1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
4 #define PROJECT_H_VERSION "$Id: project.h,v 1.72.2.2 2002/11/28 18:15:44 oes Exp $"
5 /*********************************************************************
7 * File : $Source: /cvsroot/ijbswa/current/Attic/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 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 1.72.2.2 2002/11/28 18:15:44 oes
41 * Added flag to each cgi_dispatcher that allows or denies
42 * external linking and removed const qualifier from
43 * struct list_entry.str.
45 * Revision 1.72.2.1 2002/08/10 11:25:18 oes
46 * - Include config.h for access to config data
47 * - Include <pcre*.h> depending on where they are
49 * Revision 1.72 2002/05/14 21:35:49 oes
50 * Split HELP_LINK_PREFIX into ACTIONS_HELP_PREFIX and CONFIG_HELP_PREFIX
51 * because of split in user-manual
53 * Revision 1.71 2002/05/12 21:39:36 jongfoster
54 * - Adding Doxygen-style comments to structures and #defines.
56 * Revision 1.70 2002/05/12 16:05:50 jongfoster
57 * Fixing ACTION_MASK_ALL to be unsigned long rather than
58 * just unsigned int. I don't know if anyone is porting
59 * Privoxy to 16-bit platforms, but if so, +limit-connect
60 * wouldn't have worked because of this bug.
62 * Revision 1.69 2002/05/08 16:00:16 oes
63 * Added size member to struct iob, so it can
64 * be alloced larger than needed.
66 * Revision 1.68 2002/04/26 12:56:00 oes
67 * Killed REDIRECT_URL, added USER_MANUAL_URL and HELP_LINK_PREFIX
69 * Revision 1.67 2002/04/24 02:12:43 oes
70 * - Jon's multiple AF patch:
71 * - Make csp->actions_list an array
72 * - #define MAX_ACTION_FILES
73 * - Moved CGI_PARAM_LEN_MAX (500) here
75 * Revision 1.66 2002/04/15 19:06:43 jongfoster
78 * Revision 1.65 2002/04/04 00:36:36 gliptak
79 * always use pcre for matching
81 * Revision 1.64 2002/04/03 22:28:03 gliptak
82 * Removed references to gnu_regex
84 * Revision 1.63 2002/03/31 17:19:00 jongfoster
85 * Win32 only: Enabling STRICT to fix a VC++ compile warning.
87 * Revision 1.62 2002/03/26 22:48:49 swa
90 * Revision 1.61 2002/03/26 22:29:55 swa
91 * we have a new homepage!
93 * Revision 1.60 2002/03/24 15:52:17 jongfoster
94 * Changing CGI URL prefixes for new name
96 * Revision 1.59 2002/03/24 15:23:33 jongfoster
99 * Revision 1.58 2002/03/24 13:25:43 swa
100 * name change related issues
102 * Revision 1.57 2002/03/16 20:28:34 oes
103 * Added descriptions to the filters so users will know what they select in the cgi editor
105 * Revision 1.56 2002/03/13 20:27:30 oes
106 * Fixing bug with CT_TABOO
108 * Revision 1.55 2002/03/12 01:42:50 oes
109 * Introduced modular filters
111 * Revision 1.54 2002/03/09 20:03:52 jongfoster
112 * - Making various functions return int rather than size_t.
113 * (Undoing a recent change). Since size_t is unsigned on
114 * Windows, functions like read_socket that return -1 on
115 * error cannot return a size_t.
117 * THIS WAS A MAJOR BUG - it caused frequent, unpredictable
118 * crashes, and also frequently caused JB to jump to 100%
119 * CPU and stay there. (Because it thought it had just
120 * read ((unsigned)-1) == 4Gb of data...)
122 * - The signature of write_socket has changed, it now simply
123 * returns success=0/failure=nonzero.
125 * - Trying to get rid of a few warnings --with-debug on
126 * Windows, I've introduced a new type "jb_socket". This is
127 * used for the socket file descriptors. On Windows, this
128 * is SOCKET (a typedef for unsigned). Everywhere else, it's
129 * an int. The error value can't be -1 any more, so it's
130 * now JB_INVALID_SOCKET (which is -1 on UNIX, and in
131 * Windows it maps to the #define INVALID_SOCKET.)
133 * - The signature of bind_port has changed.
135 * Revision 1.53 2002/03/08 16:48:55 oes
136 * Added FEATURE_NO_GIFS and BUILTIN_IMAGE_MIMETYPE
138 * Revision 1.52 2002/03/07 03:46:17 oes
139 * Fixed compiler warnings
141 * Revision 1.51 2002/03/05 04:52:42 oes
142 * Deleted non-errlog debugging code
144 * Revision 1.50 2002/03/04 19:32:07 oes
145 * Changed default port to 8118
147 * Revision 1.49 2002/03/04 18:28:55 oes
148 * Deleted PID_FILE_NAME
150 * Revision 1.48 2002/03/03 14:50:40 oes
151 * Fixed CLF logging: Added ocmd member for client's request to struct http_request
153 * Revision 1.47 2002/02/20 23:15:13 jongfoster
154 * Parsing functions now handle out-of-memory gracefully by returning
157 * Revision 1.46 2002/01/17 21:06:09 jongfoster
158 * Now #defining the URLs of the config interface
160 * Minor changes to struct http_request and struct url_spec due to
161 * standardizing that struct http_request is used to represent a URL, and
162 * struct url_spec is used to represent a URL pattern. (Before, URLs were
163 * represented as seperate variables and a partially-filled-in url_spec).
