[APACHE DOCUMENTATION]

Apache HTTP Server Version 1.3


Module mod_rewrite
URL Rewriting Engine

This module provides a rule-based rewriting engine to rewrite requested URLs on the fly.

Status: Extension
Source File: mod_rewrite.c
Module Identifier: rewrite_module
Compatibility: Available in Apache 1.2 and later.


Summary

``The great thing about mod_rewrite is it gives you all the configurability and flexibility of Sendmail. The downside to mod_rewrite is that it gives you all the configurability and flexibility of Sendmail.''
-- Brian Behlendorf
Apache Group
`` Despite the tons of examples and docs, mod_rewrite is voodoo. Damned cool voodoo, but still voodoo. ''
-- Brian Moore
bem@news.cmc.net
Welcome to mod_rewrite, the Swiss Army Knife of URL manipulation!

This module uses a rule-based rewriting engine (based on a regular-expression parser) to rewrite requested URLs on the fly. It supports an unlimited number of rules and an unlimited number of attached rule conditions for each rule to provide a really flexible and powerful URL manipulation mechanism. The URL manipulations can depend on various tests, for instance server variables, environment variables, HTTP headers, time stamps and even external database lookups in various formats can be used to achieve a really granular URL matching.

This module operates on the full URLs (including the path-info part) both in per-server context (httpd.conf) and per-directory context (.htaccess) and can even generate query-string parts on result. The rewritten result can lead to internal sub-processing, external request redirection or even to an internal proxy throughput.

But all this functionality and flexibility has its drawback: complexity. So don't expect to understand this entire module in just one day.

This module was invented and originally written in April 1996
and gifted exclusively to the The Apache Group in July 1997 by

Ralf S. Engelschall
rse@engelschall.com
www.engelschall.com

Table Of Contents

Internal Processing

Configuration Directives

Miscellaneous

Internal Processing

The internal processing of this module is very complex but needs to be explained once even to the average user to avoid common mistakes and to let you exploit its full functionality.

API Phases

First you have to understand that when Apache processes a HTTP request it does this in phases. A hook for each of these phases is provided by the Apache API. Mod_rewrite uses two of these hooks: the URL-to-filename translation hook which is used after the HTTP request has been read but before any authorization starts and the Fixup hook which is triggered after the authorization phases and after the per-directory config files (.htaccess) have been read, but before the content handler is activated.

So, after a request comes in and Apache has determined the corresponding server (or virtual server) the rewriting engine starts processing of all mod_rewrite directives from the per-server configuration in the URL-to-filename phase. A few steps later when the final data directories are found, the per-directory configuration directives of mod_rewrite are triggered in the Fixup phase. In both situations mod_rewrite rewrites URLs either to new URLs or to filenames, although there is no obvious distinction between them. This is a usage of the API which was not intended to be this way when the API was designed, but as of Apache 1.x this is the only way mod_rewrite can operate. To make this point more clear remember the following two points:

  1. Although mod_rewrite rewrites URLs to URLs, URLs to filenames and even filenames to filenames, the API currently provides only a URL-to-filename hook. In Apache 2.0 the two missing hooks will be added to make the processing more clear. But this point has no drawbacks for the user, it is just a fact which should be remembered: Apache does more in the URL-to-filename hook than the API intends for it.
  2. Unbelievably mod_rewrite provides URL manipulations in per-directory context, i.e., within .htaccess files, although these are reached a very long time after the URLs have been translated to filenames. It has to be this way because .htaccess files live in the filesystem, so processing has already reached this stage. In other words: According to the API phases at this time it is too late for any URL manipulations. To overcome this chicken and egg problem mod_rewrite uses a trick: When you manipulate a URL/filename in per-directory context mod_rewrite first rewrites the filename back to its corresponding URL (which is usually impossible, but see the RewriteBase directive below for the trick to achieve this) and then initiates a new internal sub-request with the new URL. This restarts processing of the API phases.

    Again mod_rewrite tries hard to make this complicated step totally transparent to the user, but you should remember here: While URL manipulations in per-server context are really fast and efficient, per-directory rewrites are slow and inefficient due to this chicken and egg problem. But on the other hand this is the only way mod_rewrite can provide (locally restricted) URL manipulations to the average user.

Don't forget these two points!

