-const char filters_rcs[] = "$Id: filters.c,v 1.13 2001/05/31 21:21:30 jongfoster Exp $";
+const char filters_rcs[] = "$Id: filters.c,v 1.107 2008/05/04 17:52:56 fabiankeil Exp $";
/*********************************************************************
*
* File : $Source: /cvsroot/ijbswa/current/filters.c,v $
* Purpose : Declares functions to parse/crunch headers and pages.
* Functions declared include:
* `acl_addr', `add_stats', `block_acl', `block_imageurl',
- * `block_url', `url_actions', `domaincmp', `dsplit',
+ * `block_url', `url_actions', `domain_split',
* `filter_popups', `forward_url', 'redirect_url',
- * `ij_untrusted_url', `intercept_url', `re_process_buffer',
- * `show_proxy_args', 'ijb_send_banner', and `trust_url'
+ * `ij_untrusted_url', `intercept_url', `pcrs_filter_respose',
+ * `ijb_send_banner', `trust_url', `gif_deanimate_response',
+ * `execute_single_pcrs_command', `rewrite_url',
+ * `get_last_url'
*
- * Copyright : Written by and Copyright (C) 2001 the SourceForge
- * IJBSWA team. http://ijbswa.sourceforge.net
+ * Copyright : Written by and Copyright (C) 2001, 2004-2008 the SourceForge
+ * Privoxy team. http://www.privoxy.org/
*
* Based on the Internet Junkbuster originally written
- * by and Copyright (C) 1997 Anonymous Coders and
+ * by and Copyright (C) 1997 Anonymous Coders and
* Junkbusters Corporation. http://www.junkbusters.com
*
- * This program is free software; you can redistribute it
+ * This program is free software; you can redistribute it
* and/or modify it under the terms of the GNU General
* Public License as published by the Free Software
* Foundation; either version 2 of the License, or (at
*
* Revisions :
* $Log: filters.c,v $
+ * Revision 1.107 2008/05/04 17:52:56 fabiankeil
+ * Adjust parse_http_url() call to new prototype.
+ *
+ * Revision 1.106 2008/05/03 16:40:44 fabiankeil
+ * Change content_filters_enabled()'s parameter from
+ * csp->action to action so it can be also used in the
+ * CGI code. Don't bother checking if there are filters
+ * loaded, as that's somewhat besides the point.
+ *
+ * Revision 1.105 2008/03/28 15:13:39 fabiankeil
+ * Remove inspect-jpegs action.
+ *
+ * Revision 1.104 2008/03/27 18:27:24 fabiankeil
+ * Remove kill-popups action.
+ *
+ * Revision 1.103 2008/03/06 16:33:45 fabiankeil
+ * If limit-connect isn't used, don't limit CONNECT requests to port 443.
+ *
+ * Revision 1.102 2008/03/01 14:00:44 fabiankeil
+ * Let the block action take the reason for the block
+ * as argument and show it on the "blocked" page.
+ *
+ * Revision 1.101 2008/02/23 16:57:12 fabiankeil
+ * Rename url_actions() to get_url_actions() and let it
+ * use the standard parameter ordering.
+ *
+ * Revision 1.100 2008/02/23 16:33:43 fabiankeil
+ * Let forward_url() use the standard parameter ordering
+ * and mark its second parameter immutable.
+ *
+ * Revision 1.99 2008/02/03 13:57:58 fabiankeil
+ * Add SOCKS5 support for forward-override{}.
+ *
+ * Revision 1.98 2008/01/04 17:43:45 fabiankeil
+ * Improve the warning messages that get logged if the action files
+ * "enable" filters but no filters of that type have been loaded.
+ *
+ * Revision 1.97 2007/11/30 15:37:03 fabiankeil
+ * Use freez instead of free.
+ *
+ * Revision 1.96 2007/10/19 16:53:28 fabiankeil
+ * Add helper function to check if any content filters are enabled.
+ *
+ * Revision 1.95 2007/10/17 19:31:20 fabiankeil
+ * Omitting the zero chunk that ends the chunk transfer encoding seems
+ * to be the new black. Log the problem and continue filtering anyway.
+ *
+ * Revision 1.94 2007/09/29 13:20:20 fabiankeil
+ * Remove two redundant and one useless log messages.
+ *
+ * Revision 1.93 2007/09/29 10:21:16 fabiankeil
+ * - Move get_filter_function() from jcc.c to filters.c
+ * so the filter functions can be static.
+ * - Don't bother filtering body-less responses.
+ *
+ * Revision 1.92 2007/09/28 16:38:55 fabiankeil
+ * - Execute content filters through execute_content_filter().
+ * - Add prepare_for_filtering() so filter functions don't have to
+ * care about de-chunking and decompression. As a side effect this enables
+ * decompression for gif_deanimate_response() and jpeg_inspect_response().
+ * - Change remove_chunked_transfer_coding()'s return type to jb_err.
+ * Some clowns feel like chunking empty responses in which case
+ * (size == 0) is valid but previously would be interpreted as error.
+ *
+ * Revision 1.91 2007/09/02 15:31:20 fabiankeil
+ * Move match_portlist() from filter.c to urlmatch.c.
+ * It's used for url matching, not for filtering.
+ *
+ * Revision 1.90 2007/09/02 12:44:17 fabiankeil
+ * Remove newline at the end of a log_error() message.
+ *
+ * Revision 1.89 2007/08/05 13:42:23 fabiankeil
+ * #1763173 from Stefan Huehner: declare some more functions static.
+ *
+ * Revision 1.88 2007/06/01 16:41:11 fabiankeil
+ * Add forward-override{} to change the forwarding settings through
+ * action sections. This is mainly interesting to forward different
+ * clients differently (for example based on User-Agent or request
+ * origin).
+ *
+ * Revision 1.87 2007/04/30 15:53:10 fabiankeil
+ * Make sure filters with dynamic jobs actually use them.
+ *
+ * Revision 1.86 2007/04/30 15:03:28 fabiankeil
+ * - Introduce dynamic pcrs jobs that can resolve variables.
+ * - Don't run redirect functions more than once,
+ * unless they are activated more than once.
+ *
+ * Revision 1.85 2007/03/21 12:24:47 fabiankeil
+ * - Log the content size after decompression in decompress_iob()
+ * instead of pcrs_filter_response().
+ *
+ * Revision 1.84 2007/03/20 15:16:34 fabiankeil
+ * Use dedicated header filter actions instead of abusing "filter".
+ * Replace "filter-client-headers" and "filter-client-headers"
+ * with "server-header-filter" and "client-header-filter".
+ *
+ * Revision 1.83 2007/03/17 15:20:05 fabiankeil
+ * New config option: enforce-blocks.
+ *
+ * Revision 1.82 2007/03/13 11:28:43 fabiankeil
+ * - Fix port handling in acl_addr() and use a temporary acl spec
+ * copy so error messages don't contain a truncated version.
+ * - Log size of iob before and after decompression.
+ *
+ * Revision 1.81 2007/03/05 14:40:53 fabiankeil
+ * - Cosmetical changes for LOG_LEVEL_RE_FILTER messages.
+ * - Hide the "Go there anyway" link for blocked CONNECT
+ * requests where going there anyway doesn't work anyway.
+ *
+ * Revision 1.80 2007/02/07 10:55:20 fabiankeil
+ * - Save the reason for generating http_responses.
+ * - Block (+block) with status code 403 instead of 404.
+ * - Use a different kludge to remember a failed decompression.
+ *
+ * Revision 1.79 2007/01/31 16:21:38 fabiankeil
+ * Search for Max-Forwards headers case-insensitive,
+ * don't generate the "501 unsupported" message for invalid
+ * Max-Forwards values and don't increase negative ones.
+ *
+ * Revision 1.78 2007/01/28 13:41:18 fabiankeil
+ * - Add HEAD support to finish_http_response.
+ * - Add error favicon to internal HTML error messages.
+ *
+ * Revision 1.77 2007/01/12 15:36:44 fabiankeil
+ * Mark *csp as immutable for is_untrusted_url()
+ * and is_imageurl(). Closes FR 1237736.
+ *
+ * Revision 1.76 2007/01/01 19:36:37 fabiankeil
+ * Integrate a modified version of Wil Mahan's
+ * zlib patch (PR #895531).
+ *
+ * Revision 1.75 2006/12/29 18:30:46 fabiankeil
+ * Fixed gcc43 conversion warnings,
+ * changed sprintf calls to snprintf.
+ *
+ * Revision 1.74 2006/12/24 17:37:38 fabiankeil
+ * Adjust comment in pcrs_filter_response()
+ * to recent pcrs changes. Hohoho.
+ *
+ * Revision 1.73 2006/12/23 16:01:02 fabiankeil
+ * Don't crash if pcre returns an error code
+ * that pcrs didn't expect. Fixes BR 1621173.
+ *
+ * Revision 1.72 2006/12/22 18:52:53 fabiankeil
+ * Modified is_untrusted_url to complain in case of
+ * write errors and to give a reason when adding new
+ * entries to the trustfile. Closes FR 1097611.
+ *
+ * Revision 1.71 2006/12/22 14:24:52 fabiankeil
+ * Skip empty filter files in pcrs_filter_response,
+ * but don't ignore the ones that come afterwards.
+ * Fixes parts of BR 1619208.
+ *
+ * Revision 1.70 2006/12/09 13:33:15 fabiankeil
+ * Added some sanity checks for get_last_url().
+ * Fixed possible segfault caused by my last commit.
+ *
+ * Revision 1.69 2006/12/08 12:39:13 fabiankeil
+ * Let get_last_url() catch https URLs as well.
+ *
+ * Revision 1.68 2006/12/05 14:45:48 fabiankeil
+ * Make sure get_last_url() behaves like advertised
+ * and fast-redirects{} can be combined with redirect{}.
+ *
+ * Revision 1.67 2006/11/28 15:19:43 fabiankeil
+ * Implemented +redirect{s@foo@bar@} to generate
+ * a redirect based on a rewritten version of the
+ * original URL.
+ *
+ * Revision 1.66 2006/09/23 13:26:38 roro
+ * Replace TABs by spaces in source code.
+ *
+ * Revision 1.65 2006/09/21 12:54:43 fabiankeil
+ * Fix +redirect{}. Didn't work with -fast-redirects.
+ *
+ * Revision 1.64 2006/08/31 10:55:49 fabiankeil
+ * Block requests for untrusted URLs with status
+ * code 403 instead of 200.
+ *
+ * Revision 1.63 2006/08/31 10:11:28 fabiankeil
+ * Don't free p which is still in use and will be later
+ * freed by free_map(). Don't claim the referrer is unknown
+ * when the client didn't set one.
+ *
+ * Revision 1.62 2006/08/14 00:27:47 david__schmidt
+ * Feature request 595948: Re-Filter logging in single line
+ *
+ * Revision 1.61 2006/08/03 02:46:41 david__schmidt
+ * Incorporate Fabian Keil's patch work:\rhttp://www.fabiankeil.de/sourcecode/privoxy/
+ *
+ * Revision 1.60 2006/07/18 14:48:46 david__schmidt
+ * Reorganizing the repository: swapping out what was HEAD (the old 3.1 branch)
+ * with what was really the latest development (the v_3_0_branch branch)
+ *
+ * Revision 1.58.2.9 2006/01/29 23:10:56 david__schmidt
+ * Multiple filter file support
+ *
+ * Revision 1.58.2.8 2005/05/07 21:50:55 david__schmidt
+ * A few memory leaks plugged (mostly on error paths)
+ *
+ * Revision 1.58.2.7 2004/10/03 12:53:32 david__schmidt
+ * Add the ability to check jpeg images for invalid
+ * lengths of comment blocks. Defensive strategy
+ * against the exploit:
+ * Microsoft Security Bulletin MS04-028
+ * Buffer Overrun in JPEG Processing (GDI+) Could
+ * Allow Code Execution (833987)
+ * Enabled with +inspect-jpegs in actions files.
+ *
+ * Revision 1.58.2.6 2003/12/06 22:18:27 gliptak
+ * Correcting compile problem with FEATURE_IMAGE_BLOCKING
+ *
+ * Revision 1.58.2.5 2003/11/11 13:10:31 oes
+ * Fixed bug #839859: "See why" link URL now gets url-encoded.
+ *
+ * Revision 1.58.2.4 2003/02/28 12:52:45 oes
+ * Fixed a typo
+ *
+ * Revision 1.58.2.3 2002/09/25 14:51:51 oes
+ * Added basic support for OPTIONS and TRACE HTTP methods:
+ * New function direct_response which handles OPTIONS and
+ * TRACE requests whose Max-Forwards header field is zero.
+ *
+ * Revision 1.58.2.2 2002/08/01 17:18:28 oes
+ * Fixed BR 537651 / SR 579724 (MSIE image detect improper for IE/Mac)
+ *
+ * Revision 1.58.2.1 2002/07/26 15:18:53 oes
+ * - Bugfix: Executing a filters without jobs no longer results in
+ * turing off *all* filters.
+ * - Security fix: Malicious web servers can't cause a seg fault
+ * through bogus chunk sizes anymore
+ *
+ * Revision 1.58 2002/04/24 02:11:17 oes
+ * Jon's multiple AF patch: url_actions now evaluates rules
+ * from all AFs.
+ *
+ * Revision 1.57 2002/04/08 20:38:34 swa
+ * fixed JB spelling
+ *
+ * Revision 1.56 2002/04/05 15:51:24 oes
+ * - bugfix: error-pages now get correct request protocol
+ * - fix for invalid HTML in trust info
+ *
+ * Revision 1.55 2002/04/02 16:13:51 oes
+ * Fix: No "Go there anyway" for SSL
+ *
+ * Revision 1.54 2002/04/02 14:55:56 oes
+ * Bugfix: is_untrusted_url() now depends on FEATURE_TRUST, not FEATURE_COOKIE_JAR
+ *
+ * Revision 1.53 2002/03/26 22:29:54 swa
+ * we have a new homepage!
+ *
+ * Revision 1.52 2002/03/24 16:35:57 jongfoster
+ * Removing logo
+ *
+ * Revision 1.51 2002/03/24 15:23:33 jongfoster
+ * Name changes
+ *
+ * Revision 1.50 2002/03/24 13:25:43 swa
+ * name change related issues
+ *
+ * Revision 1.49 2002/03/16 20:29:14 oes
+ * Cosmetics
+ *
+ * Revision 1.48 2002/03/13 20:25:34 oes
+ * Better logging for content filters
+ *
+ * Revision 1.47 2002/03/13 00:30:52 jongfoster
+ * Killing warnings
+ * Added option of always sending redirect for imageblock,
+ * currently disabled with #if 0.
