1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
4 #define PROJECT_H_VERSION "$Id: project.h,v 1.124 2008/11/16 12:43:49 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 1.124 2008/11/16 12:43:49 fabiankeil
41 * Turn keep-alive support into a runtime feature
42 * that is disabled by setting keep-alive-timeout
43 * to a negative value.
45 * Revision 1.123 2008/11/10 16:55:59 fabiankeil
46 * Fix a gcc44 warning (in filters.c).
48 * Revision 1.122 2008/10/16 07:11:34 fabiankeil
49 * Fix a bunch of gcc44 conversion warnings.
51 * Revision 1.121 2008/10/09 18:21:41 fabiankeil
52 * Flush work-in-progress changes to keep outgoing connections
53 * alive where possible. Incomplete and mostly #ifdef'd out.
55 * Revision 1.120 2008/09/21 13:36:52 fabiankeil
56 * If change-x-forwarded-for{add} is used and the client
57 * sends multiple X-Forwarded-For headers, append the client's
58 * IP address to each one of them. "Traditionally" we would
59 * lose all but the last one.
61 * Revision 1.119 2008/09/20 10:04:33 fabiankeil
62 * Remove hide-forwarded-for-headers action which has
63 * been obsoleted by change-x-forwarded-for{block}.
65 * Revision 1.118 2008/09/19 15:26:29 fabiankeil
66 * Add change-x-forwarded-for{} action to block or add
67 * X-Forwarded-For headers. Mostly based on code removed
70 * Revision 1.117 2008/08/30 12:03:07 fabiankeil
71 * Remove FEATURE_COOKIE_JAR.
73 * Revision 1.116 2008/05/20 16:05:02 fabiankeil
74 * Move parsers structure definition from project.h to parsers.h.
76 * Revision 1.115 2008/05/19 16:57:20 fabiankeil
77 * Declare all members of the parsers structure immutable.
79 * Revision 1.114 2008/04/11 16:35:39 fabiankeil
80 * Oops, I forgot to shorten the URL_SPEC_INITIALIZER in my last commit.
82 * Revision 1.113 2008/04/10 14:41:04 fabiankeil
83 * Ditch url_spec's path member now that it's no longer used.
85 * Revision 1.112 2008/04/06 15:18:34 fabiankeil
86 * Oh well, rename the --enable-pcre-host-patterns option to
87 * --enable-extended-host-patterns as it's not really PCRE syntax.
89 * Revision 1.111 2008/04/06 14:54:26 fabiankeil
90 * Use PCRE syntax in host patterns when configured
91 * with --enable-pcre-host-patterns.
93 * Revision 1.110 2008/03/29 12:13:46 fabiankeil
94 * Remove send-wafer and send-vanilla-wafer actions.
96 * Revision 1.109 2008/03/28 15:13:41 fabiankeil
97 * Remove inspect-jpegs action.
99 * Revision 1.108 2008/03/27 18:27:36 fabiankeil
100 * Remove kill-popups action.
102 * Revision 1.107 2008/03/26 18:07:08 fabiankeil
103 * Add hostname directive. Closes PR#1918189.
105 * Revision 1.106 2008/03/24 11:21:03 fabiankeil
106 * Share the action settings for multiple patterns in the same
107 * section so we waste less memory for gigantic block lists
108 * (and load them slightly faster). Reported by Franz Schwartau.
110 * Revision 1.105 2008/03/21 11:16:27 fabiankeil
111 * Garbage-collect csp->my_ip_addr_str and csp->my_hostname.
113 * Revision 1.104 2008/03/04 18:30:40 fabiankeil
114 * Remove the treat-forbidden-connects-like-blocks action. We now
115 * use the "blocked" page for forbidden CONNECT requests by default.
117 * Revision 1.103 2008/03/01 14:00:45 fabiankeil
118 * Let the block action take the reason for the block
119 * as argument and show it on the "blocked" page.
121 * Revision 1.102 2008/02/03 13:46:14 fabiankeil
122 * Add SOCKS5 support. Patch #1862863 by Eric M. Hopper with minor changes.
124 * Revision 1.101 2007/12/07 18:29:23 fabiankeil
125 * Remove now-obsolete csp member x_forwarded.
127 * Revision 1.100 2007/09/02 13:42:11 fabiankeil
128 * - Allow port lists in url patterns.
129 * - Ditch unused url_spec member pathlen.
131 * Revision 1.99 2007/07/21 11:51:36 fabiankeil
132 * As Hal noticed, checking dispatch_cgi() as the last cruncher
133 * looks like a bug if CGI requests are blocked unintentionally,
134 * so don't do it unless the user enabled the new config option
135 * "allow-cgi-request-crunching".
137 * Revision 1.98 2007/07/14 07:31:26 fabiankeil
138 * Add new csp->content_type flag (CT_DECLARED).
140 * Revision 1.97 2007/05/27 12:38:08 fabiankeil
141 * - Remove some left-overs from the switch to dedicated header filters.
142 * - Adjust "X-Filter: No" to disable dedicated header filters.
143 * - Prepare for forward-override{}
145 * Revision 1.96 2007/05/14 10:41:15 fabiankeil
146 * Ditch the csp member cookie_list[] which isn't used anymore.
148 * Revision 1.95 2007/04/30 15:02:19 fabiankeil
149 * Introduce dynamic pcrs jobs that can resolve variables.
151 * Revision 1.94 2007/04/15 16:39:21 fabiankeil
152 * Introduce tags as alternative way to specify which
153 * actions apply to a request. At the moment tags can be
154 * created based on client and server headers.
156 * Revision 1.93 2007/03/20 15:16:34 fabiankeil
157 * Use dedicated header filter actions instead of abusing "filter".
158 * Replace "filter-client-headers" and "filter-client-headers"
159 * with "server-header-filter" and "client-header-filter".
161 * Revision 1.92 2007/03/17 15:20:05 fabiankeil
162 * New config option: enforce-blocks.
164 * Revision 1.91 2007/03/05 13:28:03 fabiankeil
165 * Add some CSP_FLAGs for the header parsers.
167 * Revision 1.90 2007/02/07 10:36:16 fabiankeil
168 * Add new http_response member to save
169 * the reason why the response was generated.
171 * Revision 1.89 2007/01/27 13:09:16 fabiankeil
172 * Add new config option "templdir" to
173 * change the templates directory.
175 * Revision 1.88 2007/01/25 13:36:59 fabiankeil
176 * Add csp->error_message for failure reasons
177 * that should be embedded into the CGI pages.
179 * Revision 1.87 2007/01/01 19:36:37 fabiankeil
180 * Integrate a modified version of Wil Mahan's
181 * zlib patch (PR #895531).
183 * Revision 1.86 2006/12/31 17:56:37 fabiankeil
184 * Added config option accept-intercepted-requests
185 * and disabled it by default.
187 * Revision 1.85 2006/12/31 15:03:31 fabiankeil
188 * Fix gcc43 compiler warnings and a comment.
190 * Revision 1.84 2006/12/21 12:57:48 fabiankeil
191 * Add config option "split-large-forms"
192 * to work around the browser bug reported
195 * Revision 1.83 2006/12/06 19:26:29 fabiankeil
196 * Moved HTTP snipplets into jcc.c. They aren't
197 * used anywhere else.
199 * Revision 1.82 2006/09/20 15:50:31 fabiankeil
200 * Doubled size of HOSTENT_BUFFER_SIZE to mask
201 * problems with gethostbyname_r and some
202 * /etc/hosts configurations. Only a workaround
203 * until we get the real fix ready.
204 * Thanks Félix Rauch for reporting.
206 * Increased value of MAX_TRUSTED_REFERRERS from 64 to 512.
208 * Revision 1.81 2006/09/06 13:03:04 fabiankeil
209 * Respond with 400 and a short text message
210 * if the client tries to use Privoxy as FTP proxy.
212 * Revision 1.80 2006/09/06 10:43:32 fabiankeil
213 * Added config option enable-remote-http-toggle
214 * to specify if Privoxy should recognize special
215 * headers (currently only X-Filter) to change its
216 * behaviour. Disabled by default.
