1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
4 #define PROJECT_H_VERSION "$Id: project.h,v 1.123 2008/11/10 16:55:59 fabiankeil Exp $"
5 /*********************************************************************
7 * File : $Source: /cvsroot/ijbswa/current/project.h,v $
9 * Purpose : Defines data structures which are widely used in the
10 * project. Does not define any variables or functions
11 * (though it does declare some macros).
13 * Copyright : Written by and Copyright (C) 2001-2008 the SourceForge
14 * Privoxy team. http://www.privoxy.org/
16 * Based on the Internet Junkbuster originally written
17 * by and Copyright (C) 1997 Anonymous Coders and
18 * Junkbusters Corporation. http://www.junkbusters.com
20 * This program is free software; you can redistribute it
21 * and/or modify it under the terms of the GNU General
22 * Public License as published by the Free Software
23 * Foundation; either version 2 of the License, or (at
24 * your option) any later version.
26 * This program is distributed in the hope that it will
27 * be useful, but WITHOUT ANY WARRANTY; without even the
28 * implied warranty of MERCHANTABILITY or FITNESS FOR A
29 * PARTICULAR PURPOSE. See the GNU General Public
30 * License for more details.
32 * The GNU General Public License should be included with
33 * this file. If not, you can view it at
34 * http://www.gnu.org/copyleft/gpl.html
35 * or write to the Free Software Foundation, Inc., 59
36 * Temple Place - Suite 330, Boston, MA 02111-1307, USA.
40 * Revision 1.123 2008/11/10 16:55:59 fabiankeil
41 * Fix a gcc44 warning (in filters.c).
43 * Revision 1.122 2008/10/16 07:11:34 fabiankeil
44 * Fix a bunch of gcc44 conversion warnings.
46 * Revision 1.121 2008/10/09 18:21:41 fabiankeil
47 * Flush work-in-progress changes to keep outgoing connections
48 * alive where possible. Incomplete and mostly #ifdef'd out.
50 * Revision 1.120 2008/09/21 13:36:52 fabiankeil
51 * If change-x-forwarded-for{add} is used and the client
52 * sends multiple X-Forwarded-For headers, append the client's
53 * IP address to each one of them. "Traditionally" we would
54 * lose all but the last one.
56 * Revision 1.119 2008/09/20 10:04:33 fabiankeil
57 * Remove hide-forwarded-for-headers action which has
58 * been obsoleted by change-x-forwarded-for{block}.
60 * Revision 1.118 2008/09/19 15:26:29 fabiankeil
61 * Add change-x-forwarded-for{} action to block or add
62 * X-Forwarded-For headers. Mostly based on code removed
65 * Revision 1.117 2008/08/30 12:03:07 fabiankeil
66 * Remove FEATURE_COOKIE_JAR.
68 * Revision 1.116 2008/05/20 16:05:02 fabiankeil
69 * Move parsers structure definition from project.h to parsers.h.
71 * Revision 1.115 2008/05/19 16:57:20 fabiankeil
72 * Declare all members of the parsers structure immutable.
74 * Revision 1.114 2008/04/11 16:35:39 fabiankeil
75 * Oops, I forgot to shorten the URL_SPEC_INITIALIZER in my last commit.
77 * Revision 1.113 2008/04/10 14:41:04 fabiankeil
78 * Ditch url_spec's path member now that it's no longer used.
80 * Revision 1.112 2008/04/06 15:18:34 fabiankeil
81 * Oh well, rename the --enable-pcre-host-patterns option to
82 * --enable-extended-host-patterns as it's not really PCRE syntax.
84 * Revision 1.111 2008/04/06 14:54:26 fabiankeil
85 * Use PCRE syntax in host patterns when configured
86 * with --enable-pcre-host-patterns.
88 * Revision 1.110 2008/03/29 12:13:46 fabiankeil
89 * Remove send-wafer and send-vanilla-wafer actions.
91 * Revision 1.109 2008/03/28 15:13:41 fabiankeil
92 * Remove inspect-jpegs action.
94 * Revision 1.108 2008/03/27 18:27:36 fabiankeil
95 * Remove kill-popups action.
97 * Revision 1.107 2008/03/26 18:07:08 fabiankeil
98 * Add hostname directive. Closes PR#1918189.
100 * Revision 1.106 2008/03/24 11:21:03 fabiankeil
101 * Share the action settings for multiple patterns in the same
102 * section so we waste less memory for gigantic block lists
103 * (and load them slightly faster). Reported by Franz Schwartau.
105 * Revision 1.105 2008/03/21 11:16:27 fabiankeil
106 * Garbage-collect csp->my_ip_addr_str and csp->my_hostname.
108 * Revision 1.104 2008/03/04 18:30:40 fabiankeil
109 * Remove the treat-forbidden-connects-like-blocks action. We now
110 * use the "blocked" page for forbidden CONNECT requests by default.
112 * Revision 1.103 2008/03/01 14:00:45 fabiankeil
113 * Let the block action take the reason for the block
114 * as argument and show it on the "blocked" page.
116 * Revision 1.102 2008/02/03 13:46:14 fabiankeil
117 * Add SOCKS5 support. Patch #1862863 by Eric M. Hopper with minor changes.
119 * Revision 1.101 2007/12/07 18:29:23 fabiankeil
120 * Remove now-obsolete csp member x_forwarded.
122 * Revision 1.100 2007/09/02 13:42:11 fabiankeil
123 * - Allow port lists in url patterns.
124 * - Ditch unused url_spec member pathlen.
126 * Revision 1.99 2007/07/21 11:51:36 fabiankeil
127 * As Hal noticed, checking dispatch_cgi() as the last cruncher
128 * looks like a bug if CGI requests are blocked unintentionally,
129 * so don't do it unless the user enabled the new config option
130 * "allow-cgi-request-crunching".
132 * Revision 1.98 2007/07/14 07:31:26 fabiankeil
133 * Add new csp->content_type flag (CT_DECLARED).
135 * Revision 1.97 2007/05/27 12:38:08 fabiankeil
136 * - Remove some left-overs from the switch to dedicated header filters.
137 * - Adjust "X-Filter: No" to disable dedicated header filters.
138 * - Prepare for forward-override{}
140 * Revision 1.96 2007/05/14 10:41:15 fabiankeil
141 * Ditch the csp member cookie_list[] which isn't used anymore.
143 * Revision 1.95 2007/04/30 15:02:19 fabiankeil
144 * Introduce dynamic pcrs jobs that can resolve variables.
146 * Revision 1.94 2007/04/15 16:39:21 fabiankeil
147 * Introduce tags as alternative way to specify which
148 * actions apply to a request. At the moment tags can be
149 * created based on client and server headers.
151 * Revision 1.93 2007/03/20 15:16:34 fabiankeil
152 * Use dedicated header filter actions instead of abusing "filter".
153 * Replace "filter-client-headers" and "filter-client-headers"
154 * with "server-header-filter" and "client-header-filter".
156 * Revision 1.92 2007/03/17 15:20:05 fabiankeil
157 * New config option: enforce-blocks.
159 * Revision 1.91 2007/03/05 13:28:03 fabiankeil
160 * Add some CSP_FLAGs for the header parsers.
162 * Revision 1.90 2007/02/07 10:36:16 fabiankeil
163 * Add new http_response member to save
164 * the reason why the response was generated.
166 * Revision 1.89 2007/01/27 13:09:16 fabiankeil
167 * Add new config option "templdir" to
168 * change the templates directory.
170 * Revision 1.88 2007/01/25 13:36:59 fabiankeil
171 * Add csp->error_message for failure reasons
172 * that should be embedded into the CGI pages.
174 * Revision 1.87 2007/01/01 19:36:37 fabiankeil
175 * Integrate a modified version of Wil Mahan's
176 * zlib patch (PR #895531).
178 * Revision 1.86 2006/12/31 17:56:37 fabiankeil
179 * Added config option accept-intercepted-requests
180 * and disabled it by default.
182 * Revision 1.85 2006/12/31 15:03:31 fabiankeil
183 * Fix gcc43 compiler warnings and a comment.
