause Privoxy to reuse a server
123 socket after it got tainted by a server-header-tagger-induced
124 block that was triggered before the whole server response had
125 been read. If keep-alive was enabled and the request following
126 the blocked one was to the same host and using the same
127 forwarding settings, Privoxy would send it on the tainted server
128 socket. While the server would simply treat it as a pipelined
129 request, Privoxy would later on fail to properly parse the
130 server's response as it would try to parse the unread data from
131 the first response as server headers for the second one.
132 Regression introduced in 3.0.17.</p>
136 <p>When implying keep-alive in client_connection(), remember that
137 the client didn't. Fixes a regression introduced in 3.0.13 that
138 would cause Privoxy to wait for additional client requests after
139 receiving a HTTP/1.1 request with "Connection: close" set and
140 connection sharing enabled. With clients which terminates the
141 client connection after detecting that the whole body has been
142 received it doesn't really matter, but with clients that don't
143 the connection would be kept open until it timed out.</p>
147 <p>Fix a subtle race condition between
148 prepare_csp_for_next_request() and sweep(). A thread preparing
149 itself for the next client request could briefly appear to be
150 inactive. If all other threads were already using more recent
151 files, the thread could get its files swept away under its feet.
152 So far this has only been reproduced while stress testing in
153 valgrind while touching action files in a loop. It's unlikely to
154 have caused any actual problems in the real world.</p>
158 <p>Disable filters if SDCH compression is used unless filtering
159 is forced. If SDCH was combined with a supported compression
160 algorithm, Privoxy previously could try to decompress it and
161 ditch the Content-Encoding header even though the SDCH
162 compression wasn't dealt with. Reported by zebul666 in
167 <p>Make a copy of the --user value and only mess with that when
168 splitting user and group. On some operating systems modifying the
169 value directly is reflected in the output of ps and friends and
170 can be misleading. Reported by zepard in #3292710.</p>
174 <p>If forwarded-connect-retries is set, only retry if Privoxy is
175 actually forwarding the request. Previously direct connections
176 would be retried as well.</p>
180 <p>Fixed a small memory leak when retrying connections with IPv6
185 <p>Remove an incorrect assertion in
186 compile_dynamic_pcrs_job_list() It could be triggered by a pcrs
187 job with an invalid pcre pattern (for example one that contains a
188 lone quantifier).</p>
192 <p>If the --user argument user[.group] contains a dot, always
193 bail out if no group has been specified. Previously the intended,
194 but undocumented (and apparently untested), behaviour was to try
195 interpreting the whole argument as user name, but the detection
196 was flawed and checked for '0' instead of '\0', thus merely
197 preventing group names beginning with a zero.</p>
201 <p>In html_code_map[], use a numeric character reference instead
202 of ' which wasn't standardized before XHTML 1.0.</p>
206 <p>Fix an invalid free when compiled with
207 FEATURE_GRACEFUL_TERMINATION and shut down through
208 http://config.privoxy.org/die</p>
212 <p>In get_actions(), fix the "temporary" backwards compatibility
213 hack to accept block actions without reason. It also covered
214 other actions that should be rejected as invalid. Reported by
221 <p>General improvements:</p>
225 <p>Privoxy can (re)compress buffered content before delivering it
226 to the client. Disabled by default as most users wouldn't benefit
231 <p>The +fast-redirects{check-decoded-url} action checks URL
232 segments separately. If there are other parameters behind the
233 redirect URL, this makes it unnecessary to cut them off by
234 additionally using a +redirect{} pcrs command. Initial patch
235 submitted by Jamie Zawinski in #3429848.</p>
239 <p>When loading action sections, verify that the referenced
240 filters exist. Currently missing filters only result in an error
241 message, but eventually the severity will be upgraded to
246 <p>Allow to bind to multiple separate addresses. Patch set
247 submitted by Petr Pisar in #3354485.</p>
251 <p>Set socket_error to errno if connecting fails in
