<head>
<title>What's New in this Release</title>
<meta name="GENERATOR" content="Modular DocBook HTML Stylesheet Version 1.79">
- <link rel="HOME" title="Privoxy 3.0.28 User Manual" href="index.html">
+ <link rel="HOME" title="Privoxy 3.0.33 User Manual" href="index.html">
<link rel="PREVIOUS" title="Installation" href="installation.html">
<link rel="NEXT" title="Quickstart to Using Privoxy" href="quickstart.html">
<link rel="STYLESHEET" type="text/css" href="../p_doc.css">
<div class="NAVHEADER">
<table summary="Header navigation table" width="100%" border="0" cellpadding="0" cellspacing="0">
<tr>
- <th colspan="3" align="center">Privoxy 3.0.28 User Manual</th>
+ <th colspan="3" align="center">Privoxy 3.0.33 User Manual</th>
</tr>
<tr>
<td width="10%" align="left" valign="bottom"><a href="installation.html" accesskey="P">Prev</a></td>
</div>
<div class="SECT1">
<h1 class="SECT1"><a name="WHATSNEW" id="WHATSNEW">3. What's New in this Release</a></h1>
- <p><span class="APPLICATION">Privoxy 3.0.27</span> stable scales better in multi-user environments and brings a
- couple of tuning directives. <span class="APPLICATION">Privoxy 3.0.28</span> stable fixes two regressions
- introduced in 3.0.27.</p>
- <p>Changes in <span class="APPLICATION">Privoxy 3.0.28</span> stable:</p>
+ <p><span class="APPLICATION">Privoxy 3.0.32</span> fixes multiple DoS issues and a couple of other bugs. The issues
+ also affect earlier Privoxy releases.</p>
+ <p>Changes in <span class="APPLICATION">Privoxy 3.0.32</span> stable:</p>
<ul>
<li>
- <p>Bug fixes for regressions in 3.0.27:</p>
+ <p>Security/Reliability:</p>
<ul>
<li>
- <p>Fixed misplaced parentheses. Reported by David Binderman.</p>
+ <p>ssplit(): Remove an assertion that could be triggered with a crafted CGI request. Commit 2256d7b4d67.
+ OVE-20210203-0001. Reported by: Joshua Rogers (Opera)</p>
</li>
<li>
- <p>Changed two regression tests to depend on config directive enable-remote-toggle instead of
- FEATURE_TOGGLE.</p>
- </li>
- </ul>
- </li>
- </ul>
- <p>Changes in <span class="APPLICATION">Privoxy 3.0.27</span> stable:</p>
- <ul>
- <li>
- <p>General improvements:</p>
- <ul>
- <li>
- <p>Add a receive-buffer-size directive which can be used to set the size of the previously statically
- allocated buffer in handle_established_connection(). Increasing the 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. This is mostly relevant for fast network connections and large downloads that
- don't require filtering. Sponsored by: Robert Klemme</p>
- </li>
- <li>
- <p>Add a listen-backlog directive which specifies the backlog value passed to listen(). Sponsored by:
- Robert Klemme</p>
- </li>
- <li>
- <p>Add an enable-accept-filter directive which allows to toggle accept filter support at run time when
- compiled with FEATURE_ACCEPT_FILTER support. It makes testing more convenient and now that it's optional we
- can emit an error message if enabling the accept filter fails. Sponsored by: Robert Klemme</p>
- </li>
- <li>
- <p>Add a delay-response{} action. This is useful to tar pit JavaScript requests that are endlessly retried
- in case of blocks. It can also be used to simulate a slow Internet connection. Sponsored by: Robert
- Klemme</p>
- </li>
- <li>
- <p>Add a 'trusted-cgi-referrer' directive. It allows to configure another page or site that can be used to
- reach sensitive CGI resources. Sponsored by: Robert Klemme</p>
- </li>
- <li>
- <p>Add a --fuzz mode which exposes Privoxy internals to input from files or stdout. Mainly tested with
- American Fuzzy Lop. For details see: https://www.fabiankeil.de/talks/fuzzing-on-freebsd/ This work was
- partially funded with donations and done as part of the Privoxy month in 2015.</p>
- </li>
- <li>
- <p>Consistently use the U(ngreedy) flag in the 'img-reorder' filter.</p>
+ <p>cgi_send_banner(): Overrule invalid image types. Prevents a crash with a crafted CGI request if Privoxy
+ is toggled off. Commit e711c505c48. OVE-20210206-0001. Reported by: Joshua Rogers (Opera)</p>
</li>
<li>
- <p>listen_loop(): Reuse a single thread attribute object The object doesn't change and creating a new one
- for every thread is a waste of (CPU) time. Sponsored by: Robert Klemme</p>
+ <p>socks5_connect(): Don't try to send credentials when none are configured. Fixes a crash due to a
+ NULL-pointer dereference when the socks server misbehaves. Commit 85817cc55b9. OVE-20210207-0001. Reported
+ by: Joshua Rogers (Opera)</p>
</li>
<li>
- <p>Free csp resources in the thread that belongs to the csp instead of the main thread which has enough on
- its plate already. Sponsored by: Robert Klemme</p>
+ <p>chunked_body_is_complete(): Prevent an invalid read of size two. Commit a912ba7bc9c. OVE-20210205-0001.
