1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
4 #define PROJECT_H_VERSION "$Id: project.h,v 2.0 2002/06/04 14:34:21 jongfoster Exp $"
5 /*********************************************************************
7 * File : $Source: /cvsroot/ijbswa/current/src/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 2.0 2002/06/04 14:34:21 jongfoster
41 * Moving source files to src/
43 * Revision 1.72 2002/05/14 21:35:49 oes
44 * Split HELP_LINK_PREFIX into ACTIONS_HELP_PREFIX and CONFIG_HELP_PREFIX
45 * because of split in user-manual
47 * Revision 1.71 2002/05/12 21:39:36 jongfoster
48 * - Adding Doxygen-style comments to structures and #defines.
50 * Revision 1.70 2002/05/12 16:05:50 jongfoster
51 * Fixing ACTION_MASK_ALL to be unsigned long rather than
52 * just unsigned int. I don't know if anyone is porting
53 * Privoxy to 16-bit platforms, but if so, +limit-connect
54 * wouldn't have worked because of this bug.
56 * Revision 1.69 2002/05/08 16:00:16 oes
57 * Added size member to struct iob, so it can
58 * be alloced larger than needed.
60 * Revision 1.68 2002/04/26 12:56:00 oes
61 * Killed REDIRECT_URL, added USER_MANUAL_URL and HELP_LINK_PREFIX
63 * Revision 1.67 2002/04/24 02:12:43 oes
64 * - Jon's multiple AF patch:
65 * - Make csp->actions_list an array
66 * - #define MAX_ACTION_FILES
67 * - Moved CGI_PARAM_LEN_MAX (500) here
69 * Revision 1.66 2002/04/15 19:06:43 jongfoster
72 * Revision 1.65 2002/04/04 00:36:36 gliptak
73 * always use pcre for matching
75 * Revision 1.64 2002/04/03 22:28:03 gliptak
76 * Removed references to gnu_regex
78 * Revision 1.63 2002/03/31 17:19:00 jongfoster
79 * Win32 only: Enabling STRICT to fix a VC++ compile warning.
81 * Revision 1.62 2002/03/26 22:48:49 swa
84 * Revision 1.61 2002/03/26 22:29:55 swa
85 * we have a new homepage!
87 * Revision 1.60 2002/03/24 15:52:17 jongfoster
88 * Changing CGI URL prefixes for new name
90 * Revision 1.59 2002/03/24 15:23:33 jongfoster
93 * Revision 1.58 2002/03/24 13:25:43 swa
94 * name change related issues
96 * Revision 1.57 2002/03/16 20:28:34 oes
97 * Added descriptions to the filters so users will know what they select in the cgi editor
99 * Revision 1.56 2002/03/13 20:27:30 oes
100 * Fixing bug with CT_TABOO
102 * Revision 1.55 2002/03/12 01:42:50 oes
103 * Introduced modular filters
105 * Revision 1.54 2002/03/09 20:03:52 jongfoster
106 * - Making various functions return int rather than size_t.
107 * (Undoing a recent change). Since size_t is unsigned on
108 * Windows, functions like read_socket that return -1 on
109 * error cannot return a size_t.
111 * THIS WAS A MAJOR BUG - it caused frequent, unpredictable
112 * crashes, and also frequently caused JB to jump to 100%
113 * CPU and stay there. (Because it thought it had just
114 * read ((unsigned)-1) == 4Gb of data...)
116 * - The signature of write_socket has changed, it now simply
117 * returns success=0/failure=nonzero.
119 * - Trying to get rid of a few warnings --with-debug on
120 * Windows, I've introduced a new type "jb_socket". This is
121 * used for the socket file descriptors. On Windows, this
122 * is SOCKET (a typedef for unsigned). Everywhere else, it's
123 * an int. The error value can't be -1 any more, so it's
124 * now JB_INVALID_SOCKET (which is -1 on UNIX, and in
125 * Windows it maps to the #define INVALID_SOCKET.)
127 * - The signature of bind_port has changed.
