mod_rewrite.html revision db81e057b060e365d840d9a1d35a5797192efa81
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
<!--%hypertext -->
<!-- mod_rewrite.html -->
<!-- Documentation for the mod_rewrite Apache module -->
<html>
<head>
<title>Apache module mod_rewrite</title>
</head>
<!-- Background white, links blue (unvisited), navy (visited), red (active) -->
<BODY
BGCOLOR="#FFFFFF"
TEXT="#000000"
LINK="#0000FF"
VLINK="#000080"
ALINK="#FF0000"
>
<!--#include virtual="header.html" -->
<h1 ALIGN="CENTER">Module mod_rewrite (Version 3.0)</h1>
1.2 and later. It provides a rule-based rewriting engine to rewrite requested
URLs on the fly. <code>mod_rewrite</code> is not compiled into the server by
default. To use <code>mod_rewrite</code> you have to enable the following line
in the server build Configuration file:
<pre>
Module rewrite_module mod_rewrite.o
</pre>
<h2>Summary</h2>
This module uses a rule-based rewriting engine (based on a
regular-expression parser) to rewrite requested URLs on the fly.
<p>
It supports an unlimited number of additional rule conditions (which can
operate on a lot of variables, including HTTP headers) for granular
matching and external database lookups (either via plain text
tables, DBM hash files or external processes) for advanced URL
substitution.
<p>
It operates on the full URLs (including the PATH_INFO part) both in
per-server context (httpd.conf) and per-dir context (.htaccess) and even
can generate QUERY_STRING parts on result. The rewritten result can lead to internal sub-processing, external request redirection or to internal proxy throughput.
<p>
The latest version can be found on<br>
<a href="http://www.engelschall.com/sw/mod_rewrite/">
<p>
Copyright © 1996,1997 <b>The Apache Group</b>, All rights reserved.<br>
Copyright © 1996,1997 <i>Ralf S. Engelschall</i>, All rights reserved.
<p>
Written for <b>The Apache Group</b> by
<blockquote>
<i>Ralf S. Engelschall</i><br>
<a href="mailto:rse@engelschall.com"><tt>rse@engelschall.com</tt></a><br>
</blockquote>
<!--%hypertext -->
<HR>
<!--/%hypertext -->
<p>
<h2>Directives</h2>
<ul>
<li><a href="#RewriteEngine">RewriteEngine</a>
<li><a href="#RewriteOptions">RewriteOptions</a>
<li><a href="#RewriteLog">RewriteLog</a>
<li><a href="#RewriteLogLevel">RewriteLogLevel</a>
<li><a href="#RewriteMap">RewriteMap</a>
<li><a href="#RewriteBase">RewriteBase</a>
<li><a href="#RewriteCond">RewriteCond</a>
<li><a href="#RewriteRule">RewriteRule</a>
</ul>
<!--%hypertext -->
<hr>
<!--/%hypertext -->
<center>
<a name="Configuration">
<h1>Configuration Directives</h1>
</a>
</center>
<a name="RewriteEngine"><h3>RewriteEngine</h3></a>
<strong>Syntax:</strong> <code>RewriteEngine</code> {<code>on,off</code>}<br>
<strong>Default:</strong> <strong><code>RewriteEngine off</code></strong><br>
<strong>Context:</strong> server config, virtual host, per-directory config<br>
<p>
The <tt>RewriteEngine</tt> directive enables or disables the
runtime rewriting engine. If it is set to <code>off</code> this module does
no runtime processing at all. It does not even update the <tt>SCRIPT_URx</tt>
environment variables.
<p>
Use this directive to disable the module instead of commenting out
all <tt>RewriteRule</tt> directives!
<p>
<hr noshade size=1>
<p>
<a name="RewriteOptions"><h3>RewriteOptions</h3></a>
<strong>Syntax:</strong> <code>RewriteOptions</code> <em>Option</em> ...<br>
<strong>Default:</strong> -<em>None</em>-<br>
<strong>Context:</strong> server config, virtual host, per-directory config<br>
<p>
The <tt>RewriteOption</tt> directive sets some special options for the
current per-server or per-directory configuration. The <em>Option</em>
strings can be one of the following:
<ul>
<li>'<strong><code>inherit</code></strong>'<br>
This forces the current configuration to inherit the configuration of the
parent. In per-virtual-server context this means that the maps,
conditions and rules of the main server gets inherited. In per-directory
context this means that conditions and rules of the parent directory's
<tt>.htaccess</tt> configuration gets inherited.
<p>
</ul>
<p>
<hr noshade size=1>
<p>
<a name="RewriteLog"><h3>RewriteLog</h3></a>
<strong>Syntax:</strong> <code>RewriteLog</code> <em>Filename</em><br>
<strong>Default:</strong> -<em>None</em>-<br>
<strong>Context:</strong> server config, virtual host<br>
<p>
The <tt>RewriteLog</tt> directive sets the name of the file to which the
server logs any rewriting actions it performs. If the name does not begin
with a slash ('<tt>/</tt>') then it is assumed to be relative to the
<em>Server Root</em>. The directive should occur only once per server
config.