165 * Revision 1.45 2002/01/09 14:33:27 oes
166 * Added HOSTENT_BUFFER_SIZE
168 * Revision 1.44 2001/12/30 14:07:32 steudten
169 * - Add signal handling (unix)
170 * - Add SIGHUP handler (unix)
171 * - Add creation of pidfile (unix)
172 * - Add action 'top' in rc file (RH)
173 * - Add entry 'SIGNALS' to manpage
174 * - Add exit message to logfile (unix)
176 * Revision 1.43 2001/11/22 21:57:51 jongfoster
177 * Making action_spec->flags into an unsigned long rather than just an
179 * Adding ACTION_NO_COOKIE_KEEP
181 * Revision 1.42 2001/11/05 21:42:41 steudten
182 * Include DBG() macro.
184 * Revision 1.41 2001/10/28 19:12:06 jongfoster
185 * Adding ijb_toupper()
187 * Revision 1.40 2001/10/26 17:40:47 oes
188 * Moved ijb_isspace and ijb_tolower to project.h
189 * Removed http->user_agent, csp->referrer and csp->accept_types
191 * Revision 1.39 2001/10/25 03:45:02 david__schmidt
192 * Adding a (void*) cast to freez() because Visual Age C++ won't expand the
193 * macro when called with a cast; so moving the cast to the macro def'n
194 * seems to both eliminate compiler warnings (on darwin and OS/2, anyway) and
195 * doesn't make macro expansion complain. Hope this works for everyone else
198 * Revision 1.38 2001/10/23 21:19:04 jongfoster
199 * New error-handling support: jb_err type and JB_ERR_xxx constants
200 * CGI functions now return a jb_err, and their parameters map is const.
201 * Support for RUNTIME_FEATUREs to enable/disable config editor
202 * Adding a few comments
204 * Revision 1.37 2001/10/14 22:14:01 jongfoster
205 * Removing name_length field from struct cgi_dispatcher, as this is
206 * now calculated at runtime from the "name" field.
208 * Revision 1.36 2001/10/10 16:45:15 oes
209 * Added LIMIT_CONNECT action and string
210 * Fixed HTTP message line termination
211 * Added CFORBIDDEN HTTP message
213 * Revision 1.35 2001/10/07 18:06:43 oes
214 * Added status member to struct http_request
216 * Revision 1.34 2001/10/07 15:45:25 oes
217 * Added url member to struct http_request and commented all
222 * Added ACTION_DOWNGRADE and ACTION_NO_COMPRESSION
224 * Replaced struct client_state members rejected,
225 * force, active and toggled_on with "flags" bitmap.
227 * Added CSP_FLAG_MODIFIED and CSP_FLAG_CHUNKED
229 * Added buffer_limit to struct configuration_spec
231 * Revision 1.33 2001/09/20 13:30:08 steudten
233 * Make freez() more secure in case of: if (exp) { free(z) ; a=*z }
234 * Last case will set z to NULL in free(z) and thats bad..
236 * Revision 1.32 2001/09/16 23:02:51 jongfoster
239 * Revision 1.31 2001/09/16 13:20:29 jongfoster
240 * Rewrite of list library. Now has seperate header and list_entry
241 * structures. Also added a large sprinking of assert()s to the list
244 * Revision 1.30 2001/09/13 23:52:00 jongfoster
245 * Support for both static and dynamically generated CGI pages
247 * Revision 1.29 2001/09/13 23:29:43 jongfoster
248 * Defining FORWARD_SPEC_INITIALIZER
250 * Revision 1.28 2001/09/13 23:05:50 jongfoster
251 * Changing the string paramater to the header parsers a "const".
253 * Revision 1.27 2001/08/05 16:06:20 jongfoster
254 * Modifiying "struct map" so that there are now separate header and
255 * "map_entry" structures. This means that functions which modify a
256 * map no longer need to return a pointer to the modified map.
257 * Also, it no longer reverses the order of the entries (which may be
258 * important with some advanced template substitutions).
260 * Revision 1.26 2001/07/30 22:08:36 jongfoster
261 * Tidying up #defines:
262 * - All feature #defines are now of the form FEATURE_xxx
263 * - Permanently turned off WIN_GUI_EDIT
264 * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
266 * Revision 1.25 2001/07/29 18:43:08 jongfoster
267 * Changing #ifdef _FILENAME_H to FILENAME_H_INCLUDED, to conform to
270 * Revision 1.24 2001/07/25 17:20:27 oes
271 * Introduced http->user_agent
273 * Revision 1.23 2001/07/18 12:32:23 oes
274 * - Added ACTION_STRING_DEANIMATE
275 * - moved #define freez from jcc.h to project.h
277 * Revision 1.22 2001/07/15 17:51:41 jongfoster
278 * Renaming #define STATIC to STATIC_PCRE
280 * Revision 1.21 2001/07/13 14:03:19 oes
281 * - Reorganized regex header inclusion and #defines to
282 * comply to the scheme in configure.in
283 * - Added csp->content_type and its CT_* keys
284 * - Added ACTION_DEANIMATE
285 * - Removed all #ifdef PCRS
287 * Revision 1.20 2001/06/29 21:45:41 oes
288 * Indentation, CRLF->LF, Tab-> Space
290 * Revision 1.19 2001/06/29 13:33:36 oes
291 * - Improved comments
292 * - Introduced http_request.host_ip_addr_str
293 * - Introduced http_response.head_length
294 * - Introduced config.my_ip_addr_str, config.my_hostname,
295 * config.admin_address and config.proxy_info_url
296 * - Removed config.proxy_args_header and config.proxy_args_trailer,
297 * renamed config.proxy_args_invocation to config.proxy_args
298 * - Removed HTML snipplets and GIFs
299 * - Removed logentry from cancelled commit
301 * Revision 1.18 2001/06/09 10:57:39 jongfoster
302 * Adding definition of BUFFER_SIZE.