252 rfc2553_connect_to(). Previously rejected direct connections
253 could be incorrectly reported as DNS issues if Privoxy was
254 compiled with IPv6 support.</p>
258 <p>Adjust url_code_map[] so spaces are replaced with %20 instead
259 of '+' While '+' can be used by client's submitting form data,
260 this is not actually what Privoxy is using the lookups for. This
261 is more of a cosmetic issue and doesn't fix any known
266 <p>When compiled without FEATURE_FAST_REDIRECTS, do not silently
267 ignore +fast-redirect{} directives</p>
271 <p>Added a workaround for GNU libc's strptime() reporting
272 negative year values when the parsed year is only specified with
273 two digits. On affected systems cookies with such a date would
274 not be turned into session cookies by the +session-cookies-only
275 action. Reported by Vaeinoe in #3403560</p>
279 <p>Fixed bind failures with certain GNU libc versions if no
280 non-loopback IP address has been configured on the system. This
281 is mainly an issue if the system is using DHCP and Privoxy is
282 started before the network is completely configured. Reported by
283 Raphael Marichez in #3349356. Additional insight from Petr
288 <p>Privoxy log messages now use the ISO 8601 date format
289 %Y-%m-%d. It's only slightly longer than the old format, but
290 contains the full date including the year and allows sorting by
291 date (when grepping in multiple log files) without hassle.</p>
295 <p>In get_last_url(), do not bother trying to decode URLs that do
296 not contain at least one '%' sign. It reduces the log noise and a
297 number of unnecessary memory allocations.</p>
301 <p>In case of SOCKS5 failures, dump the socks response in the log
306 <p>Simplify the signal setup in main().</p>
310 <p>Streamline socks5_connect() slightly.</p>
314 <p>In socks5_connect(), require a complete socks response from
315 the server. Previously Privoxy didn't care how much data the
316 server response contained as long as the first two bytes
317 contained the expected values. While at it, shrink the buffer
318 size so Privoxy can't read more than a whole socks response.</p>
322 <p>In chat(), do not bother to generate a client request in case
323 of direct CONNECT requests. It will not be used anyway.</p>
327 <p>Reduce server_last_modified()'s stack size.</p>
331 <p>Shorten get_http_time() by using strftime().</p>
335 <p>Constify the known_http_methods pointers in
336 unknown_method().</p>
340 <p>Constify the time_formats pointers in parse_header_time().</p>
344 <p>Constify the formerly_valid_actions pointers in
345 action_used_to_be_valid().</p>
349 <p>Introduce a GNUMakefile MAN_PAGE variable that defaults to
350 privoxy.1. The Debian package uses section 8 for the man page and
351 this should simplify the patch.</p>
355 <p>Deduplicate the INADDR_NONE definition for Solaris by moving
356 it to jbsockets.h</p>
360 <p>In block_url(), ditch the obsolete workaround for ancient
361 Netscape versions that supposedly couldn't properly deal with
366 <p>Remove a useless NULL pointer check in load_trustfile().</p>
370 <p>Remove two useless NULL pointer checks in
371 load_one_re_filterfile().</p>
375 <p>Change url_code_map[] from an array of pointers to an array of
376 arrays It removes an unnecessary layer of indirection and on
377 64bit system reduces the size of the binary a bit.</p>
381 <p>Fix various typos. Fixes taken from Debian's 29_typos.dpatch
382 by Roland Rosenfeld.</p>
386 <p>Add a dok-tidy GNUMakefile target to clean up the messy HTML
387 generated by the other dok targets.</p>
391 <p>GNUisms in the GNUMakefile have been removed.</p>
395 <p>Change the HTTP version in static responses to 1.1</p>
399 <p>Synced config.sub and config.guess with upstream
400 2011-11-11/386c7218162c145f5f9e1ff7f558a3fbb66c37c5.</p>
404 <p>Add a dedicated function to parse the values of toggles.
405 Reduces duplicated code in load_config() and provides better
406 error handling. Invalid or missing toggle values are now a fatal
407 error instead of being silently ignored.</p>
411 <p>Terminate HTML lines in static error messages with \n instead
416 <p>Simplify cgi_error_unknown() a bit.</p>
420 <p>In LogPutString(), don't bother looking at pszText when not
421 actually logging anything.</p>
425 <p>Change ssplit()'s fourth parameter from int to size_t. Fixes a
430 <p>Add a warning that the statistics currently can't be trusted.