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
To disable the logging of rewriting actions it is not recommended
to set <em>Filename</em>
not create output to a logfile it still creates the logfile
output internally. <b>This will slow down the server with no advantage to the
administrator!</b>
To disable logging either remove or comment out the
<tt>RewriteLog</tt> directive or use <tt>RewriteLogLevel 0</tt>!
</td></tr>
</table>
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
SECURITY: See the <a
href="/misc/security_tips.html">Apache Security
Tips</a> document for details on why your security could be compromised if the
directory where logfiles are stored is writable by anyone other than the user
that starts the server.
</td></tr>
</table>
<p>
<b>Example:</b>
<blockquote>
<pre>
</pre>
</blockquote>
<p>
<hr noshade size=1>
<p>
<a name="RewriteLogLevel"><h3>RewriteLogLevel</h3></a>
<strong>Syntax:</strong> <code>RewriteLogLevel</code> <em>Level</em><br>
<strong>Default:</strong> <strong><code>RewriteLogLevel 0</code></strong><br>
<strong>Context:</strong> server config, virtual host<br>
<p>
The <tt>RewriteLogLevel</tt> directive set the verbosity level of the rewriting
logfile. The default level 0 means no logging, while 9 or more means
that practically all actions are logged.
<p>
To disable the logging of rewriting actions simply set <em>Level</em> to 0.
This disables all rewrite action logs.
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
<b>Notice:</b> Using a high value for <i>Level</i> will slow down your Apache
server dramatically! Use the rewriting logfile only for debugging or at least
at <em>Level</em> not greater than 2!
</td></tr>
</table>
<p>
<b>Example:</b>
<blockquote>
<pre>
RewriteLogLevel 3
</pre>
</blockquote>
<p>
<hr noshade size=1>
<p>
<a name="RewriteMap"><h3>RewriteMap</h3></a>
<strong>Syntax:</strong> <code>RewriteMap</code> <em>Mapname</em> <code>{txt,dbm,prg}:</code><em>Filename</em><br>
<strong>Default:</strong> not used per default<br>
<strong>Context:</strong> server config, virtual host<br>
<p>
The <tt>RewriteMap</tt> directive defines an external <em>Rewriting Map</em>
which can be used inside rule substitution strings by the mapping-functions
to insert/substitute fields through a key lookup.
<p>
The <a name="mapfunc"><em>Mapname</em></a> is the name of the map and will
be used to specify a mapping-function for the substitution strings of a
rewriting rule via
<blockquote><strong>
<code>${</code> <em>Mapname</em> <code>:</code> <em>LookupKey</em>
<code>|</code> <em>DefaultValue</em> <code>}</code>
</strong></blockquote>
When such a directive occurs the map <em>Mapname</em>
is consulted and the key <em>LookupKey</em> is looked-up. If the key is
found, the map-function directive is substituted by <em>SubstValue</em>. If
the key is not found then it is substituted by <em>DefaultValue</em>.
<p>
The <em>Filename</em> must be a valid Unix filepath, containing one
of the following formats:
<ol>
<li><b>Plain Text Format</b>
<p>
This is a ASCII file which contains either blank lines, comment lines
(starting with a '#' character) or
<blockquote><strong>
<em>MatchingKey</em> <em>SubstValue</em>
</strong></blockquote>
pairs - one per line. You can create such files either manually,
using your favorite editor, or by using the programs
<tt>mapcollect</tt> and <tt>mapmerge</tt> from the <tt>support</tt>
directory of the <b>mod_rewrite</b> distribution.
<p>
To declare such a map prefix, <em>Filename</em> with a <code>txt:</code>
string as in the following example:
<p>
<table border=2 cellspacing=1 cellpadding=5 bgcolor="#d0d0d0">
<tr><td><pre>
#
# map.real-to-user -- maps realnames to usernames
#
Ralf.S.Engelschall rse # Bastard Operator From Hell
Dr.Fred.Klabuster fred # Mr. DAU
</pre></td></tr>
</table>
<p>
<table border=2 cellspacing=1 cellpadding=5 bgcolor="#d0d0d0">
<tr><td><pre>
</pre></td></tr>
</table>
<p>
<li><b>DBM Hashfile Format</b>
<p>
This is a binary NDBM format file containing the
same contents as the <em>Plain Text Format</em> files. You can create
such a file with any NDBM tool or with the <tt>dbmmanage</tt> program
from the <tt>support</tt> directory of the Apache distribution.
<p>
To declare such a map prefix <em>Filename</em> with a <code>dbm:</code>
string.
