URL Rewrites¶

• Table of contents
• URL Rewrites
  • Description
  • Options
    • url.rewrite-once
    • url.rewrite-repeat
    • url.rewrite-[repeat-]if-not-file
  • Regular Expressions
    • Replacement Patterns
    • Extended Replacement Patterns
  • Examples
    • With mod_redirect
    • Workaround for "File name too long" on Windows
    • Passing / Matching the Query string (GET variables)

Module: mod_rewrite

Description¶

internal redirects, url rewrite

Note: regex matching is against the full REQUEST_URI, including optional path-info and query-string.

The result of a rewritten url-path must begin with '/' (enforced since 1.4.50)

Options¶

url.rewrite-once¶

Rewrites a set of URLs internally in the webserver BEFORE they are handled.

url.rewrite-once = ( "<regex>" => "<relative-uri>" )

or for multiple rules

url.rewrite-once = ( 
  "<regex1>" => "<relative-uri1>",
  "<regex2>" => "<relative-uri2>" 
)

url.rewrite-repeat¶

Rewrites a set of URLs internally in the webserver BEFORE they are handled

url.rewrite-repeat = ( "<regex>" => "<relative-uri>" )

The difference between these options is that while url.rewrite-repeat allows for applying multiple (separately defined) rewrite rules in a row, url.rewrite-once will cause further rewrite rules to be skipped if the expression was matched. As such, url.rewrite-once behaves like Apache's RewriteRule ... [L]: http://httpd.apache.org/docs/2.2/mod/mod_rewrite.html#rewriterule

url.rewrite-[repeat-]if-not-file¶

Rewrites a set of URLs internally in the webserver BEFORE they are handled unless the target exists and is a regular file (not a directory, pipe, socket, or other). This is akin to Apache´s !-f RewriteRule.

Note: All URLs with path-info appended to the end of the URL will be treated as file not existing. This is useful with frameworks where lighttpd should serve static files (which exist and are files), and everything else should be rewritten to use the dynamic backend, e.g. Drupal, or other /index.php

url.rewrite-if-not-file is a simple solution for many frameworks
e.g. have lighttpd serve static files (such as .js, .css, .jpg, etc) if file exists, or else rewrite to send request to /index.php
url.rewrite-if-not-file = ( "" => "/index.php?path=${url.path}${qsa}" ) # option, if index.php expects original URL in query string 'path=...'
url.rewrite-if-not-file = ( "" => "/index.php${url.path}${qsa}" ) # alternative, if index.php expects original URL in PATH_INFO

(As an alternative to mod_rewrite, mod_magnet lua code can be used to fully implement Apache´s -f and -d solutions.)

Regular Expressions¶

• Patterns ("wildcards") are matched against a string
• Special characters (see [http://www.regular-expressions.info/reference.html] for reference):

• . (full stop) - match any character
• * (asterisk) - match zero or more of the previous symbol
• + (plus) - match one or more of the previous symbol
• ? (question) - match zero or one of the previous symbol
• \? (backslash-something) - match special characters
• ^ (caret) - match the start of a string
• $ (dollar) - match the end of a string
• [set] - match any one of the symbols inside the square braces.
• [^set] - match any symbol that is NOT inside the square braces.
• (pattern) - grouping, remember what the pattern matched as a special variable
• {n,m} - from n to m times matching the previous character (m could be omitted to mean >=n times)
• (?!expression) - match anything BUT expression at the current position. Example: "^(/(?!(favicon.ico$|js/|images/)).*)" => "/fcgi/$1"

• Normal alphanumeric characters are treated as normal

Replacement Patterns¶

If the matched regex contains groups in parentheses, $1..$9 in the replacement refer to the captured text in the
matching group "$1" meaning the first group, "$2" the second, and so on.

Note that % replacements (like %1, %2, %0, etc.) in url.rewrite-* targets are permitted, but do not have the meaning they would have in evhost.path-pattern. If url.rewrite-* is specified within a regex conditional, % patterns are replaced by the corresponding groups from the condition regex. %1 is replaced with the first subexpression, %2 with the second, etc. %0 is replaced by the entire substring matching the regexp. See below for an example using "%0".

Extended Replacement Patterns¶

lighttpd provides ways to encode redirect and rewrite backreference substitutions (since 1.4.50) in curly-braces %{...} or ${...}

Up to nine (9) matches are saved for %{1} - %{9}. Up to nineteen (19) matches are saved for ${1} - ${19}.

In addition to %1 and $1, the following modifiers are now supported, followed by the number for the backreference, e.g. ${esc:1}.