303 * Changing struct cgi_dispatcher to use "const" strings.
305 * Revision 1.17 2001/06/07 23:15:09 jongfoster
306 * Merging ACL and forward files into config file.
307 * Moving struct gateway members into struct forward_spec
308 * Removing config->proxy_args_gateways
309 * Cosmetic: Adding a few comments
311 * Revision 1.16 2001/06/04 18:31:58 swa
312 * files are now prefixed with either `confdir' or `logdir'.
313 * `make redhat-dist' replaces both entries confdir and logdir
316 * Revision 1.15 2001/06/04 11:28:53 swa
317 * redirect did not work due to missing /
319 * Revision 1.14 2001/06/03 11:03:48 oes
321 * added struct http_response,
322 * changed struct interceptors to struct cgi_dispatcher,
323 * moved HTML stuff to cgi.h
325 * Revision 1.13 2001/06/01 20:05:36 jongfoster
326 * Support for +image-blocker{}: added ACTION_IMAGE_BLOCKER
327 * constant, and removed csp->tinygif.
329 * Revision 1.12 2001/06/01 18:49:17 jongfoster
330 * Replaced "list_share" with "list" - the tiny memory gain was not
331 * worth the extra complexity.
333 * Revision 1.11 2001/06/01 10:32:47 oes
334 * Added constants for anchoring selection bitmap
336 * Revision 1.10 2001/05/31 21:33:53 jongfoster
337 * Changes for new actions file, replacing permissionsfile
338 * and parts of the config file. Also added support for
341 * Revision 1.9 2001/05/31 17:32:31 oes
343 * - Enhanced domain part globbing with infix and prefix asterisk
344 * matching and optional unanchored operation
346 * Revision 1.8 2001/05/29 20:09:15 joergs
347 * HTTP_REDIRECT_TEMPLATE fixed.
349 * Revision 1.7 2001/05/29 09:50:24 jongfoster
350 * Unified blocklist/imagelist/actionslist.
351 * File format is still under discussion, but the internal changes
354 * Also modified interceptor behaviour:
355 * - We now intercept all URLs beginning with one of the following
356 * prefixes (and *only* these prefixes):
358 * * http://ijbswa.sf.net/config/
359 * * http://ijbswa.sourceforge.net/config/
360 * - New interceptors "home page" - go to http://i.j.b/ to see it.
361 * - Internal changes so that intercepted and fast redirect pages
362 * are not replaced with an image.
363 * - Interceptors now have the option to send a binary page direct
364 * to the client. (i.e. ijb-send-banner uses this)
365 * - Implemented show-url-info interceptor. (Which is why I needed
366 * the above interceptors changes - a typical URL is
367 * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif".
368 * The previous mechanism would not have intercepted that, and
369 * if it had been intercepted then it then it would have replaced
372 * Revision 1.6 2001/05/27 22:17:04 oes
374 * - re_process_buffer no longer writes the modified buffer
375 * to the client, which was very ugly. It now returns the
376 * buffer, which it is then written by chat.
378 * - content_length now adjusts the Content-Length: header
379 * for modified documents rather than crunch()ing it.
380 * (Length info in csp->content_length, which is 0 for
381 * unmodified documents)
383 * - For this to work, sed() is called twice when filtering.
385 * Revision 1.5 2001/05/26 00:28:36 jongfoster
386 * Automatic reloading of config file.
387 * Removed obsolete SIGHUP support (Unix) and Reload menu option (Win32).
388 * Most of the global variables have been moved to a new
389 * struct configuration_spec, accessed through csp->config->globalname
390 * Most of the globals remaining are used by the Win32 GUI.
392 * Revision 1.4 2001/05/22 18:46:04 oes
394 * - Enabled filtering banners by size rather than URL
395 * by adding patterns that replace all standard banner
396 * sizes with the "Junkbuster" gif to the re_filterfile
398 * - Enabled filtering WebBugs by providing a pattern
399 * which kills all 1x1 images
401 * - Added support for PCRE_UNGREEDY behaviour to pcrs,
402 * which is selected by the (nonstandard and therefore
403 * capital) letter 'U' in the option string.
404 * It causes the quantifiers to be ungreedy by default.
405 * Appending a ? turns back to greedy (!).
407 * - Added a new interceptor ijb-send-banner, which
408 * sends back the "Junkbuster" gif. Without imagelist or
409 * MSIE detection support, or if tinygif = 1, or the
410 * URL isn't recognized as an imageurl, a lame HTML
411 * explanation is sent instead.
413 * - Added new feature, which permits blocking remote
414 * script redirects and firing back a local redirect
416 * The feature is conditionally compiled, i.e. it
417 * can be disabled with --disable-fast-redirects,
418 * plus it must be activated by a "fast-redirects"
419 * line in the config file, has its own log level
420 * and of course wants to be displayed by show-proxy-args
421 * Note: Boy, all the #ifdefs in 1001 locations and
422 * all the fumbling with configure.in and acconfig.h
423 * were *way* more work than the feature itself :-(
425 * - Because a generic redirect template was needed for
426 * this, tinygif = 3 now uses the same.
428 * - Moved GIFs, and other static HTTP response templates
433 * - Removed some >400 CRs again (Jon, you really worked
436 * Revision 1.3 2001/05/20 01:21:20 jongfoster
437 * Version 2.9.4 checkin.
438 * - Merged popupfile and cookiefile, and added control over PCRS
439 * filtering, in new "actionsfile".
440 * - Implemented LOG_LEVEL_FATAL, so that if there is a configuration
441 * file error you now get a message box (in the Win32 GUI) rather
442 * than the program exiting with no explanation.