185 * Revision 1.84 2006/12/21 12:57:48 fabiankeil
186 * Add config option "split-large-forms"
187 * to work around the browser bug reported
190 * Revision 1.83 2006/12/06 19:26:29 fabiankeil
191 * Moved HTTP snipplets into jcc.c. They aren't
192 * used anywhere else.
194 * Revision 1.82 2006/09/20 15:50:31 fabiankeil
195 * Doubled size of HOSTENT_BUFFER_SIZE to mask
196 * problems with gethostbyname_r and some
197 * /etc/hosts configurations. Only a workaround
198 * until we get the real fix ready.
199 * Thanks Félix Rauch for reporting.
201 * Increased value of MAX_TRUSTED_REFERRERS from 64 to 512.
203 * Revision 1.81 2006/09/06 13:03:04 fabiankeil
204 * Respond with 400 and a short text message
205 * if the client tries to use Privoxy as FTP proxy.
207 * Revision 1.80 2006/09/06 10:43:32 fabiankeil
208 * Added config option enable-remote-http-toggle
209 * to specify if Privoxy should recognize special
210 * headers (currently only X-Filter) to change its
211 * behaviour. Disabled by default.
213 * Revision 1.79 2006/09/06 09:23:37 fabiankeil
214 * Make number of retries in case of forwarded-connect problems
215 * a config file option (forwarded-connect-retries) and use 0 as
218 * Revision 1.78 2006/08/31 16:25:06 fabiankeil
219 * Work around a buffer overflow that caused Privoxy to
220 * segfault if too many trusted referrers were used. Good
221 * enough for now, but should be replaced with a real
222 * solution after the next release.
224 * Revision 1.77 2006/08/21 12:50:51 david__schmidt
227 * Revision 1.76 2006/08/14 08:25:19 fabiankeil
228 * Split filter-headers{} into filter-client-headers{}
229 * and filter-server-headers{}.
230 * Added parse_header_time() to share some code.
231 * Replaced timegm() with mktime().
233 * Revision 1.75 2006/08/03 02:46:41 david__schmidt
234 * Incorporate Fabian Keil's patch work:
\r * http://www.fabiankeil.de/sourcecode/privoxy/
236 * Revision 1.74 2006/07/18 14:48:47 david__schmidt
237 * Reorganizing the repository: swapping out what was HEAD (the old 3.1 branch)
238 * with what was really the latest development (the v_3_0_branch branch)
240 * Revision 1.72.2.7 2006/01/29 23:10:56 david__schmidt
241 * Multiple filter file support
243 * Revision 1.72.2.6 2004/10/03 12:53:46 david__schmidt
244 * Add the ability to check jpeg images for invalid
245 * lengths of comment blocks. Defensive strategy
246 * against the exploit:
247 * Microsoft Security Bulletin MS04-028
248 * Buffer Overrun in JPEG Processing (GDI+) Could
249 * Allow Code Execution (833987)
250 * Enabled with +inspect-jpegs in actions files.
252 * Revision 1.72.2.5 2004/01/30 15:29:29 oes
253 * Updated the copyright note
255 * Revision 1.72.2.4 2004/01/13 16:12:14 oes
256 * Fixed double slash in USER_MANUAL_URL. Closes BR #867088.
258 * Revision 1.72.2.3 2003/03/11 11:54:37 oes
259 * Introduced RC_FLAG_* flags for use in child process return code
261 * Revision 1.72.2.2 2002/11/28 18:15:44 oes
262 * Added flag to each cgi_dispatcher that allows or denies
263 * external linking and removed const qualifier from
264 * struct list_entry.str.
266 * Revision 1.72.2.1 2002/08/10 11:25:18 oes
267 * - Include config.h for access to config data
268 * - Include <pcre*.h> depending on where they are
270 * Revision 1.72 2002/05/14 21:35:49 oes
271 * Split HELP_LINK_PREFIX into ACTIONS_HELP_PREFIX and CONFIG_HELP_PREFIX
272 * because of split in user-manual
274 * Revision 1.71 2002/05/12 21:39:36 jongfoster
275 * - Adding Doxygen-style comments to structures and #defines.
277 * Revision 1.70 2002/05/12 16:05:50 jongfoster
278 * Fixing ACTION_MASK_ALL to be unsigned long rather than
279 * just unsigned int. I don't know if anyone is porting
280 * Privoxy to 16-bit platforms, but if so, +limit-connect
281 * wouldn't have worked because of this bug.
283 * Revision 1.69 2002/05/08 16:00:16 oes
284 * Added size member to struct iob, so it can
285 * be alloced larger than needed.
287 * Revision 1.68 2002/04/26 12:56:00 oes
288 * Killed REDIRECT_URL, added USER_MANUAL_URL and HELP_LINK_PREFIX
290 * Revision 1.67 2002/04/24 02:12:43 oes
291 * - Jon's multiple AF patch:
292 * - Make csp->actions_list an array
293 * - #define MAX_AF_FILES
294 * - Moved CGI_PARAM_LEN_MAX (500) here
296 * Revision 1.66 2002/04/15 19:06:43 jongfoster
299 * Revision 1.65 2002/04/04 00:36:36 gliptak
300 * always use pcre for matching
302 * Revision 1.64 2002/04/03 22:28:03 gliptak
303 * Removed references to gnu_regex
305 * Revision 1.63 2002/03/31 17:19:00 jongfoster
306 * Win32 only: Enabling STRICT to fix a VC++ compile warning.
308 * Revision 1.62 2002/03/26 22:48:49 swa
311 * Revision 1.61 2002/03/26 22:29:55 swa
312 * we have a new homepage!
314 * Revision 1.60 2002/03/24 15:52:17 jongfoster
315 * Changing CGI URL prefixes for new name
317 * Revision 1.59 2002/03/24 15:23:33 jongfoster
320 * Revision 1.58 2002/03/24 13:25:43 swa
321 * name change related issues
323 * Revision 1.57 2002/03/16 20:28:34 oes
324 * Added descriptions to the filters so users will know what they select in the cgi editor
326 * Revision 1.56 2002/03/13 20:27:30 oes
327 * Fixing bug with CT_TABOO
329 * Revision 1.55 2002/03/12 01:42:50 oes
330 * Introduced modular filters
332 * Revision 1.54 2002/03/09 20:03:52 jongfoster
333 * - Making various functions return int rather than size_t.
334 * (Undoing a recent change). Since size_t is unsigned on
335 * Windows, functions like read_socket that return -1 on
336 * error cannot return a size_t.
338 * THIS WAS A MAJOR BUG - it caused frequent, unpredictable
339 * crashes, and also frequently caused JB to jump to 100%
340 * CPU and stay there. (Because it thought it had just
341 * read ((unsigned)-1) == 4Gb of data...)
343 * - The signature of write_socket has changed, it now simply
344 * returns success=0/failure=nonzero.
346 * - Trying to get rid of a few warnings --with-debug on
347 * Windows, I've introduced a new type "jb_socket". This is
348 * used for the socket file descriptors. On Windows, this
349 * is SOCKET (a typedef for unsigned). Everywhere else, it's
350 * an int. The error value can't be -1 any more, so it's
351 * now JB_INVALID_SOCKET (which is -1 on UNIX, and in
352 * Windows it maps to the #define INVALID_SOCKET.)
354 * - The signature of bind_port has changed.
356 * Revision 1.53 2002/03/08 16:48:55 oes
357 * Added FEATURE_NO_GIFS and BUILTIN_IMAGE_MIMETYPE
359 * Revision 1.52 2002/03/07 03:46:17 oes
360 * Fixed compiler warnings
362 * Revision 1.51 2002/03/05 04:52:42 oes
363 * Deleted non-errlog debugging code
365 * Revision 1.50 2002/03/04 19:32:07 oes
366 * Changed default port to 8118
368 * Revision 1.49 2002/03/04 18:28:55 oes
369 * Deleted PID_FILE_NAME
371 * Revision 1.48 2002/03/03 14:50:40 oes
372 * Fixed CLF logging: Added ocmd member for client's request to struct http_request
374 * Revision 1.47 2002/02/20 23:15:13 jongfoster
375 * Parsing functions now handle out-of-memory gracefully by returning
378 * Revision 1.46 2002/01/17 21:06:09 jongfoster
379 * Now #defining the URLs of the config interface
381 * Minor changes to struct http_request and struct url_spec due to
382 * standardizing that struct http_request is used to represent a URL, and
383 * struct url_spec is used to represent a URL pattern. (Before, URLs were
384 * represented as seperate variables and a partially-filled-in url_spec).