+ Reported by: Joshua Rogers (Opera)</p>
</li>
<li>
- <p>Improve 'socket timeout reached' message. Log the timeout that was triggered and downgrade the log level
- to LOG_LEVEL_CONNECT to reduce the log noise with common debug settings. The timeout isn't necessary the
- result of an error and usually merely indicates that Privoxy's socket timeout is lower than the relevant
- timeouts used by client and server. Sponsored by: Robert Klemme</p>
- </li>
- <li>
- <p>Explicitly taint the server socket in case of CONNECT requests. This doesn't fix any known problems, but
- makes some log messages less confusing.</p>
- </li>
- <li>
- <p>Let write_pid_file() terminate if the pid file can't be opened. Logging the issue at info level is
- unlikely to help.</p>
+ <p>Obsolete pcre: Prevent invalid memory accesses with an invalid pattern passed to pcre_compile(). Note
+ that the obsolete pcre code is scheduled to be removed before the 3.0.33 release. There has been a warning
+ since 2008 already. Commit 28512e5b624. OVE-20210222-0001. Reported by: Joshua Rogers (Opera)</p>
</li>
+ </ul>
+ </li>
+ <li>
+ <p>Bug fixes:</p>
+ <ul>
<li>
- <p>log_error(): Reduce the mutex-protected area by not using a heap-allocated buffer that is shared between
- all threads. This increases performance and reduces the latency with verbose debug settings and multiple
- concurrent connections. Sponsored by: Robert Klemme</p>
+ <p>Properly parse the client-tag-lifetime directive. Previously it was not accepted as an obsolete hash
+ value was being used. Reported by: Joshua Rogers (Opera)</p>
</li>
<li>
- <p>Let zalloc() use calloc() if it's available. In some situations using calloc() can be faster than
- malloc() + memset() and it should never be slower. In the real world the impact of this change is not
- expected to be noticeable. Sponsored by: Robert Klemme</p>
+ <p>decompress_iob(): Prevent reading of uninitialized data. Reported by: Joshua Rogers (Opera).</p>
</li>
<li>
- <p>Never use select() when poll() is available. On most platforms select() is limited by FD_SETSIZE while
- poll() is not. This was a scaling issue for multi-user setups. Using poll() has no downside other than the
- usual risk that code modifications may introduce new bugs that have yet to be found and fixed. At least in
- theory this commit could also reduce the latency when there are lots of connections and select() would use
- "bit fields in arrays of integers" to store file descriptors. Another side effect is that Privoxy no longer
- has to stop monitoring the client sockets when pipelined requests are waiting but can't be read yet. This
- code keeps the select()-based code behind ifdefs for now but hopefully it can be removed soonish to make
- the code more readable. Sponsored by: Robert Klemme</p>
+ <p>decompress_iob(): Don't advance cur past eod when looking for the end of the file name and comment.</p>
</li>
<li>
- <p>Add a 'reproducible-tarball-dist' target. It's currently separate from the "tarball-dist" target because
- it requires a tar implementation with mtree spec support. It's far from being perfect and does not enforce
- a reproducible mode, but it's better than nothing.</p>
+ <p>decompress_iob(): Cast value to unsigned char before shifting. Prevents a left-shift of a negative value
+ which is undefined behaviour. Reported by: Joshua Rogers (Opera)</p>
</li>
<li>
- <p>Use arc4random() if it's available. While Privoxy doesn't need high quality pseudo-random numbers
- there's no reason not to use them when we can and this silences a warning emitted by code checkers that
- can't tell whether or not the quality matters.</p>
+ <p>gif_deanimate(): Confirm that that we have enough data before doing any work. Fixes a crash when fuzzing
+ with an empty document. Reported by: Joshua Rogers (Opera).</p>
</li>
<li>
- <p>Show the FEATURE_EXTERNAL_FILTERS status on the status page. Better late than never. Previously a couple
- of tests weren't executed as Privoxy-Regression-Test couldn't detect that the FEATURE_EXTERNAL_FILTERS
- dependency was satisfied.</p>
+ <p>buf_copy(): Fail if there's no data to write or nothing to do. Prevents undefined behaviour "applying
+ zero offset to null pointer". Reported by: Joshua Rogers (Opera)</p>
</li>
<li>
- <p>Ditch FEATURE_IMAGE_DETECT_MSIE. It's an obsolete workaround we inherited from Junkbuster and was
- already disabled by default. Users that feel the urge to work around issues with image requests coming from
- an Internet Explorer version from more than 15 years ago can still do this using tags.</p>
+ <p>log_error(): Treat LOG_LEVEL_FATAL as fatal even when --stfu is being used while fuzzing. Reported by:
+ Joshua Rogers (Opera).</p>
</li>
<li>
- <p>Consistently use strdup_or_die() instead of strdup() in cases where allocation failures aren't expected.
