1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
4 #define PROJECT_H_VERSION "$Id: project.h,v 1.128 2009/03/07 13:09:17 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) 2001-2009 the
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 1.128 2009/03/07 13:09:17 fabiankeil
41 * Change csp->expected_content and_csp->expected_content_length from
42 * size_t to unsigned long long to reduce the likelihood of integer
43 * overflows that would let us close the connection prematurely.
44 * Bug found while investigating #2669131, reported by cyberpatrol.
46 * Revision 1.127 2008/12/20 14:53:55 fabiankeil
47 * Add config option socket-timeout to control the time
48 * Privoxy waits for data to arrive on a socket. Useful
49 * in case of stale ssh tunnels or when fuzz-testing.
51 * Revision 1.126 2008/12/14 17:02:54 fabiankeil
52 * Fix a cparser warning.
54 * Revision 1.125 2008/11/20 08:22:28 fabiankeil
55 * Remove an obsolete comment.
57 * Revision 1.124 2008/11/16 12:43:49 fabiankeil
58 * Turn keep-alive support into a runtime feature
59 * that is disabled by setting keep-alive-timeout
60 * to a negative value.
62 * Revision 1.123 2008/11/10 16:55:59 fabiankeil
63 * Fix a gcc44 warning (in filters.c).
65 * Revision 1.122 2008/10/16 07:11:34 fabiankeil
66 * Fix a bunch of gcc44 conversion warnings.
68 * Revision 1.121 2008/10/09 18:21:41 fabiankeil
69 * Flush work-in-progress changes to keep outgoing connections
70 * alive where possible. Incomplete and mostly #ifdef'd out.
72 * Revision 1.120 2008/09/21 13:36:52 fabiankeil
73 * If change-x-forwarded-for{add} is used and the client
74 * sends multiple X-Forwarded-For headers, append the client's
75 * IP address to each one of them. "Traditionally" we would
76 * lose all but the last one.
78 * Revision 1.119 2008/09/20 10:04:33 fabiankeil
79 * Remove hide-forwarded-for-headers action which has
80 * been obsoleted by change-x-forwarded-for{block}.
82 * Revision 1.118 2008/09/19 15:26:29 fabiankeil
83 * Add change-x-forwarded-for{} action to block or add
84 * X-Forwarded-For headers. Mostly based on code removed
87 * Revision 1.117 2008/08/30 12:03:07 fabiankeil
88 * Remove FEATURE_COOKIE_JAR.
90 * Revision 1.116 2008/05/20 16:05:02 fabiankeil
91 * Move parsers structure definition from project.h to parsers.h.
93 * Revision 1.115 2008/05/19 16:57:20 fabiankeil
94 * Declare all members of the parsers structure immutable.
96 * Revision 1.114 2008/04/11 16:35:39 fabiankeil
97 * Oops, I forgot to shorten the URL_SPEC_INITIALIZER in my last commit.
99 * Revision 1.113 2008/04/10 14:41:04 fabiankeil
100 * Ditch url_spec's path member now that it's no longer used.
102 * Revision 1.112 2008/04/06 15:18:34 fabiankeil
103 * Oh well, rename the --enable-pcre-host-patterns option to
104 * --enable-extended-host-patterns as it's not really PCRE syntax.
106 * Revision 1.111 2008/04/06 14:54:26 fabiankeil
107 * Use PCRE syntax in host patterns when configured
108 * with --enable-pcre-host-patterns.
110 * Revision 1.110 2008/03/29 12:13:46 fabiankeil
111 * Remove send-wafer and send-vanilla-wafer actions.
113 * Revision 1.109 2008/03/28 15:13:41 fabiankeil
114 * Remove inspect-jpegs action.
116 * Revision 1.108 2008/03/27 18:27:36 fabiankeil
117 * Remove kill-popups action.
119 * Revision 1.107 2008/03/26 18:07:08 fabiankeil
120 * Add hostname directive. Closes PR#1918189.
122 * Revision 1.106 2008/03/24 11:21:03 fabiankeil
123 * Share the action settings for multiple patterns in the same
124 * section so we waste less memory for gigantic block lists
125 * (and load them slightly faster). Reported by Franz Schwartau.
127 * Revision 1.105 2008/03/21 11:16:27 fabiankeil
128 * Garbage-collect csp->my_ip_addr_str and csp->my_hostname.
130 * Revision 1.104 2008/03/04 18:30:40 fabiankeil
131 * Remove the treat-forbidden-connects-like-blocks action. We now
132 * use the "blocked" page for forbidden CONNECT requests by default.
134 * Revision 1.103 2008/03/01 14:00:45 fabiankeil
135 * Let the block action take the reason for the block
136 * as argument and show it on the "blocked" page.
138 * Revision 1.102 2008/02/03 13:46:14 fabiankeil
139 * Add SOCKS5 support. Patch #1862863 by Eric M. Hopper with minor changes.
141 * Revision 1.101 2007/12/07 18:29:23 fabiankeil
142 * Remove now-obsolete csp member x_forwarded.
144 * Revision 1.100 2007/09/02 13:42:11 fabiankeil
145 * - Allow port lists in url patterns.
146 * - Ditch unused url_spec member pathlen.
148 * Revision 1.99 2007/07/21 11:51:36 fabiankeil
149 * As Hal noticed, checking dispatch_cgi() as the last cruncher
150 * looks like a bug if CGI requests are blocked unintentionally,
151 * so don't do it unless the user enabled the new config option
152 * "allow-cgi-request-crunching".
154 * Revision 1.98 2007/07/14 07:31:26 fabiankeil
155 * Add new csp->content_type flag (CT_DECLARED).
157 * Revision 1.97 2007/05/27 12:38:08 fabiankeil
158 * - Remove some left-overs from the switch to dedicated header filters.
159 * - Adjust "X-Filter: No" to disable dedicated header filters.
160 * - Prepare for forward-override{}
162 * Revision 1.96 2007/05/14 10:41:15 fabiankeil
163 * Ditch the csp member cookie_list[] which isn't used anymore.
165 * Revision 1.95 2007/04/30 15:02:19 fabiankeil
166 * Introduce dynamic pcrs jobs that can resolve variables.
168 * Revision 1.94 2007/04/15 16:39:21 fabiankeil
169 * Introduce tags as alternative way to specify which
170 * actions apply to a request. At the moment tags can be
171 * created based on client and server headers.
173 * Revision 1.93 2007/03/20 15:16:34 fabiankeil
174 * Use dedicated header filter actions instead of abusing "filter".
175 * Replace "filter-client-headers" and "filter-client-headers"
176 * with "server-header-filter" and "client-header-filter".
178 * Revision 1.92 2007/03/17 15:20:05 fabiankeil
179 * New config option: enforce-blocks.
181 * Revision 1.91 2007/03/05 13:28:03 fabiankeil
182 * Add some CSP_FLAGs for the header parsers.
184 * Revision 1.90 2007/02/07 10:36:16 fabiankeil
185 * Add new http_response member to save
186 * the reason why the response was generated.
188 * Revision 1.89 2007/01/27 13:09:16 fabiankeil
189 * Add new config option "templdir" to
190 * change the templates directory.
192 * Revision 1.88 2007/01/25 13:36:59 fabiankeil
193 * Add csp->error_message for failure reasons
194 * that should be embedded into the CGI pages.
196 * Revision 1.87 2007/01/01 19:36:37 fabiankeil
197 * Integrate a modified version of Wil Mahan's
198 * zlib patch (PR #895531).
200 * Revision 1.86 2006/12/31 17:56:37 fabiankeil
201 * Added config option accept-intercepted-requests
202 * and disabled it by default.
204 * Revision 1.85 2006/12/31 15:03:31 fabiankeil
205 * Fix gcc43 compiler warnings and a comment.
207 * Revision 1.84 2006/12/21 12:57:48 fabiankeil
208 * Add config option "split-large-forms"
209 * to work around the browser bug reported
212 * Revision 1.83 2006/12/06 19:26:29 fabiankeil
213 * Moved HTTP snipplets into jcc.c. They aren't
214 * used anywhere else.