+ *
+ * Revision 1.46 2002/03/12 01:42:49 oes
+ * Introduced modular filters
+ *
+ * Revision 1.45 2002/03/08 16:47:50 oes
+ * Added choice beween GIF and PNG built-in images
+ *
+ * Revision 1.44 2002/03/07 03:49:31 oes
+ * - Fixed compiler warnings etc
+ * - Changed built-in images from GIF to PNG
+ * (with regard to Unisys patent issue)
+ * - Added a 4x4 pattern PNG which is less intrusive
+ * than the logo but also clearly marks the deleted banners
+ *
+ * Revision 1.43 2002/01/22 23:51:59 jongfoster
+ * Replacing strsav() with the safer string_append().
+ *
+ * Adding missing html_encode() to error message generators. Where encoded
+ * and unencoded versions of a string were provided, removing the unencoded
+ * one.
+ *
+ * Revision 1.42 2002/01/17 21:00:32 jongfoster
+ * Moving all our URL and URL pattern parsing code to urlmatch.c.
+ *
+ * Using a single, simple url_match(pattern,url) function - rather than
+ * the 3-line match routine which was repeated all over the place.
+ *
+ * Renaming free_url to free_url_spec, since it frees a struct url_spec.
+ *
+ * Using parse_http_url() to parse URLs without faking a HTTP
+ * request line for parse_http_request().
+ *
+ * Revision 1.41 2001/11/13 00:14:07 jongfoster
+ * Fixing stupid bug now I've figured out what || means.
+ * (It always returns 0 or 1, not one of it's paramaters.)
+ *
+ * Revision 1.40 2001/10/26 17:37:55 oes
+ * - Re-enabled Netscape 200/404 bug workaround in block_url():
+ * - Removed OS/2 special case
+ * - Made block_url() independant from sed() having been run
+ * - Made trust_url independant from sed() having been run
+ * - Made is_imageurl independant from sed() having been run.
+ * It now checks User-Agent: and Accept: by itself.
+ *
+ *
+ * Revision 1.39 2001/10/25 03:40:48 david__schmidt
+ * Change in porting tactics: OS/2's EMX porting layer doesn't allow multiple
+ * threads to call select() simultaneously. So, it's time to do a real, live,
+ * native OS/2 port. See defines for __EMX__ (the porting layer) vs. __OS2__
+ * (native). Both versions will work, but using __OS2__ offers multi-threading.
+ *
+ * Revision 1.38 2001/10/23 21:32:33 jongfoster
+ * Adding error-checking to selected functions
+ *
+ * Revision 1.37 2001/10/22 15:33:56 david__schmidt
+ * Special-cased OS/2 out of the Netscape-abort-on-404-in-js problem in
+ * filters.c. Added a FIXME in front of the offending code. I'll gladly
+ * put in a better/more robust fix for all parties if one is presented...
+ * It seems that just returning 200 instead of 404 would pretty much fix
+ * it for everyone, but I don't know all the history of the problem.
+ *
+ * Revision 1.36 2001/10/10 16:44:16 oes
+ * Added match_portlist function
+ *
+ * Revision 1.35 2001/10/07 15:41:23 oes
+ * Replaced 6 boolean members of csp with one bitmap (csp->flags)
+ *
+ * New function remove_chunked_transfer_coding that strips chunked
+ * transfer coding to plain and is called by pcrs_filter_response
+ * and gif_deanimate_response if neccessary
+ *
+ * Improved handling of zero-change re_filter runs
+ *
+ * pcrs_filter_response and gif_deanimate_response now remove
+ * chunked transfer codeing before processing the body.
+ *
+ * Revision 1.34 2001/09/20 15:49:36 steudten
+ *
+ * Fix BUG: Change int size to size_t size in pcrs_filter_response().
+ * See cgi.c fill_template().
+ *
+ * Revision 1.33 2001/09/16 17:05:14 jongfoster
+ * Removing unused #include showarg.h
+ *
+ * Revision 1.32 2001/09/16 13:21:27 jongfoster
+ * Changes to use new list functions.
+ *
+ * Revision 1.31 2001/09/16 11:38:02 jongfoster
+ * Splitting fill_template() into 2 functions:
+ * template_load() loads the file
+ * template_fill() performs the PCRS regexps.
+ * This is because the CGI edit interface has a "table row"
+ * template which is used many times in the page - this
+ * change means it's only loaded from disk once.
+ *
+ * Revision 1.30 2001/09/16 11:00:10 jongfoster
+ * New function alloc_http_response, for symmetry with free_http_response
+ *
+ * Revision 1.29 2001/09/13 23:32:40 jongfoster
+ * Moving image data to cgi.c rather than cgi.h
+ * Fixing a GPF under Win32 (and any other OS that protects global
+ * constants from being written to).
+ *
+ * Revision 1.28 2001/09/10 10:18:51 oes
+ * Silenced compiler warnings
+ *
+ * Revision 1.27 2001/08/05 16:06:20 jongfoster
+ * Modifiying "struct map" so that there are now separate header and
+ * "map_entry" structures. This means that functions which modify a
+ * map no longer need to return a pointer to the modified map.
+ * Also, it no longer reverses the order of the entries (which may be
+ * important with some advanced template substitutions).
+ *
+ * Revision 1.26 2001/07/30 22:08:36 jongfoster
+ * Tidying up #defines:
+ * - All feature #defines are now of the form FEATURE_xxx
+ * - Permanently turned off WIN_GUI_EDIT
+ * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
+ *
+ * Revision 1.25 2001/07/26 10:09:46 oes
+ * Made browser detection a little less naive
+ *
+ * Revision 1.24 2001/07/25 17:22:51 oes
+ * Added workaround for Netscape bug that prevents display of page when loading a component fails.
+ *
+ * Revision 1.23 2001/07/23 13:40:12 oes
+ * Fixed bug that caused document body to be dropped when pcrs joblist was empty.
+ *
+ * Revision 1.22 2001/07/18 12:29:34 oes
+ * - Made gif_deanimate_response respect
+ * csp->action->string[ACTION_STRING_DEANIMATE]
+ * - Logging cosmetics
+ *
+ * Revision 1.21 2001/07/13 13:59:53 oes
+ * - Introduced gif_deanimate_response which shares the
+ * generic content modification interface of pcrs_filter_response
+ * and acts as a wrapper to deanimate.c:gif_deanimate()
+ * - Renamed re_process_buffer to pcrs_filter_response
+ * - pcrs_filter_response now returns NULL on failiure
+ * - Removed all #ifdef PCRS
+ *
+ * Revision 1.20 2001/07/01 17:01:04 oes
+ * Added comments and missing return statement in is_untrusted_url()
+ *
+ * Revision 1.19 2001/06/29 21:45:41 oes
+ * Indentation, CRLF->LF, Tab-> Space
+ *
+ * Revision 1.18 2001/06/29 13:27:38 oes
+ * - Cleaned up, renamed and reorderd functions
+ * and improved comments
+ *
+ * - block_url:
+ * - Ported to CGI platform. Now delivers
+ * http_response or NULL
+ * - Unified HTML and GIF generation (moved image detection
+ * and GIF generation here from jcc.c:chat())
+ * - Fixed HTTP status to:
+ * - 403 (Forbidden) for the "blocked" HTML message
+ * - 200 (OK) for GIF answers
+ * - 302 (Redirect) for redirect to GIF
+ *
+ * - trust_url:
+ * - Ported to CGI platform. Now delivers
+ * http_response or NULL
+ * - Separated detection of untrusted URL into
+ * (bool)is_untrusted_url
+ * - Added enforcement of untrusted requests
+ *
+ * - Moved redirect_url() from cgi.c to here
+ * and ported it to the CGI platform
+ *
+ * - Removed logentry from cancelled commit
+ *
+ * Revision 1.17 2001/06/09 10:55:28 jongfoster
+ * Changing BUFSIZ ==> BUFFER_SIZE
+ *
+ * Revision 1.16 2001/06/07 23:10:26 jongfoster
+ * Allowing unanchored domain patterns to back off and retry
+ * if they partially match. Optimized right-anchored patterns.
+ * Moving ACL and forward files into config file.
+ * Replacing struct gateway with struct forward_spec
+ *
+ * Revision 1.15 2001/06/03 19:12:00 oes
+ * extracted-CGI relevant stuff
+ *
+ * Revision 1.14 2001/06/01 10:30:55 oes
+ * Added optional left-anchoring to domaincmp
+ *
* Revision 1.13 2001/05/31 21:21:30 jongfoster
* Permissionsfile / actions file changes:
* - Changed "permission" to "action" throughout
* a lot! ;-)
*
* Revision 1.3 2001/05/20 16:44:47 jongfoster
- * Removing last hardcoded JunkBusters.com URLs.
+ * Removing last hardcoded Junkbusters.com URLs.
*
* Revision 1.2 2001/05/20 01:21:20 jongfoster
* Version 2.9.4 checkin.
#include <stdlib.h>
#include <ctype.h>
#include <string.h>
+#include <assert.h>
#ifndef _WIN32
+#ifndef __OS2__
#include <unistd.h>
+#endif /* ndef __OS2__ */
#include <netinet/in.h>
#else
#include <winsock2.h>
-#endif
+#endif /* ndef _WIN32 */
+
+#ifdef __OS2__
+#include <utils.h>
+#endif /* def __OS2__ */
#include "project.h"
#include "filters.h"
#include "encode.h"
-#include "jcc.h"
-#include "showargs.h"
#include "parsers.h"
#include "ssplit.h"
-#include "gateway.h"
-#include "jbsockets.h"
#include "errlog.h"
#include "jbsockets.h"
#include "miscutil.h"
#include "actions.h"
+#include "cgi.h"
+#include "list.h"
+#include "deanimate.h"
+#include "urlmatch.h"
+#include "loaders.h"
#ifdef _WIN32
#include "win32.h"
*/
#define ijb_isdigit(__X) isdigit((int)(unsigned char)(__X))
+static jb_err remove_chunked_transfer_coding(char *buffer, size_t *size);
+static jb_err prepare_for_filtering(struct client_state *csp);
-static const char CBLOCK[] =
-#ifdef AMIGA
- "HTTP/1.0 403 Request for blocked URL\n"
-#else /* ifndef AMIGA */
- "HTTP/1.0 202 Request for blocked URL\n"
-#endif /* ndef AMIGA */
- "Pragma: no-cache\n"
- "Last-Modified: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "Expires: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "Content-Type: text/html\n\n"
- "<html>\n"
- "<head>\n"
- "<title>Internet Junkbuster: Request for blocked URL</title>\n"
- "</head>\n"
- WHITEBG
- "<center><h1>"
- BANNER
- "</h1></center>\n"
- "<p align=center>Your request for <b>%s%s</b>\n"
- "was blocked.<br><a href=\"http://i.j.b/show-url-info?url=%s%s\">See why</a>"
-#ifdef FORCE_LOAD
- " or <a href=\"http://%s" FORCE_PREFIX "%s\">"
- "go there anyway.</a>"
-#endif /* def FORCE_LOAD */
- "</p>\n"
- "</body>\n"
- "</html>\n";
-
-#ifdef TRUST_FILES
-static const char CTRUST[] =
-#ifdef AMIGA
- "HTTP/1.0 403 Request for untrusted URL\n"
-#else /* ifndef AMIGA */
- "HTTP/1.0 202 Request for untrusted URL\n"
-#endif /* ndef AMIGA */
- "Pragma: no-cache\n"
- "Last-Modified: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "Expires: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "Content-Type: text/html\n\n"
- "<html>\n"
- "<head>\n"
- "<title>Internet Junkbuster: Request for untrusted URL</title>\n"
- "</head>\n"
- WHITEBG
- "<center>"
- "<a href=http://i.j.b/ij-untrusted-url?%s+%s+%s>"
- BANNER
- "</a>"
- "</center>"
- "</body>\n"
- "</html>\n";
-#endif /* def TRUST_FILES */
-
-
-#ifdef ACL_FILES
+#ifdef FEATURE_ACL
/*********************************************************************
*
* Function : block_acl
* Returns : 0 = FALSE (don't block) and 1 = TRUE (do block)
*
*********************************************************************/
-int block_acl(struct access_control_addr *dst,
- struct client_state *csp)
+int block_acl(const struct access_control_addr *dst, const struct client_state *csp)
{
- struct file_list *fl;
- struct access_control_list *a, *acl;
+ struct access_control_list *acl = csp->config->acl;
/* if not using an access control list, then permit the connection */
- if (((fl = csp->alist) == NULL) ||
- ((acl = (struct access_control_list *) fl->f) == NULL))
+ if (acl == NULL)
{
return(0);
}
/* search the list */
- for (a = acl->next ; a ; a = a->next)
+ while (acl != NULL)
{
- if ((csp->ip_addr_long & a->src->mask) == a->src->addr)
+ if ((csp->ip_addr_long & acl->src->mask) == acl->src->addr)
{
if (dst == NULL)
{
/* Just want to check if they have any access */
- if (a->action == ACL_PERMIT)
+ if (acl->action == ACL_PERMIT)
{
return(0);
}
}
- else if ( ((dst->addr & a->dst->mask) == a->dst->addr)
- && ((dst->port == a->dst->port) || (a->dst->port == 0)))
+ else if ( ((dst->addr & acl->dst->mask) == acl->dst->addr)
+ && ((dst->port == acl->dst->port) || (acl->dst->port == 0)))
{
- if (a->action == ACL_PERMIT)
+ if (acl->action == ACL_PERMIT)
{
return(0);
}
}
}
}
+ acl = acl->next;
}
return(1);
*
* Function : acl_addr
*
- * Description : Called from `load_aclfile' to parse an ACL address.
+ * Description : Called from `load_config' to parse an ACL address.
*
* Parameters :
* 1 : aspec = String specifying ACL address.
* Returns : 0 => Ok, everything else is an error.
*
*********************************************************************/
-int acl_addr(char *aspec, struct access_control_addr *aca)
+int acl_addr(const char *aspec, struct access_control_addr *aca)
{
- int i, masklength, port;
+ int i, masklength;
+ long port;
char *p;
+ char *acl_spec = NULL;
masklength = 32;
port = 0;
- if ((p = strchr(aspec, '/')))
+ /*
+ * Use a temporary acl spec copy so we can log
+ * the unmodified original in case of parse errors.