129 * Revision 1.53 2002/03/08 16:48:55 oes
130 * Added FEATURE_NO_GIFS and BUILTIN_IMAGE_MIMETYPE
132 * Revision 1.52 2002/03/07 03:46:17 oes
133 * Fixed compiler warnings
135 * Revision 1.51 2002/03/05 04:52:42 oes
136 * Deleted non-errlog debugging code
138 * Revision 1.50 2002/03/04 19:32:07 oes
139 * Changed default port to 8118
141 * Revision 1.49 2002/03/04 18:28:55 oes
142 * Deleted PID_FILE_NAME
144 * Revision 1.48 2002/03/03 14:50:40 oes
145 * Fixed CLF logging: Added ocmd member for client's request to struct http_request
147 * Revision 1.47 2002/02/20 23:15:13 jongfoster
148 * Parsing functions now handle out-of-memory gracefully by returning
151 * Revision 1.46 2002/01/17 21:06:09 jongfoster
152 * Now #defining the URLs of the config interface
154 * Minor changes to struct http_request and struct url_spec due to
155 * standardizing that struct http_request is used to represent a URL, and
156 * struct url_spec is used to represent a URL pattern. (Before, URLs were
157 * represented as seperate variables and a partially-filled-in url_spec).
159 * Revision 1.45 2002/01/09 14:33:27 oes
160 * Added HOSTENT_BUFFER_SIZE
162 * Revision 1.44 2001/12/30 14:07:32 steudten
163 * - Add signal handling (unix)
164 * - Add SIGHUP handler (unix)
165 * - Add creation of pidfile (unix)
166 * - Add action 'top' in rc file (RH)
167 * - Add entry 'SIGNALS' to manpage
168 * - Add exit message to logfile (unix)
170 * Revision 1.43 2001/11/22 21:57:51 jongfoster
171 * Making action_spec->flags into an unsigned long rather than just an
173 * Adding ACTION_NO_COOKIE_KEEP
175 * Revision 1.42 2001/11/05 21:42:41 steudten
176 * Include DBG() macro.
178 * Revision 1.41 2001/10/28 19:12:06 jongfoster
179 * Adding ijb_toupper()
181 * Revision 1.40 2001/10/26 17:40:47 oes
182 * Moved ijb_isspace and ijb_tolower to project.h
183 * Removed http->user_agent, csp->referrer and csp->accept_types
185 * Revision 1.39 2001/10/25 03:45:02 david__schmidt
186 * Adding a (void*) cast to freez() because Visual Age C++ won't expand the
187 * macro when called with a cast; so moving the cast to the macro def'n
188 * seems to both eliminate compiler warnings (on darwin and OS/2, anyway) and
189 * doesn't make macro expansion complain. Hope this works for everyone else
192 * Revision 1.38 2001/10/23 21:19:04 jongfoster
193 * New error-handling support: jb_err type and JB_ERR_xxx constants
194 * CGI functions now return a jb_err, and their parameters map is const.
195 * Support for RUNTIME_FEATUREs to enable/disable config editor
196 * Adding a few comments
198 * Revision 1.37 2001/10/14 22:14:01 jongfoster
199 * Removing name_length field from struct cgi_dispatcher, as this is
200 * now calculated at runtime from the "name" field.
202 * Revision 1.36 2001/10/10 16:45:15 oes
203 * Added LIMIT_CONNECT action and string
204 * Fixed HTTP message line termination
205 * Added CFORBIDDEN HTTP message
207 * Revision 1.35 2001/10/07 18:06:43 oes
208 * Added status member to struct http_request
210 * Revision 1.34 2001/10/07 15:45:25 oes
211 * Added url member to struct http_request and commented all
216 * Added ACTION_DOWNGRADE and ACTION_NO_COMPRESSION
218 * Replaced struct client_state members rejected,
219 * force, active and toggled_on with "flags" bitmap.
221 * Added CSP_FLAG_MODIFIED and CSP_FLAG_CHUNKED
223 * Added buffer_limit to struct configuration_spec
225 * Revision 1.33 2001/09/20 13:30:08 steudten
227 * Make freez() more secure in case of: if (exp) { free(z) ; a=*z }
228 * Last case will set z to NULL in free(z) and thats bad..