443 * - Made killpopup use the PCRS MIME-type checking and HTTP-header
445 * - Removed tabs from "config"
446 * - Moved duplicated url parsing code in "loaders.c" to a new funcition.
447 * - Bumped up version number.
449 * Revision 1.2 2001/05/17 23:01:01 oes
450 * - Cleaned CRLF's from the sources and related files
452 * Revision 1.1.1.1 2001/05/15 13:59:03 oes
453 * Initial import of version 2.9.3 source tree
456 *********************************************************************/
459 /* Declare struct FILE for vars and funcs. */
462 /* Need time_t for file_list */
464 /* Needed for pcre choice */
468 * Include appropriate regular expression libraries.
469 * Note that pcrs and pcre (native) are needed for cgi
470 * and are included anyway.
476 # ifdef PCRE_H_IN_SUBDIR
477 # include <pcre/pcre.h>
490 # include "pcreposix.h"
492 # ifdef PCRE_H_IN_SUBDIR
493 # include <pcre/pcreposix.h>
495 # include <pcreposix.h>
501 #endif /* def AMIGA */
505 * I don't want to have to #include all this just for the declaration
506 * of SOCKET. However, it looks like we have to...
521 typedef SOCKET jb_socket;
523 #define JB_INVALID_SOCKET INVALID_SOCKET
525 #else /* ndef _WIN32 */
528 * The type used by sockets. On UNIX it's an int. Microsoft decided to
529 * make it an unsigned.
531 typedef int jb_socket;
534 * The error value used for variables of type jb_socket. On UNIX this
535 * is -1, however Microsoft decided to make socket handles unsigned, so
536 * they use a different value.
539 #define JB_INVALID_SOCKET (-1)
541 #endif /* ndef _WIN32 */
545 * A standard error code. This should be JB_ERR_OK or one of the JB_ERR_xxx
550 #define JB_ERR_OK 0 /**< Success, no error */
551 #define JB_ERR_MEMORY 1 /**< Out of memory */
552 #define JB_ERR_CGI_PARAMS 2 /**< Missing or corrupt CGI parameters */
553 #define JB_ERR_FILE 3 /**< Error opening, reading or writing a file */
554 #define JB_ERR_PARSE 4 /**< Error parsing file */
555 #define JB_ERR_MODIFIED 5 /**< File has been modified outside of the
556 CGI actions editor. */
560 * This macro is used to free a pointer that may be NULL.
561 * It also sets the variable to NULL after it's been freed.
562 * The paramater should be a simple variable without side effects.
564 #define freez(X) { if(X) { free((void*)X); X = NULL ; } }
568 * Fix a problem with Solaris. There should be no effect on other
571 * Solaris's isspace() is a macro which uses it's argument directly
572 * as an array index. Therefore we need to make sure that high-bit
573 * characters generate +ve values, and ideally we also want to make
574 * the argument match the declared parameter type of "int".
576 * Note: Remember to #include <ctype.h> if you use these macros.
578 #define ijb_toupper(__X) toupper((int)(unsigned char)(__X))
579 #define ijb_tolower(__X) tolower((int)(unsigned char)(__X))
580 #define ijb_isspace(__X) isspace((int)(unsigned char)(__X))
583 * Use for statically allocated buffers if you have no other choice.
584 * Remember to check the length of what you write into the buffer
585 * - we don't want any buffer overflows!
587 #define BUFFER_SIZE 5000
590 * Max length of CGI parameters (arbitrary limit).
592 #define CGI_PARAM_LEN_MAX 500
595 * Buffer size for capturing struct hostent data in the
596 * gethostby(name|addr)_r library calls. Since we don't
597 * loop over gethostbyname_r, the buffer must be sufficient
598 * to accomodate multiple IN A RRs, as used in DNS round robin
599 * load balancing. W3C's wwwlib uses 1K, so that should be
600 * good enough for us, too.
602 #define HOSTENT_BUFFER_SIZE 1024
605 * Do not use. Originally this was so that you can
606 * say "while (FOREVER) { ...do something... }".
607 * However, this gives a warning with some compilers (e.g. VC++).
608 * Instead, use "for (;;) { ...do something... }".
613 * Default IP address to listen on, as a string.
614 * Set to "127.0.0.1".
616 #define HADDR_DEFAULT "127.0.0.1"
619 * Default port to listen on, as a number.
622 #define HADDR_PORT 8118
625 /* Forward def for struct client_state */
626 struct configuration_spec;
630 * Entry in a linked list of strings.
635 * The string pointer. It must point to a dynamically malloc()ed
636 * string or be NULL for the list functions to work. In the latter
637 * case, just be careful next time you iterate through the list in
642 /** Next entry in the linked list, or NULL if no more. */
643 struct list_entry *next;
647 * A header for a linked list of strings.
651 /** First entry in the list, or NULL if the list is empty. */
652 struct list_entry *first;
654 /** Last entry in the list, or NULL if the list is empty. */
655 struct list_entry *last;
660 * An entry in a map. This is a name=value pair.
664 /** The key for the map. */
666 /** The value associated with that key. */
668 /** The next map entry, or NULL if none. */
669 struct map_entry *next;
673 * A map from a string to another string.
674 * This is used for the paramaters passed in a HTTP GET request, and
675 * to store the exports when the CGI interface is filling in a template.
679 /** The first map entry, or NULL if the map is empty. */
680 struct map_entry *first;
681 /** The last map entry, or NULL if the map is empty. */
682 struct map_entry *last;
687 * A HTTP request. This includes the method (GET, POST) and
690 * This is also used whenever we want to match a URL against a
691 * URL pattern. This always contains the URL to match, and never
692 * a URL pattern. (See struct url_spec).