+ */
+ acl_spec = strdup(aspec);
+ if (acl_spec == NULL)
{
- *p++ = '\0';
+ /* XXX: This will be logged as parse error. */
+ return(-1);
+ }
+ if ((p = strchr(acl_spec, '/')) != NULL)
+ {
+ *p++ = '\0';
if (ijb_isdigit(*p) == 0)
{
+ freez(acl_spec);
return(-1);
}
masklength = atoi(p);
if ((masklength < 0) || (masklength > 32))
{
+ freez(acl_spec);
return(-1);
}
- if ((p = strchr(aspec, ':')))
+ if ((p = strchr(acl_spec, ':')) != NULL)
{
+ char *endptr;
+
*p++ = '\0';
+ port = strtol(p, &endptr, 10);
- if (ijb_isdigit(*p) == 0)
+ if (port <= 0 || port > 65535 || *endptr != '\0')
{
+ freez(acl_spec);
return(-1);
}
- port = atoi(p);
}
- aca->port = port;
+ aca->port = (unsigned long)port;
- aca->addr = ntohl(resolve_hostname_to_ip(aspec));
+ aca->addr = ntohl(resolve_hostname_to_ip(acl_spec));
+ freez(acl_spec);
- if (aca->addr == -1)
+ if (aca->addr == INADDR_NONE)
{
- log_error(LOG_LEVEL_ERROR, "can't resolve address for %s", aspec);
+ /* XXX: This will be logged as parse error. */
return(-1);
}
return(0);
}
-#endif /* def ACL_FILES */
+#endif /* def FEATURE_ACL */
/*********************************************************************
*
- * Function : block_url
+ * Function : connect_port_is_forbidden
*
- * Description : Called from `chat'. Check to see if we need to block this.
+ * Description : Check to see if CONNECT requests to the destination
+ * port of this request are forbidden. The check is
+ * independend of the actual request method.
*
* Parameters :
- * 1 : http = http_request request to "check" for blocked
- * 2 : csp = Current client state (buffers, headers, etc...)
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : NULL => unblocked, else string to HTML block description.
+ * Returns : True if yes, false otherwise.
*
*********************************************************************/
-char *block_url(struct http_request *http, struct client_state *csp)
+int connect_port_is_forbidden(const struct client_state *csp)
{
- char *p;
- int n;
- int factor = 2;
-
- if ((csp->action->flags & ACTION_BLOCK) == 0)
- {
- return(NULL);
- }
- else
- {
-#ifdef FORCE_LOAD
- factor++;
-#endif /* def FORCE_LOAD */
-
- n = strlen(CBLOCK);
- n += factor * strlen(http->hostport);
- n += factor * strlen(http->path);
-
- p = (char *)malloc(n);
-
-#ifdef FORCE_LOAD
- sprintf(p, CBLOCK, http->hostport, http->path, http->hostport, http->path,
- http->hostport, http->path);
-#else
- sprintf(p, CBLOCK, http->hostport, http->path, http->hostport, http->path);
-#endif /* def FORCE_LOAD */
-
- return(p);
- }
+ return ((csp->action->flags & ACTION_LIMIT_CONNECT) &&
+ !match_portlist(csp->action->string[ACTION_STRING_LIMIT_CONNECT],
+ csp->http->port));
}
-#ifdef IMAGE_BLOCKING
/*********************************************************************
*
- * Function : block_imageurl
+ * Function : block_url
*
- * Description : Given a URL which is blocked, decide whether to
- * send the "blocked" image or HTML.
+ * Description : Called from `chat'. Check to see if we need to block this.
*
* Parameters :
- * 1 : http = URL to check.
- * 2 : csp = Current client state (buffers, headers, etc...)
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : True (nonzero) if URL is in image list, false (0)
- * otherwise
+ * Returns : NULL => unblocked, else HTTP block response
*
*********************************************************************/
-int block_imageurl(struct http_request *http, struct client_state *csp)
+struct http_response *block_url(struct client_state *csp)
{
-#ifdef DETECT_MSIE_IMAGES
- if ((csp->accept_types
- & (ACCEPT_TYPE_IS_MSIE|ACCEPT_TYPE_MSIE_IMAGE|ACCEPT_TYPE_MSIE_HTML))
- == (ACCEPT_TYPE_IS_MSIE|ACCEPT_TYPE_MSIE_IMAGE))
+ struct http_response *rsp;
+ const char *new_content_type = NULL;
+
+ /*
+ * If it's not blocked, don't block it ;-)
+ */
+ if ((csp->action->flags & ACTION_BLOCK) == 0)
{
- return 1;
+ return NULL;
}
- else if ((csp->accept_types
- & (ACCEPT_TYPE_IS_MSIE|ACCEPT_TYPE_MSIE_IMAGE|ACCEPT_TYPE_MSIE_HTML))
- == (ACCEPT_TYPE_IS_MSIE|ACCEPT_TYPE_MSIE_HTML))
+ if (csp->action->flags & ACTION_REDIRECT)
{
- return 0;
+ log_error(LOG_LEVEL_ERROR, "redirect{} overruled by block.");
+ }
+ /*
+ * Else, prepare a response
+ */
+ if (NULL == (rsp = alloc_http_response()))
+ {
+ return cgi_error_memory();
}
-#endif
- return ((csp->action->flags & ACTION_IMAGE) != 0);
-}
-#endif /* def IMAGE_BLOCKING */
+ /*
+ * If it's an image-url, send back an image or redirect
+ * as specified by the relevant +image action
+ */
+#ifdef FEATURE_IMAGE_BLOCKING
+ if (((csp->action->flags & ACTION_IMAGE_BLOCKER) != 0)
+ && is_imageurl(csp))
+ {
+ char *p;
+ /* determine HOW images should be blocked */
+ p = csp->action->string[ACTION_STRING_IMAGE_BLOCKER];
+ if(csp->action->flags & ACTION_HANDLE_AS_EMPTY_DOCUMENT)
+ {
+ log_error(LOG_LEVEL_ERROR, "handle-as-empty-document overruled by handle-as-image.");
+ }
+#if 1 /* Two alternative strategies, use this one for now: */
-#ifdef PCRS
-/*********************************************************************
- *
- * Function : re_process_buffer
- *
- * Description : Apply all jobs from the joblist (aka. Perl regexp's) to
- * the text buffer that's been accumulated in csp->iob->buf
- * and set csp->content_length to the modified size.
- *
- * Parameters :
- * 1 : csp = Current client state (buffers, headers, etc...)
- *
- * Returns : a pointer to the (newly allocated) modified buffer.
- *
- *
- *********************************************************************/
-char *re_process_buffer(struct client_state *csp)
-{
- int hits=0;
- int size = csp->iob->eod - csp->iob->cur;
- char *old=csp->iob->cur, *new = NULL;
- pcrs_job *job, *joblist;
+ /* and handle accordingly: */
+ if ((p == NULL) || (0 == strcmpic(p, "pattern")))
+ {
+ rsp->status = strdup("403 Request blocked by Privoxy");
+ if (rsp->status == NULL)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+ rsp->body = bindup(image_pattern_data, image_pattern_length);
+ if (rsp->body == NULL)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+ rsp->content_length = image_pattern_length;
- struct file_list *fl;
- struct re_filterfile_spec *b;
+ if (enlist_unique_header(rsp->headers, "Content-Type", BUILTIN_IMAGE_MIMETYPE))
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+ }
+
+ else if (0 == strcmpic(p, "blank"))
+ {
+ rsp->status = strdup("403 Request blocked by Privoxy");
+ if (rsp->status == NULL)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+ rsp->body = bindup(image_blank_data, image_blank_length);
+ if (rsp->body == NULL)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+ rsp->content_length = image_blank_length;
+
+ if (enlist_unique_header(rsp->headers, "Content-Type", BUILTIN_IMAGE_MIMETYPE))
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+ }
+
+ else
+ {
+ rsp->status = strdup("302 Local Redirect from Privoxy");
+ if (rsp->status == NULL)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+
+ if (enlist_unique_header(rsp->headers, "Location", p))
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+ }
+
+#else /* Following code is disabled for now */
+
+ /* and handle accordingly: */
+ if ((p == NULL) || (0 == strcmpic(p, "pattern")))
+ {
+ p = CGI_PREFIX "send-banner?type=pattern";
+ }
+ else if (0 == strcmpic(p, "blank"))
+ {
+ p = CGI_PREFIX "send-banner?type=blank";
+ }
+ rsp->status = strdup("302 Local Redirect from Privoxy");
+ if (rsp->status == NULL)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
- /* Sanity first ;-) */
- if (size <= 0)
+ if (enlist_unique_header(rsp->headers, "Location", p))
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+#endif /* Preceeding code is disabled for now */
+ }
+ else if(csp->action->flags & ACTION_HANDLE_AS_EMPTY_DOCUMENT)
{
- return(strdup(""));
+ /*
+ * Send empty document.
+ */
+ new_content_type = csp->action->string[ACTION_STRING_CONTENT_TYPE];
+
+ freez(rsp->body);
+ rsp->body = strdup(" ");
+ rsp->content_length = 1;
+
+ rsp->status = strdup("403 Request blocked by Privoxy");
+ if (rsp->status == NULL)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+ if (new_content_type != 0)
+ {
+ log_error(LOG_LEVEL_HEADER, "Overwriting Content-Type with %s", new_content_type);
+ if (enlist_unique_header(rsp->headers, "Content-Type", new_content_type))
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+ }
}
+ else
+#endif /* def FEATURE_IMAGE_BLOCKING */
- if ( ( NULL == (fl = csp->rlist) ) || ( NULL == (b = fl->f) ) )
+ /*
+ * Else, generate an HTML "blocked" message:
+ */
{
- log_error(LOG_LEVEL_ERROR, "Unable to get current state of regexp filtering.");
- return(strdup(""));
- }
+ jb_err err;
+ struct map * exports;
+ char *p;
- joblist = b->joblist;
+ /*
+ * Workaround for stupid Netscape bug which prevents
+ * pages from being displayed if loading a referenced
+ * JavaScript or style sheet fails. So make it appear
+ * as if it succeeded.
+ */
+ if ( NULL != (p = get_header_value(csp->headers, "User-Agent:"))
+ && !strncmpic(p, "mozilla", 7) /* Catch Netscape but */
+ && !strstr(p, "Gecko") /* save Mozilla, */
+ && !strstr(p, "compatible") /* MSIE */
+ && !strstr(p, "Opera")) /* and Opera. */
+ {
+ rsp->status = strdup("200 Request for blocked URL");
+ }
+ else
+ {
+ rsp->status = strdup("403 Request for blocked URL");
+ }
+ if (rsp->status == NULL)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
- log_error(LOG_LEVEL_RE_FILTER, "re_filtering %s%s (size %d) ...",
- csp->http->hostport, csp->http->path, size);
+ exports = default_exports(csp, NULL);
+ if (exports == NULL)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
- /* Apply all jobs from the joblist */
- for (job = joblist; NULL != job; job = job->next)
- {
- hits += pcrs_exec_substitution(job, old, size, &new, &size);
- if (old != csp->iob->cur) free(old);
- old=new;
- }
+#ifdef FEATURE_FORCE_LOAD
+ err = map(exports, "force-prefix", 1, FORCE_PREFIX, 1);
+ /*
+ * Export the force conditional block killer if
+ *
+ * - Privoxy was compiled without FEATURE_FORCE_LOAD, or
+ * - Privoxy is configured to enforce blocks, or
+ * - it's a CONNECT request and enforcing wouldn't work anyway.
+ */
+ if ((csp->config->feature_flags & RUNTIME_FEATURE_ENFORCE_BLOCKS)
+ || (0 == strcmpic(csp->http->gpc, "connect")))
+#endif /* ndef FEATURE_FORCE_LOAD */
+ {
+ err = map_block_killer(exports, "force-support");
+ }
- log_error(LOG_LEVEL_RE_FILTER, " produced %d hits (new size %d).", hits, size);
+ if (!err) err = map(exports, "protocol", 1, csp->http->ssl ? "https://" : "http://", 1);
+ if (!err) err = map(exports, "hostport", 1, html_encode(csp->http->hostport), 0);
+ if (!err) err = map(exports, "path", 1, html_encode(csp->http->path), 0);
+ if (!err) err = map(exports, "path-ue", 1, url_encode(csp->http->path), 0);
+ if (!err)
+ {
+ const char *block_reason;
+ if (csp->action->string[ACTION_STRING_BLOCK] != NULL)
+ {
+ block_reason = csp->action->string[ACTION_STRING_BLOCK];
+ }
+ else
+ {
+ assert(connect_port_is_forbidden(csp));
+ block_reason = "Forbidden CONNECT port.";
+ }
+ err = map(exports, "block-reason", 1, html_encode(block_reason), 0);
+ }
+ if (err)
+ {
+ free_map(exports);
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
- csp->content_length = size;
+ err = template_fill_for_cgi(csp, "blocked", exports, rsp);
+ if (err)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
+ }
+ rsp->reason = RSP_REASON_BLOCKED;
- /* fwiw, reset the iob */
- IOB_RESET(csp);
- return(new);
+ return finish_http_response(csp, rsp);
}
-#endif /* def PCRS */
-#ifdef TRUST_FILES
+#ifdef FEATURE_TRUST
/*********************************************************************
*
- * Function : trust_url
+ * Function : trust_url FIXME: I should be called distrust_url
*
- * Description : Should we "trust" this URL? See "trustfile" line in config.
+ * Description : Calls is_untrusted_url to determine if the URL is trusted
+ * and if not, returns a HTTP 403 response with a reject message.
*
* Parameters :
- * 1 : http = http_request request for requested URL
- * 2 : csp = Current client state (buffers, headers, etc...)
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : NULL => trusted, else string to HTML "untrusted" description.
+ * Returns : NULL => trusted, else http_response.