Ruleset Processing

Now when mod_rewrite is triggered in these two API phases, it reads the configured rulesets from its configuration structure (which itself was either created on startup for per-server context or during the directory walk of the Apache kernel for per-directory context). Then the URL rewriting engine is started with the contained ruleset (one or more rules together with their conditions). The operation of the URL rewriting engine itself is exactly the same for both configuration contexts. Only the final result processing is different.

The order of rules in the ruleset is important because the rewriting engine processes them in a special (and not very obvious) order. The rule is this: The rewriting engine loops through the ruleset rule by rule (RewriteRule directives) and when a particular rule matches it optionally loops through existing corresponding conditions (RewriteCond directives). For historical reasons the conditions are given first, and so the control flow is a little bit long-winded. See Figure 1 for more details.

[Needs graphics capability to display]
Figure 1: The control flow through the rewriting ruleset

As you can see, first the URL is matched against the Pattern of each rule. When it fails mod_rewrite immediately stops processing this rule and continues with the next rule. If the Pattern matches, mod_rewrite looks for corresponding rule conditions. If none are present, it just substitutes the URL with a new value which is constructed from the string Substitution and goes on with its rule-looping. But if conditions exist, it starts an inner loop for processing them in the order that they are listed. For conditions the logic is different: we don't match a pattern against the current URL. Instead we first create a string TestString by expanding variables, back-references, map lookups, etc. and then we try to match CondPattern against it. If the pattern doesn't match, the complete set of conditions and the corresponding rule fails. If the pattern matches, then the next condition is processed until no more conditions are available. If all conditions match, processing is continued with the substitution of the URL with Substitution.

Quoting Special Characters

As of Apache 1.3.20, special characters in TestString and Substitution strings can be escaped (that is, treated as normal characters without their usual special meaning) by prefixing them with a slosh ('\') character. In other words, you can include an actual dollar-sign character in a Substitution string by using '\$'; this keeps mod_rewrite from trying to treat it as a backreference.

Regex Back-Reference Availability

One important thing here has to be remembered: Whenever you use parentheses in Pattern or in one of the CondPattern, back-references are internally created which can be used with the strings $N and %N (see below). These are available for creating the strings Substitution and TestString. Figure 2 shows to which locations the back-references are transfered for expansion.
[Needs graphics capability to display]
Figure 2: The back-reference flow through a rule

We know this was a crash course on mod_rewrite's internal processing. But you will benefit from this knowledge when reading the following documentation of the available directives.

Configuration Directives

RewriteEngine

Syntax: RewriteEngine on|off
Default: RewriteEngine off
Context: server config, virtual host, directory, .htaccess
Override: FileInfo
Status: Extension
Module: mod_rewrite.c
Compatibility: Apache 1.2

The RewriteEngine directive enables or disables the runtime rewriting engine. If it is set to off this module does no runtime processing at all. It does not even update the SCRIPT_URx environment variables.

Use this directive to disable the module instead of commenting out all the RewriteRule directives!

Note that, by default, rewrite configurations are not inherited. This means that you need to have a RewriteEngine on directive for each virtual host in which you wish to use it.

RewriteOptions

Syntax: RewriteOptions Option
Default: None
Context: server config, virtual host, directory, .htaccess
Override: FileInfo
Status: Extension
Module: mod_rewrite.c
Compatibility: Apache 1.2

The RewriteOptions directive sets some special options for the current per-server or per-directory configuration. The Option strings can be one of the following:

RewriteLog

Syntax: RewriteLog file-path
Default: None
Context: server config, virtual host
Override: Not applicable
Status: Extension
Module: mod_rewrite.c
Compatibility: Apache 1.2

The RewriteLog 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 ('/') then it is assumed to be relative to the Server Root. The directive should occur only once per server config.

Note: To disable the logging of rewriting actions it is not recommended to set Filename to /dev/null, because although the rewriting engine does not then output to a logfile it still creates the logfile output internally. This will slow down the server with no advantage to the administrator! To disable logging either remove or comment out the RewriteLog directive or use RewriteLogLevel 0!
Security: See the Apache Security Tips 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.

Example:

RewriteLog "/usr/local/var/apache/logs/rewrite.log"

RewriteLogLevel

Syntax: RewriteLogLevel Level
Default: RewriteLogLevel 0
Context: server config, virtual host
Override: Not applicable
Status: Extension
Module: mod_rewrite.c
Compatibility: Apache 1.2

The RewriteLogLevel directive sets 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.

To disable the logging of rewriting actions simply set Level to 0. This disables all rewrite action logs.