230 * Revision 1.32 2001/09/16 23:02:51 jongfoster
233 * Revision 1.31 2001/09/16 13:20:29 jongfoster
234 * Rewrite of list library. Now has seperate header and list_entry
235 * structures. Also added a large sprinking of assert()s to the list
238 * Revision 1.30 2001/09/13 23:52:00 jongfoster
239 * Support for both static and dynamically generated CGI pages
241 * Revision 1.29 2001/09/13 23:29:43 jongfoster
242 * Defining FORWARD_SPEC_INITIALIZER
244 * Revision 1.28 2001/09/13 23:05:50 jongfoster
245 * Changing the string paramater to the header parsers a "const".
247 * Revision 1.27 2001/08/05 16:06:20 jongfoster
248 * Modifiying "struct map" so that there are now separate header and
249 * "map_entry" structures. This means that functions which modify a
250 * map no longer need to return a pointer to the modified map.
251 * Also, it no longer reverses the order of the entries (which may be
252 * important with some advanced template substitutions).
254 * Revision 1.26 2001/07/30 22:08:36 jongfoster
255 * Tidying up #defines:
256 * - All feature #defines are now of the form FEATURE_xxx
257 * - Permanently turned off WIN_GUI_EDIT
258 * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
260 * Revision 1.25 2001/07/29 18:43:08 jongfoster
261 * Changing #ifdef _FILENAME_H to FILENAME_H_INCLUDED, to conform to
264 * Revision 1.24 2001/07/25 17:20:27 oes
265 * Introduced http->user_agent
267 * Revision 1.23 2001/07/18 12:32:23 oes
268 * - Added ACTION_STRING_DEANIMATE
269 * - moved #define freez from jcc.h to project.h
271 * Revision 1.22 2001/07/15 17:51:41 jongfoster
272 * Renaming #define STATIC to STATIC_PCRE
274 * Revision 1.21 2001/07/13 14:03:19 oes
275 * - Reorganized regex header inclusion and #defines to
276 * comply to the scheme in configure.in
277 * - Added csp->content_type and its CT_* keys
278 * - Added ACTION_DEANIMATE
279 * - Removed all #ifdef PCRS
281 * Revision 1.20 2001/06/29 21:45:41 oes
282 * Indentation, CRLF->LF, Tab-> Space
284 * Revision 1.19 2001/06/29 13:33:36 oes
285 * - Improved comments
286 * - Introduced http_request.host_ip_addr_str
287 * - Introduced http_response.head_length
288 * - Introduced config.my_ip_addr_str, config.my_hostname,
289 * config.admin_address and config.proxy_info_url
290 * - Removed config.proxy_args_header and config.proxy_args_trailer,
291 * renamed config.proxy_args_invocation to config.proxy_args
292 * - Removed HTML snipplets and GIFs
293 * - Removed logentry from cancelled commit
295 * Revision 1.18 2001/06/09 10:57:39 jongfoster
296 * Adding definition of BUFFER_SIZE.
297 * Changing struct cgi_dispatcher to use "const" strings.
299 * Revision 1.17 2001/06/07 23:15:09 jongfoster
300 * Merging ACL and forward files into config file.
301 * Moving struct gateway members into struct forward_spec
302 * Removing config->proxy_args_gateways
303 * Cosmetic: Adding a few comments
305 * Revision 1.16 2001/06/04 18:31:58 swa
306 * files are now prefixed with either `confdir' or `logdir'.
307 * `make redhat-dist' replaces both entries confdir and logdir
310 * Revision 1.15 2001/06/04 11:28:53 swa
311 * redirect did not work due to missing /
313 * Revision 1.14 2001/06/03 11:03:48 oes
315 * added struct http_response,
316 * changed struct interceptors to struct cgi_dispatcher,
317 * moved HTML stuff to cgi.h
319 * Revision 1.13 2001/06/01 20:05:36 jongfoster
320 * Support for +image-blocker{}: added ACTION_IMAGE_BLOCKER
321 * constant, and removed csp->tinygif.
323 * Revision 1.12 2001/06/01 18:49:17 jongfoster
324 * Replaced "list_share" with "list" - the tiny memory gain was not
325 * worth the extra complexity.
327 * Revision 1.11 2001/06/01 10:32:47 oes
328 * Added constants for anchoring selection bitmap
330 * Revision 1.10 2001/05/31 21:33:53 jongfoster
331 * Changes for new actions file, replacing permissionsfile
332 * and parts of the config file. Also added support for
335 * Revision 1.9 2001/05/31 17:32:31 oes
337 * - Enhanced domain part globbing with infix and prefix asterisk
338 * matching and optional unanchored operation
340 * Revision 1.8 2001/05/29 20:09:15 joergs
341 * HTTP_REDIRECT_TEMPLATE fixed.