696 char *cmd; /**< Whole command line: method, URL, Version */
697 char *ocmd; /**< Backup of original cmd for CLF logging */
698 char *gpc; /**< HTTP method: GET, POST, ... */
699 char *url; /**< The URL */
700 char *ver; /**< Protocol version */
701 int status; /**< HTTP Status */
703 char *host; /**< Host part of URL */
704 int port; /**< Port of URL or 80 (default) */
705 char *path; /**< Path of URL */
706 char *hostport; /**< host[:port] */
707 int ssl; /**< Flag if protocol is https */
709 char *host_ip_addr_str; /**< String with dotted decimal representation
710 of host's IP. NULL before connect_to() */
712 char *dbuffer; /**< Buffer with '\0'-delimited domain name. */
713 char **dvec; /**< List of pointers to the strings in dbuffer. */
714 int dcount; /**< How many parts to this domain? (length of dvec) */
719 * Response generated by CGI, blocker, or error handler
723 char *status; /**< HTTP status (string). */
724 struct list headers[1]; /**< List of header lines. */
725 char *head; /**< Formatted http response head. */
726 size_t head_length; /**< Length of http response head. */
727 char *body; /**< HTTP document body. */
728 size_t content_length; /**< Length of body, REQUIRED if binary body. */
729 int is_static; /**< Nonzero if the content will never change and
730 should be cached by the browser (e.g. images). */
738 /** The string which was parsed to produce this url_spec.
739 Used for debugging or display only. */
742 char *dbuffer; /**< Buffer with '\0'-delimited domain name, or NULL to match all hosts. */
743 char **dvec; /**< List of pointers to the strings in dbuffer. */
744 int dcount; /**< How many parts to this domain? (length of dvec) */
745 int unanchored; /**< Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT. */
747 int port; /**< The port number, or 0 to match all ports. */
749 char *path; /**< The source for the regex. */
750 int pathlen; /**< ==strlen(path). Needed for prefix matching. FIXME: Now obsolete? */
751 regex_t *preg; /**< Regex for matching path part */
755 * If you declare a static url_spec, this is the value to initialize it to zero.
757 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, 0, 0, 0, NULL, 0, NULL }
760 * Constant for host part matching in URLs. If set, indicates that the start of
761 * the pattern must match the start of the URL. E.g. this is not set for the
762 * pattern ".example.com", so that it will match both "example.com" and
763 * "www.example.com". It is set for the pattern "example.com", which makes it
764 * match "example.com" only, not "www.example.com".
766 #define ANCHOR_LEFT 1
769 * Constant for host part matching in URLs. If set, indicates that the end of
770 * the pattern must match the end of the URL. E.g. this is not set for the
771 * pattern "ad.", so that it will match any host called "ad", irrespective
772 * of how many subdomains are in the fully-qualified domain name.
774 #define ANCHOR_RIGHT 2
778 * An I/O buffer. Holds a string which can be appended to, and can have data
779 * removed from the beginning.
783 char *buf; /**< Start of buffer */
784 char *cur; /**< Start of relevant data */
785 char *eod; /**< End of relevant data */
786 size_t size; /**< Size as malloc()ed */
791 * Return the number of bytes in the I/O buffer associated with the passed
792 * client_state pointer.
795 #define IOB_PEEK(CSP) ((CSP->iob->cur > CSP->iob->eod) ? (CSP->iob->eod - CSP->iob->cur) : 0)
799 * Remove any data in the I/O buffer associated with the passed
800 * client_state pointer.
802 #define IOB_RESET(CSP) if(CSP->iob->buf) free(CSP->iob->buf); memset(CSP->iob, '\0', sizeof(CSP->iob));
804 /* Bits for csp->content_type */
805 #define CT_TEXT 1 /**< csp->content_type bitmask:
806 Suitable for pcrs filtering. */
807 #define CT_GIF 2 /**< csp->content_type bitmask:
808 Suitable for GIF filtering. */
809 #define CT_TABOO 4 /**< csp->content_type bitmask:
810 DO NOT filter, irrespective of other flags. */
813 * The mask which includes all actions.
815 #define ACTION_MASK_ALL (~0UL)
818 * The most compatible set of actions - i.e. none.