431 Mention Privoxy-Log-Parser's --statistics option as an
432 alternative for the time being.</p>
436 <p>In rfc2553_connect_to(), start setting cgi->error_message
441 <p>Change the expected status code returned for http://p.p/die
442 depending on whether or not FEATURE_GRACEFUL_TERMINATION is
447 <p>In cgi_die(), mark the client connection for closing. If the
448 client will fetch the style sheet through another connection it
449 gets the main thread out of the accept() state and should thus
450 trigger the actual shutdown.</p>
454 <p>Add a proper CGI message for cgi_die().</p>
458 <p>Don't enforce a logical line length limit in
459 read_config_line().</p>
463 <p>Slightly refactor server_last_modified() to remove useless
468 <p>In get_content_type(), also recognize '.jpeg' as JPEG
473 <p>Add '.png' to the list of recognized file extensions in
474 get_content_type().</p>
478 <p>In block_url(), consistently use the block reason "Request
479 blocked by Privoxy" In two places the reason was "Request for
480 blocked URL" which hides the fact that the request got blocked by
481 Privoxy and isn't necessarily correct as the block may be due to
486 <p>In listen_loop(), reload the configuration files after
487 accepting a new connection instead of before. Previously the
488 first connection that arrived after a configuration change would
489 still be handled with the old configuration.</p>
493 <p>In chat()'s receive-data loop, skip a client socket check if
494 the socket will be written to right away anyway. This can
495 increase the transfer speed for unfiltered content on fast
496 network connections.</p>
500 <p>The socket timeout is used for SOCKS negotiations as well
501 which previously couldn't timeout.</p>
505 <p>Don't keep the client connection alive if any configuration
506 file changed since the time the connection came in. This is
507 closer to Privoxy's behaviour before keep-alive support for
508 client connection has been added and also less confusing in
513 <p>Treat all Content-Type header values containing the pattern
514 'script' as a sign of text. Reported by pribog in #3134970.</p>
520 <p>Action file improvements:</p>
524 <p>Moved the site-specific block pattern section below the one
525 for the generic patterns so for requests that are matched in
526 both, the block reason for the domain is shown which is usually
527 more useful than showing the one for the generic pattern.</p>
531 <p>Remove -prevent-compression from the fragile alias. It's no
532 longer used anywhere by default and isn't known to break stuff
537 <p>Add a (disabled) section to block various Facebook tracking
538 URLs. Reported by Dan Stahlke in #3421764.</p>
542 <p>Add a (disabled) section to rewrite and redirect
543 click-tracking URLs used on news.google.com. Reported by Dan
544 Stahlke in #3421755.</p>
548 <p>Unblock linuxcounter.net/. Reported by Dan Stahlke in
553 <p>Block 'www91.intel.com/' which is used by Omniture. Reported
554 by Adam Piggott in #3167370.</p>
558 <p>Disable the handle-as-empty-doc-returns-ok option and mark it
559 as deprecated. Reminded by tceverling in #2790091.</p>
563 <p>Add ".ivwbox.de/" to the "Cross-site user tracking" section.
564 Reported by Nettozahler in #3172525.</p>
568 <p>Unblock and fast-redirect ".awin1.com/.*=http://". Reported by
569 Adam Piggott in #3170921.</p>
573 <p>Block "b.collective-media.net/".</p>
577 <p>Widen the Debian popcon exception to "qa.debian.org/popcon".
578 Seen in Debian's 05_default_action.dpatch by Roland
583 <p>Block ".gemius.pl/" which only seems to be used for user
584 tracking. Reported by johnd16 in #3002731. Additional input from
589 <p>Disable banners-by-size filters for '.thinkgeek.com/'. The
590 filter only seems to catch pictures of the inventory.</p>
594 <p>Block requests for 'go.idmnet.bbelements.com/please/showit/'.
595 Reported by kacperdominik in #3372959.</p>
599 <p>Unblock adainitiative.org/.</p>
603 <p>Add a fast-redirects exception for
604 '.googleusercontent.com/.*=cache'.</p>
608 <p>Add a fast-redirects exception for
609 webcache.googleusercontent.com/.</p>
613 <p>Unblock http://adassier.wordpress.com/ and
614 http://adassier.files.wordpress.com/.</p>
620 <p>Filter file improvements:</p>
624 <p>Let the yahoo filter hide '.ads'.</p>
628 <p>Let the msn filter hide overlay ads for Facebook 'likes' in
629 search results and elements with the id 's_notf_div'. They only
630 seem to be used to advertise site 'enhancements'.</p>
634 <p>Let the js-events filter additionally disarm setInterval().
635 Suggested by dg1727 in #3423775.</p>
641 <p>Documentation improvements:</p>
645 <p>Clarify the effect of compiling Privoxy with zlib support.
646 Suggested by dg1727 in #3423782.</p>
650 <p>Point out that the SourceForge messaging system works like a
651 black hole and should thus not be used to contact individual
656 <p>Mention some of the problems one can experience when not
657 explicitly configuring an IP addresses as listen address.</p>
661 <p>Explicitly mention that hostnames can be used instead of IP
662 addresses for the listen-address, that only the first address
663 returned will be used and what happens if the address is invalid.