216 * Revision 1.82 2006/09/20 15:50:31 fabiankeil
217 * Doubled size of HOSTENT_BUFFER_SIZE to mask
218 * problems with gethostbyname_r and some
219 * /etc/hosts configurations. Only a workaround
220 * until we get the real fix ready.
221 * Thanks Félix Rauch for reporting.
223 * Increased value of MAX_TRUSTED_REFERRERS from 64 to 512.
225 * Revision 1.81 2006/09/06 13:03:04 fabiankeil
226 * Respond with 400 and a short text message
227 * if the client tries to use Privoxy as FTP proxy.
229 * Revision 1.80 2006/09/06 10:43:32 fabiankeil
230 * Added config option enable-remote-http-toggle
231 * to specify if Privoxy should recognize special
232 * headers (currently only X-Filter) to change its
233 * behaviour. Disabled by default.
235 * Revision 1.79 2006/09/06 09:23:37 fabiankeil
236 * Make number of retries in case of forwarded-connect problems
237 * a config file option (forwarded-connect-retries) and use 0 as
240 * Revision 1.78 2006/08/31 16:25:06 fabiankeil
241 * Work around a buffer overflow that caused Privoxy to
242 * segfault if too many trusted referrers were used. Good
243 * enough for now, but should be replaced with a real
244 * solution after the next release.
246 * Revision 1.77 2006/08/21 12:50:51 david__schmidt
249 * Revision 1.76 2006/08/14 08:25:19 fabiankeil
250 * Split filter-headers{} into filter-client-headers{}
251 * and filter-server-headers{}.
252 * Added parse_header_time() to share some code.
253 * Replaced timegm() with mktime().
255 * Revision 1.75 2006/08/03 02:46:41 david__schmidt
256 * Incorporate Fabian Keil's patch work:
\r * http://www.fabiankeil.de/sourcecode/privoxy/
258 * Revision 1.74 2006/07/18 14:48:47 david__schmidt
259 * Reorganizing the repository: swapping out what was HEAD (the old 3.1 branch)
260 * with what was really the latest development (the v_3_0_branch branch)
262 * Revision 1.72.2.7 2006/01/29 23:10:56 david__schmidt
263 * Multiple filter file support
265 * Revision 1.72.2.6 2004/10/03 12:53:46 david__schmidt
266 * Add the ability to check jpeg images for invalid
267 * lengths of comment blocks. Defensive strategy
268 * against the exploit:
269 * Microsoft Security Bulletin MS04-028
270 * Buffer Overrun in JPEG Processing (GDI+) Could
271 * Allow Code Execution (833987)
272 * Enabled with +inspect-jpegs in actions files.
274 * Revision 1.72.2.5 2004/01/30 15:29:29 oes
275 * Updated the copyright note
277 * Revision 1.72.2.4 2004/01/13 16:12:14 oes
278 * Fixed double slash in USER_MANUAL_URL. Closes BR #867088.
280 * Revision 1.72.2.3 2003/03/11 11:54:37 oes
281 * Introduced RC_FLAG_* flags for use in child process return code
283 * Revision 1.72.2.2 2002/11/28 18:15:44 oes
284 * Added flag to each cgi_dispatcher that allows or denies
285 * external linking and removed const qualifier from
286 * struct list_entry.str.
288 * Revision 1.72.2.1 2002/08/10 11:25:18 oes
289 * - Include config.h for access to config data
290 * - Include <pcre*.h> depending on where they are
292 * Revision 1.72 2002/05/14 21:35:49 oes
293 * Split HELP_LINK_PREFIX into ACTIONS_HELP_PREFIX and CONFIG_HELP_PREFIX
294 * because of split in user-manual
296 * Revision 1.71 2002/05/12 21:39:36 jongfoster
297 * - Adding Doxygen-style comments to structures and #defines.
299 * Revision 1.70 2002/05/12 16:05:50 jongfoster
300 * Fixing ACTION_MASK_ALL to be unsigned long rather than
301 * just unsigned int. I don't know if anyone is porting
302 * Privoxy to 16-bit platforms, but if so, +limit-connect
303 * wouldn't have worked because of this bug.
305 * Revision 1.69 2002/05/08 16:00:16 oes
306 * Added size member to struct iob, so it can
307 * be alloced larger than needed.
309 * Revision 1.68 2002/04/26 12:56:00 oes
310 * Killed REDIRECT_URL, added USER_MANUAL_URL and HELP_LINK_PREFIX
312 * Revision 1.67 2002/04/24 02:12:43 oes
313 * - Jon's multiple AF patch:
314 * - Make csp->actions_list an array
315 * - #define MAX_AF_FILES
316 * - Moved CGI_PARAM_LEN_MAX (500) here
318 * Revision 1.66 2002/04/15 19:06:43 jongfoster
321 * Revision 1.65 2002/04/04 00:36:36 gliptak
322 * always use pcre for matching
324 * Revision 1.64 2002/04/03 22:28:03 gliptak
325 * Removed references to gnu_regex
327 * Revision 1.63 2002/03/31 17:19:00 jongfoster
328 * Win32 only: Enabling STRICT to fix a VC++ compile warning.
330 * Revision 1.62 2002/03/26 22:48:49 swa
333 * Revision 1.61 2002/03/26 22:29:55 swa
334 * we have a new homepage!
336 * Revision 1.60 2002/03/24 15:52:17 jongfoster
337 * Changing CGI URL prefixes for new name
339 * Revision 1.59 2002/03/24 15:23:33 jongfoster
342 * Revision 1.58 2002/03/24 13:25:43 swa
343 * name change related issues
345 * Revision 1.57 2002/03/16 20:28:34 oes
346 * Added descriptions to the filters so users will know what they select in the cgi editor
348 * Revision 1.56 2002/03/13 20:27:30 oes
349 * Fixing bug with CT_TABOO
351 * Revision 1.55 2002/03/12 01:42:50 oes
352 * Introduced modular filters
354 * Revision 1.54 2002/03/09 20:03:52 jongfoster
355 * - Making various functions return int rather than size_t.
356 * (Undoing a recent change). Since size_t is unsigned on
357 * Windows, functions like read_socket that return -1 on
358 * error cannot return a size_t.
360 * THIS WAS A MAJOR BUG - it caused frequent, unpredictable
361 * crashes, and also frequently caused JB to jump to 100%
362 * CPU and stay there. (Because it thought it had just
363 * read ((unsigned)-1) == 4Gb of data...)
365 * - The signature of write_socket has changed, it now simply
366 * returns success=0/failure=nonzero.
368 * - Trying to get rid of a few warnings --with-debug on
369 * Windows, I've introduced a new type "jb_socket". This is
370 * used for the socket file descriptors. On Windows, this
371 * is SOCKET (a typedef for unsigned). Everywhere else, it's
372 * an int. The error value can't be -1 any more, so it's
373 * now JB_INVALID_SOCKET (which is -1 on UNIX, and in
374 * Windows it maps to the #define INVALID_SOCKET.)
376 * - The signature of bind_port has changed.
378 * Revision 1.53 2002/03/08 16:48:55 oes
379 * Added FEATURE_NO_GIFS and BUILTIN_IMAGE_MIMETYPE
381 * Revision 1.52 2002/03/07 03:46:17 oes
382 * Fixed compiler warnings
384 * Revision 1.51 2002/03/05 04:52:42 oes
385 * Deleted non-errlog debugging code
387 * Revision 1.50 2002/03/04 19:32:07 oes
388 * Changed default port to 8118
390 * Revision 1.49 2002/03/04 18:28:55 oes
391 * Deleted PID_FILE_NAME
393 * Revision 1.48 2002/03/03 14:50:40 oes
394 * Fixed CLF logging: Added ocmd member for client's request to struct http_request
396 * Revision 1.47 2002/02/20 23:15:13 jongfoster
397 * Parsing functions now handle out-of-memory gracefully by returning
400 * Revision 1.46 2002/01/17 21:06:09 jongfoster
401 * Now #defining the URLs of the config interface
403 * Minor changes to struct http_request and struct url_spec due to
404 * standardizing that struct http_request is used to represent a URL, and
405 * struct url_spec is used to represent a URL pattern. (Before, URLs were
406 * represented as seperate variables and a partially-filled-in url_spec).