386 * Revision 1.45 2002/01/09 14:33:27 oes
387 * Added HOSTENT_BUFFER_SIZE
389 * Revision 1.44 2001/12/30 14:07:32 steudten
390 * - Add signal handling (unix)
391 * - Add SIGHUP handler (unix)
392 * - Add creation of pidfile (unix)
393 * - Add action 'top' in rc file (RH)
394 * - Add entry 'SIGNALS' to manpage
395 * - Add exit message to logfile (unix)
397 * Revision 1.43 2001/11/22 21:57:51 jongfoster
398 * Making action_spec->flags into an unsigned long rather than just an
400 * Adding ACTION_NO_COOKIE_KEEP
402 * Revision 1.42 2001/11/05 21:42:41 steudten
403 * Include DBG() macro.
405 * Revision 1.41 2001/10/28 19:12:06 jongfoster
406 * Adding ijb_toupper()
408 * Revision 1.40 2001/10/26 17:40:47 oes
409 * Moved ijb_isspace and ijb_tolower to project.h
410 * Removed http->user_agent, csp->referrer and csp->accept_types
412 * Revision 1.39 2001/10/25 03:45:02 david__schmidt
413 * Adding a (void*) cast to freez() because Visual Age C++ won't expand the
414 * macro when called with a cast; so moving the cast to the macro def'n
415 * seems to both eliminate compiler warnings (on darwin and OS/2, anyway) and
416 * doesn't make macro expansion complain. Hope this works for everyone else
419 * Revision 1.38 2001/10/23 21:19:04 jongfoster
420 * New error-handling support: jb_err type and JB_ERR_xxx constants
421 * CGI functions now return a jb_err, and their parameters map is const.
422 * Support for RUNTIME_FEATUREs to enable/disable config editor
423 * Adding a few comments
425 * Revision 1.37 2001/10/14 22:14:01 jongfoster
426 * Removing name_length field from struct cgi_dispatcher, as this is
427 * now calculated at runtime from the "name" field.
429 * Revision 1.36 2001/10/10 16:45:15 oes
430 * Added LIMIT_CONNECT action and string
431 * Fixed HTTP message line termination
432 * Added CFORBIDDEN HTTP message
434 * Revision 1.35 2001/10/07 18:06:43 oes
435 * Added status member to struct http_request
437 * Revision 1.34 2001/10/07 15:45:25 oes
438 * Added url member to struct http_request and commented all
443 * Added ACTION_DOWNGRADE and ACTION_NO_COMPRESSION
445 * Replaced struct client_state members rejected,
446 * force, active and toggled_on with "flags" bitmap.
448 * Added CSP_FLAG_MODIFIED and CSP_FLAG_CHUNKED
450 * Added buffer_limit to struct configuration_spec
452 * Revision 1.33 2001/09/20 13:30:08 steudten
454 * Make freez() more secure in case of: if (exp) { free(z) ; a=*z }
455 * Last case will set z to NULL in free(z) and thats bad..
457 * Revision 1.32 2001/09/16 23:02:51 jongfoster
460 * Revision 1.31 2001/09/16 13:20:29 jongfoster
461 * Rewrite of list library. Now has seperate header and list_entry
462 * structures. Also added a large sprinking of assert()s to the list
465 * Revision 1.30 2001/09/13 23:52:00 jongfoster
466 * Support for both static and dynamically generated CGI pages
468 * Revision 1.29 2001/09/13 23:29:43 jongfoster
469 * Defining FORWARD_SPEC_INITIALIZER
471 * Revision 1.28 2001/09/13 23:05:50 jongfoster
472 * Changing the string paramater to the header parsers a "const".
474 * Revision 1.27 2001/08/05 16:06:20 jongfoster
475 * Modifiying "struct map" so that there are now separate header and
476 * "map_entry" structures. This means that functions which modify a
477 * map no longer need to return a pointer to the modified map.
478 * Also, it no longer reverses the order of the entries (which may be
479 * important with some advanced template substitutions).
481 * Revision 1.26 2001/07/30 22:08:36 jongfoster
482 * Tidying up #defines:
483 * - All feature #defines are now of the form FEATURE_xxx
484 * - Permanently turned off WIN_GUI_EDIT
485 * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
487 * Revision 1.25 2001/07/29 18:43:08 jongfoster
488 * Changing #ifdef _FILENAME_H to FILENAME_H_INCLUDED, to conform to
491 * Revision 1.24 2001/07/25 17:20:27 oes
492 * Introduced http->user_agent
494 * Revision 1.23 2001/07/18 12:32:23 oes
495 * - Added ACTION_STRING_DEANIMATE
496 * - moved #define freez from jcc.h to project.h
498 * Revision 1.22 2001/07/15 17:51:41 jongfoster
499 * Renaming #define STATIC to STATIC_PCRE
501 * Revision 1.21 2001/07/13 14:03:19 oes
502 * - Reorganized regex header inclusion and #defines to
503 * comply to the scheme in configure.in
504 * - Added csp->content_type and its CT_* keys
505 * - Added ACTION_DEANIMATE
506 * - Removed all #ifdef PCRS
508 * Revision 1.20 2001/06/29 21:45:41 oes
509 * Indentation, CRLF->LF, Tab-> Space
511 * Revision 1.19 2001/06/29 13:33:36 oes
512 * - Improved comments
513 * - Introduced http_request.host_ip_addr_str
514 * - Introduced http_response.head_length
515 * - Introduced config.my_ip_addr_str, config.my_hostname,
516 * config.admin_address and config.proxy_info_url
517 * - Removed config.proxy_args_header and config.proxy_args_trailer,
518 * renamed config.proxy_args_invocation to config.proxy_args
519 * - Removed HTML snipplets and GIFs
520 * - Removed logentry from cancelled commit
522 * Revision 1.18 2001/06/09 10:57:39 jongfoster
523 * Adding definition of BUFFER_SIZE.
524 * Changing struct cgi_dispatcher to use "const" strings.
526 * Revision 1.17 2001/06/07 23:15:09 jongfoster
527 * Merging ACL and forward files into config file.
528 * Moving struct gateway members into struct forward_spec
529 * Removing config->proxy_args_gateways
530 * Cosmetic: Adding a few comments
532 * Revision 1.16 2001/06/04 18:31:58 swa
533 * files are now prefixed with either `confdir' or `logdir'.
534 * `make redhat-dist' replaces both entries confdir and logdir
537 * Revision 1.15 2001/06/04 11:28:53 swa
538 * redirect did not work due to missing /
540 * Revision 1.14 2001/06/03 11:03:48 oes
542 * added struct http_response,
543 * changed struct interceptors to struct cgi_dispatcher,
544 * moved HTML stuff to cgi.h
546 * Revision 1.13 2001/06/01 20:05:36 jongfoster
547 * Support for +image-blocker{}: added ACTION_IMAGE_BLOCKER
548 * constant, and removed csp->tinygif.
550 * Revision 1.12 2001/06/01 18:49:17 jongfoster
551 * Replaced "list_share" with "list" - the tiny memory gain was not
552 * worth the extra complexity.
554 * Revision 1.11 2001/06/01 10:32:47 oes
555 * Added constants for anchoring selection bitmap
557 * Revision 1.10 2001/05/31 21:33:53 jongfoster
558 * Changes for new actions file, replacing permissionsfile
559 * and parts of the config file. Also added support for
562 * Revision 1.9 2001/05/31 17:32:31 oes
564 * - Enhanced domain part globbing with infix and prefix asterisk
565 * matching and optional unanchored operation
567 * Revision 1.8 2001/05/29 20:09:15 joergs
568 * HTTP_REDIRECT_TEMPLATE fixed.
570 * Revision 1.7 2001/05/29 09:50:24 jongfoster
571 * Unified blocklist/imagelist/actionslist.