<p>
<li><b>Program Format</b>
<p>
This is a Unix executable, not a lookup file. To create it you can use
the language of your choice, but the result has to be a run-able Unix
binary (i.e. either object-code or a script with the
<p>
This program gets started once at startup of the Apache servers and then
communicates with the rewriting engine over its <tt>stdin</tt> and
<tt>stdout</tt> file-handles. For each map-function lookup it will
receive the key to lookup as a newline-terminated string on
<tt>stdin</tt>. It then has to give back the looked-up value as a
newline-terminated string on <tt>stdout</tt> or the four-character string
for the given key). A trivial program which will implement a 1:1 map
(i.e. key == value) could be:
<p>
<table border=2 cellspacing=1 cellpadding=5 bgcolor="#d0d0d0">
<tr><td><pre>
$| = 1;
while (<STDIN>) {
# ...here any transformations
# or lookups should occur...
print $_;
}
</pre></td></tr>
</table>
<p>
<b>But be very careful:</b><br>
<ol>
<li>``<i>Keep the program simple, stupid</i>'' (KISS), because
if this program hangs it will lead to a hang of the Apache server
when the rule occurs.
<li>Avoid one common mistake: never do buffered I/O on <tt>stdout</tt>!
This will cause a deadloop! Hence the ``<tt>$|=1</tt>'' in the above
example...
</ol>
<p>
To declare such a map prefix <em>Filename</em> with a <code>prg:</code>
string.
</ol>
The <tt>RewriteMap</tt> directive can occur more than once. For each
mapping-function use one <tt>RewriteMap</tt> directive to declare its
rewriting mapfile. While you cannot <b>declare</b> a map in per-directory
context it is of course possible to <b>use</b> this map in per-directory
context.
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
For plain text and DBM format files the looked-up keys are cached in-core
until the <tt>mtime</tt> of the mapfile changes or the server does a
restart. This way you can have map-functions in rules which are used
for <b>every</b> request. This is no problem, because the external lookup
only happens once!
</td></tr>
</table>
<p>
<hr noshade size=1>
<p>
<a name="RewriteBase"><h3>RewriteBase</h3></a>
<strong>Syntax:</strong> <code>RewriteBase</code> <em>BaseURL</em><br>
<strong>Default:</strong> <em>default is the physical directory path</em><br>
<strong>Context:</strong> per-directory config<br>
<p>
The <tt>RewriteBase</tt> directive explicitly sets the base URL for
per-directory rewrites. As you will see below, <tt>RewriteRule</tt> can be
used in per-directory config files (<tt>.htaccess</tt>). There it will act
locally, i.e. the local directory prefix is stripped at this stage of
processing and your rewriting rules act only on the remainder. At the end
it is automatically added.
<p>
When a substitution occurs for a new URL, this module has to
re-inject the URL into the server processing. To be able to do this it needs
to know what the corresponding URL-prefix or URL-base is. By default this
prefix is the corresponding filepath itself. <b>But at most websites URLs are
<b>NOT</b> directly related to physical filename paths, so this assumption
will be usually be wrong!</b> There you have to use the <tt>RewriteBase</tt>
directive to specify the correct URL-prefix.
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
So, if your webserver's URLs are <b>not</b> directly
related to physical file paths, you have to use <tt>RewriteBase</tt> in every
<tt>.htaccess</tt> files where you want to use <tt>RewriteRule</tt>
directives.
</td></tr>
</table>
<p>
<b>Example:</b>
<blockquote>
Assume the following per-directory config file:
<p>
<table border=2 cellspacing=1 cellpadding=5 bgcolor="#d0d0d0">
<tr><td><pre>
#
#
RewriteEngine On
# let the server know that we are reached via /xyz and not
RewriteBase /xyz
# now the rewriting rules
RewriteRule ^oldstuff\.html$ newstuff.html
</pre></td></tr>
</table>
<p>
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
<font size=-1>
<b>For the Apache hackers:</b><br>
The following list gives detailed information about the internal
processing steps:
<p>
<pre>
Request:
Internal Processing:
Result:
</pre>
This seems very complicated but is the correct Apache internal processing,
because the per-directory rewriting comes too late in the process. So,
when it occurs the (rewritten) request has to be re-injected into the Apache
kernel! BUT: While this seems like a serious overhead, it really isn't, because
this re-injection happens fully internal to the Apache server and the same
procedure is used by many other operations inside Apache. So, you can be
sure the design and implementation is correct.
</font>
</td></tr>
</table>
</blockquote>
<p>
<hr noshade size=1>
<p>
<a name="RewriteCond"><h3>RewriteCond</h3></a>
<strong>Syntax:</strong> <code>RewriteCond</code> <em>TestString</em> <em>CondPattern</em><br>
<strong>Default:</strong> -<em>None</em>-<br>
<strong>Context:</strong> server config, virtual host, per-directory config<br>
<p>
The <tt>RewriteCond</tt> directive defines a rule condition. Precede a
<tt>RewriteRule</tt> directive with one or more <tt>RewriteCond</tt>
directives.
The following rewriting rule is only used if its pattern matches the current
state of the URI <b>AND</b> if these additional conditions apply, too.