408 * Revision 1.45 2002/01/09 14:33:27 oes
409 * Added HOSTENT_BUFFER_SIZE
411 * Revision 1.44 2001/12/30 14:07:32 steudten
412 * - Add signal handling (unix)
413 * - Add SIGHUP handler (unix)
414 * - Add creation of pidfile (unix)
415 * - Add action 'top' in rc file (RH)
416 * - Add entry 'SIGNALS' to manpage
417 * - Add exit message to logfile (unix)
419 * Revision 1.43 2001/11/22 21:57:51 jongfoster
420 * Making action_spec->flags into an unsigned long rather than just an
422 * Adding ACTION_NO_COOKIE_KEEP
424 * Revision 1.42 2001/11/05 21:42:41 steudten
425 * Include DBG() macro.
427 * Revision 1.41 2001/10/28 19:12:06 jongfoster
428 * Adding ijb_toupper()
430 * Revision 1.40 2001/10/26 17:40:47 oes
431 * Moved ijb_isspace and ijb_tolower to project.h
432 * Removed http->user_agent, csp->referrer and csp->accept_types
434 * Revision 1.39 2001/10/25 03:45:02 david__schmidt
435 * Adding a (void*) cast to freez() because Visual Age C++ won't expand the
436 * macro when called with a cast; so moving the cast to the macro def'n
437 * seems to both eliminate compiler warnings (on darwin and OS/2, anyway) and
438 * doesn't make macro expansion complain. Hope this works for everyone else
441 * Revision 1.38 2001/10/23 21:19:04 jongfoster
442 * New error-handling support: jb_err type and JB_ERR_xxx constants
443 * CGI functions now return a jb_err, and their parameters map is const.
444 * Support for RUNTIME_FEATUREs to enable/disable config editor
445 * Adding a few comments
447 * Revision 1.37 2001/10/14 22:14:01 jongfoster
448 * Removing name_length field from struct cgi_dispatcher, as this is
449 * now calculated at runtime from the "name" field.
451 * Revision 1.36 2001/10/10 16:45:15 oes
452 * Added LIMIT_CONNECT action and string
453 * Fixed HTTP message line termination
454 * Added CFORBIDDEN HTTP message
456 * Revision 1.35 2001/10/07 18:06:43 oes
457 * Added status member to struct http_request
459 * Revision 1.34 2001/10/07 15:45:25 oes
460 * Added url member to struct http_request and commented all
465 * Added ACTION_DOWNGRADE and ACTION_NO_COMPRESSION
467 * Replaced struct client_state members rejected,
468 * force, active and toggled_on with "flags" bitmap.
470 * Added CSP_FLAG_MODIFIED and CSP_FLAG_CHUNKED
472 * Added buffer_limit to struct configuration_spec
474 * Revision 1.33 2001/09/20 13:30:08 steudten
476 * Make freez() more secure in case of: if (exp) { free(z) ; a=*z }
477 * Last case will set z to NULL in free(z) and thats bad..
479 * Revision 1.32 2001/09/16 23:02:51 jongfoster
482 * Revision 1.31 2001/09/16 13:20:29 jongfoster
483 * Rewrite of list library. Now has seperate header and list_entry
484 * structures. Also added a large sprinking of assert()s to the list
487 * Revision 1.30 2001/09/13 23:52:00 jongfoster
488 * Support for both static and dynamically generated CGI pages
490 * Revision 1.29 2001/09/13 23:29:43 jongfoster
491 * Defining FORWARD_SPEC_INITIALIZER
493 * Revision 1.28 2001/09/13 23:05:50 jongfoster
494 * Changing the string paramater to the header parsers a "const".
496 * Revision 1.27 2001/08/05 16:06:20 jongfoster
497 * Modifiying "struct map" so that there are now separate header and
498 * "map_entry" structures. This means that functions which modify a
499 * map no longer need to return a pointer to the modified map.
500 * Also, it no longer reverses the order of the entries (which may be
501 * important with some advanced template substitutions).
503 * Revision 1.26 2001/07/30 22:08:36 jongfoster
504 * Tidying up #defines:
505 * - All feature #defines are now of the form FEATURE_xxx
506 * - Permanently turned off WIN_GUI_EDIT
507 * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
509 * Revision 1.25 2001/07/29 18:43:08 jongfoster
510 * Changing #ifdef _FILENAME_H to FILENAME_H_INCLUDED, to conform to
513 * Revision 1.24 2001/07/25 17:20:27 oes
514 * Introduced http->user_agent
516 * Revision 1.23 2001/07/18 12:32:23 oes
517 * - Added ACTION_STRING_DEANIMATE
518 * - moved #define freez from jcc.h to project.h
520 * Revision 1.22 2001/07/15 17:51:41 jongfoster
521 * Renaming #define STATIC to STATIC_PCRE
523 * Revision 1.21 2001/07/13 14:03:19 oes
524 * - Reorganized regex header inclusion and #defines to
525 * comply to the scheme in configure.in
526 * - Added csp->content_type and its CT_* keys
527 * - Added ACTION_DEANIMATE
528 * - Removed all #ifdef PCRS
530 * Revision 1.20 2001/06/29 21:45:41 oes
531 * Indentation, CRLF->LF, Tab-> Space
533 * Revision 1.19 2001/06/29 13:33:36 oes
534 * - Improved comments
535 * - Introduced http_request.host_ip_addr_str
536 * - Introduced http_response.head_length
537 * - Introduced config.my_ip_addr_str, config.my_hostname,
538 * config.admin_address and config.proxy_info_url
539 * - Removed config.proxy_args_header and config.proxy_args_trailer,
540 * renamed config.proxy_args_invocation to config.proxy_args
541 * - Removed HTML snipplets and GIFs
542 * - Removed logentry from cancelled commit
544 * Revision 1.18 2001/06/09 10:57:39 jongfoster
545 * Adding definition of BUFFER_SIZE.
546 * Changing struct cgi_dispatcher to use "const" strings.
548 * Revision 1.17 2001/06/07 23:15:09 jongfoster
549 * Merging ACL and forward files into config file.
550 * Moving struct gateway members into struct forward_spec
551 * Removing config->proxy_args_gateways
552 * Cosmetic: Adding a few comments
554 * Revision 1.16 2001/06/04 18:31:58 swa
555 * files are now prefixed with either `confdir' or `logdir'.
556 * `make redhat-dist' replaces both entries confdir and logdir
559 * Revision 1.15 2001/06/04 11:28:53 swa
560 * redirect did not work due to missing /
562 * Revision 1.14 2001/06/03 11:03:48 oes
564 * added struct http_response,
565 * changed struct interceptors to struct cgi_dispatcher,
566 * moved HTML stuff to cgi.h
568 * Revision 1.13 2001/06/01 20:05:36 jongfoster
569 * Support for +image-blocker{}: added ACTION_IMAGE_BLOCKER
570 * constant, and removed csp->tinygif.
572 * Revision 1.12 2001/06/01 18:49:17 jongfoster
573 * Replaced "list_share" with "list" - the tiny memory gain was not
574 * worth the extra complexity.
576 * Revision 1.11 2001/06/01 10:32:47 oes
577 * Added constants for anchoring selection bitmap
579 * Revision 1.10 2001/05/31 21:33:53 jongfoster
580 * Changes for new actions file, replacing permissionsfile
581 * and parts of the config file. Also added support for
584 * Revision 1.9 2001/05/31 17:32:31 oes
586 * - Enhanced domain part globbing with infix and prefix asterisk
587 * matching and optional unanchored operation
589 * Revision 1.8 2001/05/29 20:09:15 joergs
590 * HTTP_REDIRECT_TEMPLATE fixed.
592 * Revision 1.7 2001/05/29 09:50:24 jongfoster
593 * Unified blocklist/imagelist/actionslist.
594 * File format is still under discussion, but the internal changes
597 * Also modified interceptor behaviour:
598 * - We now intercept all URLs beginning with one of the following
599 * prefixes (and *only* these prefixes):
601 * * http://ijbswa.sf.net/config/
602 * * http://ijbswa.sourceforge.net/config/
603 * - New interceptors "home page" - go to http://i.j.b/ to see it.