572 * File format is still under discussion, but the internal changes
575 * Also modified interceptor behaviour:
576 * - We now intercept all URLs beginning with one of the following
577 * prefixes (and *only* these prefixes):
579 * * http://ijbswa.sf.net/config/
580 * * http://ijbswa.sourceforge.net/config/
581 * - New interceptors "home page" - go to http://i.j.b/ to see it.
582 * - Internal changes so that intercepted and fast redirect pages
583 * are not replaced with an image.
584 * - Interceptors now have the option to send a binary page direct
585 * to the client. (i.e. ijb-send-banner uses this)
586 * - Implemented show-url-info interceptor. (Which is why I needed
587 * the above interceptors changes - a typical URL is
588 * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif".
589 * The previous mechanism would not have intercepted that, and
590 * if it had been intercepted then it then it would have replaced
593 * Revision 1.6 2001/05/27 22:17:04 oes
595 * - re_process_buffer no longer writes the modified buffer
596 * to the client, which was very ugly. It now returns the
597 * buffer, which it is then written by chat.
599 * - content_length now adjusts the Content-Length: header
600 * for modified documents rather than crunch()ing it.
601 * (Length info in csp->content_length, which is 0 for
602 * unmodified documents)
604 * - For this to work, sed() is called twice when filtering.
606 * Revision 1.5 2001/05/26 00:28:36 jongfoster
607 * Automatic reloading of config file.
608 * Removed obsolete SIGHUP support (Unix) and Reload menu option (Win32).
609 * Most of the global variables have been moved to a new
610 * struct configuration_spec, accessed through csp->config->globalname
611 * Most of the globals remaining are used by the Win32 GUI.
613 * Revision 1.4 2001/05/22 18:46:04 oes
615 * - Enabled filtering banners by size rather than URL
616 * by adding patterns that replace all standard banner
617 * sizes with the "Junkbuster" gif to the re_filterfile
619 * - Enabled filtering WebBugs by providing a pattern
620 * which kills all 1x1 images
622 * - Added support for PCRE_UNGREEDY behaviour to pcrs,
623 * which is selected by the (nonstandard and therefore
624 * capital) letter 'U' in the option string.
625 * It causes the quantifiers to be ungreedy by default.
626 * Appending a ? turns back to greedy (!).
628 * - Added a new interceptor ijb-send-banner, which
629 * sends back the "Junkbuster" gif. Without imagelist or
630 * MSIE detection support, or if tinygif = 1, or the
631 * URL isn't recognized as an imageurl, a lame HTML
632 * explanation is sent instead.
634 * - Added new feature, which permits blocking remote
635 * script redirects and firing back a local redirect
637 * The feature is conditionally compiled, i.e. it
638 * can be disabled with --disable-fast-redirects,
639 * plus it must be activated by a "fast-redirects"
640 * line in the config file, has its own log level
641 * and of course wants to be displayed by show-proxy-args
642 * Note: Boy, all the #ifdefs in 1001 locations and
643 * all the fumbling with configure.in and acconfig.h
644 * were *way* more work than the feature itself :-(
646 * - Because a generic redirect template was needed for
647 * this, tinygif = 3 now uses the same.
649 * - Moved GIFs, and other static HTTP response templates
654 * - Removed some >400 CRs again (Jon, you really worked
657 * Revision 1.3 2001/05/20 01:21:20 jongfoster
658 * Version 2.9.4 checkin.
659 * - Merged popupfile and cookiefile, and added control over PCRS
660 * filtering, in new "actionsfile".
661 * - Implemented LOG_LEVEL_FATAL, so that if there is a configuration
662 * file error you now get a message box (in the Win32 GUI) rather
663 * than the program exiting with no explanation.
664 * - Made killpopup use the PCRS MIME-type checking and HTTP-header
666 * - Removed tabs from "config"
667 * - Moved duplicated url parsing code in "loaders.c" to a new funcition.
668 * - Bumped up version number.
670 * Revision 1.2 2001/05/17 23:01:01 oes
671 * - Cleaned CRLF's from the sources and related files
673 * Revision 1.1.1.1 2001/05/15 13:59:03 oes
674 * Initial import of version 2.9.3 source tree
677 *********************************************************************/
680 /* Declare struct FILE for vars and funcs. */
683 /* Need time_t for file_list */
685 /* Needed for pcre choice */
689 * Include appropriate regular expression libraries.
690 * Note that pcrs and pcre (native) are needed for cgi
691 * and are included anyway.
697 # ifdef PCRE_H_IN_SUBDIR
698 # include <pcre/pcre.h>
711 # include "pcreposix.h"
713 # ifdef PCRE_H_IN_SUBDIR
714 # include <pcre/pcreposix.h>
716 # include <pcreposix.h>
722 #endif /* def AMIGA */
726 * I don't want to have to #include all this just for the declaration
727 * of SOCKET. However, it looks like we have to...
742 typedef SOCKET jb_socket;
744 #define JB_INVALID_SOCKET INVALID_SOCKET
746 #else /* ndef _WIN32 */
749 * The type used by sockets. On UNIX it's an int. Microsoft decided to
750 * make it an unsigned.
752 typedef int jb_socket;
755 * The error value used for variables of type jb_socket. On UNIX this
756 * is -1, however Microsoft decided to make socket handles unsigned, so
757 * they use a different value.
760 #define JB_INVALID_SOCKET (-1)
762 #endif /* ndef _WIN32 */
766 * A standard error code. This should be JB_ERR_OK or one of the JB_ERR_xxx
771 #define JB_ERR_OK 0 /**< Success, no error */
772 #define JB_ERR_MEMORY 1 /**< Out of memory */
773 #define JB_ERR_CGI_PARAMS 2 /**< Missing or corrupt CGI parameters */
774 #define JB_ERR_FILE 3 /**< Error opening, reading or writing a file */
775 #define JB_ERR_PARSE 4 /**< Error parsing file */
776 #define JB_ERR_MODIFIED 5 /**< File has been modified outside of the
777 CGI actions editor. */
778 #define JB_ERR_COMPRESS 6 /**< Error on decompression */
781 * This macro is used to free a pointer that may be NULL.
782 * It also sets the variable to NULL after it's been freed.
783 * The paramater should be a simple variable without side effects.
785 #define freez(X) { if(X) { free((void*)X); X = NULL ; } }
789 * Fix a problem with Solaris. There should be no effect on other
792 * Solaris's isspace() is a macro which uses it's argument directly
793 * as an array index. Therefore we need to make sure that high-bit
794 * characters generate +ve values, and ideally we also want to make
795 * the argument match the declared parameter type of "int".
797 * Note: Remember to #include <ctype.h> if you use these macros.
799 #define ijb_toupper(__X) toupper((int)(unsigned char)(__X))
800 #define ijb_tolower(__X) tolower((int)(unsigned char)(__X))
801 #define ijb_isspace(__X) isspace((int)(unsigned char)(__X))
804 * Use for statically allocated buffers if you have no other choice.
805 * Remember to check the length of what you write into the buffer
806 * - we don't want any buffer overflows!
808 #define BUFFER_SIZE 5000
811 * Max length of CGI parameters (arbitrary limit).
813 #define CGI_PARAM_LEN_MAX 500
816 * Buffer size for capturing struct hostent data in the
817 * gethostby(name|addr)_r library calls. Since we don't
818 * loop over gethostbyname_r, the buffer must be sufficient
819 * to accomodate multiple IN A RRs, as used in DNS round robin
820 * load balancing. W3C's wwwlib uses 1K, so that should be
821 * good enough for us, too.
824 * XXX: Temporary doubled, for some configurations
825 * 1K is still too small and we didn't get the
826 * real fix ready for inclusion.
828 #define HOSTENT_BUFFER_SIZE 2048
831 * Do not use. Originally this was so that you can
832 * say "while (FOREVER) { ...do something... }".
833 * However, this gives a warning with some compilers (e.g. VC++).
834 * Instead, use "for (;;) { ...do something... }".
839 * Default IP address to listen on, as a string.
840 * Set to "127.0.0.1".
842 #define HADDR_DEFAULT "127.0.0.1"
845 * Default port to listen on, as a number.
848 #define HADDR_PORT 8118
851 /* Forward def for struct client_state */
852 struct configuration_spec;
856 * Entry in a linked list of strings.