Notice: Using a high value for Level will slow down your Apache server dramatically! Use the rewriting logfile at a Level greater than 2 only for debugging!

Example:

RewriteLogLevel 3

RewriteLock

Syntax: RewriteLock file-path
Default: None
Context: server config
Override: Not applicable
Status: Extension
Module: mod_rewrite.c
Compatibility: Apache 1.3

This directive sets the filename for a synchronization lockfile which mod_rewrite needs to communicate with RewriteMap programs. Set this lockfile to a local path (not on a NFS-mounted device) when you want to use a rewriting map-program. It is not required for other types of rewriting maps.

RewriteMap

Syntax: RewriteMap MapName MapType:MapSource
Default: not used per default
Context: server config, virtual host
Override: Not applicable
Status: Extension
Module: mod_rewrite.c
Compatibility: Apache 1.2 (partially), Apache 1.3

The RewriteMap directive defines a Rewriting Map which can be used inside rule substitution strings by the mapping-functions to insert/substitute fields through a key lookup. The source of this lookup can be of various types.

The MapName is the name of the map and will be used to specify a mapping-function for the substitution strings of a rewriting rule via one of the following constructs:

${ MapName : LookupKey }
${ MapName : LookupKey | DefaultValue }
When such a construct occurs the map MapName is consulted and the key LookupKey is looked-up. If the key is found, the map-function construct is substituted by SubstValue. If the key is not found then it is substituted by DefaultValue or by the empty string if no DefaultValue was specified.

The following combinations for MapType and MapSource can be used:

The RewriteMap directive can occur more than once. For each mapping-function use one RewriteMap directive to declare its rewriting mapfile. While you cannot declare a map in per-directory context it is of course possible to use this map in per-directory context.
Note: For plain text and DBM format files the looked-up keys are cached in-core until the mtime of the mapfile changes or the server does a restart. This way you can have map-functions in rules which are used for every request. This is no problem, because the external lookup only happens once!

RewriteBase

Syntax: RewriteBase URL-path
Default: default is the physical directory path
Context: directory, .htaccess
Override: FileInfo
Status: Extension
Module: mod_rewrite.c
Compatibility: Apache 1.2

The RewriteBase directive explicitly sets the base URL for per-directory rewrites. As you will see below, RewriteRule can be used in per-directory config files (.htaccess). 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 back to the path.

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. But at most websites URLs are NOT directly related to physical filename paths, so this assumption will usually be wrong! There you have to use the RewriteBase directive to specify the correct URL-prefix.

Notice: If your webserver's URLs are not directly related to physical file paths, you have to use RewriteBase in every .htaccess files where you want to use RewriteRule directives.

Example:

Assume the following per-directory config file: 
#
#  /abc/def/.htaccess -- per-dir config file for directory /abc/def
#  Remember: /abc/def is the physical path of /xyz, i.e., the server
#            has a 'Alias /xyz /abc/def' directive e.g.
#

RewriteEngine On

#  let the server know that we were reached via /xyz and not
#  via the physical path prefix /abc/def
RewriteBase   /xyz

#  now the rewriting rules
RewriteRule   ^oldstuff\.html$  newstuff.html

In the above example, a request to /xyz/oldstuff.html gets correctly rewritten to the physical file /abc/def/newstuff.html.

Note - For Apache hackers:
The following list gives detailed information about the internal processing steps: 
Request:
  /xyz/oldstuff.html

Internal Processing:
  /xyz/oldstuff.html     -> /abc/def/oldstuff.html  (per-server Alias)
  /abc/def/oldstuff.html -> /abc/def/newstuff.html  (per-dir    RewriteRule)
  /abc/def/newstuff.html -> /xyz/newstuff.html      (per-dir    RewriteBase)
  /xyz/newstuff.html     -> /abc/def/newstuff.html  (per-server Alias)

Result:
  /abc/def/newstuff.html
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 internally 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.

RewriteCond

Syntax: RewriteCond TestString CondPattern
Default: None
Context: server config, virtual host, directory, .htaccess
Override: FileInfo
Status: Extension
Module: mod_rewrite.c
Compatibility: Apache 1.2 (partially), Apache 1.3

The RewriteCond directive defines a rule condition. Precede a RewriteRule directive with one or more RewriteCond directives. The following rewriting rule is only used if its pattern matches the current state of the URI and if these additional conditions apply too.