- Using strdup_or_die() allows to remove a couple of explicit error checks which slightly reduces the size of
- the binary.</p>
+ <p>Respect DESTDIR when considering whether or not to install config files with ".new" extension.</p>
</li>
<li>
- <p>Insert a refresh tag into the /client-tags CGI page when serving it while a client-specific tag is
- temporarily enabled. This makes it less likely that the user ends up looking at tag state that is out of
- date.</p>
+ <p>OpenSSL ssl_store_cert(): Fix two error messages.</p>
</li>
<li>
- <p>Use absolute URLs in the client-tag forms. It's more consistent with the rest of the CGI page URLs and
- makes it more convenient to copy the forms to external pages.</p>
+ <p>Fix a couple of format specifiers.</p>
</li>
<li>
- <p>cgi_error_disabled(): Use status code 403 and an appropriate response line</p>
+ <p>Silence compiler warnings when compiling with NDEBUG.</p>
</li>
<li>
- <p>Use a dedicated CGI handler to deal with tag-toggle requests As a result the /client-tags page is now
- safe to reach without trusted Referer header which makes bookmarking or linking to it more convenient.
- Finally, refreshing the /client-tags page to show the current state can no longer unintentionally repeat
- the previous toggle request.</p>
+ <p>fuzz_server_header(): Fix compiler warning.</p>
</li>
<li>
- <p>Don't add a "Connection" header for CONNECT requests. Explicitly sending "Connection: close" is not
- necessary and apparently it causes problems with some forwarding proxies that will close the connection
- prematurely. Reported by Marc Thomas.</p>
+ <p>fuzz_client_header(): Fix compiler warning.</p>
</li>
<li>
- <p>Fix compiler warnings.</p>
+ <p>cgi_send_user_manual(): Also reject requests if the user-manual directive specifies a https:// URL.
+ Previously Privoxy would try and fail to open a local file.</p>
</li>
</ul>
</li>
<li>
- <p>Bug fixes:</p>
+ <p>General improvements:</p>
<ul>
<li>
- <p>rfc2553_connect_to(): Properly detect and log when poll() reached the time out. Previously this was
- logged as: Could not connect to [...]: No error: 0. which isn't very helpful. Sponsored by: Robert
- Klemme</p>
+ <p>Log the TLS version and the the cipher when debug 2 is enabled.</p>
</li>
<li>
- <p>add_tag_for_client(): Set time_to_live properly. Previously the time_to_live was always set for the
- first tag. Attempts to temporarily enable a tag would result in enabling it permanently unless no tag was
- enabled already.</p>
+ <p>ssl_send_certificate_error(): Respect HEAD requests by not sending a body.</p>
</li>
<li>
- <p>Revert r1.165 which didn't perform as advertised. While the idea was to use "https:// when creating
- links for the user manual on the website", the actual effect was to use "https://" when Privoxy was
- supposed to serve the user manual itself. Reported by Yossi Zahn on Privoxy-devel@.</p>
+ <p>ssl_send_certificate_error(): End the body with a single new line.</p>
</li>
<li>
- <p>socks5_connect(): Fail in case of unsupported address types. Previously they would not be detected right
- away and Privoxy would fail later on with an error message that didn't make it obvious that the problem was
- socks-related. So far, no such problems have actually been reported.</p>
+ <p>serve(): Increase the chances that the host is logged when closing a server socket.</p>
</li>
<li>
- <p>socks5_connect(): Properly deal with socks replies that contain IPv6 addresses. Previously parts of the
- reply were left unread and later on treated as invalid HTTP response data. Fixes #904 reported by Danny
- Goossen who also provided the initial version of this patch.</p>
+ <p>handle_established_connection(): Add parentheses to clarify an expression Suggested by: David
+ Binderman</p>
</li>
- </ul>
- </li>
- <li>
- <p>Action file improvements:</p>
- <ul>
<li>
- <p>Unblock 'msdn.microsoft.com/'. It (presumably) isn't used to serve the kind of ads Privoxy should block
- by default but happens to serve lots of pages with URLs that are likely to result in false positives.