343 * Revision 1.7 2001/05/29 09:50:24 jongfoster
344 * Unified blocklist/imagelist/actionslist.
345 * File format is still under discussion, but the internal changes
348 * Also modified interceptor behaviour:
349 * - We now intercept all URLs beginning with one of the following
350 * prefixes (and *only* these prefixes):
352 * * http://ijbswa.sf.net/config/
353 * * http://ijbswa.sourceforge.net/config/
354 * - New interceptors "home page" - go to http://i.j.b/ to see it.
355 * - Internal changes so that intercepted and fast redirect pages
356 * are not replaced with an image.
357 * - Interceptors now have the option to send a binary page direct
358 * to the client. (i.e. ijb-send-banner uses this)
359 * - Implemented show-url-info interceptor. (Which is why I needed
360 * the above interceptors changes - a typical URL is
361 * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif".
362 * The previous mechanism would not have intercepted that, and
363 * if it had been intercepted then it then it would have replaced
366 * Revision 1.6 2001/05/27 22:17:04 oes
368 * - re_process_buffer no longer writes the modified buffer
369 * to the client, which was very ugly. It now returns the
370 * buffer, which it is then written by chat.
372 * - content_length now adjusts the Content-Length: header
373 * for modified documents rather than crunch()ing it.
374 * (Length info in csp->content_length, which is 0 for
375 * unmodified documents)
377 * - For this to work, sed() is called twice when filtering.
379 * Revision 1.5 2001/05/26 00:28:36 jongfoster
380 * Automatic reloading of config file.
381 * Removed obsolete SIGHUP support (Unix) and Reload menu option (Win32).
382 * Most of the global variables have been moved to a new
383 * struct configuration_spec, accessed through csp->config->globalname
384 * Most of the globals remaining are used by the Win32 GUI.
386 * Revision 1.4 2001/05/22 18:46:04 oes
388 * - Enabled filtering banners by size rather than URL
389 * by adding patterns that replace all standard banner
390 * sizes with the "Junkbuster" gif to the re_filterfile
392 * - Enabled filtering WebBugs by providing a pattern
393 * which kills all 1x1 images
395 * - Added support for PCRE_UNGREEDY behaviour to pcrs,
396 * which is selected by the (nonstandard and therefore
397 * capital) letter 'U' in the option string.
398 * It causes the quantifiers to be ungreedy by default.
399 * Appending a ? turns back to greedy (!).
401 * - Added a new interceptor ijb-send-banner, which
402 * sends back the "Junkbuster" gif. Without imagelist or
403 * MSIE detection support, or if tinygif = 1, or the
404 * URL isn't recognized as an imageurl, a lame HTML
405 * explanation is sent instead.
407 * - Added new feature, which permits blocking remote
408 * script redirects and firing back a local redirect
410 * The feature is conditionally compiled, i.e. it
411 * can be disabled with --disable-fast-redirects,
412 * plus it must be activated by a "fast-redirects"
413 * line in the config file, has its own log level
414 * and of course wants to be displayed by show-proxy-args
415 * Note: Boy, all the #ifdefs in 1001 locations and
416 * all the fumbling with configure.in and acconfig.h
417 * were *way* more work than the feature itself :-(
419 * - Because a generic redirect template was needed for
420 * this, tinygif = 3 now uses the same.
422 * - Moved GIFs, and other static HTTP response templates
427 * - Removed some >400 CRs again (Jon, you really worked
430 * Revision 1.3 2001/05/20 01:21:20 jongfoster
431 * Version 2.9.4 checkin.
432 * - Merged popupfile and cookiefile, and added control over PCRS
433 * filtering, in new "actionsfile".
434 * - Implemented LOG_LEVEL_FATAL, so that if there is a configuration
435 * file error you now get a message box (in the Win32 GUI) rather
436 * than the program exiting with no explanation.
437 * - Made killpopup use the PCRS MIME-type checking and HTTP-header
439 * - Removed tabs from "config"
440 * - Moved duplicated url parsing code in "loaders.c" to a new funcition.
441 * - Bumped up version number.