218 * Revision 1.79 2006/09/06 09:23:37 fabiankeil
219 * Make number of retries in case of forwarded-connect problems
220 * a config file option (forwarded-connect-retries) and use 0 as
223 * Revision 1.78 2006/08/31 16:25:06 fabiankeil
224 * Work around a buffer overflow that caused Privoxy to
225 * segfault if too many trusted referrers were used. Good
226 * enough for now, but should be replaced with a real
227 * solution after the next release.
229 * Revision 1.77 2006/08/21 12:50:51 david__schmidt
232 * Revision 1.76 2006/08/14 08:25:19 fabiankeil
233 * Split filter-headers{} into filter-client-headers{}
234 * and filter-server-headers{}.
235 * Added parse_header_time() to share some code.
236 * Replaced timegm() with mktime().
238 * Revision 1.75 2006/08/03 02:46:41 david__schmidt
239 * Incorporate Fabian Keil's patch work:
\r * http://www.fabiankeil.de/sourcecode/privoxy/
241 * Revision 1.74 2006/07/18 14:48:47 david__schmidt
242 * Reorganizing the repository: swapping out what was HEAD (the old 3.1 branch)
243 * with what was really the latest development (the v_3_0_branch branch)
245 * Revision 1.72.2.7 2006/01/29 23:10:56 david__schmidt
246 * Multiple filter file support
248 * Revision 1.72.2.6 2004/10/03 12:53:46 david__schmidt
249 * Add the ability to check jpeg images for invalid
250 * lengths of comment blocks. Defensive strategy
251 * against the exploit:
252 * Microsoft Security Bulletin MS04-028
253 * Buffer Overrun in JPEG Processing (GDI+) Could
254 * Allow Code Execution (833987)
255 * Enabled with +inspect-jpegs in actions files.
257 * Revision 1.72.2.5 2004/01/30 15:29:29 oes
258 * Updated the copyright note
260 * Revision 1.72.2.4 2004/01/13 16:12:14 oes
261 * Fixed double slash in USER_MANUAL_URL. Closes BR #867088.
263 * Revision 1.72.2.3 2003/03/11 11:54:37 oes
264 * Introduced RC_FLAG_* flags for use in child process return code
266 * Revision 1.72.2.2 2002/11/28 18:15:44 oes
267 * Added flag to each cgi_dispatcher that allows or denies
268 * external linking and removed const qualifier from
269 * struct list_entry.str.
271 * Revision 1.72.2.1 2002/08/10 11:25:18 oes
272 * - Include config.h for access to config data
273 * - Include <pcre*.h> depending on where they are
275 * Revision 1.72 2002/05/14 21:35:49 oes
276 * Split HELP_LINK_PREFIX into ACTIONS_HELP_PREFIX and CONFIG_HELP_PREFIX
277 * because of split in user-manual
279 * Revision 1.71 2002/05/12 21:39:36 jongfoster
280 * - Adding Doxygen-style comments to structures and #defines.
282 * Revision 1.70 2002/05/12 16:05:50 jongfoster
283 * Fixing ACTION_MASK_ALL to be unsigned long rather than
284 * just unsigned int. I don't know if anyone is porting
285 * Privoxy to 16-bit platforms, but if so, +limit-connect
286 * wouldn't have worked because of this bug.
288 * Revision 1.69 2002/05/08 16:00:16 oes
289 * Added size member to struct iob, so it can
290 * be alloced larger than needed.
292 * Revision 1.68 2002/04/26 12:56:00 oes
293 * Killed REDIRECT_URL, added USER_MANUAL_URL and HELP_LINK_PREFIX
295 * Revision 1.67 2002/04/24 02:12:43 oes
296 * - Jon's multiple AF patch:
297 * - Make csp->actions_list an array
298 * - #define MAX_AF_FILES
299 * - Moved CGI_PARAM_LEN_MAX (500) here
301 * Revision 1.66 2002/04/15 19:06:43 jongfoster
304 * Revision 1.65 2002/04/04 00:36:36 gliptak
305 * always use pcre for matching
307 * Revision 1.64 2002/04/03 22:28:03 gliptak
308 * Removed references to gnu_regex
310 * Revision 1.63 2002/03/31 17:19:00 jongfoster
311 * Win32 only: Enabling STRICT to fix a VC++ compile warning.
313 * Revision 1.62 2002/03/26 22:48:49 swa
316 * Revision 1.61 2002/03/26 22:29:55 swa
317 * we have a new homepage!
319 * Revision 1.60 2002/03/24 15:52:17 jongfoster
320 * Changing CGI URL prefixes for new name
322 * Revision 1.59 2002/03/24 15:23:33 jongfoster
325 * Revision 1.58 2002/03/24 13:25:43 swa
326 * name change related issues
328 * Revision 1.57 2002/03/16 20:28:34 oes
329 * Added descriptions to the filters so users will know what they select in the cgi editor
331 * Revision 1.56 2002/03/13 20:27:30 oes
332 * Fixing bug with CT_TABOO
334 * Revision 1.55 2002/03/12 01:42:50 oes
335 * Introduced modular filters
337 * Revision 1.54 2002/03/09 20:03:52 jongfoster
338 * - Making various functions return int rather than size_t.
339 * (Undoing a recent change). Since size_t is unsigned on
340 * Windows, functions like read_socket that return -1 on
341 * error cannot return a size_t.
343 * THIS WAS A MAJOR BUG - it caused frequent, unpredictable
344 * crashes, and also frequently caused JB to jump to 100%
345 * CPU and stay there. (Because it thought it had just
346 * read ((unsigned)-1) == 4Gb of data...)
348 * - The signature of write_socket has changed, it now simply
349 * returns success=0/failure=nonzero.
351 * - Trying to get rid of a few warnings --with-debug on
352 * Windows, I've introduced a new type "jb_socket". This is
353 * used for the socket file descriptors. On Windows, this
354 * is SOCKET (a typedef for unsigned). Everywhere else, it's
355 * an int. The error value can't be -1 any more, so it's
356 * now JB_INVALID_SOCKET (which is -1 on UNIX, and in
357 * Windows it maps to the #define INVALID_SOCKET.)
359 * - The signature of bind_port has changed.
361 * Revision 1.53 2002/03/08 16:48:55 oes
362 * Added FEATURE_NO_GIFS and BUILTIN_IMAGE_MIMETYPE
364 * Revision 1.52 2002/03/07 03:46:17 oes
365 * Fixed compiler warnings
367 * Revision 1.51 2002/03/05 04:52:42 oes
368 * Deleted non-errlog debugging code
370 * Revision 1.50 2002/03/04 19:32:07 oes
371 * Changed default port to 8118
373 * Revision 1.49 2002/03/04 18:28:55 oes
374 * Deleted PID_FILE_NAME
376 * Revision 1.48 2002/03/03 14:50:40 oes
377 * Fixed CLF logging: Added ocmd member for client's request to struct http_request
379 * Revision 1.47 2002/02/20 23:15:13 jongfoster
380 * Parsing functions now handle out-of-memory gracefully by returning
383 * Revision 1.46 2002/01/17 21:06:09 jongfoster
384 * Now #defining the URLs of the config interface
386 * Minor changes to struct http_request and struct url_spec due to
387 * standardizing that struct http_request is used to represent a URL, and
388 * struct url_spec is used to represent a URL pattern. (Before, URLs were
389 * represented as seperate variables and a partially-filled-in url_spec).
391 * Revision 1.45 2002/01/09 14:33:27 oes
392 * Added HOSTENT_BUFFER_SIZE
394 * Revision 1.44 2001/12/30 14:07:32 steudten
395 * - Add signal handling (unix)
396 * - Add SIGHUP handler (unix)
397 * - Add creation of pidfile (unix)
398 * - Add action 'top' in rc file (RH)
399 * - Add entry 'SIGNALS' to manpage
400 * - Add exit message to logfile (unix)
402 * Revision 1.43 2001/11/22 21:57:51 jongfoster
403 * Making action_spec->flags into an unsigned long rather than just an
405 * Adding ACTION_NO_COOKIE_KEEP
407 * Revision 1.42 2001/11/05 21:42:41 steudten
408 * Include DBG() macro.