861 * The string pointer. It must point to a dynamically malloc()ed
862 * string or be NULL for the list functions to work. In the latter
863 * case, just be careful next time you iterate through the list in
868 /** Next entry in the linked list, or NULL if no more. */
869 struct list_entry *next;
873 * A header for a linked list of strings.
877 /** First entry in the list, or NULL if the list is empty. */
878 struct list_entry *first;
880 /** Last entry in the list, or NULL if the list is empty. */
881 struct list_entry *last;
886 * An entry in a map. This is a name=value pair.
890 /** The key for the map. */
892 /** The value associated with that key. */
894 /** The next map entry, or NULL if none. */
895 struct map_entry *next;
899 * A map from a string to another string.
900 * This is used for the paramaters passed in a HTTP GET request, and
901 * to store the exports when the CGI interface is filling in a template.
905 /** The first map entry, or NULL if the map is empty. */
906 struct map_entry *first;
907 /** The last map entry, or NULL if the map is empty. */
908 struct map_entry *last;
913 * A HTTP request. This includes the method (GET, POST) and
916 * This is also used whenever we want to match a URL against a
917 * URL pattern. This always contains the URL to match, and never
918 * a URL pattern. (See struct url_spec).
922 char *cmd; /**< Whole command line: method, URL, Version */
923 char *ocmd; /**< Backup of original cmd for CLF logging */
924 char *gpc; /**< HTTP method: GET, POST, ... */
925 char *url; /**< The URL */
926 char *ver; /**< Protocol version */
927 int status; /**< HTTP Status */
929 char *host; /**< Host part of URL */
930 int port; /**< Port of URL or 80 (default) */
931 char *path; /**< Path of URL */
932 char *hostport; /**< host[:port] */
933 int ssl; /**< Flag if protocol is https */
935 char *host_ip_addr_str; /**< String with dotted decimal representation
936 of host's IP. NULL before connect_to() */
938 char *dbuffer; /**< Buffer with '\0'-delimited domain name. */
939 char **dvec; /**< List of pointers to the strings in dbuffer. */
940 int dcount; /**< How many parts to this domain? (length of dvec) */
944 * Reasons for generating a http_response instead of delivering
945 * the requested resource. Mostly ordered the way they are checked
948 #define RSP_REASON_UNSUPPORTED 1
949 #define RSP_REASON_BLOCKED 2
950 #define RSP_REASON_UNTRUSTED 3
951 #define RSP_REASON_REDIRECTED 4
952 #define RSP_REASON_CGI_CALL 5
953 #define RSP_REASON_NO_SUCH_DOMAIN 6
954 #define RSP_REASON_FORWARDING_FAILED 7
955 #define RSP_REASON_CONNECT_FAILED 8
956 #define RSP_REASON_OUT_OF_MEMORY 9
957 #define RSP_REASON_INTERNAL_ERROR 10
960 * Response generated by CGI, blocker, or error handler
964 char *status; /**< HTTP status (string). */
965 struct list headers[1]; /**< List of header lines. */
966 char *head; /**< Formatted http response head. */
967 size_t head_length; /**< Length of http response head. */
968 char *body; /**< HTTP document body. */
969 size_t content_length; /**< Length of body, REQUIRED if binary body. */
970 int is_static; /**< Nonzero if the content will never change and
971 should be cached by the browser (e.g. images). */
972 int reason; /**< Why the response was generated in the first place. */
976 * A URL or a tag pattern.
980 /** The string which was parsed to produce this url_spec.
981 Used for debugging or display only. */
984 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
985 regex_t *host_regex;/**< Regex for host matching */
987 char *dbuffer; /**< Buffer with '\0'-delimited domain name, or NULL to match all hosts. */
988 char **dvec; /**< List of pointers to the strings in dbuffer. */
989 int dcount; /**< How many parts to this domain? (length of dvec) */
990 int unanchored; /**< Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT. */
991 #endif /* defined FEATURE_EXTENDED_HOST_PATTERNS */
993 char *port_list; /**< List of acceptable ports, or NULL to match all ports */
995 regex_t *preg; /**< Regex for matching path part */
996 regex_t *tag_regex; /**< Regex for matching tags */
1000 * If you declare a static url_spec, this is the value to initialize it to zero.
1002 #ifndef FEATURE_EXTENDED_HOST_PATTERNS
1003 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, 0, 0, NULL, NULL, NULL }
1005 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, NULL, NULL }
1006 #endif /* def FEATURE_EXTENDED_HOST_PATTERNS */
1009 * Constant for host part matching in URLs. If set, indicates that the start of
1010 * the pattern must match the start of the URL. E.g. this is not set for the
1011 * pattern ".example.com", so that it will match both "example.com" and
1012 * "www.example.com". It is set for the pattern "example.com", which makes it
1013 * match "example.com" only, not "www.example.com".
1015 #define ANCHOR_LEFT 1
1018 * Constant for host part matching in URLs. If set, indicates that the end of
1019 * the pattern must match the end of the URL. E.g. this is not set for the
1020 * pattern "ad.", so that it will match any host called "ad", irrespective
1021 * of how many subdomains are in the fully-qualified domain name.
1023 #define ANCHOR_RIGHT 2
1027 * An I/O buffer. Holds a string which can be appended to, and can have data
1028 * removed from the beginning.
1032 char *buf; /**< Start of buffer */
1033 char *cur; /**< Start of relevant data */
1034 char *eod; /**< End of relevant data */
1035 size_t size; /**< Size as malloc()ed */
1040 * Return the number of bytes in the I/O buffer associated with the passed
1041 * client_state pointer.
1044 #define IOB_PEEK(CSP) ((CSP->iob->cur > CSP->iob->eod) ? (CSP->iob->eod - CSP->iob->cur) : 0)
1048 * Remove any data in the I/O buffer associated with the passed
1049 * client_state pointer.
1051 #define IOB_RESET(CSP) if(CSP->iob->buf) free(CSP->iob->buf); memset(CSP->iob, '\0', sizeof(CSP->iob));
1053 /* Bits for csp->content_type bitmask: */
1054 #define CT_TEXT 0x0001U /**< Suitable for pcrs filtering. */
1055 #define CT_GIF 0x0002U /**< Suitable for GIF filtering. */
1056 #define CT_TABOO 0x0004U /**< DO NOT filter, irrespective of other flags. */
1058 /* Although these are not, strictly speaking, content types
1059 * (they are content encodings), it is simple to handle them
1062 #define CT_GZIP 0x0010U /**< gzip-compressed data. */
1063 #define CT_DEFLATE 0x0020U /**< zlib-compressed data. */
1066 * Flag to signal that the server declared the content type,
1067 * so we can differentiate between unknown and undeclared
1070 #define CT_DECLARED 0x0040U
1073 * The mask which includes all actions.
1075 #define ACTION_MASK_ALL (~0UL)
1078 * The most compatible set of actions - i.e. none.