443 * Revision 1.2 2001/05/17 23:01:01 oes
444 * - Cleaned CRLF's from the sources and related files
446 * Revision 1.1.1.1 2001/05/15 13:59:03 oes
447 * Initial import of version 2.9.3 source tree
450 *********************************************************************/
453 /* Declare struct FILE for vars and funcs. */
456 /* Need time_t for file_list */
460 * Include appropriate regular expression libraries.
461 * Note that pcrs and pcre (native) are needed for cgi
462 * and are included anyway.
478 # include "pcreposix.h"
480 # include <pcreposix.h>
485 #endif /* def AMIGA */
489 * I don't want to have to #include all this just for the declaration
490 * of SOCKET. However, it looks like we have to...
505 typedef SOCKET jb_socket;
507 #define JB_INVALID_SOCKET INVALID_SOCKET
509 #else /* ndef _WIN32 */
512 * The type used by sockets. On UNIX it's an int. Microsoft decided to
513 * make it an unsigned.
515 typedef int jb_socket;
518 * The error value used for variables of type jb_socket. On UNIX this
519 * is -1, however Microsoft decided to make socket handles unsigned, so
520 * they use a different value.
523 #define JB_INVALID_SOCKET (-1)
525 #endif /* ndef _WIN32 */
529 * A standard error code. This should be JB_ERR_OK or one of the JB_ERR_xxx
534 #define JB_ERR_OK 0 /**< Success, no error */
535 #define JB_ERR_MEMORY 1 /**< Out of memory */
536 #define JB_ERR_CGI_PARAMS 2 /**< Missing or corrupt CGI parameters */
537 #define JB_ERR_FILE 3 /**< Error opening, reading or writing a file */
538 #define JB_ERR_PARSE 4 /**< Error parsing file */
539 #define JB_ERR_MODIFIED 5 /**< File has been modified outside of the
540 CGI actions editor. */
544 * This macro is used to free a pointer that may be NULL.
545 * It also sets the variable to NULL after it's been freed.
546 * The paramater should be a simple variable without side effects.
548 #define freez(X) { if(X) { free((void*)X); X = NULL ; } }
552 * Fix a problem with Solaris. There should be no effect on other
555 * Solaris's isspace() is a macro which uses it's argument directly
556 * as an array index. Therefore we need to make sure that high-bit
557 * characters generate +ve values, and ideally we also want to make
558 * the argument match the declared parameter type of "int".
560 * Note: Remember to #include <ctype.h> if you use these macros.
562 #define ijb_toupper(__X) toupper((int)(unsigned char)(__X))
563 #define ijb_tolower(__X) tolower((int)(unsigned char)(__X))
564 #define ijb_isspace(__X) isspace((int)(unsigned char)(__X))
567 * Use for statically allocated buffers if you have no other choice.
568 * Remember to check the length of what you write into the buffer
569 * - we don't want any buffer overflows!
571 #define BUFFER_SIZE 5000
574 * Max length of CGI parameters (arbitrary limit).
576 #define CGI_PARAM_LEN_MAX 500
579 * Buffer size for capturing struct hostent data in the
580 * gethostby(name|addr)_r library calls. Since we don't
581 * loop over gethostbyname_r, the buffer must be sufficient
582 * to accomodate multiple IN A RRs, as used in DNS round robin
583 * load balancing. W3C's wwwlib uses 1K, so that should be
584 * good enough for us, too.
586 #define HOSTENT_BUFFER_SIZE 1024
589 * Do not use. Originally this was so that you can
590 * say "while (FOREVER) { ...do something... }".
591 * However, this gives a warning with some compilers (e.g. VC++).
592 * Instead, use "for (;;) { ...do something... }".
597 * Default IP address to listen on, as a string.
598 * Set to "127.0.0.1".
600 #define HADDR_DEFAULT "127.0.0.1"
603 * Default port to listen on, as a number.
606 #define HADDR_PORT 8118
609 /* Forward def for struct client_state */
610 struct configuration_spec;
614 * Entry in a linked list of strings.
619 * The string. The "const" is only to discourage modification,
620 * you can actually change it if you *really* want to.
621 * You can even freez() it and replace it with another
622 * malloc()d string. If you replace it with NULL, the list
623 * functions will work, just be careful next time you iterate
624 * through the list in your own code.