604 * - Internal changes so that intercepted and fast redirect pages
605 * are not replaced with an image.
606 * - Interceptors now have the option to send a binary page direct
607 * to the client. (i.e. ijb-send-banner uses this)
608 * - Implemented show-url-info interceptor. (Which is why I needed
609 * the above interceptors changes - a typical URL is
610 * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif".
611 * The previous mechanism would not have intercepted that, and
612 * if it had been intercepted then it then it would have replaced
615 * Revision 1.6 2001/05/27 22:17:04 oes
617 * - re_process_buffer no longer writes the modified buffer
618 * to the client, which was very ugly. It now returns the
619 * buffer, which it is then written by chat.
621 * - content_length now adjusts the Content-Length: header
622 * for modified documents rather than crunch()ing it.
623 * (Length info in csp->content_length, which is 0 for
624 * unmodified documents)
626 * - For this to work, sed() is called twice when filtering.
628 * Revision 1.5 2001/05/26 00:28:36 jongfoster
629 * Automatic reloading of config file.
630 * Removed obsolete SIGHUP support (Unix) and Reload menu option (Win32).
631 * Most of the global variables have been moved to a new
632 * struct configuration_spec, accessed through csp->config->globalname
633 * Most of the globals remaining are used by the Win32 GUI.
635 * Revision 1.4 2001/05/22 18:46:04 oes
637 * - Enabled filtering banners by size rather than URL
638 * by adding patterns that replace all standard banner
639 * sizes with the "Junkbuster" gif to the re_filterfile
641 * - Enabled filtering WebBugs by providing a pattern
642 * which kills all 1x1 images
644 * - Added support for PCRE_UNGREEDY behaviour to pcrs,
645 * which is selected by the (nonstandard and therefore
646 * capital) letter 'U' in the option string.
647 * It causes the quantifiers to be ungreedy by default.
648 * Appending a ? turns back to greedy (!).
650 * - Added a new interceptor ijb-send-banner, which
651 * sends back the "Junkbuster" gif. Without imagelist or
652 * MSIE detection support, or if tinygif = 1, or the
653 * URL isn't recognized as an imageurl, a lame HTML
654 * explanation is sent instead.
656 * - Added new feature, which permits blocking remote
657 * script redirects and firing back a local redirect
659 * The feature is conditionally compiled, i.e. it
660 * can be disabled with --disable-fast-redirects,
661 * plus it must be activated by a "fast-redirects"
662 * line in the config file, has its own log level
663 * and of course wants to be displayed by show-proxy-args
664 * Note: Boy, all the #ifdefs in 1001 locations and
665 * all the fumbling with configure.in and acconfig.h
666 * were *way* more work than the feature itself :-(
668 * - Because a generic redirect template was needed for
669 * this, tinygif = 3 now uses the same.
671 * - Moved GIFs, and other static HTTP response templates
676 * - Removed some >400 CRs again (Jon, you really worked
679 * Revision 1.3 2001/05/20 01:21:20 jongfoster
680 * Version 2.9.4 checkin.
681 * - Merged popupfile and cookiefile, and added control over PCRS
682 * filtering, in new "actionsfile".
683 * - Implemented LOG_LEVEL_FATAL, so that if there is a configuration
684 * file error you now get a message box (in the Win32 GUI) rather
685 * than the program exiting with no explanation.
686 * - Made killpopup use the PCRS MIME-type checking and HTTP-header
688 * - Removed tabs from "config"
689 * - Moved duplicated url parsing code in "loaders.c" to a new funcition.
690 * - Bumped up version number.
692 * Revision 1.2 2001/05/17 23:01:01 oes
693 * - Cleaned CRLF's from the sources and related files
695 * Revision 1.1.1.1 2001/05/15 13:59:03 oes
696 * Initial import of version 2.9.3 source tree
699 *********************************************************************/
702 /* Declare struct FILE for vars and funcs. */
705 /* Need time_t for file_list */
707 /* Needed for pcre choice */
711 * Include appropriate regular expression libraries.
712 * Note that pcrs and pcre (native) are needed for cgi
713 * and are included anyway.
719 # ifdef PCRE_H_IN_SUBDIR
720 # include <pcre/pcre.h>
733 # include "pcreposix.h"
735 # ifdef PCRE_H_IN_SUBDIR
736 # include <pcre/pcreposix.h>
738 # include <pcreposix.h>
744 #endif /* def AMIGA */
748 * I don't want to have to #include all this just for the declaration
749 * of SOCKET. However, it looks like we have to...
764 typedef SOCKET jb_socket;
766 #define JB_INVALID_SOCKET INVALID_SOCKET
768 #else /* ndef _WIN32 */
771 * The type used by sockets. On UNIX it's an int. Microsoft decided to
772 * make it an unsigned.
774 typedef int jb_socket;
777 * The error value used for variables of type jb_socket. On UNIX this
778 * is -1, however Microsoft decided to make socket handles unsigned, so
779 * they use a different value.
782 #define JB_INVALID_SOCKET (-1)
784 #endif /* ndef _WIN32 */
788 * A standard error code. This should be JB_ERR_OK or one of the JB_ERR_xxx
793 #define JB_ERR_OK 0 /**< Success, no error */
794 #define JB_ERR_MEMORY 1 /**< Out of memory */
795 #define JB_ERR_CGI_PARAMS 2 /**< Missing or corrupt CGI parameters */
796 #define JB_ERR_FILE 3 /**< Error opening, reading or writing a file */
797 #define JB_ERR_PARSE 4 /**< Error parsing file */
798 #define JB_ERR_MODIFIED 5 /**< File has been modified outside of the
799 CGI actions editor. */
800 #define JB_ERR_COMPRESS 6 /**< Error on decompression */
803 * This macro is used to free a pointer that may be NULL.
804 * It also sets the variable to NULL after it's been freed.
805 * The paramater should be a simple variable without side effects.
807 #define freez(X) { if(X) { free((void*)X); X = NULL ; } }
811 * Fix a problem with Solaris. There should be no effect on other
814 * Solaris's isspace() is a macro which uses it's argument directly
815 * as an array index. Therefore we need to make sure that high-bit
816 * characters generate +ve values, and ideally we also want to make
817 * the argument match the declared parameter type of "int".
819 * Note: Remember to #include <ctype.h> if you use these macros.
821 #define ijb_toupper(__X) toupper((int)(unsigned char)(__X))
822 #define ijb_tolower(__X) tolower((int)(unsigned char)(__X))
823 #define ijb_isspace(__X) isspace((int)(unsigned char)(__X))
826 * Use for statically allocated buffers if you have no other choice.
827 * Remember to check the length of what you write into the buffer
828 * - we don't want any buffer overflows!
830 #define BUFFER_SIZE 5000
833 * Max length of CGI parameters (arbitrary limit).
835 #define CGI_PARAM_LEN_MAX 500U
838 * Buffer size for capturing struct hostent data in the
839 * gethostby(name|addr)_r library calls. Since we don't
840 * loop over gethostbyname_r, the buffer must be sufficient
841 * to accomodate multiple IN A RRs, as used in DNS round robin
842 * load balancing. W3C's wwwlib uses 1K, so that should be
843 * good enough for us, too.
846 * XXX: Temporary doubled, for some configurations
847 * 1K is still too small and we didn't get the
848 * real fix ready for inclusion.
850 #define HOSTENT_BUFFER_SIZE 2048
853 * Do not use. Originally this was so that you can
854 * say "while (FOREVER) { ...do something... }".
855 * However, this gives a warning with some compilers (e.g. VC++).
856 * Instead, use "for (;;) { ...do something... }".
861 * Default IP address to listen on, as a string.
862 * Set to "127.0.0.1".
864 #define HADDR_DEFAULT "127.0.0.1"
867 * Default port to listen on, as a number.
870 #define HADDR_PORT 8118
873 /* Forward def for struct client_state */
874 struct configuration_spec;
878 * Entry in a linked list of strings.