<p>
<em>TestString</em> is a string which contains server-variables of the form
<blockquote><strong>
<tt>%{</tt> <em>NAME_OF_VARIABLE</em> <tt>}</tt>
</strong></blockquote>
where <em>NAME_OF_VARIABLE</em> can be a string
of the following list:
<p>
<table bgcolor="#d0d0d0" cellspacing=0 cellpadding=5>
<tr>
<td valign=top>
<b>HTTP headers:</b><p>
<font size=-1>
HTTP_USER_AGENT<br>
HTTP_REFERER<br>
HTTP_COOKIE<br>
HTTP_FORWARDED<br>
HTTP_HOST<br>
HTTP_PROXY_CONNECTION<br>
HTTP_ACCEPT<br>
</font>
</td>
<td valign=top>
<b>connection & request:</b><p>
<font size=-1>
REMOTE_ADDR<br>
REMOTE_HOST<br>
REMOTE_USER<br>
REMOTE_IDENT<br>
REQUEST_METHOD<br>
SCRIPT_FILENAME<br>
PATH_INFO<br>
QUERY_STRING<br>
AUTH_TYPE<br>
</font>
</td>
</tr>
<tr>
<td valign=top>
<b>server internals:</b><p>
<font size=-1>
DOCUMENT_ROOT<br>
SERVER_ADMIN<br>
SERVER_NAME<br>
SERVER_PORT<br>
SERVER_PROTOCOL<br>
SERVER_SOFTWARE<br>
SERVER_VERSION<br>
</font>
</td>
<td valign=top>
<b>system stuff:</b><p>
<font size=-1>
TIME_YEAR<br>
TIME_MON<br>
TIME_DAY<br>
TIME_HOUR<br>
TIME_MIN<br>
TIME_SEC<br>
TIME_WDAY<br>
</font>
</td>
<td valign=top>
<b>specials:</b><p>
<font size=-1>
API_VERSION<br>
THE_REQUEST<br>
REQUEST_URI<br>
REQUEST_FILENAME<br>
IS_SUBREQ<br>
</font>
</td>
</tr>
</table>
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
These variables all correspond to the similar named HTTP MIME-headers, C
variables of the Apache server or <tt>struct tm</tt> fields of the Unix
system.
</td></tr>
</table>
<p>
Special Notes:
<ol>
<li>The variables SCRIPT_FILENAME and REQUEST_FILENAME contain the same
<tt>request_rec</tt> structure of the Apache server. The first name is just the
commonly known CGI variable name while the second is the consistent
counterpart to REQUEST_URI (which contains the value of the <tt>uri</tt>
field of <tt>request_rec</tt>).
<p>
<li>There is the special format: <tt>%{ENV:variable}</tt> where
<i>variable</i> can be any environment variable. This is looked-up via
internal Apache structures and (if not found there) via <tt>getenv()</tt> from
the Apache server process.
<p>
<li>There is the special format: <tt>%{HTTP:header}</tt> where
<i>header</i> can be any HTTP MIME-header name. This is looked-up
from the HTTP request. Example: <tt>%{HTTP:Proxy-Connection}</tt>
is the value of the HTTP header ``<tt>Proxy-Connection:</tt>''.
<p>
<li>There is the special format: <tt>%{LA-U:url}</tt>
for look-aheads like <tt>-U</tt>. This performs a internal sub-request to
look-ahead for the final value of <i>url</i>.
<p>
<li>There is the special format: <tt>%{LA-F:file}</tt>
for look-aheads like <tt>-F</tt>. This performs a internal sub-request to
look-ahead for the final value of <i>file</i>.
</ol>
<p>
<em>TestString</em> gets evaluated and then matched against
<em>CondPattern</em>.
<p>
<b>Remember:</b> <em>CondPattern</em> is a standard
<em>Extended Regular Expression</em> with some additions:
<ol>
<li>You can precede the pattern string with a '<tt>!</tt>' character
(exclamation mark) to specify a <b>non</b>-matching pattern.
<p>
<li>
There are some special variants of <em>CondPatterns</em>. Instead of real
regular expression strings you can also use one of the following:
<p>
<ul>
<li>'<b>-d</b>' (is <b>d</b>irectory)<br>
Treats the <i>TestString</i> as a pathname and
tests if it exists and is a directory.
<p>
<li>'<b>-f</b>' (is regular <b>f</b>ile)<br>
Treats the <i>TestString</i> as a pathname and
tests if it exists and is a regular file.
<p>
<li>'<b>-s</b>' (is regular file with <b>s</b>ize)<br>
Treats the <i>TestString</i> as a pathname and
tests if it exists and is a regular file with size greater then zero.
<p>
<li>'<b>-l</b>' (is symbolic <b>l</b>ink)<br>
Treats the <i>TestString</i> as a pathname and
tests if it exists and is a symbolic link.
<p>
<li>'<b>-F</b>' (is existing file via subrequest)<br>
Checks if <i>TestString</i> is a valid file and accessible via all the
server's currently-configured access controls for that path. This uses an
internal subrequest to determine the check, so use it with care because it
decreases your servers performance!
