Merge branch 'release/4.9.3'

README-standalone.rst

@@ -10,21 +10,30 @@ The primary package for mod_wsgi is available on the Python package index
 version of Apache pre-installed on your target system, and if you don't,
 installation of the package will fail.
 
-If you are on a UNIX like system (Linux, macOS) and need a version of
-Apache to be installed for you, you can use the ``mod_wsgi-standalone``
+If you are on a UNIX like system (Linux) and need a version of Apache
+to be installed for you, you can use the ``mod_wsgi-standalone``
 package on PyPi instead. When installing the ``mod_wsgi-standalone``
 package it will first trigger the installation of the ``mod_wsgi-httpd``
 package, which will result in a version of Apache being installed as
 part of your Python installation. Next the ``mod_wsgi`` package will be
 installed, with it using the version of Apache installed by the
 ``mod_wsgi-httpd`` package rather than any system package for Apache.
 
-Note that this method of installation is only suitable for where you want
-to use ``mod_wsgi-express``. It cannot be used to build mod_wsgi for use
-with your system Apache installation. This installation method will also
-not work on Windows.
+This method of installation is only suitable for where you want to use
+``mod_wsgi-express``. It cannot be used to build mod_wsgi for use with
+your system Apache installation. This installation method will not
+work on Windows, and also currently fails on macOS because the Apache
+Runtime (APR) library, has not been updated to latest macOS versions.
 
 When installing mod_wsgi using this method, except that you will install
 the ``mod_wsgi-standalone`` package instead of the ``mod_wsgi`` package,
 you should follow installation and usage instructions outlined on the
 PyPi page for the ``mod_wsgi`` package.
+
+**NOTE: Although this package may allow you to install a standalone Apache
+version, it is only really recommended that you use this package if you
+have absolutely no other choice for getting the Apache httpd server
+installed. Always use the Apache httpd server supplied with the operating
+system if you can. Building this package if you do choose to do so, will
+take some time. So if you you think the install is hanging, it is probably
+still busy compiling everything.**

docs/configuration-directives/WSGITrustedProxies.rst

@@ -0,0 +1,15 @@
+==================
+WSGITrustedProxies
+==================
+
+:Description: Specify a list of trusted proxies.
+:Syntax: ``WSGITrustedProxies`` *ipaddr|(ipaddr-1 ipaddr-2 ...)*
+:Context: server config, virtual host, directory, .htaccess
+:Override: ``FileInfo``
+
+Used to specify a list of IP addresses for proxies placed in front of the
+Apache instance which are trusted.
+
+This directive only has effect when used in conjunction with the
+``WSGITrustedProxyHeaders`` directive. For more details see the documentation
+for the ``WSGITrustedProxyHeaders`` directive.