883 * The string pointer. It must point to a dynamically malloc()ed
884 * string or be NULL for the list functions to work. In the latter
885 * case, just be careful next time you iterate through the list in
890 /** Next entry in the linked list, or NULL if no more. */
891 struct list_entry *next;
895 * A header for a linked list of strings.
899 /** First entry in the list, or NULL if the list is empty. */
900 struct list_entry *first;
902 /** Last entry in the list, or NULL if the list is empty. */
903 struct list_entry *last;
908 * An entry in a map. This is a name=value pair.
912 /** The key for the map. */
914 /** The value associated with that key. */
916 /** The next map entry, or NULL if none. */
917 struct map_entry *next;
921 * A map from a string to another string.
922 * This is used for the paramaters passed in a HTTP GET request, and
923 * to store the exports when the CGI interface is filling in a template.
927 /** The first map entry, or NULL if the map is empty. */
928 struct map_entry *first;
929 /** The last map entry, or NULL if the map is empty. */
930 struct map_entry *last;
935 * A HTTP request. This includes the method (GET, POST) and
938 * This is also used whenever we want to match a URL against a
939 * URL pattern. This always contains the URL to match, and never
940 * a URL pattern. (See struct url_spec).
944 char *cmd; /**< Whole command line: method, URL, Version */
945 char *ocmd; /**< Backup of original cmd for CLF logging */
946 char *gpc; /**< HTTP method: GET, POST, ... */
947 char *url; /**< The URL */
948 char *ver; /**< Protocol version */
949 int status; /**< HTTP Status */
951 char *host; /**< Host part of URL */
952 int port; /**< Port of URL or 80 (default) */
953 char *path; /**< Path of URL */
954 char *hostport; /**< host[:port] */
955 int ssl; /**< Flag if protocol is https */
957 char *host_ip_addr_str; /**< String with dotted decimal representation
958 of host's IP. NULL before connect_to() */
960 char *dbuffer; /**< Buffer with '\0'-delimited domain name. */
961 char **dvec; /**< List of pointers to the strings in dbuffer. */
962 int dcount; /**< How many parts to this domain? (length of dvec) */
966 * Reasons for generating a http_response instead of delivering
967 * the requested resource. Mostly ordered the way they are checked
970 #define RSP_REASON_UNSUPPORTED 1
971 #define RSP_REASON_BLOCKED 2
972 #define RSP_REASON_UNTRUSTED 3
973 #define RSP_REASON_REDIRECTED 4
974 #define RSP_REASON_CGI_CALL 5
975 #define RSP_REASON_NO_SUCH_DOMAIN 6
976 #define RSP_REASON_FORWARDING_FAILED 7
977 #define RSP_REASON_CONNECT_FAILED 8
978 #define RSP_REASON_OUT_OF_MEMORY 9
979 #define RSP_REASON_INTERNAL_ERROR 10
982 * Response generated by CGI, blocker, or error handler
986 char *status; /**< HTTP status (string). */
987 struct list headers[1]; /**< List of header lines. */
988 char *head; /**< Formatted http response head. */
989 size_t head_length; /**< Length of http response head. */
990 char *body; /**< HTTP document body. */
991 size_t content_length; /**< Length of body, REQUIRED if binary body. */
992 int is_static; /**< Nonzero if the content will never change and
993 should be cached by the browser (e.g. images). */
994 int reason; /**< Why the response was generated in the first place. */
998 * A URL or a tag pattern.
1002 /** The string which was parsed to produce this url_spec.
1003 Used for debugging or display only. */
1006 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
1007 regex_t *host_regex;/**< Regex for host matching */
1009 char *dbuffer; /**< Buffer with '\0'-delimited domain name, or NULL to match all hosts. */
1010 char **dvec; /**< List of pointers to the strings in dbuffer. */
1011 int dcount; /**< How many parts to this domain? (length of dvec) */
1012 int unanchored; /**< Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT. */
1013 #endif /* defined FEATURE_EXTENDED_HOST_PATTERNS */
1015 char *port_list; /**< List of acceptable ports, or NULL to match all ports */
1017 regex_t *preg; /**< Regex for matching path part */
1018 regex_t *tag_regex; /**< Regex for matching tags */
1022 * If you declare a static url_spec, this is the value to initialize it to zero.
1024 #ifndef FEATURE_EXTENDED_HOST_PATTERNS
1025 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, 0, 0, NULL, NULL, NULL }
1027 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, NULL, NULL }
1028 #endif /* def FEATURE_EXTENDED_HOST_PATTERNS */
1031 * Constant for host part matching in URLs. If set, indicates that the start of
1032 * the pattern must match the start of the URL. E.g. this is not set for the
1033 * pattern ".example.com", so that it will match both "example.com" and
1034 * "www.example.com". It is set for the pattern "example.com", which makes it
1035 * match "example.com" only, not "www.example.com".
1037 #define ANCHOR_LEFT 1
1040 * Constant for host part matching in URLs. If set, indicates that the end of
1041 * the pattern must match the end of the URL. E.g. this is not set for the
1042 * pattern "ad.", so that it will match any host called "ad", irrespective
1043 * of how many subdomains are in the fully-qualified domain name.
1045 #define ANCHOR_RIGHT 2
1049 * An I/O buffer. Holds a string which can be appended to, and can have data
1050 * removed from the beginning.
1054 char *buf; /**< Start of buffer */
1055 char *cur; /**< Start of relevant data */
1056 char *eod; /**< End of relevant data */
1057 size_t size; /**< Size as malloc()ed */
1062 * Return the number of bytes in the I/O buffer associated with the passed
1063 * client_state pointer.
1066 #define IOB_PEEK(CSP) ((CSP->iob->cur > CSP->iob->eod) ? (CSP->iob->eod - CSP->iob->cur) : 0)
1070 * Remove any data in the I/O buffer associated with the passed
1071 * client_state pointer.
1073 #define IOB_RESET(CSP) if(CSP->iob->buf) free(CSP->iob->buf); memset(CSP->iob, '\0', sizeof(CSP->iob));
1075 /* Bits for csp->content_type bitmask: */
1076 #define CT_TEXT 0x0001U /**< Suitable for pcrs filtering. */
1077 #define CT_GIF 0x0002U /**< Suitable for GIF filtering. */
1078 #define CT_TABOO 0x0004U /**< DO NOT filter, irrespective of other flags. */
1080 /* Although these are not, strictly speaking, content types
1081 * (they are content encodings), it is simple to handle them
1084 #define CT_GZIP 0x0010U /**< gzip-compressed data. */
1085 #define CT_DEFLATE 0x0020U /**< zlib-compressed data. */
1088 * Flag to signal that the server declared the content type,
1089 * so we can differentiate between unknown and undeclared
1092 #define CT_DECLARED 0x0040U
1095 * The mask which includes all actions.
1097 #define ACTION_MASK_ALL (~0UL)
1100 * The most compatible set of actions - i.e. none.