*
*********************************************************************/
-char *trust_url(struct http_request *http, struct client_state *csp)
+struct http_response *trust_url(struct client_state *csp)
{
- struct file_list *fl;
- struct block_spec *b;
- struct url_spec url[1], **tl, *t;
- char *p, *h;
- char *hostport, *path, *refer;
- struct http_request rhttp[1];
- int n;
+ struct http_response *rsp;
+ struct map * exports;
+ char buf[BUFFER_SIZE];
+ char *p;
+ struct url_spec **tl;
+ struct url_spec *t;
+ jb_err err;
- if (((fl = csp->tlist) == NULL) || ((b = fl->f) == NULL))
+ /*
+ * Don't bother to work on trusted URLs
+ */
+ if (!is_untrusted_url(csp))
{
- return(NULL);
+ return NULL;
}
- *url = dsplit(http->host);
+ /*
+ * Else, prepare a response:
+ */
+ if (NULL == (rsp = alloc_http_response()))
+ {
+ return cgi_error_memory();
+ }
- /* if splitting the domain fails, punt */
- if (url->dbuf == NULL) return(NULL);
+ rsp->status = strdup("403 Request blocked by Privoxy");
+ exports = default_exports(csp, NULL);
+ if (exports == NULL || rsp->status == NULL)
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
- memset(rhttp, '\0', sizeof(*rhttp));
+ /*
+ * Export the protocol, host, port, and referrer information
+ */
+ err = map(exports, "hostport", 1, csp->http->hostport, 1);
+ if (!err) err = map(exports, "protocol", 1, csp->http->ssl ? "https://" : "http://", 1);
+ if (!err) err = map(exports, "path", 1, csp->http->path, 1);
- for (b = b->next; b ; b = b->next)
+ if (NULL != (p = get_header_value(csp->headers, "Referer:")))
{
- if ((b->url->port == 0) || (b->url->port == http->port))
- {
- if ((b->url->domain[0] == '\0') || (domaincmp(b->url, url) == 0))
- {
- if ((b->url->path == NULL) ||
-#ifdef REGEX
- (regexec(b->url->preg, http->path, 0, NULL, 0) == 0)
-#else
- (strncmp(b->url->path, http->path, b->url->pathlen) == 0)
-#endif
- )
- {
- freez(url->dbuf);
- freez(url->dvec);
+ if (!err) err = map(exports, "referrer", 1, html_encode(p), 0);
+ }
+ else
+ {
+ if (!err) err = map(exports, "referrer", 1, "none set", 1);
+ }
- if (b->reject == 0) return(NULL);
+ if (err)
+ {
+ free_map(exports);
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
- hostport = url_encode(http->hostport);
- path = url_encode(http->path);
+ /*
+ * Export the trust list
+ */
+ p = strdup("");
+ for (tl = csp->config->trust_list; (t = *tl) != NULL ; tl++)
+ {
+ snprintf(buf, sizeof(buf), "<li>%s</li>\n", t->spec);
+ string_append(&p, buf);
+ }
+ err = map(exports, "trusted-referrers", 1, p, 0);
- if (csp->referrer)
- {
- refer = url_encode(csp->referrer);
- }
- else
- {
- refer = url_encode("undefined");
- }
-
- n = strlen(CTRUST);
- n += strlen(hostport);
- n += strlen(path);
- n += strlen(refer);
-
- p = (char *)malloc(n);
-
- sprintf(p, CTRUST, hostport, path, refer);
+ if (err)
+ {
+ free_map(exports);
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
- freez(hostport);
- freez(path);
- freez(refer);
+ /*
+ * Export the trust info, if available
+ */
+ if (csp->config->trust_info->first)
+ {
+ struct list_entry *l;
- return(p);
- }
- }
+ p = strdup("");
+ for (l = csp->config->trust_info->first; l ; l = l->next)
+ {
+ snprintf(buf, sizeof(buf), "<li> <a href=\"%s\">%s</a><br>\n", l->str, l->str);
+ string_append(&p, buf);
}
+ err = map(exports, "trust-info", 1, p, 0);
+ }
+ else
+ {
+ err = map_block_killer(exports, "have-trust-info");
}
- freez(url->dbuf);
- freez(url->dvec);
-
- if ((csp->referrer == NULL)|| (strlen(csp->referrer) <= 9))
+ if (err)
{
- /* no referrer was supplied */
- goto trust_url_not_trusted;
+ free_map(exports);
+ free_http_response(rsp);
+ return cgi_error_memory();
}
- /* forge a URL from the referrer so we can use
- * convert_url() to parse it into its components.
+ /*
+ * Export the force conditional block killer if
+ *
+ * - Privoxy was compiled without FEATURE_FORCE_LOAD, or
+ * - Privoxy is configured to enforce blocks, or
+ * - it's a CONNECT request and enforcing wouldn't work anyway.
*/
+#ifdef FEATURE_FORCE_LOAD
+ if ((csp->config->feature_flags & RUNTIME_FEATURE_ENFORCE_BLOCKS)
+ || (0 == strcmpic(csp->http->gpc, "connect")))
+ {
+ err = map_block_killer(exports, "force-support");
+ }
+ else
+ {
+ err = map(exports, "force-prefix", 1, FORCE_PREFIX, 1);
+ }
+#else /* ifndef FEATURE_FORCE_LOAD */
+ err = map_block_killer(exports, "force-support");
+#endif /* ndef FEATURE_FORCE_LOAD */
- p = NULL;
- p = strsav(p, "GET ");
- p = strsav(p, csp->referrer + 9); /* skip over "Referer: " */
- p = strsav(p, " HTTP/1.0");
-
- parse_http_request(p, rhttp, csp);
+ if (err)
+ {
+ free_map(exports);
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
- if (rhttp->cmd == NULL)
+ /*
+ * Build the response
+ */
+ err = template_fill_for_cgi(csp, "untrusted", exports, rsp);
+ if (err)
{
- freez(p);
- goto trust_url_not_trusted;
+ free_http_response(rsp);
+ return cgi_error_memory();
}
+ rsp->reason = RSP_REASON_UNTRUSTED;
- freez(p);
+ return finish_http_response(csp, rsp);
+}
+#endif /* def FEATURE_TRUST */
- *url = dsplit(rhttp->host);
- /* if splitting the domain fails, punt */
- if (url->dbuf == NULL) goto trust_url_not_trusted;
+/*********************************************************************
+ *
+ * Function : compile_dynamic_pcrs_job_list
+ *
+ * Description : Compiles a dynamic pcrs job list (one with variables
+ * resolved at request time)
+ *
+ * Parameters :
+ * 1 : csp = Current client state (buffers, headers, etc...)
+ * 2 : b = The filter list to compile
+ *
+ * Returns : NULL in case of errors, otherwise the
+ * pcrs job list.
+ *
+ *********************************************************************/
+pcrs_job *compile_dynamic_pcrs_job_list(const struct client_state *csp, const struct re_filterfile_spec *b)
+{
+ struct list_entry *pattern;
+ pcrs_job *job_list = NULL;
+ pcrs_job *dummy = NULL;
+ pcrs_job *lastjob = NULL;
+ int error = 0;
- for (tl = csp->config->trust_list; (t = *tl) ; tl++)
+ const struct pcrs_variable variables[] =
+ {
+ {"url", csp->http->url, 1},
+ {"path", csp->http->path, 1},
+ {"host", csp->http->host, 1},
+ {"origin", csp->ip_addr_str, 1},
+ {NULL, NULL, 1}
+ };
+
+ for (pattern = b->patterns->first; pattern != NULL; pattern = pattern->next)
{
- if ((t->port == 0) || (t->port == rhttp->port))
+ assert(pattern->str != NULL);
+
+ dummy = pcrs_compile_dynamic_command(pattern->str, variables, &error);
+ if (NULL == dummy)
+ {
+ assert(error < 0);
+ log_error(LOG_LEVEL_ERROR,
+ "Adding filter job \'%s\' to dynamic filter %s failed: %s",
+ pattern->str, b->name, pcrs_strerror(error));
+ continue;
+ }
+ else
{
- if ((t->domain[0] == '\0') || domaincmp(t, url) == 0)
+ if (error == PCRS_WARN_TRUNCATION)
{
- if ((t->path == NULL) ||
-#ifdef REGEX
- (regexec(t->preg, rhttp->path, 0, NULL, 0) == 0)
-#else
- (strncmp(t->path, rhttp->path, t->pathlen) == 0)
-#endif
- )
- {
- /* if the URL's referrer is from a trusted referrer, then
- * add the target spec to the trustfile as an unblocked
- * domain and return NULL (which means it's OK).
- */
-
- FILE *fp;
+ log_error(LOG_LEVEL_ERROR,
+ "At least one of the variables in \'%s\' had to "
+ "be truncated before compilation", pattern->str);
+ }
+ if (job_list == NULL)
+ {
+ job_list = dummy;
+ }
+ else
+ {
+ lastjob->next = dummy;
+ }
+ lastjob = dummy;
+ }
+ }
- freez(url->dbuf);
- freez(url->dvec);
+ return job_list;
+}
- if ((fp = fopen(csp->config->trustfile, "a")))
- {
- h = NULL;
- h = strsav(h, "~");
- h = strsav(h, http->hostport);
+/*********************************************************************
+ *
+ * Function : rewrite_url
+ *
+ * Description : Rewrites a URL with a single pcrs command
+ * and returns the result if it differs from the
+ * original and isn't obviously invalid.
+ *
+ * Parameters :
+ * 1 : old_url = URL to rewrite.
+ * 2 : pcrs_command = pcrs command formatted as string (s@foo@bar@)
+ *
+ *
+ * Returns : NULL if the pcrs_command didn't change the url, or
+ * the result of the modification.
+ *
+ *********************************************************************/
+char *rewrite_url(char *old_url, const char *pcrs_command)
+{
+ char *new_url = NULL;
+ int hits;
- p = http->path;
- if ((*p++ == '/')
- && (*p++ == '~'))
- {
- /* since this path points into a user's home space
- * be sure to include this spec in the trustfile.
- */
- if ((p = strchr(p, '/')))
- {
- *p = '\0';
- h = strsav(h, http->path);
- h = strsav(h, "/");
- }
- }
+ assert(old_url);
+ assert(pcrs_command);
- free_http_request(rhttp);
+ new_url = pcrs_execute_single_command(old_url, pcrs_command, &hits);
- fprintf(fp, "%s\n", h);
- freez(h);
- fclose(fp);
- }
- return(NULL);
- }
- }
- }
+ if (hits == 0)
+ {
+ log_error(LOG_LEVEL_REDIRECTS,
+ "pcrs command \"%s\" didn't change \"%s\".",
+ pcrs_command, old_url);
+ freez(new_url);
}
-
-trust_url_not_trusted:
- free_http_request(rhttp);
-
- hostport = url_encode(http->hostport);
- path = url_encode(http->path);
-
- if (csp->referrer)
+ else if (hits < 0)
{
- refer = url_encode(csp->referrer);
+ log_error(LOG_LEVEL_REDIRECTS,
+ "executing pcrs command \"%s\" to rewrite %s failed: %s",
+ pcrs_command, old_url, pcrs_strerror(hits));
+ freez(new_url);
+ }
+ else if (strncmpic(new_url, "http://", 7) && strncmpic(new_url, "https://", 8))
+ {
+ log_error(LOG_LEVEL_ERROR,
+ "pcrs command \"%s\" changed \"%s\" to \"%s\" (%u hi%s), "
+ "but the result doesn't look like a valid URL and will be ignored.",
+ pcrs_command, old_url, new_url, hits, (hits == 1) ? "t" : "ts");
+ freez(new_url);
}
else
{
- refer = url_encode("undefined");
+ log_error(LOG_LEVEL_REDIRECTS,
+ "pcrs command \"%s\" changed \"%s\" to \"%s\" (%u hi%s).",
+ pcrs_command, old_url, new_url, hits, (hits == 1) ? "t" : "ts");
}
- n = strlen(CTRUST);
- n += strlen(hostport);
- n += strlen(path);
- n += strlen(refer);
-
- p = (char *)malloc(n);
- sprintf(p, CTRUST, hostport, path, refer);
-
- freez(hostport);
- freez(path);
- freez(refer);
-
- return(p);
+ return new_url;
}
-#endif /* def TRUST_FILES */
-
-
-static const char C_HOME_PAGE[] =
- "HTTP/1.0 200 OK\n"
- "Pragma: no-cache\n"
- "Expires: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "Content-Type: text/html\n\n"
- "<html>\n"
- "<head>\n"
- "<title>Internet Junkbuster: Information</title>\n"
- "</head>\n"
- BODY
- "<h1><center>"
- BANNER
- "</h1></center>\n"
- "<p><a href=\"" HOME_PAGE_URL "\">JunkBuster web site</a></p>\n"
- "<p><a href=\"http://i.j.b/show-proxy-arg\">Proxy configuration</a></p>\n"
- "<p><a href=\"http://i.j.b/show-url-info\">Look up a URL</a></p>\n"
- "</body>\n"
- "</html>\n";
+#ifdef FEATURE_FAST_REDIRECTS
/*********************************************************************
*
- * Function : intercept_url
+ * Function : get_last_url
*
- * Description : checks the URL `basename' against a list of URLs to
- * snarf. If it matches, it calls the associated function
- * which returns an HTML page to send back to the client.
- * Right now, we snarf:
- * "show-proxy-args", and
- * "ij-untrusted-url" (optional w/TRUST_FILES)
+ * Description : Search for the last URL inside a string.
+ * If the string already is a URL, it will
+ * be the first URL found.
*
* Parameters :
- * 1 : http = http_request request, check `basename's of blocklist
- * 2 : csp = Current client state (buffers, headers, etc...)
+ * 1 : subject = the string to check
+ * 2 : redirect_mode = +fast-redirect{} mode
*
- * Returns : 1 if it intercepts & handles the request.
+ * Returns : NULL if no URL was found, or
+ * the last URL found.
*
*********************************************************************/
-int intercept_url(struct http_request *http, struct client_state *csp)
+char *get_last_url(char *subject, const char *redirect_mode)
{
- char *basename = NULL;
- const struct interceptors *v;
+ char *new_url = NULL;
+ char *tmp;
- if (0 == strcmpic(http->host,"i.j.b"))
+ assert(subject);
+ assert(redirect_mode);
+
+ subject = strdup(subject);
+ if (subject == NULL)
{
- /*
- * Catch http://i.j.b/...