docs/configuration-directives/WSGITrustedProxyHeaders.rst

@@ -0,0 +1,231 @@
+=======================
+WSGITrustedProxyHeaders
+=======================
+
+:Description: Specify a list of trusted proxy headers.
+:Syntax: ``WSGITrustedProxyHeaders`` *header|(header-1 header-2 ...)*
+:Context: server config, virtual host, directory, .htaccess
+:Override: ``FileInfo``
+
+When trusted proxies are designated, this is used to specify the headers
+which are used to convey information from a proxy to a web server behind the
+proxy that are to be trusted.
+
+The IP addresses of the proxies to be trusted should be specified using the
+``WSGITrustedProxies`` directive.
+
+As there are multiple conventions for what headers are used to convey
+information from the proxy to the web server you need to specify the specific
+header from a supported list of headers for a particular purpose that you want
+to trust using the ``WSGITrustedProxyHeaders`` directive.
+
+When a request is then received from a trusted proxy, only the header from
+the set of headers for that particular purpose is passed through to the WSGI
+application and all others will be dropped. If a request was instead from an
+IP address which isn't a trusted proxy, then all headers in that set of headers
+will be dropped and not passed through.
+
+Depending on the purpose of the header, modifications will be made to other
+special variables passed through to the WSGI application. It is these other
+variables which is what the WSGI application should consult and the original
+header should never be consulted, with it only being provided as an indication
+of which header was used to set the special variable.
+
+The different sets of supported headers used by proxies are as follows.
+
+For passing through the IP address of the remote HTTP client the supported
+headers are:
+
+* X-Forwarded-For
+* X-Client-IP
+* X-Real-IP
+
+You should select only one of these headers as the authoritative source for
+the IP address of the remote HTTP client as sent by the proxy. Never select
+multiple headers because if you do which will be used is indeterminate.
+
+The de-facto standard for this type of header is ``X-Forwarded-For`` and it
+is recommended that it be used if your proxy supports it.
+
+The configuration might therefore be::
+
+    WSGITrustedProxies 1.2.3.4
+    WSGITrustedProxyHeaders X-Forwarded-For
+
+With this configuration, when a request is received from the trusted proxy only
+the ``X-Forwarded-For`` header will be passed through to the WSGI application.
+This will be done following CGI convention as used by WSGI, namely in the
+``HTTP_X_FORWARDED_FOR`` variable.
+
+For this set of headers, the ``REMOTE_ADDR`` CGI variable as used by WSGI will
+be modified and set to the IP address of the remote HTTP client. A WSGI
+application in this case should always use ``REMOTE_ADDR`` and never consult
+the original header files.
+
+For passing through the protocol of the original request received by the
+trusted proxy the supported headers are:
+
+* X-Forwarded-HTTPS
+* X-Forwarded-Proto
+* X-Forwarded-Scheme
+* X-Forwarded-SSL
+* X-HTTPS
+* X-Scheme
+
+You should select only one of these headers as the authoritative source for what
+protocol was used by the remote HTTP client as sent by the proxy. Never select
+multiple headers because if you do which will be used is indeterminate.
+
+The de-facto standard for this type of header is ``X-Forwarded-Proto`` and it
+is recommended that it be used if your proxy supports it.
+
+The configuration might therefore be::
+
+    WSGITrustedProxies 1.2.3.4
+    WSGITrustedProxyHeaders X-Forwarded-Proto
+
+With this configuration, when a request is received from the trusted proxy only
+the ``X-Forwarded-Proto`` header will be passed through to the WSGI application.
+This will be done following CGI convention as used by WSGI, namely in the
+``HTTP_X_FORWARDED_PROTO`` variable.
+
+For this set of headers, the ``wsgi.url_scheme`` variable passed to the WSGI
+application will be modified to indicate whether the original request used the
+``https`` protocol. Note that although it is a convention when using CGI
+scripts with Apache, the mod_wsgi module removes the ``HTTPS`` variable from
+the set of variables passed to the WSGI application. You should always use
+the ``wsgi.url_scheme`` variable in a WSGI application.
+
+For passing through the host name targeted by the original request received by
+the trusted proxy the supported headers are:
+
+* X-Forwarded-Host
+* X-Host
+
+You should select only one of these headers as the authoritative source for the
+host targeted by the original request as sent by the proxy. Never select
+multiple headers because if you do which will be used is indeterminate.
+
+The de-facto standard for this type of header is ``X-Forwarded-Host`` and it
+is recommended that it be used if your proxy supports it.
+
+The configuration might therefore be::
+
+    WSGITrustedProxies 1.2.3.4
+    WSGITrustedProxyHeaders X-Forwarded-Host
+
+With this configuration, when a request is received from the trusted proxy only
+the ``X-Forwarded-Host`` header will be passed through to the WSGI application.
+This will be done following CGI convention as used by WSGI, namely in the