1102 #define ACTION_MOST_COMPATIBLE 0x00000000UL
1104 /** Action bitmap: Block the request. */
1105 #define ACTION_BLOCK 0x00000001UL
1106 /** Action bitmap: Deanimate if it's a GIF. */
1107 #define ACTION_DEANIMATE 0x00000002UL
1108 /** Action bitmap: Downgrade HTTP/1.1 to 1.0. */
1109 #define ACTION_DOWNGRADE 0x00000004UL
1110 /** Action bitmap: Fast redirects. */
1111 #define ACTION_FAST_REDIRECTS 0x00000008UL
1112 /** Action bitmap: Remove or add "X-Forwarded-For" header. */
1113 #define ACTION_CHANGE_X_FORWARDED_FOR 0x00000010UL
1114 /** Action bitmap: Hide "From" header. */
1115 #define ACTION_HIDE_FROM 0x00000020UL
1116 /** Action bitmap: Hide "Referer" header. (sic - follow HTTP, not English). */
1117 #define ACTION_HIDE_REFERER 0x00000040UL
1118 /** Action bitmap: Hide "User-Agent" and similar headers. */
1119 #define ACTION_HIDE_USER_AGENT 0x00000080UL
1120 /** Action bitmap: This is an image. */
1121 #define ACTION_IMAGE 0x00000100UL
1122 /** Action bitmap: Sets the image blocker. */
1123 #define ACTION_IMAGE_BLOCKER 0x00000200UL
1124 /** Action bitmap: Prevent compression. */
1125 #define ACTION_NO_COMPRESSION 0x00000400UL
1126 /** Action bitmap: Change cookies to session only cookies. */
1127 #define ACTION_NO_COOKIE_KEEP 0x00000800UL
1128 /** Action bitmap: Block rending cookies. */
1129 #define ACTION_NO_COOKIE_READ 0x00001000UL
1130 /** Action bitmap: Block setting cookies. */
1131 #define ACTION_NO_COOKIE_SET 0x00002000UL
1132 /** Action bitmap: Override the forward settings in the config file */
1133 #define ACTION_FORWARD_OVERRIDE 0x00004000UL
1134 /** Action bitmap: Block as empty document */
1135 #define ACTION_HANDLE_AS_EMPTY_DOCUMENT 0x00008000UL
1136 /** Action bitmap: Limit CONNECT requests to safe ports. */
1137 #define ACTION_LIMIT_CONNECT 0x00010000UL
1138 /** Action bitmap: Redirect request. */
1139 #define ACTION_REDIRECT 0x00020000UL
1140 /** Action bitmap: Crunch or modify "if-modified-since" header. */
1141 #define ACTION_HIDE_IF_MODIFIED_SINCE 0x00040000UL
1142 /** Action bitmap: Overwrite Content-Type header. */
1143 #define ACTION_CONTENT_TYPE_OVERWRITE 0x00080000UL
1144 /** Action bitmap: Crunch specified server header. */
1145 #define ACTION_CRUNCH_SERVER_HEADER 0x00100000UL
1146 /** Action bitmap: Crunch specified client header */
1147 #define ACTION_CRUNCH_CLIENT_HEADER 0x00200000UL
1148 /** Action bitmap: Enable text mode by force */
1149 #define ACTION_FORCE_TEXT_MODE 0x00400000UL
1150 /** Action bitmap: Enable text mode by force */
1151 #define ACTION_CRUNCH_IF_NONE_MATCH 0x00800000UL
1152 /** Action bitmap: Enable content-dispostion crunching */
1153 #define ACTION_HIDE_CONTENT_DISPOSITION 0x01000000UL
1154 /** Action bitmap: Replace or block Last-Modified header */
1155 #define ACTION_OVERWRITE_LAST_MODIFIED 0x02000000UL
1156 /** Action bitmap: Replace or block Accept-Language header */
1157 #define ACTION_HIDE_ACCEPT_LANGUAGE 0x04000000UL
1160 /** Action string index: How to deanimate GIFs */
1161 #define ACTION_STRING_DEANIMATE 0
1162 /** Action string index: Replacement for "From:" header */
1163 #define ACTION_STRING_FROM 1
1164 /** Action string index: How to block images */
1165 #define ACTION_STRING_IMAGE_BLOCKER 2
1166 /** Action string index: Replacement for "Referer:" header */
1167 #define ACTION_STRING_REFERER 3
1168 /** Action string index: Replacement for "User-Agent:" header */
1169 #define ACTION_STRING_USER_AGENT 4
1170 /** Action string index: Legal CONNECT ports. */
1171 #define ACTION_STRING_LIMIT_CONNECT 5
1172 /** Action string index: Server headers containing this pattern are crunched*/
1173 #define ACTION_STRING_SERVER_HEADER 6
1174 /** Action string index: Client headers containing this pattern are crunched*/
1175 #define ACTION_STRING_CLIENT_HEADER 7
1176 /** Action string index: Replacement for the "Accept-Language:" header*/
1177 #define ACTION_STRING_LANGUAGE 8
1178 /** Action string index: Replacement for the "Content-Type:" header*/
1179 #define ACTION_STRING_CONTENT_TYPE 9
1180 /** Action string index: Replacement for the "content-dispostion:" header*/
1181 #define ACTION_STRING_CONTENT_DISPOSITION 10
1182 /** Action string index: Replacement for the "If-Modified-Since:" header*/
1183 #define ACTION_STRING_IF_MODIFIED_SINCE 11
1184 /** Action string index: Replacement for the "Last-Modified:" header. */
1185 #define ACTION_STRING_LAST_MODIFIED 12
1186 /** Action string index: Redirect URL */
1187 #define ACTION_STRING_REDIRECT 13
1188 /** Action string index: Decode before redirect? */
1189 #define ACTION_STRING_FAST_REDIRECTS 14
1190 /** Action string index: Overriding forward rule. */
1191 #define ACTION_STRING_FORWARD_OVERRIDE 15
1192 /** Action string index: Reason for the block. */
1193 #define ACTION_STRING_BLOCK 16
1194 /** Action string index: what to do with the "X-Forwarded-For" header. */
1195 #define ACTION_STRING_CHANGE_X_FORWARDED_FOR 17
1196 /** Number of string actions. */
1197 #define ACTION_STRING_COUNT 18
1200 /* To make the ugly hack in sed easier to understand */
1201 #define CHECK_EVERY_HEADER_REMAINING 0
1204 /** Index into current_action_spec::multi[] for headers to add. */
1205 #define ACTION_MULTI_ADD_HEADER 0
1206 /** Index into current_action_spec::multi[] for content filters to apply. */
1207 #define ACTION_MULTI_FILTER 1
1208 /** Index into current_action_spec::multi[] for server-header filters to apply. */
1209 #define ACTION_MULTI_SERVER_HEADER_FILTER 2
1210 /** Index into current_action_spec::multi[] for client-header filters to apply. */
1211 #define ACTION_MULTI_CLIENT_HEADER_FILTER 3
1212 /** Index into current_action_spec::multi[] for client-header tags to apply. */
1213 #define ACTION_MULTI_CLIENT_HEADER_TAGGER 4
1214 /** Index into current_action_spec::multi[] for server-header tags to apply. */
1215 #define ACTION_MULTI_SERVER_HEADER_TAGGER 5
1216 /** Number of multi-string actions. */
1217 #define ACTION_MULTI_COUNT 6
1221 * This structure contains a list of actions to apply to a URL.
1222 * It only contains positive instructions - no "-" options.
1223 * It is not used to store the actions list itself, only for
1224 * url_actions() to return the current values.
1226 struct current_action_spec
1228 /** Actions to apply. A bit set to "1" means perform the action. */
1229 unsigned long flags;
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 for multi-string actions. */
1239 struct list multi[ACTION_MULTI_COUNT][1];
1244 * This structure contains a set of changes to actions.
1245 * It can contain both positive and negative instructions.
1246 * It is used to store an entry in the actions list.
1250 unsigned long mask; /**< Actions to keep. A bit set to "0" means remove action. */
1251 unsigned long add; /**< Actions to add. A bit set to "1" means add action. */
1254 * Paramaters for those actions that require them.
1255 * Each entry is valid if & only if the corresponding entry in "flags" is
1258 char * string[ACTION_STRING_COUNT];
1260 /** Lists of strings to remove, for multi-string actions. */
1261 struct list multi_remove[ACTION_MULTI_COUNT][1];
1263 /** If nonzero, remove *all* strings from the multi-string action. */
1264 int multi_remove_all[ACTION_MULTI_COUNT];
1266 /** Lists of strings to add, for multi-string actions. */
1267 struct list multi_add[ACTION_MULTI_COUNT][1];
1272 * This structure is used to store action files.
1274 * It contains an URL or tag pattern, and the changes to
1275 * the actions. It's a linked list and should only be
1276 * free'd through unload_actions_file() unless there's
1277 * only a single entry.
1281 struct url_spec url[1]; /**< The URL or tag pattern. */
1283 struct action_spec *action; /**< Action settings that might be shared with
1284 the list entry before or after the current
1285 one and can't be free'd willy nilly. */
1287 struct url_actions *next; /**< Next action section in file, or NULL. */
1292 * Flags for use in csp->flags
1296 * Flag for csp->flags: Set if this client is processing data.
1297 * Cleared when the thread associated with this structure dies.
1299 #define CSP_FLAG_ACTIVE 0x01U
1302 * Flag for csp->flags: Set if the server's reply is in "chunked"
1305 #define CSP_FLAG_CHUNKED 0x02U
1308 * Flag for csp->flags: Set if this request was enforced, although it would
1309 * normally have been blocked.