- */
- basename = http->path;
+ log_error(LOG_LEVEL_ERROR, "Out of memory while searching for redirects.");
+ return NULL;
}
- else if ( ( (0 == strcmpic(http->host,"ijbswa.sourceforge.net"))
- || (0 == strcmpic(http->host,"ijbswa.sf.net")) )
- && (0 == strncmpic(http->path,"/config", 7))
- && ((http->path[7] == '/') || (http->path[7] == '\0')))
+
+ if (0 == strcmpic(redirect_mode, "check-decoded-url"))
+ {
+ log_error(LOG_LEVEL_REDIRECTS, "Decoding \"%s\" if necessary.", subject);
+ new_url = url_decode(subject);
+ if (new_url != NULL)
+ {
+ freez(subject);
+ subject = new_url;
+ }
+ else
+ {
+ log_error(LOG_LEVEL_ERROR, "Unable to decode \"%s\".", subject);
+ }
+ }
+
+ log_error(LOG_LEVEL_REDIRECTS, "Checking \"%s\" for redirects.", subject);
+
+ /*
+ * Find the last URL encoded in the request
+ */
+ tmp = subject;
+ while ((tmp = strstr(tmp, "http://")) != NULL)
+ {
+ new_url = tmp++;
+ }
+ tmp = (new_url != NULL) ? new_url : subject;
+ while ((tmp = strstr(tmp, "https://")) != NULL)
+ {
+ new_url = tmp++;
+ }
+
+ if ((new_url != NULL)
+ && ( (new_url != subject)
+ || (0 == strncmpic(subject, "http://", 7))
+ || (0 == strncmpic(subject, "https://", 8))
+ ))
{
/*
- * Catch http://ijbswa.sourceforge.net/config/...
- * and http://ijbswa.sf.net/config/...
+ * Return new URL if we found a redirect
+ * or if the subject already was a URL.
+ *
+ * The second case makes sure that we can
+ * chain get_last_url after another redirection check
+ * (like rewrite_url) without losing earlier redirects.
*/
- basename = http->path + 7;
+ new_url = strdup(new_url);
+ freez(subject);
+ return new_url;
}
- if (!basename)
+ freez(subject);
+ return NULL;
+
+}
+#endif /* def FEATURE_FAST_REDIRECTS */
+
+
+/*********************************************************************
+ *
+ * Function : redirect_url
+ *
+ * Description : Checks if Privoxy should answer the request with
+ * a HTTP redirect and generates the redirect if
+ * necessary.
+ *
+ * Parameters :
+ * 1 : csp = Current client state (buffers, headers, etc...)
+ *
+ * Returns : NULL if the request can pass, HTTP redirect otherwise.
+ *
+ *********************************************************************/
+struct http_response *redirect_url(struct client_state *csp)
+{
+ struct http_response *rsp;
+#ifdef FEATURE_FAST_REDIRECTS
+ /*
+ * XXX: Do we still need FEATURE_FAST_REDIRECTS
+ * as compile-time option? The user can easily disable
+ * it in his action file.
+ */
+ char * redirect_mode;
+#endif /* def FEATURE_FAST_REDIRECTS */
+ char *old_url = NULL;
+ char *new_url = NULL;
+ char *redirection_string;
+
+ if ((csp->action->flags & ACTION_REDIRECT))
{
- /* Don't want to intercept */
- return(0);
- }
+ redirection_string = csp->action->string[ACTION_STRING_REDIRECT];
- /* We have intercepted it. */
+ /*
+ * If the redirection string begins with 's',
+ * assume it's a pcrs command, otherwise treat it as
+ * properly formatted URL and use it for the redirection
+ * directly.
+ *
+ * According to RFC 2616 section 14.30 the URL
+ * has to be absolute and if the user tries:
+ * +redirect{shit/this/will/be/parsed/as/pcrs_command.html}
+ * she would get undefined results anyway.
+ *
+ */
- /* remove any leading slash */
- if (*basename == '/')
+ if (*redirection_string == 's')
+ {
+ old_url = csp->http->url;
+ new_url = rewrite_url(old_url, redirection_string);
+ }
+ else
+ {
+ log_error(LOG_LEVEL_REDIRECTS,
+ "No pcrs command recognized, assuming that \"%s\" is already properly formatted.",
+ redirection_string);
+ new_url = strdup(redirection_string);
+ }
+ }
+
+#ifdef FEATURE_FAST_REDIRECTS
+ if ((csp->action->flags & ACTION_FAST_REDIRECTS))
{
- basename++;
+ redirect_mode = csp->action->string[ACTION_STRING_FAST_REDIRECTS];
+
+ /*
+ * If it exists, use the previously rewritten URL as input
+ * otherwise just use the old path.
+ */
+ old_url = (new_url != NULL) ? new_url : strdup(csp->http->path);
+ new_url = get_last_url(old_url, redirect_mode);
+ freez(old_url);
}
- log_error(LOG_LEVEL_GPC, "%s%s intercepted!", http->hostport, http->path);
- log_error(LOG_LEVEL_CLF, "%s - - [%T] \"%s\" 200 3",
- csp->ip_addr_str, http->cmd);
+ /*
+ * Disable redirect checkers, so that they
+ * will be only run more than once if the user
+ * also enables them through tags.
+ *
+ * From a performance point of view
+ * it doesn't matter, but the duplicated
+ * log messages are annoying.
+ */
+ csp->action->flags &= ~ACTION_FAST_REDIRECTS;
+#endif /* def FEATURE_FAST_REDIRECTS */
+ csp->action->flags &= ~ACTION_REDIRECT;
- for (v = intercept_patterns; v->str; v++)
+ /* Did any redirect action trigger? */
+ if (new_url)
{
- if (strncmp(basename, v->str, v->len) == 0)
+ if (0 == strcmpic(new_url, csp->http->url))
{
- char * p = ((v->interceptor)(http, csp));
+ log_error(LOG_LEVEL_ERROR,
+ "New URL \"%s\" and old URL \"%s\" are the same. Redirection loop prevented.",
+ csp->http->url, new_url);
+ freez(new_url);
+ }
+ else
+ {
+ log_error(LOG_LEVEL_REDIRECTS, "New URL is: %s", new_url);
- if (p != NULL)
+ if (NULL == (rsp = alloc_http_response()))
{
- /* Send HTML redirection result */
- write_socket(csp->cfd, p, strlen(p));
+ freez(new_url);
+ return cgi_error_memory();
+ }
- freez(p);
+ if ( enlist_unique_header(rsp->headers, "Location", new_url)
+ || (NULL == (rsp->status = strdup("302 Local Redirect from Privoxy"))) )
+ {
+ freez(new_url);
+ free_http_response(rsp);
+ return cgi_error_memory();
}
- return(1);
+ rsp->reason = RSP_REASON_REDIRECTED;
+ freez(new_url);
+
+ return finish_http_response(csp, rsp);
}
}
- write_socket(csp->cfd, C_HOME_PAGE, strlen(C_HOME_PAGE));
+ /* Only reached if no redirect is required */
+ return NULL;
- return(1);
}
-#ifdef FAST_REDIRECTS
+
+#ifdef FEATURE_IMAGE_BLOCKING
/*********************************************************************
*
- * Function : redirect_url
+ * Function : is_imageurl
*
- * Description : Checks for redirection URLs and returns a HTTP redirect
- * to the destination URL.
+ * Description : Given a URL, decide whether it is an image or not,
+ * using either the info from a previous +image action
+ * or, #ifdef FEATURE_IMAGE_DETECT_MSIE, and the browser
+ * is MSIE and not on a Mac, tell from the browser's accept
+ * header.
*
* Parameters :
- * 1 : http = http_request request, check `basename's of blocklist
- * 2 : csp = Current client state (buffers, headers, etc...)
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : NULL if URL was clean, HTTP redirect otherwise.
+ * Returns : True (nonzero) if URL is an image, false (0)
+ * otherwise
*
*********************************************************************/
-char *redirect_url(struct http_request *http, struct client_state *csp)
+int is_imageurl(const struct client_state *csp)
{
- char *p, *q;
-
- p = q = csp->http->path;
- log_error(LOG_LEVEL_REDIRECTS, "checking path: %s", p);
+#ifdef FEATURE_IMAGE_DETECT_MSIE
+ char *tmp;
- /* find the last URL encoded in the request */
- while (p = strstr(p, "http://"))
+ tmp = get_header_value(csp->headers, "User-Agent:");
+ if (tmp && strstr(tmp, "MSIE") && !strstr(tmp, "Mac_"))
{
- q = p++;
+ tmp = get_header_value(csp->headers, "Accept:");
+ if (tmp && strstr(tmp, "image/gif"))
+ {
+ /* Client will accept HTML. If this seems counterintuitive,
+ * blame Microsoft.
+ */
+ return(0);
+ }
+ else
+ {
+ return(1);
+ }
}
+#endif /* def FEATURE_IMAGE_DETECT_MSIE */
- /* if there was any, generate and return a HTTP redirect */
- if (q != csp->http->path)
- {
- log_error(LOG_LEVEL_REDIRECTS, "redirecting to: %s", q);
-
- p = (char *)malloc(strlen(HTTP_REDIRECT_TEMPLATE) + strlen(q));
- sprintf(p, HTTP_REDIRECT_TEMPLATE, q);
- return(p);
- }
- else
- {
- return(NULL);
- }
+ return ((csp->action->flags & ACTION_IMAGE) != 0);
}
-#endif /* def FAST_REDIRECTS */
+#endif /* def FEATURE_IMAGE_BLOCKING */
+
+#ifdef FEATURE_TRUST
/*********************************************************************
*
- * Function : url_actions
+ * Function : is_untrusted_url
*
- * Description : Gets the actions for this URL.
+ * Description : Should we "distrust" this URL (and block it)?
+ *
+ * Yes if it matches a line in the trustfile, or if the
+ * referrer matches a line starting with "+" in the
+ * trustfile.
+ * No otherwise.
*
* Parameters :
- * 1 : http = http_request request for blocked URLs
- * 2 : csp = Current client state (buffers, headers, etc...)
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : N/A
+ * Returns : 0 => trusted, 1 => untrusted
*
*********************************************************************/
-void url_actions(struct http_request *http,
- struct client_state *csp)
+int is_untrusted_url(const struct client_state *csp)
{
struct file_list *fl;
- struct url_actions *b;
+ struct block_spec *b;
+ struct url_spec **trusted_url;
+ struct http_request rhttp[1];
+ const char * referer;
+ jb_err err;
- init_current_action(csp->action);
+ /*
+ * If we don't have a trustlist, we trust everybody
+ */
+ if (((fl = csp->tlist) == NULL) || ((b = fl->f) == NULL))
+ {
+ return 0;
+ }
- if (((fl = csp->actions_list) == NULL) || ((b = fl->f) == NULL))
+ memset(rhttp, '\0', sizeof(*rhttp));
+
+ /*
+ * Do we trust the request URL itself?
+ */
+ for (b = b->next; b ; b = b->next)
{
- return;
+ if (url_match(b->url, csp->http))
+ {
+ return b->reject;
+ }
+ }
+
+ if (NULL == (referer = get_header_value(csp->headers, "Referer:")))
+ {
+ /* no referrer was supplied */
+ return 1;
+ }
+
+
+ /*
+ * If not, do we maybe trust its referrer?
+ */
+ err = parse_http_url(referer, rhttp, REQUIRE_PROTOCOL);
+ if (err)
+ {
+ return 1;
}
- apply_url_actions(csp->action, http, b);
+ for (trusted_url = csp->config->trust_list; *trusted_url != NULL; trusted_url++)
+ {
+ if (url_match(*trusted_url, rhttp))
+ {
+ /* if the URL's referrer is from a trusted referrer, then
+ * add the target spec to the trustfile as an unblocked
+ * domain and return 0 (which means it's OK).
+ */
+
+ FILE *fp;
+
+ if (NULL != (fp = fopen(csp->config->trustfile, "a")))
+ {
+ char * path;
+ char * path_end;
+ char * new_entry = strdup("~");
+
+ string_append(&new_entry, csp->http->hostport);
+
+ path = csp->http->path;
+ if ( (path[0] == '/')
+ && (path[1] == '~')
+ && ((path_end = strchr(path + 2, '/')) != NULL))
+ {
+ /* since this path points into a user's home space
+ * be sure to include this spec in the trustfile.
+ */
+ int path_len = path_end - path; /* save offset */
+ path = strdup(path); /* Copy string */
+ if (path != NULL)
+ {
+ path_end = path + path_len; /* regenerate ptr to new buffer */
+ *(path_end + 1) = '\0'; /* Truncate path after '/' */
+ }
+ string_join(&new_entry, path);
+ }
+
+ /*
+ * Give a reason for generating this entry.
+ */
+ string_append(&new_entry, " # Trusted referrer was: ");
+ string_append(&new_entry, referer);
+
+ if (new_entry != NULL)
+ {
+ if (-1 == fprintf(fp, "%s\n", new_entry))
+ {
+ log_error(LOG_LEVEL_ERROR, "Failed to append \'%s\' to trustfile \'%s\': %E",
+ new_entry, csp->config->trustfile);
+ }
+ freez(new_entry);
+ }
+ else
+ {
+ /* FIXME: No way to handle out-of memory, so mostly ignoring it */
+ log_error(LOG_LEVEL_ERROR, "Out of memory adding pattern to trust file");
+ }
+
+ fclose(fp);
+ }
+ else
+ {
+ log_error(LOG_LEVEL_ERROR, "Failed to append new entry for \'%s\' to trustfile \'%s\': %E",
+ csp->http->hostport, csp->config->trustfile);
+ }
+ return 0;
+ }
+ }
+
+ return 1;
}
+#endif /* def FEATURE_TRUST */
/*********************************************************************
*
- * Function : apply_url_actions
+ * Function : pcrs_filter_response
*
- * Description : Applies a list of URL actions.
+ * Description : Execute all text substitutions from all applying
+ * +filter actions on the text buffer that's been
+ * accumulated in csp->iob->buf.
*
* Parameters :
- * 1 : action = Destination.
- * 2 : http = Current URL
- * 3 : b = list of URL actions to apply
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : N/A
+ * Returns : a pointer to the (newly allocated) modified buffer.
+ * or NULL if there were no hits or something went wrong
*
*********************************************************************/
-void apply_url_actions(struct current_action_spec *action,
- struct http_request *http,
- struct url_actions *b)
+static char *pcrs_filter_response(struct client_state *csp)
{
- struct url_spec url[1];
+ int hits=0;
+ size_t size, prev_size;
- if (b == NULL)
+ char *old = NULL;
+ char *new = NULL;
+ pcrs_job *job;
+
+ struct file_list *fl;
+ struct re_filterfile_spec *b;
+ struct list_entry *filtername;
+
+ int i, found_filters = 0;
+
+ /*
+ * Sanity first
+ */
+ if (csp->iob->cur >= csp->iob->eod)
{
- /* Should never happen */
- return;
+ return(NULL);
}
- *url = dsplit(http->host);
+ /*
+ * Need to check the set of re_filterfiles...