1080 #define ACTION_MOST_COMPATIBLE 0x00000000UL
1082 /** Action bitmap: Block the request. */
1083 #define ACTION_BLOCK 0x00000001UL
1084 /** Action bitmap: Deanimate if it's a GIF. */
1085 #define ACTION_DEANIMATE 0x00000002UL
1086 /** Action bitmap: Downgrade HTTP/1.1 to 1.0. */
1087 #define ACTION_DOWNGRADE 0x00000004UL
1088 /** Action bitmap: Fast redirects. */
1089 #define ACTION_FAST_REDIRECTS 0x00000008UL
1090 /** Action bitmap: Remove or add "X-Forwarded-For" header. */
1091 #define ACTION_CHANGE_X_FORWARDED_FOR 0x00000010UL
1092 /** Action bitmap: Hide "From" header. */
1093 #define ACTION_HIDE_FROM 0x00000020UL
1094 /** Action bitmap: Hide "Referer" header. (sic - follow HTTP, not English). */
1095 #define ACTION_HIDE_REFERER 0x00000040UL
1096 /** Action bitmap: Hide "User-Agent" and similar headers. */
1097 #define ACTION_HIDE_USER_AGENT 0x00000080UL
1098 /** Action bitmap: This is an image. */
1099 #define ACTION_IMAGE 0x00000100UL
1100 /** Action bitmap: Sets the image blocker. */
1101 #define ACTION_IMAGE_BLOCKER 0x00000200UL
1102 /** Action bitmap: Prevent compression. */
1103 #define ACTION_NO_COMPRESSION 0x00000400UL
1104 /** Action bitmap: Change cookies to session only cookies. */
1105 #define ACTION_NO_COOKIE_KEEP 0x00000800UL
1106 /** Action bitmap: Block rending cookies. */
1107 #define ACTION_NO_COOKIE_READ 0x00001000UL
1108 /** Action bitmap: Block setting cookies. */
1109 #define ACTION_NO_COOKIE_SET 0x00002000UL
1110 /** Action bitmap: Override the forward settings in the config file */
1111 #define ACTION_FORWARD_OVERRIDE 0x00004000UL
1112 /** Action bitmap: Block as empty document */
1113 #define ACTION_HANDLE_AS_EMPTY_DOCUMENT 0x00008000UL
1114 /** Action bitmap: Limit CONNECT requests to safe ports. */
1115 #define ACTION_LIMIT_CONNECT 0x00010000UL
1116 /** Action bitmap: Redirect request. */
1117 #define ACTION_REDIRECT 0x00020000UL
1118 /** Action bitmap: Crunch or modify "if-modified-since" header. */
1119 #define ACTION_HIDE_IF_MODIFIED_SINCE 0x00040000UL
1120 /** Action bitmap: Overwrite Content-Type header. */
1121 #define ACTION_CONTENT_TYPE_OVERWRITE 0x00080000UL
1122 /** Action bitmap: Crunch specified server header. */
1123 #define ACTION_CRUNCH_SERVER_HEADER 0x00100000UL
1124 /** Action bitmap: Crunch specified client header */
1125 #define ACTION_CRUNCH_CLIENT_HEADER 0x00200000UL
1126 /** Action bitmap: Enable text mode by force */
1127 #define ACTION_FORCE_TEXT_MODE 0x00400000UL
1128 /** Action bitmap: Enable text mode by force */
1129 #define ACTION_CRUNCH_IF_NONE_MATCH 0x00800000UL
1130 /** Action bitmap: Enable content-dispostion crunching */
1131 #define ACTION_HIDE_CONTENT_DISPOSITION 0x01000000UL
1132 /** Action bitmap: Replace or block Last-Modified header */
1133 #define ACTION_OVERWRITE_LAST_MODIFIED 0x02000000UL
1134 /** Action bitmap: Replace or block Accept-Language header */
1135 #define ACTION_HIDE_ACCEPT_LANGUAGE 0x04000000UL
1138 /** Action string index: How to deanimate GIFs */
1139 #define ACTION_STRING_DEANIMATE 0
1140 /** Action string index: Replacement for "From:" header */
1141 #define ACTION_STRING_FROM 1
1142 /** Action string index: How to block images */
1143 #define ACTION_STRING_IMAGE_BLOCKER 2
1144 /** Action string index: Replacement for "Referer:" header */
1145 #define ACTION_STRING_REFERER 3
1146 /** Action string index: Replacement for "User-Agent:" header */
1147 #define ACTION_STRING_USER_AGENT 4
1148 /** Action string index: Legal CONNECT ports. */
1149 #define ACTION_STRING_LIMIT_CONNECT 5
1150 /** Action string index: Server headers containing this pattern are crunched*/
1151 #define ACTION_STRING_SERVER_HEADER 6
1152 /** Action string index: Client headers containing this pattern are crunched*/
1153 #define ACTION_STRING_CLIENT_HEADER 7
1154 /** Action string index: Replacement for the "Accept-Language:" header*/
1155 #define ACTION_STRING_LANGUAGE 8
1156 /** Action string index: Replacement for the "Content-Type:" header*/
1157 #define ACTION_STRING_CONTENT_TYPE 9
1158 /** Action string index: Replacement for the "content-dispostion:" header*/
1159 #define ACTION_STRING_CONTENT_DISPOSITION 10
1160 /** Action string index: Replacement for the "If-Modified-Since:" header*/
1161 #define ACTION_STRING_IF_MODIFIED_SINCE 11
1162 /** Action string index: Replacement for the "Last-Modified:" header. */
1163 #define ACTION_STRING_LAST_MODIFIED 12
1164 /** Action string index: Redirect URL */
1165 #define ACTION_STRING_REDIRECT 13
1166 /** Action string index: Decode before redirect? */
1167 #define ACTION_STRING_FAST_REDIRECTS 14
1168 /** Action string index: Overriding forward rule. */
1169 #define ACTION_STRING_FORWARD_OVERRIDE 15
1170 /** Action string index: Reason for the block. */
1171 #define ACTION_STRING_BLOCK 16
1172 /** Action string index: what to do with the "X-Forwarded-For" header. */
1173 #define ACTION_STRING_CHANGE_X_FORWARDED_FOR 17
1174 /** Number of string actions. */
1175 #define ACTION_STRING_COUNT 18
1178 /* To make the ugly hack in sed easier to understand */
1179 #define CHECK_EVERY_HEADER_REMAINING 0
1182 /** Index into current_action_spec::multi[] for headers to add. */
1183 #define ACTION_MULTI_ADD_HEADER 0
1184 /** Index into current_action_spec::multi[] for content filters to apply. */
1185 #define ACTION_MULTI_FILTER 1
1186 /** Index into current_action_spec::multi[] for server-header filters to apply. */
1187 #define ACTION_MULTI_SERVER_HEADER_FILTER 2
1188 /** Index into current_action_spec::multi[] for client-header filters to apply. */
1189 #define ACTION_MULTI_CLIENT_HEADER_FILTER 3
1190 /** Index into current_action_spec::multi[] for client-header tags to apply. */
1191 #define ACTION_MULTI_CLIENT_HEADER_TAGGER 4
1192 /** Index into current_action_spec::multi[] for server-header tags to apply. */
1193 #define ACTION_MULTI_SERVER_HEADER_TAGGER 5
1194 /** Number of multi-string actions. */
1195 #define ACTION_MULTI_COUNT 6
1199 * This structure contains a list of actions to apply to a URL.
1200 * It only contains positive instructions - no "-" options.
1201 * It is not used to store the actions list itself, only for
1202 * url_actions() to return the current values.
1204 struct current_action_spec
1206 /** Actions to apply. A bit set to "1" means perform the action. */
1207 unsigned long flags;
1210 * Paramaters for those actions that require them.
1211 * Each entry is valid if & only if the corresponding entry in "flags" is
1214 char * string[ACTION_STRING_COUNT];
1216 /** Lists of strings for multi-string actions. */
1217 struct list multi[ACTION_MULTI_COUNT][1];
1222 * This structure contains a set of changes to actions.
1223 * It can contain both positive and negative instructions.
1224 * It is used to store an entry in the actions list.
1228 unsigned long mask; /**< Actions to keep. A bit set to "0" means remove action. */
1229 unsigned long add; /**< Actions to add. A bit set to "1" means add action. */
1232 * Paramaters for those actions that require them.
1233 * Each entry is valid if & only if the corresponding entry in "flags" is
1236 char * string[ACTION_STRING_COUNT];
1238 /** Lists of strings to remove, for multi-string actions. */
1239 struct list multi_remove[ACTION_MULTI_COUNT][1];
1241 /** If nonzero, remove *all* strings from the multi-string action. */
1242 int multi_remove_all[ACTION_MULTI_COUNT];
1244 /** Lists of strings to add, for multi-string actions. */
1245 struct list multi_add[ACTION_MULTI_COUNT][1];
1250 * This structure is used to store action files.
1252 * It contains an URL or tag pattern, and the changes to
1253 * the actions. It's a linked list and should only be
1254 * free'd through unload_actions_file() unless there's
1255 * only a single entry.
1259 struct url_spec url[1]; /**< The URL or tag pattern. */
1261 struct action_spec *action; /**< Action settings that might be shared with
1262 the list entry before or after the current
1263 one and can't be free'd willy nilly. */
1265 struct url_actions *next; /**< Next action section in file, or NULL. */
1270 * Flags for use in csp->flags
1274 * Flag for csp->flags: Set if this client is processing data.
1275 * Cleared when the thread associated with this structure dies.