<p>
<li>'<b>-U</b>' (is existing URL via subrequest)<br>
Checks if <i>TestString</i> is a valid URL and accessible via all the server's
currently-configured access controls for that path. This uses an internal
subrequest to determine the check, so use it with care because it decreases
your servers performance!
</ul>
<p>
Notice: All of these tests can also be prefixed by a not ('!') character
to negate their meaning.
</ol>
<p>
Additionally you can set special flags for <em>CondPattern</em> by appending
<blockquote><strong>
<code>[</code><em>flags</em><code>]</code>
</strong></blockquote>
as the third argument to the <tt>RewriteCond</tt> directive. <em>Flags</em>
is a comma-separated list of the following flags:
<ul>
<li>'<strong><code>nocase|NC</code></strong>' (<b>n</b>o <b>c</b>ase)<br>
This makes the condition test case-insensitive, i.e. there is
no difference between 'A-Z' and 'a-z' both in the expanded
<em>TestString</em> and the <em>CondPattern</em>.
<p>
<li>'<strong><code>ornext|OR</code></strong>' (<b>or</b> next condition)<br>
Use this to combine rule conditions with a local OR instead of the
implicit AND. Typical example:
<p>
<blockquote><pre>
RewriteCond %{REMOTE_HOST} ^host1.* [OR]
RewriteCond %{REMOTE_HOST} ^host2.* [OR]
RewriteCond %{REMOTE_HOST} ^host3.*
RewriteRule ...some special stuff for any of these hosts...
</pre></blockquote>
<p>
</ul>
<p>
<b>Example:</b>
<blockquote>
To rewrite the Homepage of a site according to the ``<tt>User-Agent:</tt>''
header of the request, you can use the following:
<blockquote><pre>
RewriteCond %{HTTP_USER_AGENT} ^Mozilla.*
RewriteRule ^/$ /homepage.max.html [L]
RewriteCond %{HTTP_USER_AGENT} ^Lynx.*
RewriteRule ^/$ /homepage.min.html [L]
RewriteRule ^/$ /homepage.std.html [L]
</pre></blockquote>
Interpretation: If you use Netscape Navigator as your browser (which identifies
itself as 'Mozilla'), then you get the max homepage, which includes
Frames, etc. If you use the Lynx browser (which is Terminal-based), then you
get the min homepage, which contains no images, no tables, etc. If you
use any other browser you get the standard homepage.
</blockquote>
<p>
<p>
<hr noshade size=1>
<p>
<a name="RewriteRule"><h3>RewriteRule</h3></a>
<strong>Syntax:</strong> <code>RewriteRule</code> <em>Pattern</em> <em>Substitution</em><br>
<strong>Default:</strong> -<em>None</em>-<br>
<strong>Context:</strong> server config, virtual host, per-directory config<br>
<p>
The <tt>RewriteRule</tt> directive is the real rewriting workhorse. The
directive can occur more than once. Each directive then defines one single
rewriting rule. The <b>definition order</b> of these rules is
<b>important</b>, because this order is used when applying the rules at
run-time.
<p>
which gets applied to the current URL. Here ``current'' means the value of the
URL when this rule gets applied. This may not be the original requested
URL, because there could be any number of rules before which already matched
and made alterations to it.
<p>
Some hints about the syntax of regular expressions:
<p>
<table bgcolor="#d0d0d0" cellspacing=0 cellpadding=5>
<tr>
<td valign=top>
<pre>
<strong><code>^</code></strong> Start of line
<strong><code>$</code></strong> End of line
<strong><code>.</code></strong> Any single character
<strong><code>[</code></strong>chars<strong><code>]</code></strong> One of chars
<strong><code>[^</code></strong>chars<strong><code>]</code></strong> None of chars
<strong><code>?</code></strong> 0 or 1 of the preceding char
<strong><code>*</code></strong> 0 or N of the preceding char
<strong><code>+</code></strong> 1 or N of the preceding char
<strong><code>\</code></strong>char escape that specific char
<strong><code>(</code></strong>string<strong><code>)</code></strong> Grouping of chars (the <b>N</b>th group can be used on the RHS with <code>$</code><b>N</b>)
</pre>
</td>
</tr>
</table>
<p>
Additionally the NOT character ('<tt>!</tt>') is a possible pattern
prefix. This gives you the ability to negate a pattern; to say, for instance: ``<i>if
the current URL does <b>NOT</b> match to this pattern</i>''. This can be used
for special cases where it is better to match the negative pattern or as a
last default rule.
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
<b>Notice!</b> When using the NOT character to negate a pattern you cannot
have grouped wildcard parts in the pattern. This is impossible because when
the pattern does NOT match, there are no contents for the groups. In
consequence, if negated patterns are used, you cannot use <tt>$N</tt> in the
substitution string!