820 #define ACTION_MOST_COMPATIBLE 0x00000000UL
822 /** Action bitmap: Block the request. */
823 #define ACTION_BLOCK 0x00000001UL
824 /** Action bitmap: Deanimate if it's a GIF. */
825 #define ACTION_DEANIMATE 0x00000002UL
826 /** Action bitmap: Downgrade HTTP/1.1 to 1.0. */
827 #define ACTION_DOWNGRADE 0x00000004UL
828 /** Action bitmap: Fast redirects. */
829 #define ACTION_FAST_REDIRECTS 0x00000008UL
830 /** Action bitmap: Remove existing "Forwarded" header, and do not add another. */
831 #define ACTION_HIDE_FORWARDED 0x00000010UL
832 /** Action bitmap: Hide "From" header. */
833 #define ACTION_HIDE_FROM 0x00000020UL
834 /** Action bitmap: Hide "Referer" header. (sic - follow HTTP, not English). */
835 #define ACTION_HIDE_REFERER 0x00000040UL
836 /** Action bitmap: Hide "User-Agent" and similar headers. */
837 #define ACTION_HIDE_USER_AGENT 0x00000080UL
838 /** Action bitmap: This is an image. */
839 #define ACTION_IMAGE 0x00000100UL
840 /** Action bitmap: Sets the image blocker. */
841 #define ACTION_IMAGE_BLOCKER 0x00000200UL
842 /** Action bitmap: Prevent compression. */
843 #define ACTION_NO_COMPRESSION 0x00000400UL
844 /** Action bitmap: Change cookies to session only cookies. */
845 #define ACTION_NO_COOKIE_KEEP 0x00000800UL
846 /** Action bitmap: Block rending cookies. */
847 #define ACTION_NO_COOKIE_READ 0x00001000UL
848 /** Action bitmap: Block setting cookies. */
849 #define ACTION_NO_COOKIE_SET 0x00002000UL
850 /** Action bitmap: Filter out popups. */
851 #define ACTION_NO_POPUPS 0x00004000UL
852 /** Action bitmap: Send a vanilla wafer. */
853 #define ACTION_VANILLA_WAFER 0x00008000UL
854 /** Action bitmap: Limit CONNECT requests to safe ports. */
855 #define ACTION_LIMIT_CONNECT 0x00010000UL
857 /** Action string index: How to deanimate GIFs */
858 #define ACTION_STRING_DEANIMATE 0
859 /** Action string index: Replacement for "From:" header */
860 #define ACTION_STRING_FROM 1
861 /** Action string index: How to block images */
862 #define ACTION_STRING_IMAGE_BLOCKER 2
863 /** Action string index: Replacement for "Referer:" header */
864 #define ACTION_STRING_REFERER 3
865 /** Action string index: Replacement for "User-Agent:" header */
866 #define ACTION_STRING_USER_AGENT 4
867 /** Action string index: Legal CONNECT ports. */
868 #define ACTION_STRING_LIMIT_CONNECT 5
869 /** Number of string actions. */
870 #define ACTION_STRING_COUNT 6
872 /** Index into current_action_spec::multi[] for headers to add. */
873 #define ACTION_MULTI_ADD_HEADER 0
874 /** Index into current_action_spec::multi[] for headers to add. */
875 #define ACTION_MULTI_WAFER 1
876 /** Index into current_action_spec::multi[] for filters to apply. */
877 #define ACTION_MULTI_FILTER 2
878 /** Number of multi-string actions. */
879 #define ACTION_MULTI_COUNT 3
883 * This structure contains a list of actions to apply to a URL.
884 * It only contains positive instructions - no "-" options.
885 * It is not used to store the actions list itself, only for
886 * url_actions() to return the current values.
888 struct current_action_spec
890 /** Actions to apply. A bit set to "1" means perform the action. */
894 * Paramaters for those actions that require them.
895 * Each entry is valid if & only if the corresponding entry in "flags" is
898 char * string[ACTION_STRING_COUNT];
900 /** Lists of strings for multi-string actions. */
901 struct list multi[ACTION_MULTI_COUNT][1];
906 * This structure contains a set of changes to actions.
907 * It can contain both positive and negative instructions.
908 * It is used to store an entry in the actions list.
912 unsigned long mask; /**< Actions to keep. A bit set to "0" means remove action. */
913 unsigned long add; /**< Actions to add. A bit set to "1" means add action. */
916 * Paramaters for those actions that require them.
917 * Each entry is valid if & only if the corresponding entry in "flags" is
920 char * string[ACTION_STRING_COUNT];
922 /** Lists of strings to remove, for multi-string actions. */
923 struct list multi_remove[ACTION_MULTI_COUNT][1];
925 /** If nonzero, remove *all* strings from the multi-string action. */
926 int multi_remove_all[ACTION_MULTI_COUNT];
928 /** Lists of strings to add, for multi-string actions. */
929 struct list multi_add[ACTION_MULTI_COUNT][1];
934 * This structure is used to store the actions list.
936 * It contains a URL pattern, and the chages to the actions.
937 * It is a linked list.
941 struct url_spec url[1]; /**< URL pattern. */
943 struct action_spec action[1]; /**< Actions. */
945 struct url_actions * next; /**< Next action in file, or NULL. */
950 * Flags for use in csp->flags
954 * Flag for csp->flags: Set if this client is processing data.
955 * Cleared when the thread associated with this structure dies.
957 #define CSP_FLAG_ACTIVE 0x01
960 * Flag for csp->flags: Set if the server's reply is in "chunked"
963 #define CSP_FLAG_CHUNKED 0x02
966 * Flag for csp->flags: Set if this request was enforced, although it would
967 * normally have been blocked.
969 #define CSP_FLAG_FORCED 0x04
972 * Flag for csp->flags: Set if any modification to the body was done.
974 #define CSP_FLAG_MODIFIED 0x08
977 * Flag for csp->flags: Set if request was blocked.
979 #define CSP_FLAG_REJECTED 0x10
982 * Flag for csp->flags: Set if we are toggled on (FEATURE_TOGGLE).
984 #define CSP_FLAG_TOGGLED_ON 0x20
988 * Flags for use in return codes of child processes
992 * Flag for process return code: Set if exiting porcess has been toggled
993 * during its lifetime.
995 #define RC_FLAG_TOGGLED 0x10
998 * Flag for process return code: Set if exiting porcess has blocked its
1001 #define RC_FLAG_BLOCKED 0x20
1004 * Maximum number of actions files. This limit is arbitrary - it's just used
1007 #define MAX_ACTION_FILES 10
1010 * The state of a Privoxy processing thread.
1014 /** The proxy's configuration */
1015 struct configuration_spec * config;
1017 /** The actions to perform on the current request */
1018 struct current_action_spec action[1];
1020 /** socket to talk to client (web browser) */
1023 /** socket to talk to server (web server or proxy) */
1026 /** Multi-purpose flag container, see CSP_FLAG_* above */
1027 unsigned short int flags;
1029 /** Client PC's IP address, as reported by the accept() function.
1032 /** Client PC's IP address, as reported by the accept() function.