1277 #define CSP_FLAG_ACTIVE 0x01U
1280 * Flag for csp->flags: Set if the server's reply is in "chunked"
1283 #define CSP_FLAG_CHUNKED 0x02U
1286 * Flag for csp->flags: Set if this request was enforced, although it would
1287 * normally have been blocked.
1289 #define CSP_FLAG_FORCED 0x04U
1292 * Flag for csp->flags: Set if any modification to the body was done.
1294 #define CSP_FLAG_MODIFIED 0x08U
1297 * Flag for csp->flags: Set if request was blocked.
1299 #define CSP_FLAG_REJECTED 0x10U
1302 * Flag for csp->flags: Set if we are toggled on (FEATURE_TOGGLE).
1304 #define CSP_FLAG_TOGGLED_ON 0x20U
1307 * Flag for csp->flags: Set if an acceptable Connection header
1310 #define CSP_FLAG_CLIENT_CONNECTION_HEADER_SET 0x00000040UL
1313 * Flag for csp->flags: Set if adding the 'Connection: close' header
1314 * for the server isn't necessary.
1316 #define CSP_FLAG_SERVER_CONNECTION_CLOSE_SET 0x00000080UL
1319 * Flag for csp->flags: Signals header parsers whether they
1320 * are parsing server or client headers.
1322 #define CSP_FLAG_CLIENT_HEADER_PARSING_DONE 0x00000100UL
1325 * Flag for csp->flags: Set if adding the Host: header
1328 #define CSP_FLAG_HOST_HEADER_IS_SET 0x00000200UL
1331 * Flag for csp->flags: Set if filtering is disabled by X-Filter: No
1332 * XXX: As we now have tags we might as well ditch this.
1334 #define CSP_FLAG_NO_FILTERING 0x00000400UL
1337 * Flag for csp->flags: Set the client IP has appended to
1338 * an already existing X-Forwarded-For header in which case
1339 * no new header has to be generated.
1341 #define CSP_FLAG_X_FORWARDED_FOR_APPENDED 0x00000800UL
1344 * Flag for csp->flags: Set if the server wants to keep
1345 * the connection alive.
1347 * XXX: Incomplete implementation, we currently only
1348 * look for "Connection: keep-alive".
1350 #define CSP_FLAG_SERVER_CONNECTION_KEEP_ALIVE 0x00001000UL
1352 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
1354 * Flag for csp->flags: Set if the server specified the
1357 #define CSP_FLAG_CONTENT_LENGTH_SET 0x00002000UL
1358 #endif /* def FEATURE_CONNECTION_KEEP_ALIVE */
1361 * Flags for use in return codes of child processes
1365 * Flag for process return code: Set if exiting porcess has been toggled
1366 * during its lifetime.
1368 #define RC_FLAG_TOGGLED 0x10
1371 * Flag for process return code: Set if exiting porcess has blocked its
1374 #define RC_FLAG_BLOCKED 0x20
1377 * Maximum number of actions/filter files. This limit is arbitrary - it's just used
1380 #define MAX_AF_FILES 10
1383 * The state of a Privoxy processing thread.
1387 /** The proxy's configuration */
1388 struct configuration_spec * config;
1390 /** The actions to perform on the current request */
1391 struct current_action_spec action[1];
1393 /** socket to talk to client (web browser) */
1396 /** socket to talk to server (web server or proxy) */
1399 /** Multi-purpose flag container, see CSP_FLAG_* above */
1402 /** Client PC's IP address, as reported by the accept() function.
1405 /** Client PC's IP address, as reported by the accept() function.
1407 unsigned long ip_addr_long;
1409 /** The URL that was requested */
1410 struct http_request http[1];
1413 * The final forwarding settings.
1414 * XXX: Currently this is only used for forward-override,
1415 * so we can free the space in sweep.
1417 struct forward_spec * fwd;
1419 /** An I/O buffer used for buffering data read from the network */
1422 /** List of all headers for this request */
1423 struct list headers[1];
1425 /** List of all tags that apply to this request */
1426 struct list tags[1];
1428 /** MIME-Type key, see CT_* above */
1429 unsigned int content_type;
1431 /** Actions files associated with this client */
1432 struct file_list *actions_list[MAX_AF_FILES];
1434 /** pcrs job files. */
1435 struct file_list *rlist[MAX_AF_FILES];
1437 /** Length after content modification. */
1438 size_t content_length;
1440 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
1441 /** Expected length of content after which we
1442 * should stop reading from the server socket.
1444 /* XXX: is this the right location? */
1445 size_t expected_content_length;
1446 #endif /* def FEATURE_CONNECTION_KEEP_ALIVE */
1448 #ifdef FEATURE_TRUST
1451 struct file_list *tlist;
1453 #endif /* def FEATURE_TRUST */
1456 * Failure reason to embedded in the CGI error page,
1457 * or NULL. Currently only used for socks errors.
1459 char *error_message;
1461 /** Next thread in linked list. Only read or modify from the main thread! */
1462 struct client_state *next;
1467 * A function to add a header
1469 typedef jb_err (*add_header_func_ptr)(struct client_state *);
1472 * A function to process a header
1474 typedef jb_err (*parser_func_ptr )(struct client_state *, char **);
1478 * List of available CGI functions.
1480 struct cgi_dispatcher
1482 /** The URL of the CGI, relative to the CGI root. */
1483 const char * const name;
1485 /** The handler function for the CGI */
1486 jb_err (* const handler)(struct client_state *csp, struct http_response *rsp, const struct map *parameters);
1488 /** The description of the CGI, to appear on the main menu, or NULL to hide it. */
1489 const char * const description;
1491 /** A flag that indicates whether unintentional calls to this CGI can cause damage */
1497 * A data file used by Privoxy. Kept in a linked list.
1502 * This is a pointer to the data structures associated with the file.
1503 * Read-only once the structure has been created.
1508 * The unloader function.
1509 * Normally NULL. When we are finished with file (i.e. when we have
1510 * loaded a new one), set to a pointer to an unloader function.
1511 * Unloader will be called by sweep() (called from main loop) when
1512 * all clients using this file are done. This prevents threading
1515 void (*unloader)(void *);
1518 * Used internally by sweep(). Do not access from elsewhere.
1523 * File last-modified time, so we can check if file has been changed.
1524 * Read-only once the structure has been created.
1526 time_t lastmodified;
1529 * The full filename.
1534 * Pointer to next entry in the linked list of all "file_list"s.
1535 * This linked list is so that sweep() can navigate it.
1536 * Since sweep() can remove items from the list, we must be careful
1537 * to only access this value from main thread (when we know sweep
1538 * won't be running).
1540 struct file_list *next;
1544 #ifdef FEATURE_TRUST
1547 * The format of a trust file when loaded into memory.
1551 struct url_spec url[1]; /**< The URL pattern */
1552 int reject; /**< FIXME: Please document this! */
1553 struct block_spec *next; /**< Next entry in linked list */
1557 * Arbitrary limit for the number of trusted referrers.
1559 #define MAX_TRUSTED_REFERRERS 512
1561 #endif /* def FEATURE_TRUST */
1564 #define SOCKS_NONE 0 /**< Don't use a SOCKS server */
1565 #define SOCKS_4 40 /**< original SOCKS 4 protocol */
1566 #define SOCKS_4A 41 /**< as modified for hosts w/o external DNS */
1567 #define SOCKS_5 50 /**< as modified for hosts w/o external DNS */
1571 * How to forward a connection to a parent proxy.
1575 /** URL pattern that this forward_spec is for. */
1576 struct url_spec url[1];
1578 /** Connection type. Must be SOCKS_NONE, SOCKS_4, SOCKS_4A or SOCKS_5. */
1581 /** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
1584 /** SOCKS server port. */
1587 /** Parent HTTP proxy hostname, or NULL for none. */
1590 /** Parent HTTP proxy port. */
1593 /** Next entry in the linked list. */
1594 struct forward_spec *next;
1599 * Initializer for a static struct forward_spec.