410 * Revision 1.41 2001/10/28 19:12:06 jongfoster
411 * Adding ijb_toupper()
413 * Revision 1.40 2001/10/26 17:40:47 oes
414 * Moved ijb_isspace and ijb_tolower to project.h
415 * Removed http->user_agent, csp->referrer and csp->accept_types
417 * Revision 1.39 2001/10/25 03:45:02 david__schmidt
418 * Adding a (void*) cast to freez() because Visual Age C++ won't expand the
419 * macro when called with a cast; so moving the cast to the macro def'n
420 * seems to both eliminate compiler warnings (on darwin and OS/2, anyway) and
421 * doesn't make macro expansion complain. Hope this works for everyone else
424 * Revision 1.38 2001/10/23 21:19:04 jongfoster
425 * New error-handling support: jb_err type and JB_ERR_xxx constants
426 * CGI functions now return a jb_err, and their parameters map is const.
427 * Support for RUNTIME_FEATUREs to enable/disable config editor
428 * Adding a few comments
430 * Revision 1.37 2001/10/14 22:14:01 jongfoster
431 * Removing name_length field from struct cgi_dispatcher, as this is
432 * now calculated at runtime from the "name" field.
434 * Revision 1.36 2001/10/10 16:45:15 oes
435 * Added LIMIT_CONNECT action and string
436 * Fixed HTTP message line termination
437 * Added CFORBIDDEN HTTP message
439 * Revision 1.35 2001/10/07 18:06:43 oes
440 * Added status member to struct http_request
442 * Revision 1.34 2001/10/07 15:45:25 oes
443 * Added url member to struct http_request and commented all
448 * Added ACTION_DOWNGRADE and ACTION_NO_COMPRESSION
450 * Replaced struct client_state members rejected,
451 * force, active and toggled_on with "flags" bitmap.
453 * Added CSP_FLAG_MODIFIED and CSP_FLAG_CHUNKED
455 * Added buffer_limit to struct configuration_spec
457 * Revision 1.33 2001/09/20 13:30:08 steudten
459 * Make freez() more secure in case of: if (exp) { free(z) ; a=*z }
460 * Last case will set z to NULL in free(z) and thats bad..
462 * Revision 1.32 2001/09/16 23:02:51 jongfoster
465 * Revision 1.31 2001/09/16 13:20:29 jongfoster
466 * Rewrite of list library. Now has seperate header and list_entry
467 * structures. Also added a large sprinking of assert()s to the list
470 * Revision 1.30 2001/09/13 23:52:00 jongfoster
471 * Support for both static and dynamically generated CGI pages
473 * Revision 1.29 2001/09/13 23:29:43 jongfoster
474 * Defining FORWARD_SPEC_INITIALIZER
476 * Revision 1.28 2001/09/13 23:05:50 jongfoster
477 * Changing the string paramater to the header parsers a "const".
479 * Revision 1.27 2001/08/05 16:06:20 jongfoster
480 * Modifiying "struct map" so that there are now separate header and
481 * "map_entry" structures. This means that functions which modify a
482 * map no longer need to return a pointer to the modified map.
483 * Also, it no longer reverses the order of the entries (which may be
484 * important with some advanced template substitutions).
486 * Revision 1.26 2001/07/30 22:08:36 jongfoster
487 * Tidying up #defines:
488 * - All feature #defines are now of the form FEATURE_xxx
489 * - Permanently turned off WIN_GUI_EDIT
490 * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
492 * Revision 1.25 2001/07/29 18:43:08 jongfoster
493 * Changing #ifdef _FILENAME_H to FILENAME_H_INCLUDED, to conform to
496 * Revision 1.24 2001/07/25 17:20:27 oes
497 * Introduced http->user_agent
499 * Revision 1.23 2001/07/18 12:32:23 oes
500 * - Added ACTION_STRING_DEANIMATE
501 * - moved #define freez from jcc.h to project.h
503 * Revision 1.22 2001/07/15 17:51:41 jongfoster
504 * Renaming #define STATIC to STATIC_PCRE
506 * Revision 1.21 2001/07/13 14:03:19 oes
507 * - Reorganized regex header inclusion and #defines to
508 * comply to the scheme in configure.in
509 * - Added csp->content_type and its CT_* keys
510 * - Added ACTION_DEANIMATE
511 * - Removed all #ifdef PCRS
513 * Revision 1.20 2001/06/29 21:45:41 oes
514 * Indentation, CRLF->LF, Tab-> Space
516 * Revision 1.19 2001/06/29 13:33:36 oes
517 * - Improved comments
518 * - Introduced http_request.host_ip_addr_str
519 * - Introduced http_response.head_length
520 * - Introduced config.my_ip_addr_str, config.my_hostname,
521 * config.admin_address and config.proxy_info_url
522 * - Removed config.proxy_args_header and config.proxy_args_trailer,
523 * renamed config.proxy_args_invocation to config.proxy_args
524 * - Removed HTML snipplets and GIFs
525 * - Removed logentry from cancelled commit
527 * Revision 1.18 2001/06/09 10:57:39 jongfoster
528 * Adding definition of BUFFER_SIZE.
529 * Changing struct cgi_dispatcher to use "const" strings.
531 * Revision 1.17 2001/06/07 23:15:09 jongfoster
532 * Merging ACL and forward files into config file.
533 * Moving struct gateway members into struct forward_spec
534 * Removing config->proxy_args_gateways
535 * Cosmetic: Adding a few comments
537 * Revision 1.16 2001/06/04 18:31:58 swa
538 * files are now prefixed with either `confdir' or `logdir'.
539 * `make redhat-dist' replaces both entries confdir and logdir
542 * Revision 1.15 2001/06/04 11:28:53 swa
543 * redirect did not work due to missing /
545 * Revision 1.14 2001/06/03 11:03:48 oes
547 * added struct http_response,
548 * changed struct interceptors to struct cgi_dispatcher,
549 * moved HTML stuff to cgi.h
551 * Revision 1.13 2001/06/01 20:05:36 jongfoster
552 * Support for +image-blocker{}: added ACTION_IMAGE_BLOCKER
553 * constant, and removed csp->tinygif.
555 * Revision 1.12 2001/06/01 18:49:17 jongfoster
556 * Replaced "list_share" with "list" - the tiny memory gain was not
557 * worth the extra complexity.
559 * Revision 1.11 2001/06/01 10:32:47 oes
560 * Added constants for anchoring selection bitmap
562 * Revision 1.10 2001/05/31 21:33:53 jongfoster
563 * Changes for new actions file, replacing permissionsfile
564 * and parts of the config file. Also added support for
567 * Revision 1.9 2001/05/31 17:32:31 oes
569 * - Enhanced domain part globbing with infix and prefix asterisk
570 * matching and optional unanchored operation
572 * Revision 1.8 2001/05/29 20:09:15 joergs
573 * HTTP_REDIRECT_TEMPLATE fixed.
575 * Revision 1.7 2001/05/29 09:50:24 jongfoster
576 * Unified blocklist/imagelist/actionslist.
577 * File format is still under discussion, but the internal changes
580 * Also modified interceptor behaviour:
581 * - We now intercept all URLs beginning with one of the following
582 * prefixes (and *only* these prefixes):
584 * * http://ijbswa.sf.net/config/
585 * * http://ijbswa.sourceforge.net/config/
586 * - New interceptors "home page" - go to http://i.j.b/ to see it.
587 * - Internal changes so that intercepted and fast redirect pages
588 * are not replaced with an image.
589 * - Interceptors now have the option to send a binary page direct
590 * to the client. (i.e. ijb-send-banner uses this)
591 * - Implemented show-url-info interceptor. (Which is why I needed
592 * the above interceptors changes - a typical URL is
593 * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif".