- Reported by bugreporter1694 in AF#939.</p>
+ <p>continue_https_chat(): Explicitly unset CSP_FLAG_CLIENT_CONNECTION_KEEP_ALIVE if
+ process_encrypted_request() fails. This makes it more obvious that the connection will not be reused.
+ Previously serve() relied on CSP_FLAG_SERVER_CONTENT_LENGTH_SET and CSP_FLAG_CHUNKED being unset. Inspired
+ by a patch from Joshua Rogers (Opera).</p>
</li>
<li>
- <p>Disable gif deanimation for requests tagged with CSS-REQUEST. The action will ignore content that isn't
- considered text anyway and explicitly disabling it makes this more obvious if "action" debugging (debug
- 65536) is enabled while "gif deanimation" debugging (debug 256) isn't.</p>
+ <p>decompress_iob(): Add periods to a couple of log messages</p>
</li>
<li>
- <p>Explicitly disable HTML filters for requests with CSS-REQUEST tag. The filters are unlikely to break CSS
- files but executing them without (intentionally) getting any hits is a waste of cpu time and makes the log
- more noisy when running with "debug 64".</p>
+ <p>Terminate the body of the HTTP snipplets with a single new line instead of "\r\n".</p>
</li>
<li>
- <p>Unblock 'adventofcode.com/'. Reported by Clint Adams in Debian bug #848211. Fixes Roland's AF#937.</p>
+ <p>configure: Add --with-assertions option and only enable assertions when it is used</p>
</li>
<li>
- <p>Unblock 'adlibris.com'. Reported by Wyrex in #935</p>
+ <p>windows build: Use --with-brotli and --with-mbedtls by default and enable dynamic error checking.</p>
</li>
<li>
- <p>Unblock .golang.org/</p>
+ <p>gif_deanimate(): Confirm we've got an image before trying to write it Saves a pointless buf_copy()
+ call.</p>
</li>
<li>
- <p>Add fast-redirects exception for '.youtube.com/.*origin=http'</p>
+ <p>OpenSSL ssl_store_cert(): Remove a superfluous space before the serial number.</p>
</li>
</ul>
</li>
<li>
- <p>Privoxy-Log-Parser:</p>
+ <p>Action file improvements:</p>
<ul>
<li>
- <p>Don't gather host and resource statistics if they aren't requested. While the performance impact seems
- negligible this significantly reduces the memory usage if there are lots of requests.</p>
+ <p>Disable fast-redirects for .golem.de/</p>
</li>
<li>
- <p>Bump version as the behaviour (slightly) changed.</p>
+ <p>Unblock requests to adri*.</p>
</li>
<li>
- <p>Count connection failures as well in statistics mode. Sponsored by: Robert Klemme</p>
+ <p>Block requests for trc*.taboola.com/</p>
</li>
<li>
- <p>Count connection timeouts as well in statistics mode. Sponsored by: Robert Klemme</p>
- </li>
- <li>
- <p>Fix an 'uninitialized value' warning when generating statistics for a log file without response headers.
- While privoxy-log-parser was supposed to detect this already, the check was flawed and the message the user
- didn't see was somewhat confusing anyway. Now the message is less confusing, more helpful and actually
- printed. Reported by: Robert Klemme</p>
+ <p>Disable fast-redirects for .linkedin.com/</p>
</li>
</ul>
</li>
<li>
- <p>Documentation improvements:</p>
+ <p>Filter file improvements:</p>
<ul>
<li>
- <p>Refer to the git sources instead of CVS.</p>
- </li>
- <li>
- <p>Use GNU/Linux when referring to the OS instead of the kernel.</p>
- </li>
- <li>
- <p>Add FAQ entry for what to do if editing the config file is access denied.</p>
- </li>
- <li>
- <p>Add brief HTTP/2 FAQ.</p>
- </li>
- <li>
- <p>Add a small fuzzing section to the developer documentation.</p>
- </li>
- <li>
- <p>Add a client-header-tagger{client-ip-address} example.</p>
- </li>
- <li>
- <p>Stop suggesting that Privoxy is an anonymizing proxy. The term could lead to Privoxy users
- overestimating what it can do on its own (without Tor).</p>
- </li>
- <li>
- <p>Make it more obvious that SPI accepts Paypal, too. Currently most donations are made through the Paypal
- account managed by Zwiebelfreunde e.V. and a more even distribution would be useful.</p>
- </li>
- <li>
- <p>Suggest to log applying actions as well when reproducing problems.</p>
- </li>
- <li>
- <p>Explicitly mention that Privoxy binaries are built by individuals on their own systems. Buyer
- beware!</p>
- </li>
- <li>
- <p>Mention the release feed on the homepage.</p>
- </li>
- <li>
- <p>Remove a mysterious comment with a GNU FDL link as it isn't useful and could confuse license scanners.