+ */
+ for (i = 0; i < MAX_AF_FILES; i++)
+ {
+ fl = csp->rlist[i];
+ if (NULL != fl)
+ {
+ if (NULL != fl->f)
+ {
+ found_filters = 1;
+ break;
+ }
+ }
+ }
- /* if splitting the domain fails, punt */
- if (url->dbuf == NULL)
+ if (0 == found_filters)
{
- return;
+ log_error(LOG_LEVEL_ERROR, "Inconsistent configuration: "
+ "content filtering enabled, but no content filters available.");
+ return(NULL);
}
- for (b = b->next; NULL != b; b = b->next)
+ size = (size_t)(csp->iob->eod - csp->iob->cur);
+ old = csp->iob->cur;
+
+ for (i = 0; i < MAX_AF_FILES; i++)
{
- if ((b->url->port == 0) || (b->url->port == http->port))
+ fl = csp->rlist[i];
+ if ((NULL == fl) || (NULL == fl->f))
+ {
+ /*
+ * Either there are no filter files
+ * left, or this filter file just
+ * contains no valid filters.
+ *
+ * Continue to be sure we don't miss
+ * valid filter files that are chained
+ * after empty or invalid ones.
+ */
+ continue;
+ }
+ /*
+ * For all applying +filter actions, look if a filter by that
+ * name exists and if yes, execute it's pcrs_joblist on the
+ * buffer.
+ */
+ for (b = fl->f; b; b = b->next)
+ {
+ if (b->type != FT_CONTENT_FILTER)
+ {
+ /* Skip header filters */
+ continue;
+ }
+
+ for (filtername = csp->action->multi[ACTION_MULTI_FILTER]->first;
+ filtername ; filtername = filtername->next)
{
- if ((b->url->domain[0] == '\0') || (domaincmp(b->url, url) == 0))
+ if (strcmp(b->name, filtername->str) == 0)
{
- if ((b->url->path == NULL) ||
-#ifdef REGEX
- (regexec(b->url->preg, http->path, 0, NULL, 0) == 0)
-#else
- (strncmp(b->url->path, http->path, b->url->pathlen) == 0)
-#endif
- )
+ int current_hits = 0; /* Number of hits caused by this filter */
+ int job_number = 0; /* Which job we're currently executing */
+ int job_hits = 0; /* How many hits the current job caused */
+ pcrs_job *joblist = b->joblist;
+
+ if (b->dynamic) joblist = compile_dynamic_pcrs_job_list(csp, b);
+
+ if (NULL == joblist)
{
- merge_current_action(action, b->action);
+ log_error(LOG_LEVEL_RE_FILTER, "Filter %s has empty joblist. Nothing to do.", b->name);
+ continue;
+ }
+
+ prev_size = size;
+ /* Apply all jobs from the joblist */
+ for (job = joblist; NULL != job; job = job->next)
+ {
+ job_number++;
+ job_hits = pcrs_execute(job, old, size, &new, &size);
+
+ if (job_hits >= 0)
+ {
+ /*
+ * That went well. Continue filtering
+ * and use the result of this job as
+ * input for the next one.
+ */
+ current_hits += job_hits;
+ if (old != csp->iob->cur)
+ {
+ freez(old);
+ }
+ old = new;
+ }
+ else
+ {
+ /*
+ * This job caused an unexpected error. Inform the user
+ * and skip the rest of the jobs in this filter. We could
+ * continue with the next job, but usually the jobs
+ * depend on each other or are similar enough to
+ * fail for the same reason.
+ *
+ * At the moment our pcrs expects the error codes of pcre 3.4,
+ * but newer pcre versions can return additional error codes.
+ * As a result pcrs_strerror()'s error message might be
+ * "Unknown error ...", therefore we print the numerical value
+ * as well.
+ *
+ * XXX: Is this important enough for LOG_LEVEL_ERROR or
+ * should we use LOG_LEVEL_RE_FILTER instead?
+ */
+ log_error(LOG_LEVEL_ERROR, "Skipped filter \'%s\' after job number %u: %s (%d)",
+ b->name, job_number, pcrs_strerror(job_hits), job_hits);
+ break;
+ }
}
+
+ if (b->dynamic) pcrs_free_joblist(joblist);
+
+ log_error(LOG_LEVEL_RE_FILTER,
+ "filtering %s%s (size %d) with \'%s\' produced %d hits (new size %d).",
+ csp->http->hostport, csp->http->path, prev_size, b->name, current_hits, size);
+
+ hits += current_hits;
}
}
}
+ }
+
+ /*
+ * If there were no hits, destroy our copy and let
+ * chat() use the original in csp->iob
+ */
+ if (!hits)
+ {
+ freez(new);
+ return(NULL);
+ }
+
+ csp->flags |= CSP_FLAG_MODIFIED;
+ csp->content_length = size;
+ IOB_RESET(csp);
+
+ return(new);
- freez(url->dbuf);
- freez(url->dvec);
}
/*********************************************************************
*
- * Function : forward_url
+ * Function : gif_deanimate_response
*
- * Description : Should we forward this to another proxy?
+ * Description : Deanimate the GIF image that has been accumulated in
+ * csp->iob->buf, set csp->content_length to the modified
+ * size and raise the CSP_FLAG_MODIFIED flag.
*
* Parameters :
- * 1 : http = http_request request for current URL
- * 2 : csp = Current client state (buffers, headers, etc...)
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : Return gw_default for no forward match,
- * else a gateway pointer to a specific forwarding proxy.
+ * Returns : a pointer to the (newly allocated) modified buffer.
+ * or NULL in case something went wrong.
*
*********************************************************************/
-const struct gateway *forward_url(struct http_request *http, struct client_state *csp)
+static char *gif_deanimate_response(struct client_state *csp)
{
- struct file_list *fl;
- struct forward_spec *b;
- struct url_spec url[1];
+ struct binbuffer *in, *out;
+ char *p;
+ size_t size;
+
+ size = (size_t)(csp->iob->eod - csp->iob->cur);
- if (((fl = csp->flist) == NULL) || ((b = fl->f) == NULL))
+ if ( (NULL == (in = (struct binbuffer *)zalloc(sizeof *in )))
+ || (NULL == (out = (struct binbuffer *)zalloc(sizeof *out))) )
{
- return(gw_default);
+ log_error(LOG_LEVEL_DEANIMATE, "failed! (no mem)");
+ return NULL;
}
- *url = dsplit(http->host);
+ in->buffer = csp->iob->cur;
+ in->size = size;
- /* if splitting the domain fails, punt */
- if (url->dbuf == NULL) return(gw_default);
-
- for (b = b->next; b ; b = b->next)
+ if (gif_deanimate(in, out, strncmp("last", csp->action->string[ACTION_STRING_DEANIMATE], 4)))
+ {
+ log_error(LOG_LEVEL_DEANIMATE, "failed! (gif parsing)");
+ freez(in);
+ buf_free(out);
+ return(NULL);
+ }
+ else
{
- if ((b->url->port == 0) || (b->url->port == http->port))
+ if ((int)size == out->offset)
{
- if ((b->url->domain[0] == '\0') || (domaincmp(b->url, url) == 0))
- {
- if ((b->url->path == NULL) ||
-#ifdef REGEX
- (regexec(b->url->preg, http->path, 0, NULL, 0) == 0)
-#else
- (strncmp(b->url->path, http->path, b->url->pathlen) == 0)
-#endif
- )
- {
- freez(url->dbuf);
- freez(url->dvec);
- return(b->gw);
- }
- }
+ log_error(LOG_LEVEL_DEANIMATE, "GIF not changed.");
+ }
+ else
+ {
+ log_error(LOG_LEVEL_DEANIMATE, "Success! GIF shrunk from %d bytes to %d.", size, out->offset);
}
+ csp->content_length = out->offset;
+ csp->flags |= CSP_FLAG_MODIFIED;
+ p = out->buffer;
+ freez(in);
+ freez(out);
+ return(p);
}
- freez(url->dbuf);
- freez(url->dvec);
- return(gw_default);
-
}
/*********************************************************************
*
- * Function : dsplit
+ * Function : get_filter_function
*
- * Description : Takes a domain and returns a pointer to a url_spec
- * structure populated with dbuf, dcnt and dvec. The
- * other fields in the structure that is returned are zero.
+ * Description : Decides which content filter function has
+ * to be applied (if any).
*
* Parameters :
- * 1 : domain = a URL address
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : url_spec structure populated with dbuf, dcnt and dvec.
+ * Returns : The content filter function to run, or
+ * NULL if no content filter is active
*
*********************************************************************/
-struct url_spec dsplit(char *domain)
+filter_function_ptr get_filter_function(struct client_state *csp)
{
- struct url_spec ret[1];
- char *v[BUFSIZ];
- int size;
- char *p;
+ filter_function_ptr filter_function = NULL;
- memset(ret, '\0', sizeof(*ret));
-
- if (domain[strlen(domain) - 1] == '.')
- {
- ret->unanchored |= ANCHOR_RIGHT;
- }
- if (domain[0] == '.')
+ /*
+ * Are we enabling text mode by force?
+ */
+ if (csp->action->flags & ACTION_FORCE_TEXT_MODE)
{
- ret->unanchored |= ANCHOR_LEFT;
- }
-
- ret->dbuf = strdup(domain);
-
- /* map to lower case */
- for (p = ret->dbuf; *p ; p++) *p = tolower(*p);
-
- /* split the domain name into components */
- ret->dcnt = ssplit(ret->dbuf, ".", v, SZ(v), 1, 1);
+ /*
+ * Do we really have to?
+ */
+ if (csp->content_type & CT_TEXT)
+ {
+ log_error(LOG_LEVEL_HEADER, "Text mode is already enabled.");
+ }
+ else
+ {
+ csp->content_type |= CT_TEXT;
+ log_error(LOG_LEVEL_HEADER, "Text mode enabled by force. Take cover!");
+ }
+ }
- if (ret->dcnt <= 0)
+ if (!(csp->content_type & CT_DECLARED))
{
- memset(ret, '\0', sizeof(ret));
- return(*ret);
+ /*
+ * The server didn't bother to declare a MIME-Type.
+ * Assume it's text that can be filtered.
+ *
+ * This also regulary happens with 304 responses,
+ * therefore logging anything here would cause
+ * too much noise.
+ */
+ csp->content_type |= CT_TEXT;
}
- /* save a copy of the pointers in dvec */
- size = ret->dcnt * sizeof(*ret->dvec);
-
- if ((ret->dvec = (char **)malloc(size)))
+ /*
+ * Choose the applying filter function based on
+ * the content type and action settings.
+ */
+ if ((csp->content_type & CT_TEXT) &&
+ (csp->rlist != NULL) &&
+ (!list_is_empty(csp->action->multi[ACTION_MULTI_FILTER])))
{
- memcpy(ret->dvec, v, size);
+ filter_function = pcrs_filter_response;
+ }
+ else if ((csp->content_type & CT_GIF) &&
+ (csp->action->flags & ACTION_DEANIMATE))
+ {
+ filter_function = gif_deanimate_response;
}
-
- return(*ret);
-
+ return filter_function;
}
/*********************************************************************
*
- * Function : domaincmp
+ * Function : remove_chunked_transfer_coding
*
- * Description : Domain-wise Compare fqdn's. Governed by the bimap in
- * pattern->unachored, the comparison is un-, left-,
- * right-anchored, or both.
- * The individual domain names are compared with
- * trivialmatch().
+ * Description : In-situ remove the "chunked" transfer coding as defined
+ * in rfc2616 from a buffer.
*
* Parameters :
- * 1 : pattern = a domain that may contain a '*' as a wildcard.
- * 2 : fqdn = domain name against which the patterns are compared.
+ * 1 : buffer = Pointer to the text buffer
+ * 2 : size = In: Number of bytes to be processed,
+ * Out: Number of bytes after de-chunking.
+ * (undefined in case of errors)
*
- * Returns : 0 => domains are equivalent, else no match.
+ * Returns : JB_ERR_OK for success,
+ * JB_ERR_PARSE otherwise
*
*********************************************************************/
-int domaincmp(struct url_spec *pattern, struct url_spec *fqdn)
+static jb_err remove_chunked_transfer_coding(char *buffer, size_t *size)
{
- char **pv, **fv; /* vectors */
- int pn, fn; /* counters */
- char *p, *f; /* chars */
+ size_t newsize = 0;
+ unsigned int chunksize = 0;
+ char *from_p, *to_p;
- pv = pattern->dvec;
- fv = fqdn->dvec;
- fn = pn = 0;
+ assert(buffer);
+ from_p = to_p = buffer;
- while (fn < fqdn->dcnt && pn < pattern->dcnt)
+ if (sscanf(buffer, "%x", &chunksize) != 1)
{
- p = pv[pn];
- f = fv[fn];
+ log_error(LOG_LEVEL_ERROR, "Invalid first chunksize while stripping \"chunked\" transfer coding");
+ return JB_ERR_PARSE;
+ }
- if (simplematch(p, f))
+ while (chunksize > 0)
+ {
+ if (NULL == (from_p = strstr(from_p, "\r\n")))
{
- if(pn || !(pattern->unanchored & ANCHOR_LEFT))
- {
- return 1;
- }
+ log_error(LOG_LEVEL_ERROR, "Parse error while stripping \"chunked\" transfer coding");
+ return JB_ERR_PARSE;
}
- else
+
+ if ((newsize += chunksize) >= *size)
+ {
+ log_error(LOG_LEVEL_ERROR, "Chunksize exceeds buffer in \"chunked\" transfer coding");
+ return JB_ERR_PARSE;
+ }
+ from_p += 2;
+
+ memmove(to_p, from_p, (size_t) chunksize);
+ to_p = buffer + newsize;
+ from_p += chunksize + 2;
+
+ if (sscanf(from_p, "%x", &chunksize) != 1)
{
- pn++;
+ log_error(LOG_LEVEL_INFO, "Invalid \"chunked\" transfer encoding detected and ignored.");
+ break;
}
- fn++;
}
+
+ /* XXX: Should get its own loglevel. */
+ log_error(LOG_LEVEL_RE_FILTER, "De-chunking successful. Shrunk from %d to %d", *size, newsize);
- return ((pn < pattern->dcnt) || ((fn < fqdn->dcnt) && !(pattern->unanchored & ANCHOR_RIGHT)));
+ *size = newsize;
-}
+ return JB_ERR_OK;
+}
-/* intercept functions */
/*********************************************************************
*
- * Function : show_proxy_args
+ * Function : prepare_for_filtering
*
- * Description : This "crunch"es "http:/any.thing/show-proxy-args" and
- * returns a web page describing the current status of IJB.