626 * FIXME: Should we remove the "const"?
630 /** Next entry in the linked list, or NULL if no more. */
631 struct list_entry *next;
635 * A header for a linked list of strings.
639 /** First entry in the list, or NULL if the list is empty. */
640 struct list_entry *first;
642 /** Last entry in the list, or NULL if the list is empty. */
643 struct list_entry *last;
648 * An entry in a map. This is a name=value pair.
652 /** The key for the map. */
654 /** The value associated with that key. */
656 /** The next map entry, or NULL if none. */
657 struct map_entry *next;
661 * A map from a string to another string.
662 * This is used for the paramaters passed in a HTTP GET request, and
663 * to store the exports when the CGI interface is filling in a template.
667 /** The first map entry, or NULL if the map is empty. */
668 struct map_entry *first;
669 /** The last map entry, or NULL if the map is empty. */
670 struct map_entry *last;
675 * A HTTP request. This includes the method (GET, POST) and
678 * This is also used whenever we want to match a URL against a
679 * URL pattern. This always contains the URL to match, and never
680 * a URL pattern. (See struct url_spec).
684 char *cmd; /**< Whole command line: method, URL, Version */
685 char *ocmd; /**< Backup of original cmd for CLF logging */
686 char *gpc; /**< HTTP method: GET, POST, ... */
687 char *url; /**< The URL */
688 char *ver; /**< Protocol version */
689 int status; /**< HTTP Status */
691 char *host; /**< Host part of URL */
692 int port; /**< Port of URL or 80 (default) */
693 char *path; /**< Path of URL */
694 char *hostport; /**< host[:port] */
695 int ssl; /**< Flag if protocol is https */
697 char *host_ip_addr_str; /**< String with dotted decimal representation
698 of host's IP. NULL before connect_to() */
700 char *dbuffer; /**< Buffer with '\0'-delimited domain name. */
701 char **dvec; /**< List of pointers to the strings in dbuffer. */
702 int dcount; /**< How many parts to this domain? (length of dvec) */
707 * Response generated by CGI, blocker, or error handler
711 char *status; /**< HTTP status (string). */
712 struct list headers[1]; /**< List of header lines. */
713 char *head; /**< Formatted http response head. */
714 size_t head_length; /**< Length of http response head. */
715 char *body; /**< HTTP document body. */
716 size_t content_length; /**< Length of body, REQUIRED if binary body. */
717 int is_static; /**< Nonzero if the content will never change and
718 should be cached by the browser (e.g. images). */
726 /** The string which was parsed to produce this url_spec.
727 Used for debugging or display only. */
730 char *dbuffer; /**< Buffer with '\0'-delimited domain name, or NULL to match all hosts. */
731 char **dvec; /**< List of pointers to the strings in dbuffer. */
732 int dcount; /**< How many parts to this domain? (length of dvec) */
733 int unanchored; /**< Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT. */
735 int port; /**< The port number, or 0 to match all ports. */
737 char *path; /**< The source for the regex. */
738 int pathlen; /**< ==strlen(path). Needed for prefix matching. FIXME: Now obsolete? */
739 regex_t *preg; /**< Regex for matching path part */
743 * If you declare a static url_spec, this is the value to initialize it to zero.
745 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, 0, 0, 0, NULL, 0, NULL }
748 * Constant for host part matching in URLs. If set, indicates that the start of
749 * the pattern must match the start of the URL. E.g. this is not set for the
750 * pattern ".example.com", so that it will match both "example.com" and
751 * "www.example.com". It is set for the pattern "example.com", which makes it
752 * match "example.com" only, not "www.example.com".
754 #define ANCHOR_LEFT 1
757 * Constant for host part matching in URLs. If set, indicates that the end of
758 * the pattern must match the end of the URL. E.g. this is not set for the
759 * pattern "ad.", so that it will match any host called "ad", irrespective
760 * of how many subdomains are in the fully-qualified domain name.
762 #define ANCHOR_RIGHT 2
766 * An I/O buffer. Holds a string which can be appended to, and can have data
767 * removed from the beginning.
771 char *buf; /**< Start of buffer */
772 char *cur; /**< Start of relevant data */
773 char *eod; /**< End of relevant data */
774 size_t size; /**< Size as malloc()ed */
779 * Return the number of bytes in the I/O buffer associated with the passed
780 * client_state pointer.