594 * The previous mechanism would not have intercepted that, and
595 * if it had been intercepted then it then it would have replaced
598 * Revision 1.6 2001/05/27 22:17:04 oes
600 * - re_process_buffer no longer writes the modified buffer
601 * to the client, which was very ugly. It now returns the
602 * buffer, which it is then written by chat.
604 * - content_length now adjusts the Content-Length: header
605 * for modified documents rather than crunch()ing it.
606 * (Length info in csp->content_length, which is 0 for
607 * unmodified documents)
609 * - For this to work, sed() is called twice when filtering.
611 * Revision 1.5 2001/05/26 00:28:36 jongfoster
612 * Automatic reloading of config file.
613 * Removed obsolete SIGHUP support (Unix) and Reload menu option (Win32).
614 * Most of the global variables have been moved to a new
615 * struct configuration_spec, accessed through csp->config->globalname
616 * Most of the globals remaining are used by the Win32 GUI.
618 * Revision 1.4 2001/05/22 18:46:04 oes
620 * - Enabled filtering banners by size rather than URL
621 * by adding patterns that replace all standard banner
622 * sizes with the "Junkbuster" gif to the re_filterfile
624 * - Enabled filtering WebBugs by providing a pattern
625 * which kills all 1x1 images
627 * - Added support for PCRE_UNGREEDY behaviour to pcrs,
628 * which is selected by the (nonstandard and therefore
629 * capital) letter 'U' in the option string.
630 * It causes the quantifiers to be ungreedy by default.
631 * Appending a ? turns back to greedy (!).
633 * - Added a new interceptor ijb-send-banner, which
634 * sends back the "Junkbuster" gif. Without imagelist or
635 * MSIE detection support, or if tinygif = 1, or the
636 * URL isn't recognized as an imageurl, a lame HTML
637 * explanation is sent instead.
639 * - Added new feature, which permits blocking remote
640 * script redirects and firing back a local redirect
642 * The feature is conditionally compiled, i.e. it
643 * can be disabled with --disable-fast-redirects,
644 * plus it must be activated by a "fast-redirects"
645 * line in the config file, has its own log level
646 * and of course wants to be displayed by show-proxy-args
647 * Note: Boy, all the #ifdefs in 1001 locations and
648 * all the fumbling with configure.in and acconfig.h
649 * were *way* more work than the feature itself :-(
651 * - Because a generic redirect template was needed for
652 * this, tinygif = 3 now uses the same.
654 * - Moved GIFs, and other static HTTP response templates
659 * - Removed some >400 CRs again (Jon, you really worked
662 * Revision 1.3 2001/05/20 01:21:20 jongfoster
663 * Version 2.9.4 checkin.
664 * - Merged popupfile and cookiefile, and added control over PCRS
665 * filtering, in new "actionsfile".
666 * - Implemented LOG_LEVEL_FATAL, so that if there is a configuration
667 * file error you now get a message box (in the Win32 GUI) rather
668 * than the program exiting with no explanation.
669 * - Made killpopup use the PCRS MIME-type checking and HTTP-header
671 * - Removed tabs from "config"
672 * - Moved duplicated url parsing code in "loaders.c" to a new funcition.
673 * - Bumped up version number.
675 * Revision 1.2 2001/05/17 23:01:01 oes
676 * - Cleaned CRLF's from the sources and related files
678 * Revision 1.1.1.1 2001/05/15 13:59:03 oes
679 * Initial import of version 2.9.3 source tree
682 *********************************************************************/
685 /* Declare struct FILE for vars and funcs. */
688 /* Need time_t for file_list */
690 /* Needed for pcre choice */
694 * Include appropriate regular expression libraries.
695 * Note that pcrs and pcre (native) are needed for cgi
696 * and are included anyway.
702 # ifdef PCRE_H_IN_SUBDIR
703 # include <pcre/pcre.h>
716 # include "pcreposix.h"
718 # ifdef PCRE_H_IN_SUBDIR
719 # include <pcre/pcreposix.h>
721 # include <pcreposix.h>
727 #endif /* def AMIGA */
731 * I don't want to have to #include all this just for the declaration
732 * of SOCKET. However, it looks like we have to...
747 typedef SOCKET jb_socket;
749 #define JB_INVALID_SOCKET INVALID_SOCKET
751 #else /* ndef _WIN32 */
754 * The type used by sockets. On UNIX it's an int. Microsoft decided to
755 * make it an unsigned.
757 typedef int jb_socket;
760 * The error value used for variables of type jb_socket. On UNIX this
761 * is -1, however Microsoft decided to make socket handles unsigned, so
762 * they use a different value.
765 #define JB_INVALID_SOCKET (-1)
767 #endif /* ndef _WIN32 */
771 * A standard error code. This should be JB_ERR_OK or one of the JB_ERR_xxx
776 #define JB_ERR_OK 0 /**< Success, no error */
777 #define JB_ERR_MEMORY 1 /**< Out of memory */
778 #define JB_ERR_CGI_PARAMS 2 /**< Missing or corrupt CGI parameters */
779 #define JB_ERR_FILE 3 /**< Error opening, reading or writing a file */
780 #define JB_ERR_PARSE 4 /**< Error parsing file */
781 #define JB_ERR_MODIFIED 5 /**< File has been modified outside of the
782 CGI actions editor. */
783 #define JB_ERR_COMPRESS 6 /**< Error on decompression */
786 * This macro is used to free a pointer that may be NULL.
787 * It also sets the variable to NULL after it's been freed.
788 * The paramater should be a simple variable without side effects.
790 #define freez(X) { if(X) { free((void*)X); X = NULL ; } }
794 * Fix a problem with Solaris. There should be no effect on other
797 * Solaris's isspace() is a macro which uses it's argument directly
798 * as an array index. Therefore we need to make sure that high-bit
799 * characters generate +ve values, and ideally we also want to make
800 * the argument match the declared parameter type of "int".
802 * Note: Remember to #include <ctype.h> if you use these macros.
804 #define ijb_toupper(__X) toupper((int)(unsigned char)(__X))
805 #define ijb_tolower(__X) tolower((int)(unsigned char)(__X))
806 #define ijb_isspace(__X) isspace((int)(unsigned char)(__X))
809 * Use for statically allocated buffers if you have no other choice.
810 * Remember to check the length of what you write into the buffer
811 * - we don't want any buffer overflows!
813 #define BUFFER_SIZE 5000
816 * Max length of CGI parameters (arbitrary limit).
818 #define CGI_PARAM_LEN_MAX 500
821 * Buffer size for capturing struct hostent data in the
822 * gethostby(name|addr)_r library calls. Since we don't
823 * loop over gethostbyname_r, the buffer must be sufficient
824 * to accomodate multiple IN A RRs, as used in DNS round robin
825 * load balancing. W3C's wwwlib uses 1K, so that should be
826 * good enough for us, too.
829 * XXX: Temporary doubled, for some configurations
830 * 1K is still too small and we didn't get the
831 * real fix ready for inclusion.
833 #define HOSTENT_BUFFER_SIZE 2048
836 * Do not use. Originally this was so that you can
837 * say "while (FOREVER) { ...do something... }".
838 * However, this gives a warning with some compilers (e.g. VC++).
839 * Instead, use "for (;;) { ...do something... }".
844 * Default IP address to listen on, as a string.
845 * Set to "127.0.0.1".
847 #define HADDR_DEFAULT "127.0.0.1"
850 * Default port to listen on, as a number.
853 #define HADDR_PORT 8118
856 /* Forward def for struct client_state */
857 struct configuration_spec;
861 * Entry in a linked list of strings.
866 * The string pointer. It must point to a dynamically malloc()ed
867 * string or be NULL for the list functions to work. In the latter
868 * case, just be careful next time you iterate through the list in
873 /** Next entry in the linked list, or NULL if no more. */
874 struct list_entry *next;
878 * A header for a linked list of strings.
882 /** First entry in the list, or NULL if the list is empty. */
883 struct list_entry *first;
885 /** Last entry in the list, or NULL if the list is empty. */
886 struct list_entry *last;
891 * An entry in a map. This is a name=value pair.