+ * Description : If necessary, de-chunks and decompresses
+ * the content so it can get filterd.
*
* Parameters :
- * 1 : http = ignored
- * 2 : csp = Current client state (buffers, headers, etc...)
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : A string that contains the current status of IJB.
+ * Returns : JB_ERR_OK for success,
+ * JB_ERR_PARSE otherwise
*
*********************************************************************/
-char *show_proxy_args(struct http_request *http, struct client_state *csp)
+static jb_err prepare_for_filtering(struct client_state *csp)
{
- char *s = NULL;
-
-#ifdef SPLIT_PROXY_ARGS
- FILE * fp;
- char buf[BUFSIZ];
- char * p;
- const char * filename = NULL;
- const char * file_description = NULL;
- char * query_string = strrchr(http->path, '?');
- char which_file = '\0';
+ jb_err err = JB_ERR_OK;
-
- if (query_string != NULL)
- {
- /* first char past the last '?' (maybe '\0')*/
- which_file = query_string[1];
- }
- switch (which_file)
+ /*
+ * If the body has a "chunked" transfer-encoding,
+ * get rid of it, adjusting size and iob->eod
+ */
+ if (csp->flags & CSP_FLAG_CHUNKED)
{
- case 'p':
- if (csp->actions_list)
+ size_t size = (size_t)(csp->iob->eod - csp->iob->cur);
+
+ log_error(LOG_LEVEL_RE_FILTER, "Need to de-chunk first");
+ err = remove_chunked_transfer_coding(csp->iob->cur, &size);
+ if (JB_ERR_OK == err)
{
- filename = csp->actions_list->filename;
- file_description = "Actions List";
+ csp->iob->eod = csp->iob->cur + size;
+ csp->flags |= CSP_FLAG_MODIFIED;
}
- break;
- case 'f':
- if (csp->flist)
+ else
{
- filename = csp->flist->filename;
- file_description = "Forward List";
+ return JB_ERR_PARSE;
}
- break;
+ }
-#ifdef ACL_FILES
- case 'a':
- if (csp->alist)
+#ifdef FEATURE_ZLIB
+ /*
+ * If the body has a supported transfer-encoding,
+ * decompress it, adjusting size and iob->eod.
+ */
+ if (csp->content_type & (CT_GZIP|CT_DEFLATE))
+ {
+ if (0 == csp->iob->eod - csp->iob->cur)
{
- filename = csp->alist->filename;
- file_description = "Access Control List";
+ /* Nothing left after de-chunking. */
+ return JB_ERR_OK;
}
- break;
-#endif /* def ACL_FILES */
-#ifdef PCRS
- case 'r':
- if (csp->rlist)
- {
- filename = csp->rlist->filename;
- file_description = "RE Filter List";
- }
- break;
-#endif /* def PCRS */
+ err = decompress_iob(csp);
-#ifdef TRUST_FILES
- case 't':
- if (csp->tlist)
- {
- filename = csp->tlist->filename;
- file_description = "Trust List";
- }
- break;
-#endif /* def TRUST_FILES */
- }
-
- if (filename)
- {
- /* Display specified file */
- /* FIXME: Add HTTP headers so this isn't cached */
- s = strsav(s,
- "HTTP/1.0 200 OK\n"
- "Server: IJ/" VERSION "\n"
- "Content-type: text/html\n"
- "Pragma: no-cache\n"
- "Last-Modified: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "Expires: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "\n"
-
- "<html>"
- "<head>"
- "<title>Internet Junkbuster Proxy Status - ");
- s = strsav(s, file_description);
- s = strsav(s,
- "</title>"
- "</head>\n"
- "<body bgcolor=\"#f8f8f0\" link=\"#000078\" alink=\"#ff0022\" vlink=\"#787878\">\n"
- "<center>\n"
- "<h1>" BANNER "\n");
- s = strsav(s, file_description);
- s = strsav(s,
- "</h1></center>\n"
- "<p><a href=\"show-proxy-args\">Back to proxy status</a></p>\n"
- "<h2>");
- s = strsav(s, file_description);
- s = strsav(s,
- "</h2>\n"
- "Contents of file "<code>");
- p = html_encode(filename);
- s = strsav(s, p);
- freez(p);
- s = strsav(s,
- "</code>":<br>\n"
- "</p>\n"
- "<pre>");
-
- if ((fp = fopen(filename, "r")) == NULL)
+ if (JB_ERR_OK == err)
{
- s = strsav(s, "</pre><h1>ERROR OPENING FILE!</h1><pre>");
+ csp->flags |= CSP_FLAG_MODIFIED;
+ csp->content_type &= ~CT_TABOO;
}
else
{
- while (fgets(buf, sizeof(buf), fp))
- {
- p = html_encode(buf);
- if (p)
- {
- s = strsav(s, p);
- freez(p);
- s = strsav(s, "<br>");
- }
- }
- fclose(fp);
+ /*
+ * Unset CT_GZIP and CT_DEFLATE to remember not
+ * to modify the Content-Encoding header later.
+ */
+ csp->content_type &= ~CT_GZIP;
+ csp->content_type &= ~CT_DEFLATE;
}
+ }
+#endif
- s = strsav(s,
- "</pre>\n"
- "<br>\n"
- "<p><a href=\"show-proxy-args\">Back to proxy status</a></p>\n"
- "<br>\n"
- "<small><small><p>\n"
- "The " BANNER " Proxy - \n"
- "<a href=\"" HOME_PAGE_URL "\">" HOME_PAGE_URL "</a>\n"
- "</small></small>"
- "</body></html>\n");
- return(s);
- }
-#endif /* def SPLIT_PROXY_ARGS */
-
- s = strsav(s, csp->config->proxy_args_header);
- s = strsav(s, csp->config->proxy_args_invocation);
-#ifdef STATISTICS
- s = add_stats(s);
-#endif /* def STATISTICS */
- s = strsav(s, csp->config->proxy_args_gateways);
+ return err;
+}
-#ifdef SPLIT_PROXY_ARGS
- s = strsav(s,
- "<h2>The following files are in use:</h2>\n"
- "<p>(Click a filename to view it)</p>\n"
- "<ul>\n");
- if (csp->actions_list)
+/*********************************************************************
+ *
+ * Function : execute_content_filter
+ *
+ * Description : Executes a given content filter.
+ *
+ * Parameters :
+ * 1 : csp = Current client state (buffers, headers, etc...)
+ * 2 : content_filter = The filter function to execute
+ *
+ * Returns : Pointer to the modified buffer, or
+ * NULL if filtering failed or wasn't necessary.
+ *
+ *********************************************************************/
+char *execute_content_filter(struct client_state *csp, filter_function_ptr content_filter)
+{
+ if (0 == csp->iob->eod - csp->iob->cur)
{
- s = strsav(s, "<li>Actions List: <a href=\"show-proxy-args?permit\"><code>");
- s = strsav(s, csp->actions_list->filename);
- s = strsav(s, "</code></a></li>\n");
+ /*
+ * No content (probably status code 301, 302 ...),
+ * no filtering necessary.
+ */
+ return NULL;
}
- if (csp->flist)
+ if (JB_ERR_OK != prepare_for_filtering(csp))
{
- s = strsav(s, "<li>Forward List: <a href=\"show-proxy-args?forward\"><code>");
- s = strsav(s, csp->flist->filename);
- s = strsav(s, "</code></a></li>\n");
+ /*
+ * failed to de-chunk or decompress.
+ */
+ return NULL;
}
-#ifdef ACL_FILES
- if (csp->alist)
+ if (0 == csp->iob->eod - csp->iob->cur)
{
- s = strsav(s, "<li>Access Control List: <a href=\"show-proxy-args?acl\"><code>");
- s = strsav(s, csp->alist->filename);
- s = strsav(s, "</code></a></li>\n");
+ /*
+ * Clown alarm: chunked and/or compressed nothing delivered.
+ */
+ return NULL;
}
-#endif /* def ACL_FILES */
-#ifdef PCRS
- if (csp->rlist)
- {
- s = strsav(s, "<li>RE Filter List: <a href=\"show-proxy-args?re\"><code>");
- s = strsav(s, csp->rlist->filename);
- s = strsav(s, "</code></a></li>\n");
- }
-#endif /* def PCRS */
+ return ((*content_filter)(csp));
+}
-#ifdef TRUST_FILES
- if (csp->tlist)
- {
- s = strsav(s, "<li>Trust List: <a href=\"show-proxy-args?trust\"><code>");
- s = strsav(s, csp->tlist->filename);
- s = strsav(s, "</code></a></li>\n");
- }
-#endif /* def TRUST_FILES */
- s = strsav(s, "</ul>");
+/*********************************************************************
+ *
+ * Function : get_url_actions
+ *
+ * Description : Gets the actions for this URL.
+ *
+ * Parameters :
+ * 1 : csp = Current client state (buffers, headers, etc...)
+ * 2 : http = http_request request for blocked URLs
+ *
+ * Returns : N/A
+ *
+ *********************************************************************/
+void get_url_actions(struct client_state *csp, struct http_request *http)
+{
+ struct file_list *fl;
+ struct url_actions *b;
+ int i;
+
+ init_current_action(csp->action);
-#else /* ifndef SPLIT_PROXY_ARGS */
- if (csp->clist)
+ for (i = 0; i < MAX_AF_FILES; i++)
{
- s = strsav(s, csp->clist->proxy_args);
- }
+ if (((fl = csp->actions_list[i]) == NULL) || ((b = fl->f) == NULL))
+ {
+ return;
+ }
- if (csp->flist)
- {
- s = strsav(s, csp->flist->proxy_args);
+ apply_url_actions(csp->action, http, b);
}
-#ifdef ACL_FILES
- if (csp->alist)
- {
- s = strsav(s, csp->alist->proxy_args);
- }
-#endif /* def ACL_FILES */
+ return;
+}
+
-#ifdef PCRS
- if (csp->rlist)
+/*********************************************************************
+ *
+ * Function : apply_url_actions
+ *
+ * Description : Applies a list of URL actions.
+ *
+ * Parameters :
+ * 1 : action = Destination.
+ * 2 : http = Current URL
+ * 3 : b = list of URL actions to apply
+ *
+ * Returns : N/A
+ *
+ *********************************************************************/
+void apply_url_actions(struct current_action_spec *action,
+ struct http_request *http,
+ struct url_actions *b)
+{
+ if (b == NULL)
{
- s = strsav(s, csp->rlist->proxy_args);
+ /* Should never happen */
+ return;
}
-#endif /* def PCRS */
-#ifdef TRUST_FILES
- if (csp->tlist)
+ for (b = b->next; NULL != b; b = b->next)
{
- s = strsav(s, csp->tlist->proxy_args);
+ if (url_match(b->url, http))
+ {
+ merge_current_action(action, b->action);
+ }
}
-#endif /* def TRUST_FILES */
-
-#endif /* ndef SPLIT_PROXY_ARGS */
-
- s = strsav(s, csp->config->proxy_args_trailer);
-
- return(s);
-
}
-static const char C_URL_INFO_HEADER[] =
- "HTTP/1.0 200 OK\n"
- "Pragma: no-cache\n"
- "Expires: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "Content-Type: text/html\n\n"
- "<html>\n"
- "<head>\n"
- "<title>Internet Junkbuster: URL Info</title>\n"
- "</head>\n"
- BODY
- "<h1><center>"
- BANNER
- "</h1></center>\n"
- "<p>Information for: <a href=\"http://%s\">http://%s</a></p>\n";
-static const char C_URL_INFO_FOOTER[] =
- "\n</p>\n"
- "</body>\n"
- "</html>\n";
-
-static const char C_URL_INFO_FORM[] =
- "HTTP/1.0 200 OK\n"
- "Pragma: no-cache\n"
- "Expires: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "Content-Type: text/html\n\n"
- "<html>\n"
- "<head>\n"
- "<title>Internet Junkbuster: URL Info</title>\n"
- "</head>\n"
- BODY
- "<h1><center>"
- BANNER
- "</h1></center>\n"
- "<form method=\"GET\" action=\"http://i.j.b/show-url-info\">\n"
- "<p>Please enter a URL, without the leading "http://":</p>"
- "<p><input type=\"text\" name=\"url\" size=\"80\">"
- "<input type=\"submit\" value=\"Info\"></p>\n"
- "</form>\n"
- "</body>\n"
- "</html>\n";
-
-
- /*********************************************************************
- *
- * Function : ijb_show_url_info
- *
- * Description : (please fill me in)
+/*********************************************************************
+ *
+ * Function : get_forward_override_settings
+ *
+ * Description : Returns forward settings as specified with the
+ * forward-override{} action. forward-override accepts
+ * forward lines similar to the one used in the
+ * configuration file, but without the URL pattern.
+ *
+ * For example:
+ *
+ * forward / .
+ *
+ * in the configuration file can be replaced with
+ * the action section:
+ *
+ * {+forward-override{forward .}}
+ * /
*
* Parameters :
- * 1 : http = http_request request for crunched URL
- * 2 : csp = Current client state (buffers, headers, etc...)
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : ???FIXME
+ * Returns : Pointer to forwarding structure in case of success.
+ * Invalid syntax is fatal.
*
*********************************************************************/
-char *ijb_show_url_info(struct http_request *http, struct client_state *csp)
+const static struct forward_spec *get_forward_override_settings(struct client_state *csp)
{
- char * query_string = strchr(http->path, '?');
- char * host = NULL;
-
- if (query_string != NULL)
+ const char *forward_override_line = csp->action->string[ACTION_STRING_FORWARD_OVERRIDE];
+ char forward_settings[BUFFER_SIZE];
+ char *http_parent = NULL;
+ /* variable names were chosen for consistency reasons. */
+ struct forward_spec *fwd = NULL;
+ int vec_count;
+ char *vec[3];
+
+ assert(csp->action->flags & ACTION_FORWARD_OVERRIDE);
+ /* Should be enforced by load_one_actions_file() */
+ assert(strlen(forward_override_line) < sizeof(forward_settings) - 1);
+
+ /* Create a copy ssplit can modify */
+ strlcpy(forward_settings, forward_override_line, sizeof(forward_settings));
+
+ if (NULL != csp->fwd)
{
- query_string = url_decode(query_string + 1);
- if (strncmpic(query_string, "url=", 4) == 0)
- {
- host = strdup(query_string + 4);
- }
- freez(query_string);
+ /*
+ * XXX: Currently necessary to prevent memory
+ * leaks when the show-url-info cgi page is visited.