+``HTTP_X_FORWARDED_HOST`` variable.
+
+For this set of headers, the ``HTTP_HOST`` variable passed to the WSGI
+application will be overridden with the value from the header supplied by the
+proxy. That is, the value from the proxy for the original request will even
+override any explicit ``Host`` header supplied in the request from the proxy,
+which in normal cases would be the host of the web server. A WSGI application
+should always consult the ``HTTP_HOST`` variable and not the separate header
+supplied by the proxy.
+
+For passing through the port targeted by the original request received by the
+trusted proxy, the only supported header is:
+
+* X-Forwarded-Port
+
+Although it is the only supported header, you still must select if as a trusted
+header to have it processed in the same way as other trusted headers.
+
+The configuration might therefore be::
+
+    WSGITrustedProxies 1.2.3.4
+    WSGITrustedProxyHeaders X-Forwarded-Port
+
+With this configuration, when a request is received from the trusted proxy only
+the ``X-Forwarded-Port`` header will be passed through to the WSGI application.
+This will be done following CGI convention as used by WSGI, namely in the
+``HTTP_X_FORWARDED_PORT`` variable.
+
+For this header, the ``SERVER_PORT`` variable passed to the WSGI application
+will be overridden with the value from the header supplied by the proxy. A WSGI
+application should always consult the ``SERVER_PORT`` variable and not the
+separate header supplied by the proxy.
+
+For passing through the host name of any proxy, to use in overriding the host
+name of the web server, the only supported header is:
+
+* X-Forwarded-Server
+
+Although it is the only supported header, you still must select if as a trusted
+header to have it processed in the same way as other trusted headers.
+
+The configuration might therefore be::
+
+    WSGITrustedProxies 1.2.3.4
+    WSGITrustedProxyHeaders X-Forwarded-Server
+
+With this configuration, when a request is received from the trusted proxy only
+the ``X-Forwarded-Server`` header will be passed through to the WSGI application.
+This will be done following CGI convention as used by WSGI, namely in the
+``HTTP_X_FORWARDED_SERVER`` variable.
+
+For this header, the ``SERVER_NAME`` variable passed to the WSGI application
+will be overridden with the value from the header supplied by the proxy. A WSGI
+application should always consult the ``SERVER_NAME`` variable and not the
+separate header supplied by the proxy.
+
+For passing through the apparent URL sub path of a web application, as mapped
+by the trusted proxy, the supported headers are:
+
+* X-Script-Name
+* X-Forwarded-Script-Name
+
+You should select only one of these headers as the authoritative source for the
+host targeted by the original request as sent by the proxy. Never select
+multiple headers because if you do which will be used is indeterminate.
+
+The configuration might therefore be::
+
+    WSGITrustedProxies 1.2.3.4
+    WSGITrustedProxyHeaders X-Script-Name
+
+With this configuration, when a request is received from the trusted proxy only
+the ``X-Script-Name`` header will be passed through to the WSGI application.
+This will be done following CGI convention as used by WSGI, namely in the
+``HTTP_X_SCRIPT_NAME`` variable.
+
+For this header, the ``SCRIPT_NAME`` variable passed to the WSGI application
+will be overridden with the value from the header supplied by the proxy. A WSGI
+application should always consult the ``SCRIPT_NAME`` variable and not the
+separate header supplied by the proxy.
+
+Examples above show using a single header of a specific purpose at one time.
+When you need to trust multiple headers for different purposes, you can list
+them separated by spaces using one instance of ``WSGITrustedProxyHeaders``::
+
+    WSGITrustedProxyHeaders X-Forwarded-For X-Forwarded-Host X-Forwarded-Port
+
+or in separate directives::
+
+    WSGITrustedProxyHeaders X-Forwarded-For
+    WSGITrustedProxyHeaders X-Forwarded-Host
+    WSGITrustedProxyHeaders X-Forwarded-Port
+
+As already highlighted you should only list one header for a specific purpose
+when there are multiple conventions for what header to use. Which you use will
+depend on the configuration of your proxy. You should only trust headers which
+are always set by the proxy, never trust headers which are optionally set by
+proxies because if not overridden by a proxy, a remote client could still
+supply the header.
+
+Also remember that in general you should not consult the proxied headers
+themselves, but instead consult the special variables set from those headers
+which are passed to the WSGI application and which are defined as being special
+to WSGI. As illustration of how such special variables are used, consider
+for example the notes in the WSGI specification around URL reconstruction.
+
+* https://peps.python.org/pep-3333/#url-reconstruction
+
+Finally, if using this feature to trust proxies and designated headers, do not
+enable in any WSGI framework or application separate functionality it may have
+for also processing the proxy headers. You should only rely on what mod_wsgi
+has done to update variables special to WSGI.