895 /** The key for the map. */
897 /** The value associated with that key. */
899 /** The next map entry, or NULL if none. */
900 struct map_entry *next;
904 * A map from a string to another string.
905 * This is used for the paramaters passed in a HTTP GET request, and
906 * to store the exports when the CGI interface is filling in a template.
910 /** The first map entry, or NULL if the map is empty. */
911 struct map_entry *first;
912 /** The last map entry, or NULL if the map is empty. */
913 struct map_entry *last;
918 * A HTTP request. This includes the method (GET, POST) and
921 * This is also used whenever we want to match a URL against a
922 * URL pattern. This always contains the URL to match, and never
923 * a URL pattern. (See struct url_spec).
927 char *cmd; /**< Whole command line: method, URL, Version */
928 char *ocmd; /**< Backup of original cmd for CLF logging */
929 char *gpc; /**< HTTP method: GET, POST, ... */
930 char *url; /**< The URL */
931 char *ver; /**< Protocol version */
932 int status; /**< HTTP Status */
934 char *host; /**< Host part of URL */
935 int port; /**< Port of URL or 80 (default) */
936 char *path; /**< Path of URL */
937 char *hostport; /**< host[:port] */
938 int ssl; /**< Flag if protocol is https */
940 char *host_ip_addr_str; /**< String with dotted decimal representation
941 of host's IP. NULL before connect_to() */
943 char *dbuffer; /**< Buffer with '\0'-delimited domain name. */
944 char **dvec; /**< List of pointers to the strings in dbuffer. */
945 int dcount; /**< How many parts to this domain? (length of dvec) */
949 * Reasons for generating a http_response instead of delivering
950 * the requested resource. Mostly ordered the way they are checked
953 #define RSP_REASON_UNSUPPORTED 1
954 #define RSP_REASON_BLOCKED 2
955 #define RSP_REASON_UNTRUSTED 3
956 #define RSP_REASON_REDIRECTED 4
957 #define RSP_REASON_CGI_CALL 5
958 #define RSP_REASON_NO_SUCH_DOMAIN 6
959 #define RSP_REASON_FORWARDING_FAILED 7
960 #define RSP_REASON_CONNECT_FAILED 8
961 #define RSP_REASON_OUT_OF_MEMORY 9
962 #define RSP_REASON_INTERNAL_ERROR 10
965 * Response generated by CGI, blocker, or error handler
969 char *status; /**< HTTP status (string). */
970 struct list headers[1]; /**< List of header lines. */
971 char *head; /**< Formatted http response head. */
972 size_t head_length; /**< Length of http response head. */
973 char *body; /**< HTTP document body. */
974 size_t content_length; /**< Length of body, REQUIRED if binary body. */
975 int is_static; /**< Nonzero if the content will never change and
976 should be cached by the browser (e.g. images). */
977 int reason; /**< Why the response was generated in the first place. */
981 * A URL or a tag pattern.
985 /** The string which was parsed to produce this url_spec.
986 Used for debugging or display only. */
989 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
990 regex_t *host_regex;/**< Regex for host matching */
992 char *dbuffer; /**< Buffer with '\0'-delimited domain name, or NULL to match all hosts. */
993 char **dvec; /**< List of pointers to the strings in dbuffer. */
994 int dcount; /**< How many parts to this domain? (length of dvec) */
995 int unanchored; /**< Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT. */
996 #endif /* defined FEATURE_EXTENDED_HOST_PATTERNS */
998 char *port_list; /**< List of acceptable ports, or NULL to match all ports */
1000 regex_t *preg; /**< Regex for matching path part */
1001 regex_t *tag_regex; /**< Regex for matching tags */
1005 * If you declare a static url_spec, this is the value to initialize it to zero.
1007 #ifndef FEATURE_EXTENDED_HOST_PATTERNS
1008 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, 0, 0, NULL, NULL, NULL }
1010 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, NULL, NULL }
1011 #endif /* def FEATURE_EXTENDED_HOST_PATTERNS */
1014 * Constant for host part matching in URLs. If set, indicates that the start of
1015 * the pattern must match the start of the URL. E.g. this is not set for the
1016 * pattern ".example.com", so that it will match both "example.com" and
1017 * "www.example.com". It is set for the pattern "example.com", which makes it
1018 * match "example.com" only, not "www.example.com".
1020 #define ANCHOR_LEFT 1
1023 * Constant for host part matching in URLs. If set, indicates that the end of
1024 * the pattern must match the end of the URL. E.g. this is not set for the
1025 * pattern "ad.", so that it will match any host called "ad", irrespective
1026 * of how many subdomains are in the fully-qualified domain name.
1028 #define ANCHOR_RIGHT 2
1032 * An I/O buffer. Holds a string which can be appended to, and can have data
1033 * removed from the beginning.
1037 char *buf; /**< Start of buffer */
1038 char *cur; /**< Start of relevant data */
1039 char *eod; /**< End of relevant data */
1040 size_t size; /**< Size as malloc()ed */
1045 * Return the number of bytes in the I/O buffer associated with the passed
1046 * client_state pointer.
1049 #define IOB_PEEK(CSP) ((CSP->iob->cur > CSP->iob->eod) ? (CSP->iob->eod - CSP->iob->cur) : 0)
1053 * Remove any data in the I/O buffer associated with the passed
1054 * client_state pointer.
1056 #define IOB_RESET(CSP) if(CSP->iob->buf) free(CSP->iob->buf); memset(CSP->iob, '\0', sizeof(CSP->iob));
1058 /* Bits for csp->content_type bitmask: */
1059 #define CT_TEXT 0x0001U /**< Suitable for pcrs filtering. */
1060 #define CT_GIF 0x0002U /**< Suitable for GIF filtering. */
1061 #define CT_TABOO 0x0004U /**< DO NOT filter, irrespective of other flags. */
1063 /* Although these are not, strictly speaking, content types
1064 * (they are content encodings), it is simple to handle them
1067 #define CT_GZIP 0x0010U /**< gzip-compressed data. */
1068 #define CT_DEFLATE 0x0020U /**< zlib-compressed data. */
1071 * Flag to signal that the server declared the content type,
1072 * so we can differentiate between unknown and undeclared
1075 #define CT_DECLARED 0x0040U
1078 * The mask which includes all actions.
1080 #define ACTION_MASK_ALL (~0UL)
1083 * The most compatible set of actions - i.e. none.