- In May 2002 it was briefly claimed that "this document" was covered by the GNU FDL. The commit message
- (r1.5) doesn't explain the motivation or whether all copyright holders were actually asked and agreed to
- the declared license change. It's thus hard to tell whether or not the license change was legit, but
- luckily two days later the "doc license" was "put" "back to GPL" anyway (r1.6). At the same time the
- offending comment with a link to the FDL (not the GPL) was added for no obvious reason. Now it's gone
- again.</p>
+ <p>Make the second pcrs job of the img-reorder filter greedy again. The ungreedy version broke the img tags
+ on: https://bulk.fefe.de/scalability/.</p>
</li>
</ul>
</li>
<li>
- <p>Regression tests:</p>
+ <p>Privoxy-Log-Parser:</p>
<ul>
<li>
- <p>Bump for-privoxy-version to 3.0.27 as we now rely on untrusted CGI request being rejected with status
- code 403 (instead of 200).</p>
+ <p>Highlight a few more messages.</p>
+ </li>
+ <li>
+ <p>Clarify the --statistics output. The shown "Reused connections" are server connections so name them
+ appropriately.</p>
</li>
<li>
- <p>Update test for /send-stylesheet and add another one</p>
+ <p>Bump version to 0.9.3.</p>
</li>
</ul>
</li>
<li>
- <p>Templates:</p>
+ <p>Privoxy-Regression-Test:</p>
<ul>
<li>
- <p>Consistently use https:// when linking to the Privoxy website.</p>
+ <p>Add the --check-bad-ssl option to the --help output.</p>
</li>
<li>
- <p>Remove SourceForge references in Copyright header.</p>
- </li>
- <li>
- <p>Remove a couple of SourceForge references in a comment. While at it, fix the grammar.</p>
- </li>
- <li>
- <p>Move the site-specific documentation block before the generic one. While most Privoxy installations
- don't have a site-specific documentation block, in cases were it exists it's likely to be more relevant
- than the generic one. Showing it first makes it less likely that users stop reading before they reach it,
- especially on pages that don't fit on the screen.</p>
+ <p>Bump version to 0.7.3.</p>
</li>
</ul>
</li>
<li>
- <p>Build system improvements:</p>
+ <p>Documentation:</p>
<ul>
<li>
- <p>Prefer openjade to jade. On some systems Jade produces HTML with unescaped ampersands in URLs.</p>
- </li>
- <li>
- <p>Prefer OpenSP to SP to be consistent.</p>
- </li>
- <li>
- <p>Have Docbook generated HTML files be straight ASCII. Dealing with a mixture of ISO-8859 and UTF-8 files
- is problematic.</p>
- </li>
- <li>
- <p>Echo the filename to stderr for 'make dok-tidy'. Make it a bit easier to find errors in docbook
- generated HTML.</p>
- </li>
- <li>
- <p>Warn when still using select().</p>
- </li>
- <li>
- <p>Warn when compiling without calloc().</p>
- </li>
- <li>
- <p>Make it more obvious that the --with-fdsetsize configure switch is pointless if poll() is available.</p>
- </li>
- <li>
- <p>Remove support for AmigaOS.</p>
+ <p>Add pushing the created tag to the release steps in the developer manual.</p>
</li>
<li>
- <p>Update windows build system to use supported software. The cygwin gcc -mno-cygwin option is no longer
- supported, so convert the windows build system to use the cygwin cross-compiler to build "native" code.</p>
+ <p>Clarify that 'debug 32768' should be used in addition to the other debug directives when reporting
+ problems.</p>
</li>
<li>
- <p>Add --enable-static-linking option for configure does the same thing as LDFLAGS=-static; ./configure but
- nicer than mixing evars and configure options.</p>
+ <p>Add a 'Third-party licenses and copyrights' section to the user manual.</p>
</li>
</ul>
</li>