• ${noesc:...} no escaping
• ${esc:...} escape all non-alphanumeric - . _ ~ incl double-escape %
• ${escape:...} escape all non-alphanumeric - . _ ~ incl double-escape %
• ${escnde:...} escape all non-alphanumeric - . _ ~ but no double-esc %
• ${escpsnde:...} escape all non-alphanumeric - . _ ~ / but no double-esc % (preserves literal /)
• ${tolower:...}
• ${toupper:...}
• ${encb64u:...} encode to base64url characters (no-padding)
• ${decb64u:...} decode from base64url characters

• %{noesc:...}
• %{esc:...}
• %{escape:...}
• %{escnde:...}
• %{escpsnde:...}
• %{tolower:...}
• %{toupper:...}
• %{encb64u:...}
• %{decb64u:...}

lighttpd provides ways to substitute URI parts without needing a regex match (since 1.4.50) (and can be preceded by encoding modifier, e.g. ${tolower:url.authority})

• ${url.scheme}
• ${url.authority}
• ${url.port}
• ${url.path}
• ${url.query}

• ${qsa} appends query string, if not empty

Examples¶

Note: regex matching is against the full REQUEST_URI, including optional path-info and query-string.

# the following example, is, however just simulating vhost by rewrite
# * you can never change document-root by mod_rewrite
# use mod_*host instead to make real mass-vhost

server.document-root = "/www/htdocs/" 
$HTTP["host"] =~ "^.*\.([^.]+\.com)$" {
  url.rewrite-once = ( "^/(.*)" => "/%0/$1" )
}

# request:        http://any.domain.com/url/ 
# before rewrite: REQUEST_URI="/www/htdocs/url/" 
# and DOCUMENT_ROOT="/www/htdocs/" %0="any.domain.com" $1="url/" 
# after rewrite:  REQUEST_URI="/www/htdocs/any.domain.com/url/" 
# still, you have DOCUMENT_ROOT=/www/htdocs/

# please note, that we have two regular expressions: the one which 
# $HTTP["host"] is been compared with, and the one of the rewrite rule.
# the numbered subexpressions available to build the relative uri are
# being prefixed by '%' for subexpressions of the first regular expression 
# match and by '$' for subexpressions of the second one.
# subexpression 0 interpolates the whole matching string: %0 for the whole
# string matching the conditional, and $0 for the whole string matching the
# rewrite rule.

# if the rewrite rule is not included in a conditional 
# block, only the '$' prefixed variables are available.

url.rewrite-once = ( "^/id/([0-9]+)$" => "/index.php?id=$1",
                     "^/link/([a-zA-Z]+)" => "/index.php?link=$1" )

With mod_redirect¶

Rewrite rules always execute before redirect rules. This is true regardless of the order of module loading or the order of rules in the configuration (lighttpd v1.4.13). However, mod_rewrite provides a mechanism to pass URLs through unmangled: specify "$0" as the rule target, but be sure that the rule matches the entire string since $0 is the entire matched string.

url.rewrite-once = (
    "^/foo"  => "$0",
    "^/(.*)" => "/handler/$1" 
)

url.redirect = (
    "^/foo"  => "http://foo.bar/" 
)

Since version 1.4.40, an alternative is to specify a blank target to the rewrite rule. This will cause the matched rule to leave the url unmodified, and will skip any further rewrite rules.

url.rewrite-once = (
    "^/foo"  => "",   # instead of (nonsensical) blank url, the url will not be modified
    "^/(.*)" => "/handler/$1" 
)

Workaround for "File name too long" on Windows¶

While running Lighttpd on Windows you may get 500 Internal Server Error if computed filename is longer than 255 characters.
In error log it will be (response.c.537) file not found ... or so: File name too long /very_looooong_path ->.
As workaround you can use mod_rewrite to avoid this error.

server.modules += ("mod_rewrite")
url.rewrite-once = ( ".{250,}" => "/toolong.php" )

If error handler is PHP, $_SERVER['REQUEST_URI'] will contain full URI.

Passing / Matching the Query string (GET variables)¶

If you want to pass the Query String (?foo=bar) to the rewrite destination you can use ${qsa} with modern versions of lighttpd ("qsa" is short for "query-string-append")
url.rewrite-once = ( "^/news/([^\?]+)" => "/news.php?title=$1${qsa}" )
With older versions of lighttpd, you may have to explicitly match the query string:
url.rewrite-once = ( "^/news/([^\?]+)(?:\?(.*))?" => "/news.php?title=$1&$2" )