1085 #define ACTION_MOST_COMPATIBLE 0x00000000UL
1087 /** Action bitmap: Block the request. */
1088 #define ACTION_BLOCK 0x00000001UL
1089 /** Action bitmap: Deanimate if it's a GIF. */
1090 #define ACTION_DEANIMATE 0x00000002UL
1091 /** Action bitmap: Downgrade HTTP/1.1 to 1.0. */
1092 #define ACTION_DOWNGRADE 0x00000004UL
1093 /** Action bitmap: Fast redirects. */
1094 #define ACTION_FAST_REDIRECTS 0x00000008UL
1095 /** Action bitmap: Remove or add "X-Forwarded-For" header. */
1096 #define ACTION_CHANGE_X_FORWARDED_FOR 0x00000010UL
1097 /** Action bitmap: Hide "From" header. */
1098 #define ACTION_HIDE_FROM 0x00000020UL
1099 /** Action bitmap: Hide "Referer" header. (sic - follow HTTP, not English). */
1100 #define ACTION_HIDE_REFERER 0x00000040UL
1101 /** Action bitmap: Hide "User-Agent" and similar headers. */
1102 #define ACTION_HIDE_USER_AGENT 0x00000080UL
1103 /** Action bitmap: This is an image. */
1104 #define ACTION_IMAGE 0x00000100UL
1105 /** Action bitmap: Sets the image blocker. */
1106 #define ACTION_IMAGE_BLOCKER 0x00000200UL
1107 /** Action bitmap: Prevent compression. */
1108 #define ACTION_NO_COMPRESSION 0x00000400UL
1109 /** Action bitmap: Change cookies to session only cookies. */
1110 #define ACTION_NO_COOKIE_KEEP 0x00000800UL
1111 /** Action bitmap: Block rending cookies. */
1112 #define ACTION_NO_COOKIE_READ 0x00001000UL
1113 /** Action bitmap: Block setting cookies. */
1114 #define ACTION_NO_COOKIE_SET 0x00002000UL
1115 /** Action bitmap: Override the forward settings in the config file */
1116 #define ACTION_FORWARD_OVERRIDE 0x00004000UL
1117 /** Action bitmap: Block as empty document */
1118 #define ACTION_HANDLE_AS_EMPTY_DOCUMENT 0x00008000UL
1119 /** Action bitmap: Limit CONNECT requests to safe ports. */
1120 #define ACTION_LIMIT_CONNECT 0x00010000UL
1121 /** Action bitmap: Redirect request. */
1122 #define ACTION_REDIRECT 0x00020000UL
1123 /** Action bitmap: Crunch or modify "if-modified-since" header. */
1124 #define ACTION_HIDE_IF_MODIFIED_SINCE 0x00040000UL
1125 /** Action bitmap: Overwrite Content-Type header. */
1126 #define ACTION_CONTENT_TYPE_OVERWRITE 0x00080000UL
1127 /** Action bitmap: Crunch specified server header. */
1128 #define ACTION_CRUNCH_SERVER_HEADER 0x00100000UL
1129 /** Action bitmap: Crunch specified client header */
1130 #define ACTION_CRUNCH_CLIENT_HEADER 0x00200000UL
1131 /** Action bitmap: Enable text mode by force */
1132 #define ACTION_FORCE_TEXT_MODE 0x00400000UL
1133 /** Action bitmap: Enable text mode by force */
1134 #define ACTION_CRUNCH_IF_NONE_MATCH 0x00800000UL
1135 /** Action bitmap: Enable content-dispostion crunching */
1136 #define ACTION_HIDE_CONTENT_DISPOSITION 0x01000000UL
1137 /** Action bitmap: Replace or block Last-Modified header */
1138 #define ACTION_OVERWRITE_LAST_MODIFIED 0x02000000UL
1139 /** Action bitmap: Replace or block Accept-Language header */
1140 #define ACTION_HIDE_ACCEPT_LANGUAGE 0x04000000UL
1143 /** Action string index: How to deanimate GIFs */
1144 #define ACTION_STRING_DEANIMATE 0
1145 /** Action string index: Replacement for "From:" header */
1146 #define ACTION_STRING_FROM 1
1147 /** Action string index: How to block images */
1148 #define ACTION_STRING_IMAGE_BLOCKER 2
1149 /** Action string index: Replacement for "Referer:" header */
1150 #define ACTION_STRING_REFERER 3
1151 /** Action string index: Replacement for "User-Agent:" header */
1152 #define ACTION_STRING_USER_AGENT 4
1153 /** Action string index: Legal CONNECT ports. */
1154 #define ACTION_STRING_LIMIT_CONNECT 5
1155 /** Action string index: Server headers containing this pattern are crunched*/
1156 #define ACTION_STRING_SERVER_HEADER 6
1157 /** Action string index: Client headers containing this pattern are crunched*/
1158 #define ACTION_STRING_CLIENT_HEADER 7
1159 /** Action string index: Replacement for the "Accept-Language:" header*/
1160 #define ACTION_STRING_LANGUAGE 8
1161 /** Action string index: Replacement for the "Content-Type:" header*/
1162 #define ACTION_STRING_CONTENT_TYPE 9
1163 /** Action string index: Replacement for the "content-dispostion:" header*/
1164 #define ACTION_STRING_CONTENT_DISPOSITION 10
1165 /** Action string index: Replacement for the "If-Modified-Since:" header*/
1166 #define ACTION_STRING_IF_MODIFIED_SINCE 11
1167 /** Action string index: Replacement for the "Last-Modified:" header. */
1168 #define ACTION_STRING_LAST_MODIFIED 12
1169 /** Action string index: Redirect URL */
1170 #define ACTION_STRING_REDIRECT 13
1171 /** Action string index: Decode before redirect? */
1172 #define ACTION_STRING_FAST_REDIRECTS 14
1173 /** Action string index: Overriding forward rule. */
1174 #define ACTION_STRING_FORWARD_OVERRIDE 15
1175 /** Action string index: Reason for the block. */
1176 #define ACTION_STRING_BLOCK 16
1177 /** Action string index: what to do with the "X-Forwarded-For" header. */
1178 #define ACTION_STRING_CHANGE_X_FORWARDED_FOR 17
1179 /** Number of string actions. */
1180 #define ACTION_STRING_COUNT 18
1183 /* To make the ugly hack in sed easier to understand */
1184 #define CHECK_EVERY_HEADER_REMAINING 0
1187 /** Index into current_action_spec::multi[] for headers to add. */
1188 #define ACTION_MULTI_ADD_HEADER 0
1189 /** Index into current_action_spec::multi[] for content filters to apply. */
1190 #define ACTION_MULTI_FILTER 1
1191 /** Index into current_action_spec::multi[] for server-header filters to apply. */
1192 #define ACTION_MULTI_SERVER_HEADER_FILTER 2
1193 /** Index into current_action_spec::multi[] for client-header filters to apply. */
1194 #define ACTION_MULTI_CLIENT_HEADER_FILTER 3
1195 /** Index into current_action_spec::multi[] for client-header tags to apply. */
1196 #define ACTION_MULTI_CLIENT_HEADER_TAGGER 4
1197 /** Index into current_action_spec::multi[] for server-header tags to apply. */
1198 #define ACTION_MULTI_SERVER_HEADER_TAGGER 5
1199 /** Number of multi-string actions. */
1200 #define ACTION_MULTI_COUNT 6
1204 * This structure contains a list of actions to apply to a URL.
1205 * It only contains positive instructions - no "-" options.
1206 * It is not used to store the actions list itself, only for
1207 * url_actions() to return the current values.
1209 struct current_action_spec
1211 /** Actions to apply. A bit set to "1" means perform the action. */
1212 unsigned long flags;
1215 * Paramaters for those actions that require them.
1216 * Each entry is valid if & only if the corresponding entry in "flags" is
1219 char * string[ACTION_STRING_COUNT];
1221 /** Lists of strings for multi-string actions. */
1222 struct list multi[ACTION_MULTI_COUNT][1];
1227 * This structure contains a set of changes to actions.
1228 * It can contain both positive and negative instructions.
1229 * It is used to store an entry in the actions list.
1233 unsigned long mask; /**< Actions to keep. A bit set to "0" means remove action. */
1234 unsigned long add; /**< Actions to add. A bit set to "1" means add action. */
1237 * Paramaters for those actions that require them.
1238 * Each entry is valid if & only if the corresponding entry in "flags" is
1241 char * string[ACTION_STRING_COUNT];
1243 /** Lists of strings to remove, for multi-string actions. */
1244 struct list multi_remove[ACTION_MULTI_COUNT][1];
1246 /** If nonzero, remove *all* strings from the multi-string action. */
1247 int multi_remove_all[ACTION_MULTI_COUNT];
1249 /** Lists of strings to add, for multi-string actions. */
1250 struct list multi_add[ACTION_MULTI_COUNT][1];
1255 * This structure is used to store action files.
1257 * It contains an URL or tag pattern, and the changes to
1258 * the actions. It's a linked list and should only be
1259 * free'd through unload_actions_file() unless there's
1260 * only a single entry.
1264 struct url_spec url[1]; /**< The URL or tag pattern. */
1266 struct action_spec *action; /**< Action settings that might be shared with
1267 the list entry before or after the current
1268 one and can't be free'd willy nilly. */
1270 struct url_actions *next; /**< Next action section in file, or NULL. */
1275 * Flags for use in csp->flags
1279 * Flag for csp->flags: Set if this client is processing data.
1280 * Cleared when the thread associated with this structure dies.
1282 #define CSP_FLAG_ACTIVE 0x01U
1285 * Flag for csp->flags: Set if the server's reply is in "chunked"
1288 #define CSP_FLAG_CHUNKED 0x02U
1291 * Flag for csp->flags: Set if this request was enforced, although it would
1292 * normally have been blocked.