1311 #define CSP_FLAG_FORCED 0x04U
1314 * Flag for csp->flags: Set if any modification to the body was done.
1316 #define CSP_FLAG_MODIFIED 0x08U
1319 * Flag for csp->flags: Set if request was blocked.
1321 #define CSP_FLAG_REJECTED 0x10U
1324 * Flag for csp->flags: Set if we are toggled on (FEATURE_TOGGLE).
1326 #define CSP_FLAG_TOGGLED_ON 0x20U
1329 * Flag for csp->flags: Set if an acceptable Connection header
1332 #define CSP_FLAG_CLIENT_CONNECTION_HEADER_SET 0x00000040U
1335 * Flag for csp->flags: Set if adding the 'Connection: close' header
1336 * for the server isn't necessary.
1338 #define CSP_FLAG_SERVER_CONNECTION_CLOSE_SET 0x00000080U
1341 * Flag for csp->flags: Signals header parsers whether they
1342 * are parsing server or client headers.
1344 #define CSP_FLAG_CLIENT_HEADER_PARSING_DONE 0x00000100U
1347 * Flag for csp->flags: Set if adding the Host: header
1350 #define CSP_FLAG_HOST_HEADER_IS_SET 0x00000200U
1353 * Flag for csp->flags: Set if filtering is disabled by X-Filter: No
1354 * XXX: As we now have tags we might as well ditch this.
1356 #define CSP_FLAG_NO_FILTERING 0x00000400U
1359 * Flag for csp->flags: Set the client IP has appended to
1360 * an already existing X-Forwarded-For header in which case
1361 * no new header has to be generated.
1363 #define CSP_FLAG_X_FORWARDED_FOR_APPENDED 0x00000800U
1366 * Flag for csp->flags: Set if the server wants to keep
1367 * the connection alive.
1369 #define CSP_FLAG_SERVER_CONNECTION_KEEP_ALIVE 0x00001000U
1371 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
1373 * Flag for csp->flags: Set if the server specified the
1376 #define CSP_FLAG_CONTENT_LENGTH_SET 0x00002000U
1377 #endif /* def FEATURE_CONNECTION_KEEP_ALIVE */
1380 * Flags for use in return codes of child processes
1384 * Flag for process return code: Set if exiting porcess has been toggled
1385 * during its lifetime.
1387 #define RC_FLAG_TOGGLED 0x10
1390 * Flag for process return code: Set if exiting porcess has blocked its
1393 #define RC_FLAG_BLOCKED 0x20
1396 * Maximum number of actions/filter files. This limit is arbitrary - it's just used
1399 #define MAX_AF_FILES 10
1402 * The state of a Privoxy processing thread.
1406 /** The proxy's configuration */
1407 struct configuration_spec * config;
1409 /** The actions to perform on the current request */
1410 struct current_action_spec action[1];
1412 /** socket to talk to client (web browser) */
1415 /** socket to talk to server (web server or proxy) */
1418 /** Multi-purpose flag container, see CSP_FLAG_* above */
1421 /** Client PC's IP address, as reported by the accept() function.
1424 /** Client PC's IP address, as reported by the accept() function.
1426 unsigned long ip_addr_long;
1428 /** The URL that was requested */
1429 struct http_request http[1];
1432 * The final forwarding settings.
1433 * XXX: Currently this is only used for forward-override,
1434 * so we can free the space in sweep.
1436 struct forward_spec * fwd;
1438 /** An I/O buffer used for buffering data read from the network */
1441 /** List of all headers for this request */
1442 struct list headers[1];
1444 /** List of all tags that apply to this request */
1445 struct list tags[1];
1447 /** MIME-Type key, see CT_* above */
1448 unsigned int content_type;
1450 /** Actions files associated with this client */
1451 struct file_list *actions_list[MAX_AF_FILES];
1453 /** pcrs job files. */
1454 struct file_list *rlist[MAX_AF_FILES];
1456 /** Length after content modification. */
1457 unsigned long long content_length;
1459 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
1460 /** Expected length of content after which we
1461 * should stop reading from the server socket.
1463 /* XXX: is this the right location? */
1464 unsigned long long expected_content_length;
1465 #endif /* def FEATURE_CONNECTION_KEEP_ALIVE */
1467 #ifdef FEATURE_TRUST
1470 struct file_list *tlist;
1472 #endif /* def FEATURE_TRUST */
1475 * Failure reason to embedded in the CGI error page,
1476 * or NULL. Currently only used for socks errors.
1478 char *error_message;
1480 /** Next thread in linked list. Only read or modify from the main thread! */
1481 struct client_state *next;
1486 * A function to add a header
1488 typedef jb_err (*add_header_func_ptr)(struct client_state *);
1491 * A function to process a header
1493 typedef jb_err (*parser_func_ptr )(struct client_state *, char **);
1497 * List of available CGI functions.
1499 struct cgi_dispatcher
1501 /** The URL of the CGI, relative to the CGI root. */
1502 const char * const name;
1504 /** The handler function for the CGI */
1505 jb_err (* const handler)(struct client_state *csp, struct http_response *rsp, const struct map *parameters);
1507 /** The description of the CGI, to appear on the main menu, or NULL to hide it. */
1508 const char * const description;
1510 /** A flag that indicates whether unintentional calls to this CGI can cause damage */
1516 * A data file used by Privoxy. Kept in a linked list.
1521 * This is a pointer to the data structures associated with the file.
1522 * Read-only once the structure has been created.
1527 * The unloader function.
1528 * Normally NULL. When we are finished with file (i.e. when we have
1529 * loaded a new one), set to a pointer to an unloader function.
1530 * Unloader will be called by sweep() (called from main loop) when
1531 * all clients using this file are done. This prevents threading
1534 void (*unloader)(void *);
1537 * Used internally by sweep(). Do not access from elsewhere.
1542 * File last-modified time, so we can check if file has been changed.
1543 * Read-only once the structure has been created.
1545 time_t lastmodified;
1548 * The full filename.
1553 * Pointer to next entry in the linked list of all "file_list"s.
1554 * This linked list is so that sweep() can navigate it.
1555 * Since sweep() can remove items from the list, we must be careful
1556 * to only access this value from main thread (when we know sweep
1557 * won't be running).
1559 struct file_list *next;
1563 #ifdef FEATURE_TRUST
1566 * The format of a trust file when loaded into memory.
1570 struct url_spec url[1]; /**< The URL pattern */
1571 int reject; /**< FIXME: Please document this! */
1572 struct block_spec *next; /**< Next entry in linked list */
1576 * Arbitrary limit for the number of trusted referrers.
1578 #define MAX_TRUSTED_REFERRERS 512
1580 #endif /* def FEATURE_TRUST */
1583 #define SOCKS_NONE 0 /**< Don't use a SOCKS server */
1584 #define SOCKS_4 40 /**< original SOCKS 4 protocol */
1585 #define SOCKS_4A 41 /**< as modified for hosts w/o external DNS */
1586 #define SOCKS_5 50 /**< as modified for hosts w/o external DNS */
1590 * How to forward a connection to a parent proxy.
1594 /** URL pattern that this forward_spec is for. */
1595 struct url_spec url[1];
1597 /** Connection type. Must be SOCKS_NONE, SOCKS_4, SOCKS_4A or SOCKS_5. */
1600 /** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
1603 /** SOCKS server port. */
1606 /** Parent HTTP proxy hostname, or NULL for none. */
1609 /** Parent HTTP proxy port. */
1612 /** Next entry in the linked list. */
1613 struct forward_spec *next;
1618 * Initializer for a static struct forward_spec.
1620 #define FORWARD_SPEC_INITIALIZER { { URL_SPEC_INITIALIZER }, 0, NULL, 0, NULL, 0, NULL }
1622 /* Supported filter types */
1623 #define FT_CONTENT_FILTER 0
1624 #define FT_CLIENT_HEADER_FILTER 1
1625 #define FT_SERVER_HEADER_FILTER 2
1626 #define FT_CLIENT_HEADER_TAGGER 3
1627 #define FT_SERVER_HEADER_TAGGER 4
1629 #define MAX_FILTER_TYPES 5
1632 * This struct represents one filter (one block) from
1633 * the re_filterfile. If there is more than one filter
1634 * in the file, the file will be represented by a
1635 * chained list of re_filterfile specs.