+ */
+ unload_forward_spec(csp->fwd);
}
- if (host != NULL)
+
+ /*
+ * allocate a new forward node, valid only for
+ * the lifetime of this request. Save its location
+ * in csp as well, so sweep() can free it later on.
+ */
+ fwd = csp->fwd = zalloc(sizeof(*fwd));
+ if (NULL == fwd)
{
- char * result;
- char * path;
- char * s;
- int port = 80;
- struct file_list *fl;
- struct url_actions *b;
- struct url_spec url[1];
- struct current_action_spec action[1];
+ log_error(LOG_LEVEL_FATAL,
+ "can't allocate memory for forward-override{%s}", forward_override_line);
+ /* Never get here - LOG_LEVEL_FATAL causes program exit */
+ }
- init_current_action(action);
+ vec_count = ssplit(forward_settings, " \t", vec, SZ(vec), 1, 1);
+ if ((vec_count == 2) && !strcasecmp(vec[0], "forward"))
+ {
+ fwd->type = SOCKS_NONE;
- result = (char *)malloc(sizeof(C_URL_INFO_HEADER) + 2 * strlen(host));
- sprintf(result, C_URL_INFO_HEADER, host, host);
+ /* Parse the parent HTTP proxy host:port */
+ http_parent = vec[1];
- s = current_action_to_text(action);
- result = strsav(result, "<h3>Defaults:</h3>\n<p><b>{");
- result = strsav(result, s);
- result = strsav(result, " }</b></p>\n<h3>Patterns affecting the URL:</h3>\n<p>\n");
- freez(s);
+ }
+ else if (vec_count == 3)
+ {
+ char *socks_proxy = NULL;
- s = strchr(host, '/');
- if (s != NULL)
+ if (!strcasecmp(vec[0], "forward-socks4"))
{
- path = strdup(s);
- *s = '\0';
+ fwd->type = SOCKS_4;
+ socks_proxy = vec[1];
}
- else
+ else if (!strcasecmp(vec[0], "forward-socks4a"))
{
- path = strdup("");
+ fwd->type = SOCKS_4A;
+ socks_proxy = vec[1];
}
- s = strchr(host, ':');
- if (s != NULL)
+ else if (!strcasecmp(vec[0], "forward-socks5"))
{
- *s++ = '\0';
- port = atoi(s);
+ fwd->type = SOCKS_5;
+ socks_proxy = vec[1];
}
- if (((fl = csp->actions_list) == NULL) || ((b = fl->f) == NULL))
+ if (NULL != socks_proxy)
{
- freez(host);
- freez(path);
- result = strsav(result, C_URL_INFO_FOOTER);
- return result;
- }
-
- *url = dsplit(host);
+ /* Parse the SOCKS proxy host[:port] */
+ fwd->gateway_host = strdup(socks_proxy);
- /* if splitting the domain fails, punt */
- if (url->dbuf == NULL)
- {
- freez(host);
- freez(path);
- result = strsav(result, C_URL_INFO_FOOTER);
- return result;
- }
+ if (NULL != (socks_proxy = strchr(fwd->gateway_host, ':')))
+ {
+ *socks_proxy++ = '\0';
+ fwd->gateway_port = strtol(socks_proxy, NULL, 0);
+ }
- for (b = b->next; NULL != b; b = b->next)
- {
- if ((b->url->port == 0) || (b->url->port == port))
+ if (fwd->gateway_port <= 0)
{
- if ((b->url->domain[0] == '\0') || (domaincmp(b->url, url) == 0))
- {
- if ((b->url->path == NULL) ||
-#ifdef REGEX
- (regexec(b->url->preg, path, 0, NULL, 0) == 0)
-#else
- (strncmp(b->url->path, path, b->url->pathlen) == 0)
-#endif
- )
- {
- s = actions_to_text(b->action);
- result = strsav(result, "<b>{");
- result = strsav(result, s);
- result = strsav(result, " }</b><br>\n<code>");
- result = strsav(result, b->url->spec);
- result = strsav(result, "</code><br>\n<br>\n");
- freez(s);
-
- merge_current_action(action, b->action);
- }
- }
+ fwd->gateway_port = 1080;
}
- }
- freez(url->dbuf);
- freez(url->dvec);
+ http_parent = vec[2];
+ }
+ }
- freez(host);
- freez(path);
+ if (NULL == http_parent)
+ {
+ log_error(LOG_LEVEL_FATAL,
+ "Invalid forward-override syntax in: %s", forward_override_line);
+ /* Never get here - LOG_LEVEL_FATAL causes program exit */
+ }
- s = current_action_to_text(action);
- result = strsav(result, "</p>\n<h2>Final Results:</h2>\n<p><b>{");
- result = strsav(result, s);
- result = strsav(result, " }</b><br>\n<br>\n");
- freez(s);
+ /* Parse http forwarding settings */
+ if (strcmp(http_parent, ".") != 0)
+ {
+ fwd->forward_host = strdup(http_parent);
- free_current_action(action);
+ if (NULL != (http_parent = strchr(fwd->forward_host, ':')))
+ {
+ *http_parent++ = '\0';
+ fwd->forward_port = strtol(http_parent, NULL, 0);
+ }
- result = strsav(result, C_URL_INFO_FOOTER);
- return result;
- }
- else
- {
- return strdup(C_URL_INFO_FORM);
+ if (fwd->forward_port <= 0)
+ {
+ fwd->forward_port = 8000;
+ }
}
-}
+ assert (NULL != fwd);
-/*********************************************************************
- *
- * Function : ijb_send_banner
- *
- * Description : This "crunch"es "http://i.j.b/ijb-send-banner and
- * sends the image.
- *
- * Parameters :
- * 1 : http = http_request request for crunched URL
- * 2 : csp = Current client state (buffers, headers, etc...)
- *
- * Returns : NULL, indicating that it has already sent the data.
- *
- *********************************************************************/
-char *ijb_send_banner(struct http_request *http, struct client_state *csp)
-{
- write_socket(csp->cfd, JBGIF, sizeof(JBGIF)-1);
-
- return(NULL);
+ log_error(LOG_LEVEL_CONNECT,
+ "Overriding forwarding settings based on \'%s\'", forward_override_line);
+
+ return fwd;
}
-#ifdef TRUST_FILES
+
/*********************************************************************
*
- * Function : ij_untrusted_url
+ * Function : forward_url
*
- * Description : This "crunch"es "http:/any.thing/ij-untrusted-url" and
- * returns a web page describing why it was untrusted.
+ * Description : Should we forward this to another proxy?
*
* Parameters :
- * 1 : http = http_request request for crunched URL
- * 2 : csp = Current client state (buffers, headers, etc...)
+ * 1 : csp = Current client state (buffers, headers, etc...)
+ * 2 : http = http_request request for current URL
*
- * Returns : A string that contains why this was untrusted.
+ * Returns : Pointer to forwarding information.
*
*********************************************************************/
-char *ij_untrusted_url(struct http_request *http, struct client_state *csp)
+const struct forward_spec *forward_url(struct client_state *csp,
+ const struct http_request *http)
{
- int n;
- char *hostport, *path, *refer, *p, *v[9];
- char buf[BUFSIZ];
- struct url_spec **tl, *t;
-
-
- static const char format[] =
- "HTTP/1.0 200 OK\r\n"
- "Pragma: no-cache\n"
- "Last-Modified: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "Expires: Thu Jul 31, 1997 07:42:22 pm GMT\n"
- "Content-Type: text/html\n\n"
- "<html>\n"
- "<head>\n"
- "<title>Internet Junkbuster: Request for untrusted URL</title>\n"
- "</head>\n"
- BODY
- "<center><h1>"
- BANNER
- "</h1></center>"
- "The " BANNER " Proxy "
- "<A href=\"" HOME_PAGE_URL "\">"
- "(" HOME_PAGE_URL ") </A>"
- "intercepted the request for %s%s\n"
- "because the URL is not trusted.\n"
- "<br><br>\n";
-
- if ((n = ssplit(http->path, "?+", v, SZ(v), 0, 0)) == 4)
- {
- hostport = url_decode(v[1]);
- path = url_decode(v[2]);
- refer = url_decode(v[3]);
- }
- else
- {
- hostport = strdup("undefined_host");
- path = strdup("/undefined_path");
- refer = strdup("undefined");
- }
+ static const struct forward_spec fwd_default[1] = { FORWARD_SPEC_INITIALIZER };
+ struct forward_spec *fwd = csp->config->forward;
- n = sizeof(format);
- n += strlen(hostport);
- n += strlen(path );
-
- if ((p = (char *)malloc(n)))
+ if (csp->action->flags & ACTION_FORWARD_OVERRIDE)
{
- sprintf(p, format, hostport, path);
+ return get_forward_override_settings(csp);
}
- strsav(p, "The referrer in this request was <strong>");
- strsav(p, refer);
- strsav(p, "</strong><br>\n");
-
- freez(hostport);
- freez(path );
- freez(refer );
-
- p = strsav(p, "<h3>The following referrers are trusted</h3>\n");
-
- for (tl = csp->config->trust_list; (t = *tl) ; tl++)
+ if (fwd == NULL)
{
- sprintf(buf, "%s<br>\n", t->spec);
- p = strsav(p, buf);
+ return fwd_default;
}
- if (csp->config->trust_info->next)
+ while (fwd != NULL)
{
- struct list *l;
-
- strcpy(buf,
- "<p>"
- "You can learn more about what this means "
- "and what you may be able to do about it by "
- "reading the following documents:<br>\n"
- "<ol>\n"
- );
-
- p = strsav(p, buf);
-
- for (l = csp->config->trust_info->next; l ; l = l->next)
+ if (url_match(fwd->url, http))
{
- sprintf(buf,
- "<li> <a href=%s>%s</a><br>\n",
- l->str, l->str);
- p = strsav(p, buf);
+ return fwd;
}
-
- p = strsav(p, "</ol>\n");
+ fwd = fwd->next;
}
- p = strsav(p, "</body>\n" "</html>\n");
-
- return(p);
-
+ return fwd_default;
}
-#endif /* def TRUST_FILES */
-#ifdef STATISTICS
/*********************************************************************
*
- * Function : add_stats
+ * Function : direct_response
*
- * Description : Statistics function of JB. Called by `show_proxy_args'.
+ * Description : Check if Max-Forwards == 0 for an OPTIONS or TRACE
+ * request and if so, return a HTTP 501 to the client.
*
- * Parameters :
- * 1 : s = string that holds the proxy args description page
+ * FIXME: I have a stupid name and I should handle the
+ * requests properly. Still, what we do here is rfc-
+ * compliant, whereas ignoring or forwarding are not.
+ *
+ * Parameters :
+ * 1 : csp = Current client state (buffers, headers, etc...)
*
- * Returns : A pointer to the descriptive status web page.
+ * Returns : http_response if , NULL if nonmatch or handler fail
*
*********************************************************************/
-char *add_stats(char *s)
+struct http_response *direct_response(struct client_state *csp)
{
- /*
- * Output details of the number of requests rejected and
- * accepted. This is switchable in the junkbuster config.
- * Does nothing if this option is not enabled.
- */
-
- float perc_rej; /* Percentage of http requests rejected */
- char out_str[81];
- int local_urls_read = urls_read;
- int local_urls_rejected = urls_rejected;
+ struct http_response *rsp;
+ struct list_entry *p;
- /*
- * Need to alter the stats not to include the fetch of this
- * page.
- *
- * Can't do following thread safely! doh!
- *
- * urls_read--;
- * urls_rejected--; * This will be incremented subsequently *
- */
-
- s = strsav(s,"<h2>Statistics for this " BANNER ":</h2>\n");
-
- if (local_urls_read == 0)
+ if ((0 == strcmpic(csp->http->gpc, "trace"))
+ || (0 == strcmpic(csp->http->gpc, "options")))
{
+ for (p = csp->headers->first; (p != NULL) ; p = p->next)
+ {
+ if (!strncmpic("Max-Forwards:", p->str, 13))
+ {
+ unsigned int max_forwards;
- s = strsav(s,"No activity so far!\n");
-
- }
- else
- {
+ /*
+ * If it's a Max-Forwards value of zero,
+ * we have to intercept the request.
+ */
+ if (1 == sscanf(p->str+12, ": %u", &max_forwards) && max_forwards == 0)
+ {
+ /*
+ * FIXME: We could handle at least TRACE here,
+ * but that would require a verbatim copy of
+ * the request which we don't have anymore
+ */
+ log_error(LOG_LEVEL_HEADER,
+ "Detected header \'%s\' in OPTIONS or TRACE request. Returning 501.",
+ p->str);
- perc_rej = (float)local_urls_rejected * 100.0F /
- (float)local_urls_read;
+ /* Get mem for response or fail*/
+ if (NULL == (rsp = alloc_http_response()))
+ {
+ return cgi_error_memory();
+ }
+
+ if (NULL == (rsp->status = strdup("501 Not Implemented")))
+ {
+ free_http_response(rsp);
+ return cgi_error_memory();
+ }
- sprintf(out_str,
- "%d requests received, %d filtered "
- "(%6.2f %%).",
- local_urls_read,
- local_urls_rejected, perc_rej);
+ rsp->is_static = 1;
+ rsp->reason = RSP_REASON_UNSUPPORTED;
- s = strsav(s,out_str);
+ return(finish_http_response(csp, rsp));
+ }
+ }
+ }
}
-
- return(s);
+ return NULL;
}
-#endif /* def STATISTICS */
+/*********************************************************************
+ *
+ * Function : content_filters_enabled
+ *
+ * Description : Checks whether there are any content filters
+ * enabled for the current request.
+ *
+ * Parameters :
+ * 1 : action = Action spec to check.
+ *
+ * Returns : TRUE for yes, FALSE otherwise
+ *
+ *********************************************************************/
+int content_filters_enabled(const struct current_action_spec *action)
+{
+ return ((action->flags & ACTION_DEANIMATE) ||
+ !list_is_empty(action->multi[ACTION_MULTI_FILTER]));
+}
+
/*
Local Variables:
tab-width: 3