1294 #define CSP_FLAG_FORCED 0x04U
1297 * Flag for csp->flags: Set if any modification to the body was done.
1299 #define CSP_FLAG_MODIFIED 0x08U
1302 * Flag for csp->flags: Set if request was blocked.
1304 #define CSP_FLAG_REJECTED 0x10U
1307 * Flag for csp->flags: Set if we are toggled on (FEATURE_TOGGLE).
1309 #define CSP_FLAG_TOGGLED_ON 0x20U
1312 * Flag for csp->flags: Set if an acceptable Connection header
1315 #define CSP_FLAG_CLIENT_CONNECTION_HEADER_SET 0x00000040UL
1318 * Flag for csp->flags: Set if adding the 'Connection: close' header
1319 * for the server isn't necessary.
1321 #define CSP_FLAG_SERVER_CONNECTION_CLOSE_SET 0x00000080UL
1324 * Flag for csp->flags: Signals header parsers whether they
1325 * are parsing server or client headers.
1327 #define CSP_FLAG_CLIENT_HEADER_PARSING_DONE 0x00000100UL
1330 * Flag for csp->flags: Set if adding the Host: header
1333 #define CSP_FLAG_HOST_HEADER_IS_SET 0x00000200UL
1336 * Flag for csp->flags: Set if filtering is disabled by X-Filter: No
1337 * XXX: As we now have tags we might as well ditch this.
1339 #define CSP_FLAG_NO_FILTERING 0x00000400UL
1342 * Flag for csp->flags: Set the client IP has appended to
1343 * an already existing X-Forwarded-For header in which case
1344 * no new header has to be generated.
1346 #define CSP_FLAG_X_FORWARDED_FOR_APPENDED 0x00000800UL
1349 * Flag for csp->flags: Set if the server wants to keep
1350 * the connection alive.
1352 #define CSP_FLAG_SERVER_CONNECTION_KEEP_ALIVE 0x00001000UL
1354 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
1356 * Flag for csp->flags: Set if the server specified the
1359 #define CSP_FLAG_CONTENT_LENGTH_SET 0x00002000UL
1360 #endif /* def FEATURE_CONNECTION_KEEP_ALIVE */
1363 * Flags for use in return codes of child processes
1367 * Flag for process return code: Set if exiting porcess has been toggled
1368 * during its lifetime.
1370 #define RC_FLAG_TOGGLED 0x10
1373 * Flag for process return code: Set if exiting porcess has blocked its
1376 #define RC_FLAG_BLOCKED 0x20
1379 * Maximum number of actions/filter files. This limit is arbitrary - it's just used
1382 #define MAX_AF_FILES 10
1385 * The state of a Privoxy processing thread.
1389 /** The proxy's configuration */
1390 struct configuration_spec * config;
1392 /** The actions to perform on the current request */
1393 struct current_action_spec action[1];
1395 /** socket to talk to client (web browser) */
1398 /** socket to talk to server (web server or proxy) */
1401 /** Multi-purpose flag container, see CSP_FLAG_* above */
1404 /** Client PC's IP address, as reported by the accept() function.
1407 /** Client PC's IP address, as reported by the accept() function.
1409 unsigned long ip_addr_long;
1411 /** The URL that was requested */
1412 struct http_request http[1];
1415 * The final forwarding settings.
1416 * XXX: Currently this is only used for forward-override,
1417 * so we can free the space in sweep.
1419 struct forward_spec * fwd;
1421 /** An I/O buffer used for buffering data read from the network */
1424 /** List of all headers for this request */
1425 struct list headers[1];
1427 /** List of all tags that apply to this request */
1428 struct list tags[1];
1430 /** MIME-Type key, see CT_* above */
1431 unsigned int content_type;
1433 /** Actions files associated with this client */
1434 struct file_list *actions_list[MAX_AF_FILES];
1436 /** pcrs job files. */
1437 struct file_list *rlist[MAX_AF_FILES];
1439 /** Length after content modification. */
1440 size_t content_length;
1442 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
1443 /** Expected length of content after which we
1444 * should stop reading from the server socket.
1446 /* XXX: is this the right location? */
1447 size_t expected_content_length;
1448 #endif /* def FEATURE_CONNECTION_KEEP_ALIVE */
1450 #ifdef FEATURE_TRUST
1453 struct file_list *tlist;
1455 #endif /* def FEATURE_TRUST */
1458 * Failure reason to embedded in the CGI error page,
1459 * or NULL. Currently only used for socks errors.
1461 char *error_message;
1463 /** Next thread in linked list. Only read or modify from the main thread! */
1464 struct client_state *next;
1469 * A function to add a header
1471 typedef jb_err (*add_header_func_ptr)(struct client_state *);
1474 * A function to process a header
1476 typedef jb_err (*parser_func_ptr )(struct client_state *, char **);
1480 * List of available CGI functions.
1482 struct cgi_dispatcher
1484 /** The URL of the CGI, relative to the CGI root. */
1485 const char * const name;
1487 /** The handler function for the CGI */
1488 jb_err (* const handler)(struct client_state *csp, struct http_response *rsp, const struct map *parameters);
1490 /** The description of the CGI, to appear on the main menu, or NULL to hide it. */
1491 const char * const description;
1493 /** A flag that indicates whether unintentional calls to this CGI can cause damage */
1499 * A data file used by Privoxy. Kept in a linked list.
1504 * This is a pointer to the data structures associated with the file.
1505 * Read-only once the structure has been created.
1510 * The unloader function.
1511 * Normally NULL. When we are finished with file (i.e. when we have
1512 * loaded a new one), set to a pointer to an unloader function.
1513 * Unloader will be called by sweep() (called from main loop) when
1514 * all clients using this file are done. This prevents threading
1517 void (*unloader)(void *);
1520 * Used internally by sweep(). Do not access from elsewhere.
1525 * File last-modified time, so we can check if file has been changed.
1526 * Read-only once the structure has been created.
1528 time_t lastmodified;
1531 * The full filename.
1536 * Pointer to next entry in the linked list of all "file_list"s.
1537 * This linked list is so that sweep() can navigate it.
1538 * Since sweep() can remove items from the list, we must be careful
1539 * to only access this value from main thread (when we know sweep
1540 * won't be running).
1542 struct file_list *next;
1546 #ifdef FEATURE_TRUST
1549 * The format of a trust file when loaded into memory.
1553 struct url_spec url[1]; /**< The URL pattern */
1554 int reject; /**< FIXME: Please document this! */
1555 struct block_spec *next; /**< Next entry in linked list */
1559 * Arbitrary limit for the number of trusted referrers.
1561 #define MAX_TRUSTED_REFERRERS 512
1563 #endif /* def FEATURE_TRUST */
1566 #define SOCKS_NONE 0 /**< Don't use a SOCKS server */
1567 #define SOCKS_4 40 /**< original SOCKS 4 protocol */
1568 #define SOCKS_4A 41 /**< as modified for hosts w/o external DNS */
1569 #define SOCKS_5 50 /**< as modified for hosts w/o external DNS */
1573 * How to forward a connection to a parent proxy.
1577 /** URL pattern that this forward_spec is for. */
1578 struct url_spec url[1];
1580 /** Connection type. Must be SOCKS_NONE, SOCKS_4, SOCKS_4A or SOCKS_5. */
1583 /** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
1586 /** SOCKS server port. */
1589 /** Parent HTTP proxy hostname, or NULL for none. */
1592 /** Parent HTTP proxy port. */
1595 /** Next entry in the linked list. */
1596 struct forward_spec *next;
1601 * Initializer for a static struct forward_spec.
1603 #define FORWARD_SPEC_INITIALIZER { { URL_SPEC_INITIALIZER }, 0, NULL, 0, NULL, 0, NULL }
1605 /* Supported filter types */
1606 #define FT_CONTENT_FILTER 0
1607 #define FT_CLIENT_HEADER_FILTER 1
1608 #define FT_SERVER_HEADER_FILTER 2
1609 #define FT_CLIENT_HEADER_TAGGER 3
1610 #define FT_SERVER_HEADER_TAGGER 4
1612 #define MAX_FILTER_TYPES 5
1615 * This struct represents one filter (one block) from
1616 * the re_filterfile. If there is more than one filter
1617 * in the file, the file will be represented by a
1618 * chained list of re_filterfile specs.