1637 struct re_filterfile_spec
1639 char *name; /**< Name from FILTER: statement in re_filterfile. */
1640 char *description; /**< Description from FILTER: statement in re_filterfile. */
1641 struct list patterns[1]; /**< The patterns from the re_filterfile. */
1642 pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
1643 int type; /**< Filter type (content, client-header, server-header). */
1644 int dynamic; /**< Set to one if the pattern might contain variables
1645 and has to be recompiled for every request. */
1646 struct re_filterfile_spec *next; /**< The pointer for chaining. */
1652 #define ACL_PERMIT 1 /**< Accept connection request */
1653 #define ACL_DENY 2 /**< Reject connection request */
1656 * An IP address pattern. Used to specify networks in the ACL.
1658 struct access_control_addr
1660 unsigned long addr; /**< The IP address as an integer. */
1661 unsigned long mask; /**< The network mask as an integer. */
1662 unsigned long port; /**< The port number. */
1666 * An access control list (ACL) entry.
1668 * This is a linked list.
1670 struct access_control_list
1672 struct access_control_addr src[1]; /**< Client IP address */
1673 struct access_control_addr dst[1]; /**< Website or parent proxy IP address */
1675 short action; /**< ACL_PERMIT or ACL_DENY */
1676 struct access_control_list *next; /**< The next entry in the ACL. */
1679 #endif /* def FEATURE_ACL */
1682 /** Maximum number of loaders (actions, re_filter, ...) */
1686 /** configuration_spec::feature_flags: CGI actions editor. */
1687 #define RUNTIME_FEATURE_CGI_EDIT_ACTIONS 1U
1689 /** configuration_spec::feature_flags: Web-based toggle. */
1690 #define RUNTIME_FEATURE_CGI_TOGGLE 2U
1692 /** configuration_spec::feature_flags: HTTP-header-based toggle. */
1693 #define RUNTIME_FEATURE_HTTP_TOGGLE 4U
1695 /** configuration_spec::feature_flags: Split large forms to limit the number of GET arguments. */
1696 #define RUNTIME_FEATURE_SPLIT_LARGE_FORMS 8U
1698 /** configuration_spec::feature_flags: Check the host header for requests with host-less request lines. */
1699 #define RUNTIME_FEATURE_ACCEPT_INTERCEPTED_REQUESTS 16U
1701 /** configuration_spec::feature_flags: Don't allow to circumvent blocks with the force prefix. */
1702 #define RUNTIME_FEATURE_ENFORCE_BLOCKS 32U
1704 /** configuration_spec::feature_flags: Allow to block or redirect CGI requests. */
1705 #define RUNTIME_FEATURE_CGI_CRUNCHING 64U
1707 /** configuration_spec::feature_flags: Try to keep the connection to the server alive. */
1708 #define RUNTIME_FEATURE_CONNECTION_KEEP_ALIVE 128U
1711 * Data loaded from the configuration file.
1713 * (Anomaly: toggle is still handled through a global, not this structure)
1715 struct configuration_spec
1720 /** Nonzero to enable multithreading. */
1724 * Bitmask of features that can be enabled/disabled through the config
1725 * file. Currently defined bits:
1727 * - RUNTIME_FEATURE_CGI_EDIT_ACTIONS
1728 * - RUNTIME_FEATURE_CGI_TOGGLE
1729 * - RUNTIME_FEATURE_HTTP_TOGGLE
1730 * - RUNTIME_FEATURE_SPLIT_LARGE_FORMS
1732 unsigned feature_flags;
1734 /** The log file name. */
1735 const char *logfile;
1737 /** The config file directory. */
1738 const char *confdir;
1740 /** The directory for customized CGI templates. */
1741 const char *templdir;
1743 /** The log file directory. */
1746 /** The full paths to the actions files. */
1747 const char *actions_file[MAX_AF_FILES];
1749 /** The short names of the actions files. */
1750 const char *actions_file_short[MAX_AF_FILES];
1752 /** The administrator's email address */
1753 char *admin_address;
1755 /** A URL with info on this proxy */
1756 char *proxy_info_url;
1758 /** URL to the user manual (on our website or local copy) */
1761 /** The file names of the pcre filter files. */
1762 const char *re_filterfile[MAX_AF_FILES];
1764 /** The short names of the pcre filter files. */
1765 const char *re_filterfile_short[MAX_AF_FILES];
1767 /** The hostname to show on CGI pages, or NULL to use the real one. */
1768 const char *hostname;
1770 /** IP address to bind to. Defaults to HADDR_DEFAULT == 127.0.0.1. */
1773 /** Port to bind to. Defaults to HADDR_PORT == 8118. */
1776 /** Size limit for IOB */
1777 size_t buffer_limit;
1779 #ifdef FEATURE_TRUST
1781 /** The file name of the trust file. */
1782 const char * trustfile;
1784 /** FIXME: DOCME: Document this. */
1785 struct list trust_info[1];
1787 /** FIXME: DOCME: Document this. */
1788 struct url_spec *trust_list[MAX_TRUSTED_REFERRERS];
1790 #endif /* def FEATURE_TRUST */
1794 /** The access control list (ACL). */
1795 struct access_control_list *acl;
1797 #endif /* def FEATURE_ACL */
1799 /** Information about parent proxies (forwarding). */
1800 struct forward_spec *forward;
1802 /** Number of retries in case a forwarded connection attempt fails */
1803 int forwarded_connect_retries;
1805 /* Timeout when waiting on sockets for data to become available. */
1808 /** All options from the config file, HTML-formatted. */
1811 /** The configuration file object. */
1812 struct file_list *config_file_list;
1814 /** List of loaders */
1815 int (*loaders[NLOADERS])(struct client_state *);
1817 /** Nonzero if we need to bind() to the new port. */
1821 /** Calculates the number of elements in an array, using sizeof. */
1822 #define SZ(X) (sizeof(X) / sizeof(*X))
1824 #ifdef FEATURE_FORCE_LOAD
1825 /** The force load URL prefix. */
1826 #define FORCE_PREFIX "/PRIVOXY-FORCE"
1827 #endif /* def FEATURE_FORCE_LOAD */
1829 #ifdef FEATURE_NO_GIFS
1830 /** The MIME type for images ("image/png" or "image/gif"). */
1831 #define BUILTIN_IMAGE_MIMETYPE "image/png"
1833 #define BUILTIN_IMAGE_MIMETYPE "image/gif"
1834 #endif /* def FEATURE_NO_GIFS */
1841 /** URL for the Privoxy home page. */
1842 #define HOME_PAGE_URL "http://www.privoxy.org/"
1844 /** URL for the Privoxy user manual. */
1845 #define USER_MANUAL_URL HOME_PAGE_URL VERSION "/user-manual/"
1847 /** Prefix for actions help links (append to USER_MANUAL_URL). */
1848 #define ACTIONS_HELP_PREFIX "actions-file.html#"
1850 /** Prefix for config option help links (append to USER_MANUAL_URL). */
1851 #define CONFIG_HELP_PREFIX "config.html#"
1854 * The "hosts" to intercept and display CGI pages.
1855 * First one is a hostname only, second one can specify host and path.
1858 * 1) Do not specify the http: prefix
1859 * 2) CGI_SITE_2_PATH must not end with /, one will be added automatically.
1860 * 3) CGI_SITE_2_PATH must start with /, unless it is the empty string.
1862 #define CGI_SITE_1_HOST "p.p"
1863 #define CGI_SITE_2_HOST "config.privoxy.org"
1864 #define CGI_SITE_2_PATH ""
1867 * The prefix for CGI pages. Written out in generated HTML.
1868 * INCLUDES the trailing slash.
1870 #define CGI_PREFIX "http://" CGI_SITE_2_HOST CGI_SITE_2_PATH "/"
1876 #endif /* ndef PROJECT_H_INCLUDED */