docs/configuration.rst

@@ -31,3 +31,5 @@ Configuration
    configuration-directives/WSGIScriptAliasMatch
    configuration-directives/WSGIScriptReloading
    configuration-directives/WSGISocketPrefix
+   configuration-directives/WSGITrustedProxies
+   configuration-directives/WSGITrustedProxyHeaders

docs/release-notes.rst

@@ -5,6 +5,7 @@ Release Notes
 .. toctree::
    :maxdepth: 2
 
+   release-notes/version-4.9.3
    release-notes/version-4.9.2
    release-notes/version-4.9.1
    release-notes/version-4.9.0

docs/release-notes/version-4.9.3.rst

@@ -0,0 +1,95 @@
+=============
+Version 4.9.3
+=============
+
+Version 4.9.3 of mod_wsgi can be obtained from:
+
+  https://codeload.github.com/GrahamDumpleton/mod_wsgi/tar.gz/4.9.3
+
+Bugs Fixed
+----------
+
+* When using ``WSGITrustedProxies`` and ``WSGITrustedProxyHeaders`` in the
+  Apache configuration, or ``--trust-proxy`` and ``--trust-proxy-header``
+  options with ``mod_wsgi-express``, if you trusted the ``X-Client-IP``
+  header and a request was received from an untrusted client, the header
+  was not being correctly removed from the set of headers passed through to
+  the WSGI application.
+
+  This only occurred with the ``X-Client-IP`` header and the same problem was
+  not present if trusting the ``X-Real-IP`` or ``X-Forwarded-For`` headers.
+
+  The purpose of this feature for trusting a front end proxy was in this
+  case for the headers:
+
+    * ``X-Client-IP``
+    * ``X-Real-IP``
+    * ``X-Forwarded-For``
+
+  and was designed to allow the value of ``REMOTE_ADDR`` passed to the WSGI
+  application to be rewritten to the IP address that a trusted proxy said
+  was the real remote address of the client.
+
+  In other words, if a request was received from a proxy the IP address
+  of which was trusted, ``REMOTE_ADDR`` would be set to the value of the
+  single designated header out of those listed above which was to be
+  trusted.
+
+  In the case where the proxy was trusted, in addition to ``REMOTE_ADDR``
+  being rewritten, only the trusted header would be passed through. That is,
+  if ``X-Real-IP`` was the trusted header, then ``HTTP_X_REAL_IP`` would
+  be passed to the WSGI application, but ``HTTP_X_CLIENT_IP`` and
+  ``HTTP_X_FORWARDED_FOR`` would be dropped if corresponding headers had
+  also been supplied. That the header used to rewrite ``REMOTE_ADDR`` was
+  passed through still was only intended for the purpose of documenting
+  where the value of ``REMOTE_ADDR`` came from. A WSGI application when
+  relying on this feature should only ever use the value of ``REMOTE_ADDR``
+  and should ignore the header passed through.
+
+  The behaviour as described was therefore based on a WSGI application
+  not at the same time enabling any WSGI or web framework middleware to
+  try and process any proxy headers a second time and ``REMOTE_ADDR``
+  should be the single source of truth. Albeit the headers which were
+  passed through should have resulted in the same result for ``REMOTE_ADDR``
+  if the proxy headers were processed a second time.
+
+  Now in the case of the client a request was received from not being a
+  trusted proxy, then ``REMOTE_ADDR`` would not be rewritten, and would
+  be left as the IP of the client, and none of the headers listed above
+  were supposed to be passed through.
+
+  That ``REMOTE_ADDR`` is not rewritten is implemented correctly when the
+  client is not a trusted proxy, but of the three headers listed above,
+  ``HTTP_X_CLIENT_ID`` was not being dropped if the corresponding header
+  was supplied.
+
+  If the WSGI application followed best practice and only relied on the
+  value of ``REMOTE_ADDR`` as the source of truth for the remote client
+  address, then that ``HTTP_X_CLIENT_ID`` was not being dropped should
+  pose no security risk. There would however be a problem if a WSGI
+  application was still enabling a WSGI or web framework specific middleware
+  to process the proxy headers a second time even though not required. In this
+  case, the middleware used by the WSGI application may still trust the
+  ``X-Client-IP`` header and rewrite ``REMOTE_ADDR`` allowing a malicious
+  client to pretend to have a different IP address.
+
+  In addition to the WSGI application having redundant checks for the proxy
+  headers, to take advantage of this, a client would also need direct access
+  to the Apache/mod_wsgi server instance.
+
+  In the case that only clients on your private network behind your proxy
+  could access the Apache/mod_wsgi server instance, that would imply any
+  malicious actor already had access to your private network and had access
+  to hosts in that private network or could attach their own device to that
+  private network.
+
+  In the case where your Apache/mod_wsgi server instance could be accessed
+  from the same external networks as a proxy forwarding requests to it, such
+  as may occur if making use of a CDN proxy cache, a client would still need
+  to know the direct address used by the Apache/mod_wsgi server instance.
+
+  Note that only one proxy header for designating the IP of a client should
+  ever be trusted. If you trust more than one, then which will be used if
+  both are present is undefined as it is dependent on the order that Apache
+  processes headers. This hasn't changed and as before to avoid ambiguity you
+  should only trust one of the proxy headers recognised for this purpose.

pyproject.toml.in

@@ -1,3 +1,3 @@
 [build-system]
-requires = ["setuptools>=40.8.0", "wheel", "mod_wsgi-httpd==2.4.48.1"]
+requires = ["setuptools>=40.8.0", "wheel", "mod_wsgi-httpd==2.4.54.1"]
 build-backend = "setuptools.build_meta:__legacy__"