1620 struct re_filterfile_spec
1622 char *name; /**< Name from FILTER: statement in re_filterfile. */
1623 char *description; /**< Description from FILTER: statement in re_filterfile. */
1624 struct list patterns[1]; /**< The patterns from the re_filterfile. */
1625 pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
1626 int type; /**< Filter type (content, client-header, server-header). */
1627 int dynamic; /**< Set to one if the pattern might contain variables
1628 and has to be recompiled for every request. */
1629 struct re_filterfile_spec *next; /**< The pointer for chaining. */
1635 #define ACL_PERMIT 1 /**< Accept connection request */
1636 #define ACL_DENY 2 /**< Reject connection request */
1639 * An IP address pattern. Used to specify networks in the ACL.
1641 struct access_control_addr
1643 unsigned long addr; /**< The IP address as an integer. */
1644 unsigned long mask; /**< The network mask as an integer. */
1645 unsigned long port; /**< The port number. */
1649 * An access control list (ACL) entry.
1651 * This is a linked list.
1653 struct access_control_list
1655 struct access_control_addr src[1]; /**< Client IP address */
1656 struct access_control_addr dst[1]; /**< Website or parent proxy IP address */
1658 short action; /**< ACL_PERMIT or ACL_DENY */
1659 struct access_control_list *next; /**< The next entry in the ACL. */
1662 #endif /* def FEATURE_ACL */
1665 /** Maximum number of loaders (actions, re_filter, ...) */
1669 /** configuration_spec::feature_flags: CGI actions editor. */
1670 #define RUNTIME_FEATURE_CGI_EDIT_ACTIONS 1U
1672 /** configuration_spec::feature_flags: Web-based toggle. */
1673 #define RUNTIME_FEATURE_CGI_TOGGLE 2U
1675 /** configuration_spec::feature_flags: HTTP-header-based toggle. */
1676 #define RUNTIME_FEATURE_HTTP_TOGGLE 4U
1678 /** configuration_spec::feature_flags: Split large forms to limit the number of GET arguments. */
1679 #define RUNTIME_FEATURE_SPLIT_LARGE_FORMS 8U
1681 /** configuration_spec::feature_flags: Check the host header for requests with host-less request lines. */
1682 #define RUNTIME_FEATURE_ACCEPT_INTERCEPTED_REQUESTS 16U
1684 /** configuration_spec::feature_flags: Don't allow to circumvent blocks with the force prefix. */
1685 #define RUNTIME_FEATURE_ENFORCE_BLOCKS 32U
1687 /** configuration_spec::feature_flags: Allow to block or redirect CGI requests. */
1688 #define RUNTIME_FEATURE_CGI_CRUNCHING 64U
1690 /** configuration_spec::feature_flags: Try to keep the connection to the server alive. */
1691 #define RUNTIME_FEATURE_CONNECTION_KEEP_ALIVE 128U
1694 * Data loaded from the configuration file.
1696 * (Anomaly: toggle is still handled through a global, not this structure)
1698 struct configuration_spec
1703 /** Nonzero to enable multithreading. */
1707 * Bitmask of features that can be enabled/disabled through the config
1708 * file. Currently defined bits:
1710 * - RUNTIME_FEATURE_CGI_EDIT_ACTIONS
1711 * - RUNTIME_FEATURE_CGI_TOGGLE
1712 * - RUNTIME_FEATURE_HTTP_TOGGLE
1713 * - RUNTIME_FEATURE_SPLIT_LARGE_FORMS
1715 unsigned feature_flags;
1717 /** The log file name. */
1718 const char *logfile;
1720 /** The config file directory. */
1721 const char *confdir;
1723 /** The directory for customized CGI templates. */
1724 const char *templdir;
1726 /** The log file directory. */
1729 /** The full paths to the actions files. */
1730 const char *actions_file[MAX_AF_FILES];
1732 /** The short names of the actions files. */
1733 const char *actions_file_short[MAX_AF_FILES];
1735 /** The administrator's email address */
1736 char *admin_address;
1738 /** A URL with info on this proxy */
1739 char *proxy_info_url;
1741 /** URL to the user manual (on our website or local copy) */
1744 /** The file names of the pcre filter files. */
1745 const char *re_filterfile[MAX_AF_FILES];
1747 /** The short names of the pcre filter files. */
1748 const char *re_filterfile_short[MAX_AF_FILES];
1750 /** The hostname to show on CGI pages, or NULL to use the real one. */
1751 const char *hostname;
1753 /** IP address to bind to. Defaults to HADDR_DEFAULT == 127.0.0.1. */
1756 /** Port to bind to. Defaults to HADDR_PORT == 8118. */
1759 /** Size limit for IOB */
1760 size_t buffer_limit;
1762 #ifdef FEATURE_TRUST
1764 /** The file name of the trust file. */
1765 const char * trustfile;
1767 /** FIXME: DOCME: Document this. */
1768 struct list trust_info[1];
1770 /** FIXME: DOCME: Document this. */
1771 struct url_spec *trust_list[MAX_TRUSTED_REFERRERS];
1773 #endif /* def FEATURE_TRUST */
1777 /** The access control list (ACL). */
1778 struct access_control_list *acl;
1780 #endif /* def FEATURE_ACL */
1782 /** Information about parent proxies (forwarding). */
1783 struct forward_spec *forward;
1785 /** Number of retries in case a forwarded connection attempt fails */
1786 int forwarded_connect_retries;
1788 /** All options from the config file, HTML-formatted. */
1791 /** The configuration file object. */
1792 struct file_list *config_file_list;
1794 /** List of loaders */
1795 int (*loaders[NLOADERS])(struct client_state *);
1797 /** Nonzero if we need to bind() to the new port. */
1801 /** Calculates the number of elements in an array, using sizeof. */
1802 #define SZ(X) (sizeof(X) / sizeof(*X))
1804 #ifdef FEATURE_FORCE_LOAD
1805 /** The force load URL prefix. */
1806 #define FORCE_PREFIX "/PRIVOXY-FORCE"
1807 #endif /* def FEATURE_FORCE_LOAD */
1809 #ifdef FEATURE_NO_GIFS
1810 /** The MIME type for images ("image/png" or "image/gif"). */
1811 #define BUILTIN_IMAGE_MIMETYPE "image/png"
1813 #define BUILTIN_IMAGE_MIMETYPE "image/gif"
1814 #endif /* def FEATURE_NO_GIFS */
1821 /** URL for the Privoxy home page. */
1822 #define HOME_PAGE_URL "http://www.privoxy.org/"
1824 /** URL for the Privoxy user manual. */
1825 #define USER_MANUAL_URL HOME_PAGE_URL VERSION "/user-manual/"
1827 /** Prefix for actions help links (append to USER_MANUAL_URL). */
1828 #define ACTIONS_HELP_PREFIX "actions-file.html#"
1830 /** Prefix for config option help links (append to USER_MANUAL_URL). */
1831 #define CONFIG_HELP_PREFIX "config.html#"
1834 * The "hosts" to intercept and display CGI pages.
1835 * First one is a hostname only, second one can specify host and path.
1838 * 1) Do not specify the http: prefix
1839 * 2) CGI_SITE_2_PATH must not end with /, one will be added automatically.
1840 * 3) CGI_SITE_2_PATH must start with /, unless it is the empty string.
1842 #define CGI_SITE_1_HOST "p.p"
1843 #define CGI_SITE_2_HOST "config.privoxy.org"
1844 #define CGI_SITE_2_PATH ""
1847 * The prefix for CGI pages. Written out in generated HTML.
1848 * INCLUDES the trailing slash.
1850 #define CGI_PREFIX "http://" CGI_SITE_2_HOST CGI_SITE_2_PATH "/"
1856 #endif /* ndef PROJECT_H_INCLUDED */