664 Requested by Calestyo in #3302213.</p>
670 <p>Log message improvements:</p>
674 <p>If only the server connection is kept alive, do not pretend to
675 wait for a new client request.</p>
679 <p>Remove a superfluous log message in forget_connection().</p>
683 <p>In chat(), properly report missing server responses as such
684 instead of calling them empty.</p>
688 <p>In forwarded_connect(), fix a log message nobody should ever
693 <p>Fix a log message in socks5_connect(), a failed write
694 operation was logged as failed read operation.</p>
698 <p>Let load_one_actions_file() properly complain about a missing
699 '{' at the beginning of the file. Simply stating that a line is
700 invalid isn't particularly helpful.</p>
704 <p>Do not claim to listen on a socket until Privoxy actually
705 does. Patch submitted by Petr Pisar #3354485</p>
709 <p>Prevent a duplicated LOG_LEVEL_CLF message when sending out
710 the "no-server-data" response.</p>
714 <p>Also log the client socket when dropping a connection.</p>
718 <p>Include the destination host in the 'Request ... marked for
719 blocking. limit-connect{...} doesn't allow CONNECT ...' message
720 Patch submitted by Saperski in #3296250.</p>
724 <p>Prevent a duplicated log message if none of the resolved IP
725 addresses were reachable.</p>
729 <p>In connect_to(), do not pretend to retry if
730 forwarded-connect-retries is zero or unset.</p>
734 <p>When a specified user or group can't be found, put the name in
735 single-quotes when logging it.</p>
739 <p>In rfc2553_connect_to(), explain getnameinfo() errors
744 <p>Remove a useless log message in chat().</p>
748 <p>When retrying to connect, also log the maximum number of
749 connection attempts.</p>
753 <p>Rephrase a log message in compile_dynamic_pcrs_job_list().
754 Divide the error code and its meaning with a colon. Call the pcrs
755 job dynamic and not the filter. Filters may contain dynamic and
756 non-dynamic pcrs jobs at the same time. Only mention the name of
757 the filter or tagger, but don't claim it's a filter when it could
762 <p>In a fatal error message in load_one_actions_file(), cover
763 both URL and TAG patterns.</p>
767 <p>In pcrs_strerror(), properly report unknown positive error
768 code values as such. Previously they were handled like 0 (no
773 <p>In compile_dynamic_pcrs_job_list(), also log the actual error
774 code as pcrs_strerror() doesn't handle all errors reported by
779 <p>Don't bother trying to continue chatting if the client didn't
780 ask for it. Reduces log noise a bit.</p>
784 <p>Make two fatal error message in load_one_actions_file() more
789 <p>In cgi_send_user_manual(), log when rejecting a file name due
794 <p>In load_file(), log a message if opening a file failed. The
795 CGI error message alone isn't too helpful.</p>
799 <p>In connection_destination_matches(), improve two log messages
800 to help understand why the destinations don't match.</p>
804 <p>Rephrase a log message in serve(). Client request arrival
805 should be differentiated from closed client connections now.</p>
809 <p>In serve(), log if a client connection isn't reused due to a
810 configuration file change.</p>
814 <p>Let mark_server_socket_tainted() always mark the server socket
815 tainted, just don't talk about it in cases where it has no
816 effect. It doesn't change Privoxy's behaviour, but makes
817 understanding the log file easier.</p>
827 <p>Added a --disable-ipv6-support switch for platforms where
828 support is detected but doesn't actually work.</p>
832 <p>Do not check for the existence of strerror() and memmove()
837 <p>Remove a useless test for setpgrp(2). Privoxy doesn't need it
838 and it can cause problems when cross-compiling.</p>
842 <p>Rename the --disable-acl-files switch to
843 --disable-acl-support. Since about 2001, ACL directives are
844 specified in the standard config file.</p>
848 <p>Update the URL of the 'Removing outdated PCRE version after
849 the next stable release' posting. The old URL stopped working
850 after one of SF's recent site "optimizations". Reported by Han
857 <p>Privoxy-Regression-Test:</p>
861 <p>Added --shuffle-tests option to increase the chances of
862 detection race conditions.</p>
866 <p>Added a --local-test-file option that allows to use
867 Privoxy-Regression-Test without Privoxy.</p>