</td></tr>
</table>
<p>
<a name="rhs"><em>Substitution</em></a> of a rewriting rule is the string
which is substituted for (or replaces) the original URL for which
<em>Pattern</em> matched. Beside plain text you can use
<ol>
<li>pattern-group back-references (<code>$N</code>)
<li>server-variables as in rule condition test-strings (<code>%{VARNAME}</code>)
<li><a href="#mapfunc">mapping-function</a> calls (<code>${mapname:key|default}</code>)
</ol>
Back-references are <code>$</code><b>N</b> (<b>N</b>=1..9) identifiers which
will be replaced by the contents of the <b>N</b>th group of the matched
<em>Pattern</em>. The server-variables are the same as for the
<em>TestString</em> of a <tt>RewriteCond</tt> directive. The
mapping-functions come from the <tt>RewriteMap</tt> directive and are
explained there. These three types of variables are expanded in the order of
the above list.
<p>
As already mentioned above, all the rewriting rules are applied to the
<em>Substitution</em> (in the order of definition in the config file). The
URL is <b>completely replaced</b> by the <em>Substitution</em> and the
rewriting process goes on until there are no more rules (unless explicitly
terminated by a <code><b>L</b></code> flag - see below).
<p>
There is a special substitution string named '<tt>-</tt>' which means:
<b>NO substitution</b>! Sounds silly? No, it is useful to provide rewriting
conjunction with the <b>C</b> (chain) flag to be able to have more than one
pattern to be applied before a substitution occurs.
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
<b>Notice</b>: There is a special feature. When you prefix a substitution
field with <tt>http://</tt><em>thishost</em>[<em>:thisport</em>] then
<b>mod_rewrite</b> automatically strips it out. This auto-reduction on
implicit external redirect URLs is a useful and important feature when
used in combination with a mapping-function which generates the hostname
part. Have a look at the first example in the example section below to
understand this.
<p>
<b>Remember:</b> An unconditional external redirect to your own server will
To achieve such a self-redirect, you have to use the <b>R</b>-flag (see
below).
</td></tr>
</table>
<p>
Additionally you can set special flags for <em>Substitution</em> by appending
<blockquote><strong>
<code>[</code><em>flags</em><code>]</code>
</strong></blockquote>
as the third argument to the <tt>RewriteRule</tt> directive. <em>Flags</em> is a
comma-separated list of the following flags:
<ul>
<li>'<strong><code>redirect|R</code>[=<i>code</i>]</strong>' (force <a name="redirect"><b>r</b>edirect</a>)<br>
Prefix <em>Substitution</em>
force a external redirection. If no <i>code</i> is given a HTTP response
of 302 (MOVED TEMPORARILY) is used. If you want to use other response
codes in the range 300-400 just specify them as a number or use
one of the following symbolic names: <tt>temp</tt> (default), <tt>permanent</tt>,
<tt>seeother</tt>.
Use it for rules which should
canonicalize the URL and gives it back to the client, e.g. translate
``<code>/~</code>'' into ``<code>/u/</code>'' or always append a slash to
<code>/u/</code><em>user</em>, etc.<br>
<p>
<b>Notice:</b> When you use this flag, make sure that the
substitution field is a valid URL! If not, you are redirecting to an
invalid location! And remember that this flag itself only prefixes the
Usually you also want to stop and do the redirection immediately. To stop
the rewriting you also have to provide the 'L' flag.
<p>
<li>'<strong><code>forbidden|F</code></strong>' (force URL to be <b>f</b>orbidden)<br>
This forces the current URL to be forbidden, i.e. it immediately sends
back a HTTP response of 403 (FORBIDDEN). Use this flag in conjunction with
appropriate RewriteConds to conditionally block some URLs.
<p>
<li>'<strong><code>gone|G</code></strong>' (force URL to be <b>g</b>one)<br>
This forces the current URL to be gone, i.e. it immediately sends back a
HTTP response of 410 (GONE). Use this flag to mark no longer existing
pages as gone.
<p>
<li>'<strong><code>proxy|P</code></strong>' (force <b>p</b>roxy)<br>
This flag forces the substitution part to be internally forced as a proxy
request and immediately (i.e. rewriting rule processing stops here) put
through the proxy module. You have to make sure that the substitution
be handled by the Apache proxy module. If not you get an error from
the proxy module. Use this flag to achieve a more powerful implementation
of the <tt>mod_proxy</tt> directive <tt>ProxyPass</tt>, to map
some remote stuff into the namespace of the local server.