1036 /** Our IP address. I.e. the IP address that the client used to reach us,
1038 char *my_ip_addr_str;
1040 /** Our hostname. I.e. the reverse DNS of the IP address that the client
1041 used to reach us, as a string. */
1044 /** The URL that was requested */
1045 struct http_request http[1];
1047 /** An I/O buffer used for buffering data read from the network */
1050 /** List of all headers for this request */
1051 struct list headers[1];
1053 /** List of all cookies for this request */
1054 struct list cookie_list[1];
1056 /** MIME-Type key, see CT_* above */
1057 unsigned short int content_type;
1059 /** The "X-Forwarded-For:" header sent by the client */
1062 /** Actions files associated with this client */
1063 struct file_list *actions_list[MAX_ACTION_FILES];
1065 /** pcrs job file. */
1066 struct file_list *rlist;
1068 /** Length after content modification. */
1069 size_t content_length;
1071 #ifdef FEATURE_TRUST
1074 struct file_list *tlist;
1076 #endif /* def FEATURE_TRUST */
1078 /** Next thread in linked list. Only read or modify from the main thread! */
1079 struct client_state *next;
1084 * A function to add a header
1086 typedef jb_err (*add_header_func_ptr)(struct client_state *);
1089 * A function to process a header
1091 typedef jb_err (*parser_func_ptr )(struct client_state *, char **);
1095 * List of functions to run on a list of headers
1099 /** The header prefix to match */
1102 /** The length of the prefix to match */
1105 /** The function to apply to this line */
1106 parser_func_ptr parser;
1111 * List of available CGI functions.
1113 struct cgi_dispatcher
1115 /** The URL of the CGI, relative to the CGI root. */
1116 const char * const name;
1118 /** The handler function for the CGI */
1119 jb_err (* const handler)(struct client_state *csp, struct http_response *rsp, const struct map *parameters);
1121 /** The description of the CGI, to appear on the main menu, or NULL to hide it. */
1122 const char * const description;
1124 /** A flag that indicates whether unintentional calls to this CGI can cause damage */
1130 * A data file used by Privoxy. Kept in a linked list.
1135 * This is a pointer to the data structures associated with the file.
1136 * Read-only once the structure has been created.
1141 * The unloader function.
1142 * Normally NULL. When we are finished with file (i.e. when we have
1143 * loaded a new one), set to a pointer to an unloader function.
1144 * Unloader will be called by sweep() (called from main loop) when
1145 * all clients using this file are done. This prevents threading
1148 void (*unloader)(void *);
1151 * Used internally by sweep(). Do not access from elsewhere.
1156 * File last-modified time, so we can check if file has been changed.
1157 * Read-only once the structure has been created.
1159 time_t lastmodified;
1162 * The full filename.
1167 * Pointer to next entry in the linked list of all "file_list"s.
1168 * This linked list is so that sweep() can navigate it.
1169 * Since sweep() can remove items from the list, we must be careful
1170 * to only access this value from main thread (when we know sweep
1171 * won't be running).
1173 struct file_list *next;
1177 #ifdef FEATURE_TRUST
1180 * The format of a trust file when loaded into memory.
1184 struct url_spec url[1]; /**< The URL pattern */
1185 int reject; /**< FIXME: Please document this! */
1186 struct block_spec *next; /**< Next entry in linked list */
1189 #endif /* def FEATURE_TRUST */
1192 #define SOCKS_NONE 0 /**< Don't use a SOCKS server */
1193 #define SOCKS_4 40 /**< original SOCKS 4 protocol */
1194 #define SOCKS_4A 41 /**< as modified for hosts w/o external DNS */
1198 * How to forward a connection to a parent proxy.
1202 /** URL pattern that this forward_spec is for. */
1203 struct url_spec url[1];
1205 /** Connection type. Must be SOCKS_NONE, SOCKS_4, or SOCKS_4A. */
1208 /** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
1211 /** SOCKS server port. */
1214 /** Parent HTTP proxy hostname, or NULL for none. */
1217 /** Parent HTTP proxy port. */
1220 /** Next entry in the linked list. */
1221 struct forward_spec *next;
1226 * Initializer for a static struct forward_spec.
1228 #define FORWARD_SPEC_INITIALIZER { { URL_SPEC_INITIALIZER }, 0, NULL, 0, NULL, 0, NULL }
1232 * This struct represents one filter (one block) from
1233 * the re_filterfile. If there is more than one filter
1234 * in the file, the file will be represented by a
1235 * chained list of re_filterfile specs.
1237 struct re_filterfile_spec
1239 char *name; /**< Name from FILTER: statement in re_filterfile. */
1240 char *description; /**< Description from FILTER: statement in re_filterfile. */
1241 struct list patterns[1]; /**< The patterns from the re_filterfile. */
1242 pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
1243 struct re_filterfile_spec *next; /**< The pointer for chaining. */
1249 #define ACL_PERMIT 1 /**< Accept connection request */
1250 #define ACL_DENY 2 /**< Reject connection request */
1253 * An IP address pattern. Used to specify networks in the ACL.
1255 struct access_control_addr
1257 unsigned long addr; /**< The IP address as an integer. */
1258 unsigned long mask; /**< The network mask as an integer. */
1259 unsigned long port; /**< The port number. */
1263 * An access control list (ACL) entry.
1265 * This is a linked list.