871 <p>Added tests for missing socks4 and socks4a forwarders.</p>
875 <p>The --privoxy-address option now works with IPv6 addresses
876 containing brackets, too.</p>
880 <p>Perform limited sanity checks for parameters that are supposed
881 to have numerical values.</p>
885 <p>Added a --sleep-time option to specify a number of seconds to
886 sleep between tests, defaults to 0.</p>
890 <p>Disable the range-requests tagger for tests that break if it's
895 <p>Log messages use the ISO 8601 date format %Y-%m-%d.</p>
899 <p>Fix spelling in two error messages.</p>
903 <p>In the --help output, include a list of supported tests and
904 their default levels.</p>
908 <p>Adjust the tests to properly deal with FEATURE_TOGGLE being
915 <p>Privoxy-Log-Parser:</p>
919 <p>Perform limited sanity checks for command line parameters that
920 are supposed to have numerical values.</p>
924 <p>Implement a --unbreak-lines-only option to try to revert MUA
929 <p>Accept and highlight: Added header: Content-Encoding:
934 <p>Accept and highlight: Compressed content from 29258 to 8630
939 <p>Accept and highlight: Client request arrived in time on socket
944 <p>Highlight: Didn't receive data in time: a.fsdn.com:443</p>
948 <p>Accept log messages with ISO 8601 time stamps, too.</p>
958 <p>Bump generated Firefox version to 8.0.</p>
962 <p>Only randomize the release date if the new
963 --randomize-release-date option is enabled. Firefox versions
964 after 4 use a fixed date string without meaning.</p>
971 <h2 class="SECT2"><a name="UPGRADERSNOTE" id="UPGRADERSNOTE">3.1. Note
972 to Upgraders</a></h2>
974 <p>A quick list of things to be aware of before upgrading from earlier
975 versions of <span class="APPLICATION">Privoxy</span>:</p>
979 <p>The recommended way to upgrade <span class=
980 "APPLICATION">Privoxy</span> is to backup your old configuration
981 files, install the new ones, verify that <span class=
982 "APPLICATION">Privoxy</span> is working correctly and finally merge
983 back your changes using <span class="APPLICATION">diff</span> and
984 maybe <span class="APPLICATION">patch</span>.</p>
986 <p>There are a number of new features in each <span class=
987 "APPLICATION">Privoxy</span> release and most of them have to be
988 explicitly enabled in the configuration files. Old configuration
989 files obviously don't do that and due to syntax changes using old
990 configuration files with a new <span class=
991 "APPLICATION">Privoxy</span> isn't always possible anyway.</p>
995 <p>Note that some installers remove earlier versions completely,
996 including configuration files, therefore you should really save any
997 important configuration files!</p>
1001 <p>On the other hand, other installers don't overwrite existing
1002 configuration files, thinking you will want to do that
1007 <p><tt class="FILENAME">standard.action</tt> has been merged into
1008 the <tt class="FILENAME">default.action</tt> file.</p>
1012 <p>In the default configuration only fatal errors are logged now.
1013 You can change that in the <a href="config.html#DEBUG">debug
1014 section</a> of the configuration file. You may also want to enable
1015 more verbose logging until you verified that the new <span class=
1016 "APPLICATION">Privoxy</span> version is working as expected.</p>
1020 <p>Three other config file settings are now off by default:
1021 <a href="config.html#ENABLE-REMOTE-TOGGLE">enable-remote-toggle</a>,
1023 "config.html#ENABLE-REMOTE-HTTP-TOGGLE">enable-remote-http-toggle</a>,
1025 "config.html#ENABLE-EDIT-ACTIONS">enable-edit-actions</a>. If you
1026 use or want these, you will need to explicitly enable them, and be
1027 aware of the security issues involved.</p>
1033 <div class="NAVFOOTER">
1034 <hr class="c1" width="100%">
1036 <table summary="Footer navigation table" width="100%" border="0"
1037 cellpadding="0" cellspacing="0">
1039 <td width="33%" align="left" valign="top"><a href="installation.html"
1040 accesskey="P">Prev</a></td>
1042 <td width="34%" align="center" valign="top"><a href="index.html"
1043 accesskey="H">Home</a></td>
1045 <td width="33%" align="right" valign="top"><a href="quickstart.html"
1046 accesskey="N">Next</a></td>
1050 <td width="33%" align="left" valign="top">Installation</td>
1052 <td width="34%" align="center" valign="top"> </td>
1054 <td width="33%" align="right" valign="top">Quickstart to Using