<p>
Notice: <b>You really have to put <tt>ProxyRequests On</tt> into your
server configuration to prevent proxy requests from leading to core-dumps
inside the Apache kernel. If you have not compiled in the proxy module,
then there is no core-dump problem, because mod_rewrite checks for
existence of the proxy module and if lost forbids proxy URLs. </b>
<p>
<li>'<strong><code>last|L</code></strong>' (<b>l</b>ast rule)<br>
Stop the rewriting process here and
don't apply any more rewriting rules. This corresponds to the Perl
<code>last</code> command or the <code>break</code> command from the C
language. Use this flag to prevent the currently rewritten URL from being
rewritten further by following rules which may be wrong. For
example, use it to rewrite the root-path URL ('<code>/</code>') to a real
<p>
<li>'<strong><code>next|N</code></strong>' (<b>n</b>ext round)<br>
Re-run the rewriting process (starting again with the first rewriting
rule). Here the URL to match is again not the original URL but the URL
from the last rewriting rule. This corresponds to the Perl
<code>next</code> command or the <code>continue</code> command from the C
language. Use this flag to restart the rewriting process, i.e. to
immediately go to the top of the loop. <br>
<b>But be careful not to create a deadloop!</b>
<p>
<li>'<strong><code>chain|C</code></strong>' (<b>c</b>hained with next rule)<br>
This flag chains the current rule with the next rule (which itself can
also be chained with its following rule, etc.). This has the following
effect: if a rule matches, then processing continues as usual, i.e. the
flag has no effect. If the rule does <b>not</b> match, then all following
chained rules are skipped. For instance, use it to remove the
``<tt>.www</tt>'' part inside a per-directory rule set when you let an
external redirect happen (where the ``<tt>.www</tt>'' part should not to
occur!).
<p>
<li>'<strong><code>type|T</code></strong>=<em>mime-type</em>' (force MIME <b>t</b>ype)<br>
Force the MIME-type of the target file to be <em>mime-type</em>. For
instance, this can be used to simulate the old <tt>mod_alias</tt>
directive <tt>ScriptAlias</tt> which internally forces all files inside
the mapped directory to have a MIME type of
<p>
<li>'<strong><code>nosubreq|NS</code></strong>' (used only if <b>n</b>o internal <b>s</b>ub-request)<br>
This flag forces the rewriting engine to skip a rewriting rule if the
current request is an internal sub-request. For instance, sub-requests
occur internally in Apache when <tt>mod_include</tt> tries to find out
On sub-requests it is not always useful and even sometimes causes a failure to
if the complete set of rules are applied. Use this flag to exclude some rules.<br>
<p>
Use the following rule for your decision: whenever you prefix some URLs
with CGI-scripts to force them to be processed by the CGI-script, the
chance is high that you will run into problems (or even overhead) on sub-requests.
In these cases, use this flag.
<p>
<li>'<strong><code>passthrough|PT</code></strong>' (<b>p</b>ass <b>t</b>hrough to next handler)<br>
This flag forces the rewriting engine to set the <code>uri</code> field
of the internal <code>request_rec</code> structure to the value
of the <code>filename</code> field. This flag is just a hack to be able
to post-process the output of <tt>RewriteRule</tt> directives by
<tt>Alias</tt>, <tt>ScriptAlias</tt>, <tt>Redirect</tt>, etc. directives
from other URI-to-filename translators. A trivial example to show the
semantics:
If you want to rewrite <tt>/abc</tt> to <tt>/def</tt> via the rewriting
engine of <tt>mod_rewrite</tt> and then <tt>/def</tt> to <tt>/ghi</tt>
with <tt>mod_alias</tt>:
<pre>
RewriteRule ^/abc(.*) /def$1 [PT]
Alias /def /ghi
</pre>
If you omit the <tt>PT</tt> flag then <tt>mod_rewrite</tt>
<tt>filename=/def/...</tt> as a full API-compliant URI-to-filename
translator should do. Then <tt>mod_alias</tt> comes and tries to do a
URI-to-filename transition which will not work.
<p>
Notice: <b>You have to use this flag if you want to intermix directives
of different modules which contain URL-to-filename translators</b>. The
typical example is the use of <tt>mod_alias</tt> and
<tt>mod_rewrite</tt>..
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
<font size=-1>
<b>For the Apache hackers:</b><br>
If the current Apache API had a
filename-to-filename hook additionally to the URI-to-filename hook then
we wouldn't need this flag! But without such a hook this flag is the
only solution. The Apache Group has discussed this problem and will
add such hooks into Apache version 2.0.
</font>
</td></tr>
</table>
<p>
<li>'<strong><code>skip|S</code></strong>=<em>num</em>' (<b>s</b>kip next rule(s))<br>
This flag forces the rewriting engine to skip the next <em>num</em> rules
in sequence when the current rule matches. Use this to make pseudo
if-then-else constructs: The last rule of the then-clause becomes
a <tt>skip=N</tt> where N is the number of rules in the else-clause.
(This is <b>not</b> the same as the 'chain|C' flag!)
<p>
<li>'<strong><code>env|E=</code></strong><i>VAR</i>:<i>VAL</i>' (set <b>e</b>nvironment variable)<br>
This forces an environment variable named <i>VAR</i> to be set to the value
<i>VAL</i>, where <i>VAL</i> can contain regexp backreferences <tt>$N</tt>
which will be expanded. You can use this flag more than once to set more
than one variable. The variables can be later dereferenced at a lot of
situations, but the usual location will be from within XSSI (via
But additionally you can also dereference it in a following RewriteCond
pattern via <tt>%{ENV:VAR}</tt>. Use this to strip but remember
information from URLs.