TestString is a string which can contains the following expanded constructs in addition to plain text:

Special Notes:

  1. The variables SCRIPT_FILENAME and REQUEST_FILENAME contain the same value, i.e., the value of the filename field of the internal request_rec 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 uri field of request_rec).
  3. There is the special format: %{ENV:variable} where variable can be any environment variable. This is looked-up via internal Apache structures and (if not found there) via getenv() from the Apache server process.
  4. There is the special format: %{HTTP:header} where header can be any HTTP MIME-header name. This is looked-up from the HTTP request. Example: %{HTTP:Proxy-Connection} is the value of the HTTP header ``Proxy-Connection:''.
  5. There is the special format %{LA-U:variable} for look-aheads which perform an internal (URL-based) sub-request to determine the final value of variable. Use this when you want to use a variable for rewriting which is actually set later in an API phase and thus is not available at the current stage. For instance when you want to rewrite according to the REMOTE_USER variable from within the per-server context (httpd.conf file) you have to use %{LA-U:REMOTE_USER} because this variable is set by the authorization phases which come after the URL translation phase where mod_rewrite operates. On the other hand, because mod_rewrite implements its per-directory context (.htaccess file) via the Fixup phase of the API and because the authorization phases come before this phase, you just can use %{REMOTE_USER} there.
  6. There is the special format: %{LA-F:variable} which performs an internal (filename-based) sub-request to determine the final value of variable. Most of the time this is the same as LA-U above.

CondPattern is the condition pattern, i.e., a regular expression which is applied to the current instance of the TestString, i.e., TestString is evaluated and then matched against CondPattern.

Remember: CondPattern is a standard Extended Regular Expression with some additions:

  1. You can prefix the pattern string with a '!' character (exclamation mark) to specify a non-matching pattern.
  2. There are some special variants of CondPatterns. Instead of real regular expression strings you can also use one of the following:
    • '<CondPattern' (is lexically lower)
      Treats the CondPattern as a plain string and compares it lexically to TestString. True if TestString is lexically lower than CondPattern.
    • '>CondPattern' (is lexically greater)
      Treats the CondPattern as a plain string and compares it lexically to TestString. True if TestString is lexically greater than CondPattern.
    • '=CondPattern' (is lexically equal)
      Treats the CondPattern as a plain string and compares it lexically to TestString. True if TestString is lexically equal to CondPattern, i.e the two strings are exactly equal (character by character). If CondPattern is just "" (two quotation marks) this compares TestString to the empty string.
    • '-d' (is directory)
      Treats the TestString as a pathname and tests if it exists and is a directory.
    • '-f' (is regular file)
      Treats the TestString as a pathname and tests if it exists and is a regular file.
    • '-s' (is regular file with size)
      Treats the TestString as a pathname and tests if it exists and is a regular file with size greater than zero.
    • '-l' (is symbolic link)
      Treats the TestString as a pathname and tests if it exists and is a symbolic link.
    • '-F' (is existing file via subrequest)
      Checks if TestString 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!
    • '-U' (is existing URL via subrequest)
      Checks if TestString 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 server's performance!
    Notice: All of these tests can also be prefixed by an exclamation mark ('!') to negate their meaning.

Additionally you can set special flags for CondPattern by appending

[flags]
as the third argument to the RewriteCond directive. Flags is a comma-separated list of the following flags:

Example:

To rewrite the Homepage of a site according to the ``User-Agent:'' header of the request, you can use the following: 
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]
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.

RewriteRule

Syntax: RewriteRule Pattern Substitution
Default: None
Context: server config, virtual host, directory, .htaccess
Override: FileInfo
Status: Extension
Module: mod_rewrite.c
Compatibility: Apache 1.2 (partially), Apache 1.3

The RewriteRule directive is the real rewriting workhorse. The directive can occur more than once. Each directive then defines one single rewriting rule. The definition order of these rules is important, because this order is used when applying the rules at run-time.

Pattern can be (for Apache 1.1.x a System V8 and for Apache 1.2.x and later a POSIX) regular expression 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 originally requested URL, because any number of rules may already have matched and made alterations to it.