1601 #define FORWARD_SPEC_INITIALIZER { { URL_SPEC_INITIALIZER }, 0, NULL, 0, NULL, 0, NULL }
1603 /* Supported filter types */
1604 #define FT_CONTENT_FILTER 0
1605 #define FT_CLIENT_HEADER_FILTER 1
1606 #define FT_SERVER_HEADER_FILTER 2
1607 #define FT_CLIENT_HEADER_TAGGER 3
1608 #define FT_SERVER_HEADER_TAGGER 4
1610 #define MAX_FILTER_TYPES 5
1613 * This struct represents one filter (one block) from
1614 * the re_filterfile. If there is more than one filter
1615 * in the file, the file will be represented by a
1616 * chained list of re_filterfile specs.
1618 struct re_filterfile_spec
1620 char *name; /**< Name from FILTER: statement in re_filterfile. */
1621 char *description; /**< Description from FILTER: statement in re_filterfile. */
1622 struct list patterns[1]; /**< The patterns from the re_filterfile. */
1623 pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
1624 int type; /**< Filter type (content, client-header, server-header). */
1625 int dynamic; /**< Set to one if the pattern might contain variables
1626 and has to be recompiled for every request. */
1627 struct re_filterfile_spec *next; /**< The pointer for chaining. */
1633 #define ACL_PERMIT 1 /**< Accept connection request */
1634 #define ACL_DENY 2 /**< Reject connection request */
1637 * An IP address pattern. Used to specify networks in the ACL.
1639 struct access_control_addr
1641 unsigned long addr; /**< The IP address as an integer. */
1642 unsigned long mask; /**< The network mask as an integer. */
1643 unsigned long port; /**< The port number. */
1647 * An access control list (ACL) entry.
1649 * This is a linked list.
1651 struct access_control_list
1653 struct access_control_addr src[1]; /**< Client IP address */
1654 struct access_control_addr dst[1]; /**< Website or parent proxy IP address */
1656 short action; /**< ACL_PERMIT or ACL_DENY */
1657 struct access_control_list *next; /**< The next entry in the ACL. */
1660 #endif /* def FEATURE_ACL */
1663 /** Maximum number of loaders (actions, re_filter, ...) */
1667 /** configuration_spec::feature_flags: CGI actions editor. */
1668 #define RUNTIME_FEATURE_CGI_EDIT_ACTIONS 1U
1670 /** configuration_spec::feature_flags: Web-based toggle. */
1671 #define RUNTIME_FEATURE_CGI_TOGGLE 2U
1673 /** configuration_spec::feature_flags: HTTP-header-based toggle. */
1674 #define RUNTIME_FEATURE_HTTP_TOGGLE 4U
1676 /** configuration_spec::feature_flags: Split large forms to limit the number of GET arguments. */
1677 #define RUNTIME_FEATURE_SPLIT_LARGE_FORMS 8U
1679 /** configuration_spec::feature_flags: Check the host header for requests with host-less request lines. */
1680 #define RUNTIME_FEATURE_ACCEPT_INTERCEPTED_REQUESTS 16U
1682 /** configuration_spec::feature_flags: Don't allow to circumvent blocks with the force prefix. */
1683 #define RUNTIME_FEATURE_ENFORCE_BLOCKS 32U
1685 /** configuration_spec::feature_flags: Allow to block or redirect CGI requests. */
1686 #define RUNTIME_FEATURE_CGI_CRUNCHING 64U
1688 /** configuration_spec::feature_flags: Try to keep the connection to the server alive. */
1689 #define RUNTIME_FEATURE_CONNECTION_KEEP_ALIVE 128U
1692 * Data loaded from the configuration file.
1694 * (Anomaly: toggle is still handled through a global, not this structure)
1696 struct configuration_spec
1701 /** Nonzero to enable multithreading. */
1705 * Bitmask of features that can be enabled/disabled through the config
1706 * file. Currently defined bits:
1708 * - RUNTIME_FEATURE_CGI_EDIT_ACTIONS
1709 * - RUNTIME_FEATURE_CGI_TOGGLE
1710 * - RUNTIME_FEATURE_HTTP_TOGGLE
1711 * - RUNTIME_FEATURE_SPLIT_LARGE_FORMS
1713 unsigned feature_flags;
1715 /** The log file name. */
1716 const char *logfile;
1718 /** The config file directory. */
1719 const char *confdir;
1721 /** The directory for customized CGI templates. */
1722 const char *templdir;
1724 /** The log file directory. */
1727 /** The full paths to the actions files. */
1728 const char *actions_file[MAX_AF_FILES];
1730 /** The short names of the actions files. */
1731 const char *actions_file_short[MAX_AF_FILES];
1733 /** The administrator's email address */
1734 char *admin_address;
1736 /** A URL with info on this proxy */
1737 char *proxy_info_url;
1739 /** URL to the user manual (on our website or local copy) */
1742 /** The file names of the pcre filter files. */
1743 const char *re_filterfile[MAX_AF_FILES];
1745 /** The short names of the pcre filter files. */
1746 const char *re_filterfile_short[MAX_AF_FILES];
1748 /** The hostname to show on CGI pages, or NULL to use the real one. */
1749 const char *hostname;
1751 /** IP address to bind to. Defaults to HADDR_DEFAULT == 127.0.0.1. */
1754 /** Port to bind to. Defaults to HADDR_PORT == 8118. */
1757 /** Size limit for IOB */
1758 size_t buffer_limit;
1760 #ifdef FEATURE_TRUST
1762 /** The file name of the trust file. */
1763 const char * trustfile;
1765 /** FIXME: DOCME: Document this. */
1766 struct list trust_info[1];
1768 /** FIXME: DOCME: Document this. */
1769 struct url_spec *trust_list[MAX_TRUSTED_REFERRERS];
1771 #endif /* def FEATURE_TRUST */
1775 /** The access control list (ACL). */
1776 struct access_control_list *acl;
1778 #endif /* def FEATURE_ACL */
1780 /** Information about parent proxies (forwarding). */
1781 struct forward_spec *forward;
1783 /** Number of retries in case a forwarded connection attempt fails */
1784 int forwarded_connect_retries;
1786 /** All options from the config file, HTML-formatted. */
1789 /** The configuration file object. */
1790 struct file_list *config_file_list;
1792 /** List of loaders */
1793 int (*loaders[NLOADERS])(struct client_state *);
1795 /** Nonzero if we need to bind() to the new port. */
1799 /** Calculates the number of elements in an array, using sizeof. */
1800 #define SZ(X) (sizeof(X) / sizeof(*X))
1802 #ifdef FEATURE_FORCE_LOAD
1803 /** The force load URL prefix. */
1804 #define FORCE_PREFIX "/PRIVOXY-FORCE"
1805 #endif /* def FEATURE_FORCE_LOAD */
1807 #ifdef FEATURE_NO_GIFS
1808 /** The MIME type for images ("image/png" or "image/gif"). */
1809 #define BUILTIN_IMAGE_MIMETYPE "image/png"
1811 #define BUILTIN_IMAGE_MIMETYPE "image/gif"
1812 #endif /* def FEATURE_NO_GIFS */
1819 /** URL for the Privoxy home page. */
1820 #define HOME_PAGE_URL "http://www.privoxy.org/"
1822 /** URL for the Privoxy user manual. */
1823 #define USER_MANUAL_URL HOME_PAGE_URL VERSION "/user-manual/"
1825 /** Prefix for actions help links (append to USER_MANUAL_URL). */
1826 #define ACTIONS_HELP_PREFIX "actions-file.html#"
1828 /** Prefix for config option help links (append to USER_MANUAL_URL). */
1829 #define CONFIG_HELP_PREFIX "config.html#"
1832 * The "hosts" to intercept and display CGI pages.
1833 * First one is a hostname only, second one can specify host and path.
1836 * 1) Do not specify the http: prefix
1837 * 2) CGI_SITE_2_PATH must not end with /, one will be added automatically.
1838 * 3) CGI_SITE_2_PATH must start with /, unless it is the empty string.
1840 #define CGI_SITE_1_HOST "p.p"
1841 #define CGI_SITE_2_HOST "config.privoxy.org"
1842 #define CGI_SITE_2_PATH ""
1845 * The prefix for CGI pages. Written out in generated HTML.
1846 * INCLUDES the trailing slash.
1848 #define CGI_PREFIX "http://" CGI_SITE_2_HOST CGI_SITE_2_PATH "/"
1854 #endif /* ndef PROJECT_H_INCLUDED */