</ul>
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
Remember: Never forget that <em>Pattern</em> gets applied to a complete URL
in per-server configuration files. <b>But in per-directory configuration
files, the per-directory prefix (which always is the same for a specific
directory!) gets automatically <em>removed</em> for the pattern matching and
automatically <em>added</em> after the substitution has been done.</b> This feature is
essential for many sorts of rewriting, because without this prefix stripping
you have to match the parent directory which is not always possible.
<p>
There is one exception: If a substitution string starts with
``<tt>http://</tt>'' then the directory prefix will be <b>not</b> added and a
external redirect or proxy throughput (if flag <b>P</b> is used!) is forced!
</td></tr>
</table>
<p>
<table width="70%" border=2 bgcolor="#c0c0e0" cellspacing=0 cellpadding=10>
<tr><td>
Notice! To enable the rewriting engine for per-directory configuration files
you need to set ``<tt>RewriteEngine On</tt>'' in these files <b>and</b>
``<tt>Option FollowSymLinks</tt>'' enabled. If your administrator has
disabled override of <tt>FollowSymLinks</tt> for a user's directory, then
you cannot use the rewriting engine. This restriction is needed for
security reasons.
</td></tr>
</table>
<p>
Here are all possible substitution combinations and their meanings:
<p>
<p>
<table bgcolor="#d0d0d0" cellspacing=0 cellpadding=5>
<tr>
<td>
<pre>
<b>Given Rule</b> <b>Resulting Substitution</b>
---------------------------------------------- ----------------------------------
^/somepath(.*) otherpath$1 not supported, because invalid!
^/somepath(.*) otherpath$1 [R] not supported, because invalid!
^/somepath(.*) otherpath$1 [P] not supported, because invalid!
---------------------------------------------- ----------------------------------
^/somepath(.*) /otherpath$1 [R] http://thishost/otherpath/pathinfo
via external redirection
^/somepath(.*) /otherpath$1 [P] not supported, because silly!
---------------------------------------------- ----------------------------------
^/somepath(.*) http://thishost/otherpath$1 [R] http://thishost/otherpath/pathinfo
via external redirection
^/somepath(.*) http://thishost/otherpath$1 [P] not supported, because silly!
---------------------------------------------- ----------------------------------
^/somepath(.*) http://otherhost/otherpath$1 http://otherhost/otherpath/pathinfo
via external redirection
^/somepath(.*) http://otherhost/otherpath$1 [R] http://otherhost/otherpath/pathinfo
via external redirection
(the [R] flag is redundant)
^/somepath(.*) http://otherhost/otherpath$1 [P] http://otherhost/otherpath/pathinfo
via internal proxy
</pre>
</td>
</tr>
</table>
<p>
<b>Inside per-directory configuration for <tt>/somepath</tt><br>
<tt>RewriteBase /somepath</tt>)<br> for
<p>
<table bgcolor="#d0d0d0" cellspacing=0 cellpadding=5>
<tr>
<td>
<pre>
<b>Given Rule</b> <b>Resulting Substitution</b>
---------------------------------------------- ----------------------------------
^localpath(.*) otherpath$1 [R] http://thishost/somepath/otherpath/pathinfo
via external redirection
^localpath(.*) otherpath$1 [P] not supported, because silly!
---------------------------------------------- ----------------------------------
^localpath(.*) /otherpath$1 [R] http://thishost/otherpath/pathinfo
via external redirection
^localpath(.*) /otherpath$1 [P] not supported, because silly!
---------------------------------------------- ----------------------------------
^localpath(.*) http://thishost/otherpath$1 [R] http://thishost/otherpath/pathinfo
via external redirection
^localpath(.*) http://thishost/otherpath$1 [P] not supported, because silly!
---------------------------------------------- ----------------------------------
^localpath(.*) http://otherhost/otherpath$1 http://otherhost/otherpath/pathinfo
via external redirection
^localpath(.*) http://otherhost/otherpath$1 [R] http://otherhost/otherpath/pathinfo
via external redirection
(the [R] flag is redundant)
^localpath(.*) http://otherhost/otherpath$1 [P] http://otherhost/otherpath/pathinfo
via internal proxy
</pre>
</td>
</tr>
</table>
<p>
<b>Example:</b>
<p>
<blockquote>
We want to rewrite URLs of the form
<blockquote>
<code>/</code> <em>Language</em>
<code>/~</code> <em>Realname</em>
<code>/.../</code> <em>File</em>
</blockquote>
into
<blockquote>
<code>/u/</code> <em>Username</em>
<code>/.../</code> <em>File</em>
<code>.</code> <em>Language</em>
</blockquote>
<p>
We take the rewrite mapfile from above and save it under
following lines to the Apache server configuration file:
<blockquote>
<pre>
RewriteLog /anywhere/rewrite.log
RewriteMap real-to-user txt:/anywhere/map.real-to-host
RewriteRule ^/([^/]+)/~([^/]+)/(.*)$ /u/${real-to-user:$2|nobody}/$3.$1
</pre>
</blockquote>
</blockquote>
<!--#include virtual="footer.html" -->
</BODY>
</HTML>
<!--/%hypertext -->