Apache::JumpByReferer - Jump or block by Referer header field
You need to compile mod_perl with PERL_ACCESS (or EVERYTHING) enabled. And write the setting like below:
<Directory /protected/directory/> PerlAccessHandler Apache::JumpByReferer PerlSetVar RefererListFile conf/jump.conf PerlSetVar JumpByReferer On </Directory>
in RefererListFile (conf/jump.conf)
# Syntax: # Referer Regex URL to Jump (or forbidden) http://malicious\.site\.example\.com/ http://goodbye.example.com/ http://another\.malicious\.site/ forbidden http://ime\.nu/ forbidden http://[^.]+\.google\.([^/]+)/ /hello_googler.html http://[^.]+\.yahoo\.([^/]+)/ /do_you_yahoo/? "Field blocked by" /do/not/block/the/field/
Apache::JumpByReferer is an access phase handler of Apache + mod_perl. You can block or let the user jump to another URL if the user was coming from your specified web site.
This handler will cache the settings at the first time calling, and check the last-modified time of the
RefererListFile, and will re-cache the settings when the file was modified after the last cached time. That is to say, you can always rewrite to change your settings.
Write a regular expressions (REGEX) of your specified URL in the
RefererListFile, and join a URL to jump behind the REGEX in the blank(s) if you want to let jump the user. Write and join
Forbidden (case insensitive) string behind the REGEX in the blank(s) if you don't want to access to the directory.
JumpByReferer is a switch to work of this handler. You must write
Off (case insensitive) value to this directive. It returns DECLINED and will not work if the directive is not set to
You must write your settings to
RefererListFile. The file must be readable for the user and group of settings of Apache
You should write a valid REGEX from line-head. The REGEX will be tested to check it is valid, and cached on this namespace with timestamp of the modified time. But won't be cached if the REGEX is invalid.
and write the jumping URL or
Forbidden behind the REGEX.
If you want to write a REGEX which is including some space character, you have to quote it. If you don't quote it, the parsing of the text is failed.
This handler will do
internal_redirect() if the URL is internal of own server. And this handler will print a
Location header with
REDIRECT status if the URL is external. The judgement of it, an external URL is started from
http(s)://, or an internal URI is started from others.
The comment line is started by
# character, and it will be ignored a blank line.
This handler applies as for the initial request, namely does not work for sub request and internal redirect. Because the server may be fallen into endless loop if it applies for them. Almost every non-initial requests are had same Referer header field as initial request. This handler calls the internal redirect, and this handler will redirect to current URI again and again when the handler applies for non-initial request too and the directory setting of redirection URI is inherited the parent settings. Because this case is inside of the effective range of the JumpByReferer setting. Perhaps your system resources are run through in an instant by the handler when the handler works for non-initial request too. Therefore this function (does not work for non-initial request) exists for self-defense. There is probably no problem in this setting. But, understand these things fully when you use this module, please.
* I should know how to use Apache::test for testing this module.
* I should know whether it's the best way to put the handler under
* Probably, I should study English more to write the document neatly.
mod_perl(1), Apache(3), Text::ParseWords
Koichi Taniguchi <email@example.com>
Copyright (c) 2003 Koichi Taniguchi. Japan. All rights reserved.
This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.