1267 struct access_control_list
1269 struct access_control_addr src[1]; /**< Client IP address */
1270 struct access_control_addr dst[1]; /**< Website or parent proxy IP address */
1272 short action; /**< ACL_PERMIT or ACL_DENY */
1273 struct access_control_list *next; /**< The next entry in the ACL. */
1276 #endif /* def FEATURE_ACL */
1279 /** Maximum number of loaders (actions, re_filter, ...) */
1283 /** configuration_spec::feature_flags: CGI actions editor. */
1284 #define RUNTIME_FEATURE_CGI_EDIT_ACTIONS 1
1286 /** configuration_spec::feature_flags: Web-based toggle. */
1287 #define RUNTIME_FEATURE_CGI_TOGGLE 2
1291 * Data loaded from the configuration file.
1293 * (Anomaly: toggle is still handled through a global, not this structure)
1295 struct configuration_spec
1300 /** Nonzero to enable multithreading. */
1304 * Bitmask of features that can be enabled/disabled through the config
1305 * file. Currently defined bits:
1307 * - RUNTIME_FEATURE_CGI_EDIT_ACTIONS
1308 * - RUNTIME_FEATURE_CGI_TOGGLE
1310 unsigned feature_flags;
1312 /** The log file name. */
1313 const char *logfile;
1315 /** The config file directory. */
1316 const char *confdir;
1318 /** The log file directory. */
1321 /** The full paths to the actions files. */
1322 const char *actions_file[MAX_ACTION_FILES];
1324 /** The short names of the actions files. */
1325 const char *actions_file_short[MAX_ACTION_FILES];
1327 /** The administrator's email address */
1328 char *admin_address;
1330 /** A URL with info on this proxy */
1331 char *proxy_info_url;
1333 /** URL to the user manual (on our website or local copy) */
1336 /** The file name of the pcre filter file */
1337 const char *re_filterfile;
1339 #ifdef FEATURE_COOKIE_JAR
1341 /** The file name of the cookie jar file */
1342 const char * jarfile;
1344 /** The handle to the cookie jar file */
1347 #endif /* def FEATURE_COOKIE_JAR */
1349 /** IP address to bind to. Defaults to HADDR_DEFAULT == 127.0.0.1. */
1352 /** Port to bind to. Defaults to HADDR_PORT == 8118. */
1355 /** Size limit for IOB */
1356 size_t buffer_limit;
1358 #ifdef FEATURE_TRUST
1360 /** The file name of the trust file. */
1361 const char * trustfile;
1363 /** FIXME: DOCME: Document this. */
1364 struct list trust_info[1];
1366 /** FIXME: DOCME: Document this. */
1367 struct url_spec *trust_list[64];
1369 #endif /* def FEATURE_TRUST */
1373 /** The access control list (ACL). */
1374 struct access_control_list *acl;
1376 #endif /* def FEATURE_ACL */
1378 /** Information about parent proxies (forwarding). */
1379 struct forward_spec *forward;
1381 /** All options from the config file, HTML-formatted. */
1384 /** The configuration file object. */
1385 struct file_list *config_file_list;
1387 /** List of loaders */
1388 int (*loaders[NLOADERS])(struct client_state *);
1390 /** Nonzero if we need to bind() to the new port. */
1394 /** Calculates the number of elements in an array, using sizeof. */
1395 #define SZ(X) (sizeof(X) / sizeof(*X))
1397 #ifdef FEATURE_FORCE_LOAD
1398 /** The force load URL prefix. */
1399 #define FORCE_PREFIX "/PRIVOXY-FORCE"
1400 #endif /* def FEATURE_FORCE_LOAD */
1402 #ifdef FEATURE_NO_GIFS
1403 /** The MIME type for images ("image/png" or "image/gif"). */
1404 #define BUILTIN_IMAGE_MIMETYPE "image/png"
1406 #define BUILTIN_IMAGE_MIMETYPE "image/gif"
1407 #endif /* def FEATURE_NO_GIFS */
1414 /** URL for the Privoxy home page. */
1415 #define HOME_PAGE_URL "http://www.privoxy.org/"
1417 /** URL for the Privoxy user manual. */
1418 #define USER_MANUAL_URL HOME_PAGE_URL "/" VERSION "/user-manual/"
1420 /** Prefix for actions help links (append to USER_MANUAL_URL). */
1421 #define ACTIONS_HELP_PREFIX "actions-file.html#"
1423 /** Prefix for config option help links (append to USER_MANUAL_URL). */
1424 #define CONFIG_HELP_PREFIX "config.html#"
1427 * The "hosts" to intercept and display CGI pages.
1428 * First one is a hostname only, second one can specify host and path.
1431 * 1) Do not specify the http: prefix
1432 * 2) CGI_SITE_2_PATH must not end with /, one will be added automatically.
1433 * 3) CGI_SITE_2_PATH must start with /, unless it is the empty string.
1435 #define CGI_SITE_1_HOST "p.p"
1436 #define CGI_SITE_2_HOST "config.privoxy.org"
1437 #define CGI_SITE_2_PATH ""
1440 * The prefix for CGI pages. Written out in generated HTML.
1441 * INCLUDES the trailing slash.
1443 #define CGI_PREFIX "http://" CGI_SITE_2_HOST CGI_SITE_2_PATH "/"
1448 * FIXME: This is very inefficient. There could be one copy of these strings
1449 * for each .c file!! They should be "extern", not "static".
1451 static const char CSUCCEED[] =
1452 "HTTP/1.0 200 Connection established\n"
1453 "Proxy-Agent: Privoxy/" VERSION "\r\n\r\n";
1455 static const char CHEADER[] =
1456 "HTTP/1.0 400 Invalid header received from browser\r\n\r\n";
1458 static const char CFORBIDDEN[] =
1459 "HTTP/1.0 403 Connection not allowable\r\nX-Hint: If you read this message interactively, then you know why this happens ,-)\r\n\r\n";
1465 #endif /* ndef PROJECT_H_INCLUDED */