Some hints about the syntax of regular expressions:

Text:
.           Any single character
[chars]     Character class: One  of chars
[^chars]    Character class: None of chars
  text1|text2 Alternative: text1 or text2

Quantifiers:
?           0 or 1 of the preceding text
*           0 or N of the preceding text (N > 0)
+           1 or N of the preceding text (N > 1)

Grouping:
(text)      Grouping of text
              (either to set the borders of an alternative or
              for making backreferences where the Nth group can 
              be used on the RHS of a RewriteRule with $N)

Anchors:
^           Start of line anchor
$           End   of line anchor

Escaping:
\char       escape that particular char
              (for instance to specify the chars ".[]()" etc.)

For more information about regular expressions either have a look at your local regex(3) manpage or its src/regex/regex.3 copy in the Apache 1.3 distribution. If you are interested in more detailed information about regular expressions and their variants (POSIX regex, Perl regex, etc.) have a look at the following dedicated book on this topic:

Mastering Regular Expressions
Jeffrey E.F. Friedl
Nutshell Handbook Series
O'Reilly & Associates, Inc. 1997
ISBN 1-56592-257-3

Additionally in mod_rewrite the NOT character ('!') is a possible pattern prefix. This gives you the ability to negate a pattern; to say, for instance: ``if the current URL does NOT match this pattern''. This can be used for exceptional cases, where it is easier to match the negative pattern, or as a last default rule.

Notice: 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 $N in the substitution string!

Substitution of a rewriting rule is the string which is substituted for (or replaces) the original URL for which Pattern matched. Beside plain text you can use

  1. back-references $N to the RewriteRule pattern
  2. back-references %N to the last matched RewriteCond pattern
  3. server-variables as in rule condition test-strings (%{VARNAME})
  4. mapping-function calls (${mapname:key|default})
Back-references are $N (N=0..9) identifiers which will be replaced by the contents of the Nth group of the matched Pattern. The server-variables are the same as for the TestString of a RewriteCond directive. The mapping-functions come from the RewriteMap directive and are explained there. These three types of variables are expanded in the order of the above list.

As already mentioned above, all the rewriting rules are applied to the Substitution (in the order of definition in the config file). The URL is completely replaced by the Substitution and the rewriting process goes on until there are no more rules unless explicitly terminated by a L flag - see below.

There is a special substitution string named '-' which means: NO substitution! Sounds silly? No, it is useful to provide rewriting rules which only match some URLs but do no substitution, e.g., in conjunction with the C (chain) flag to be able to have more than one pattern to be applied before a substitution occurs.

One more note: You can even create URLs in the substitution string containing a query string part. Just use a question mark inside the substitution string to indicate that the following stuff should be re-injected into the QUERY_STRING. When you want to erase an existing query string, end the substitution string with just the question mark.

Note: There is a special feature: When you prefix a substitution field with http://thishost[:thisport] then mod_rewrite 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.
Remember: An unconditional external redirect to your own server will not work with the prefix http://thishost because of this feature. To achieve such a self-redirect, you have to use the R-flag (see below).

Additionally you can set special flags for Substitution by appending

[flags]
as the third argument to the RewriteRule directive. Flags is a comma-separated list of the following flags:
Note: Never forget that Pattern is applied to a complete URL in per-server configuration files. But in per-directory configuration files, the per-directory prefix (which always is the same for a specific directory!) is automatically removed for the pattern matching and automatically added after the substitution has been done. 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.

There is one exception: If a substitution string starts with ``http://'' then the directory prefix will not be added and an external redirect or proxy throughput (if flag P is used!) is forced!

Note: To enable the rewriting engine for per-directory configuration files you need to set ``RewriteEngine On'' in these files and ``Options FollowSymLinks'' must be enabled. If your administrator has disabled override of FollowSymLinks for a user's directory, then you cannot use the rewriting engine. This restriction is needed for security reasons.

Here are all possible substitution combinations and their meanings:

Inside per-server configuration (httpd.conf)
for request ``GET /somepath/pathinfo'':

Given Rule                                      Resulting Substitution
----------------------------------------------  ----------------------------------
^/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                     /otherpath/pathinfo

^/somepath(.*) /otherpath$1 [R]                 http://thishost/otherpath/pathinfo
                                                via external redirection

^/somepath(.*) /otherpath$1 [P]                 not supported, because silly!
----------------------------------------------  ----------------------------------
^/somepath(.*) http://thishost/otherpath$1      /otherpath/pathinfo

^/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

Inside per-directory configuration for /somepath
(i.e., file .htaccess in dir /physical/path/to/somepath containing RewriteBase /somepath)
for request ``GET /somepath/localpath/pathinfo'':

Given Rule                                      Resulting Substitution