783 #define IOB_PEEK(CSP) ((CSP->iob->cur > CSP->iob->eod) ? (CSP->iob->eod - CSP->iob->cur) : 0)
787 * Remove any data in the I/O buffer associated with the passed
788 * client_state pointer.
790 #define IOB_RESET(CSP) if(CSP->iob->buf) free(CSP->iob->buf); memset(CSP->iob, '\0', sizeof(CSP->iob));
792 /* Bits for csp->content_type */
793 #define CT_TEXT 1 /**< csp->content_type bitmask:
794 Suitable for pcrs filtering. */
795 #define CT_GIF 2 /**< csp->content_type bitmask:
796 Suitable for GIF filtering. */
797 #define CT_TABOO 4 /**< csp->content_type bitmask:
798 DO NOT filter, irrespective of other flags. */
801 * The mask which includes all actions.
803 #define ACTION_MASK_ALL (~0UL)
806 * The most compatible set of actions - i.e. none.
808 #define ACTION_MOST_COMPATIBLE 0x00000000UL
810 /** Action bitmap: Block the request. */
811 #define ACTION_BLOCK 0x00000001UL
812 /** Action bitmap: Deanimate if it's a GIF. */
813 #define ACTION_DEANIMATE 0x00000002UL
814 /** Action bitmap: Downgrade HTTP/1.1 to 1.0. */
815 #define ACTION_DOWNGRADE 0x00000004UL
816 /** Action bitmap: Fast redirects. */
817 #define ACTION_FAST_REDIRECTS 0x00000008UL
818 /** Action bitmap: Remove existing "Forwarded" header, and do not add another. */
819 #define ACTION_HIDE_FORWARDED 0x00000010UL
820 /** Action bitmap: Hide "From" header. */
821 #define ACTION_HIDE_FROM 0x00000020UL
822 /** Action bitmap: Hide "Referer" header. (sic - follow HTTP, not English). */
823 #define ACTION_HIDE_REFERER 0x00000040UL
824 /** Action bitmap: Hide "User-Agent" and similar headers. */
825 #define ACTION_HIDE_USER_AGENT 0x00000080UL
826 /** Action bitmap: This is an image. */
827 #define ACTION_IMAGE 0x00000100UL
828 /** Action bitmap: Sets the image blocker. */
829 #define ACTION_IMAGE_BLOCKER 0x00000200UL
830 /** Action bitmap: Prevent compression. */
831 #define ACTION_NO_COMPRESSION 0x00000400UL
832 /** Action bitmap: Change cookies to session only cookies. */
833 #define ACTION_NO_COOKIE_KEEP 0x00000800UL
834 /** Action bitmap: Block rending cookies. */
835 #define ACTION_NO_COOKIE_READ 0x00001000UL
836 /** Action bitmap: Block setting cookies. */
837 #define ACTION_NO_COOKIE_SET 0x00002000UL
838 /** Action bitmap: Filter out popups. */
839 #define ACTION_NO_POPUPS 0x00004000UL
840 /** Action bitmap: Send a vanilla wafer. */
841 #define ACTION_VANILLA_WAFER 0x00008000UL
842 /** Action bitmap: Limit CONNECT requests to safe ports. */
843 #define ACTION_LIMIT_CONNECT 0x00010000UL
845 /** Action string index: How to deanimate GIFs */
846 #define ACTION_STRING_DEANIMATE 0
847 /** Action string index: Replacement for "From:" header */
848 #define ACTION_STRING_FROM 1
849 /** Action string index: How to block images */
850 #define ACTION_STRING_IMAGE_BLOCKER 2
851 /** Action string index: Replacement for "Referer:" header */
852 #define ACTION_STRING_REFERER 3
853 /** Action string index: Replacement for "User-Agent:" header */
854 #define ACTION_STRING_USER_AGENT 4
855 /** Action string index: Legal CONNECT ports. */
856 #define ACTION_STRING_LIMIT_CONNECT 5
857 /** Number of string actions. */
858 #define ACTION_STRING_COUNT 6
860 /** Index into current_action_spec::multi[] for headers to add. */
861 #define ACTION_MULTI_ADD_HEADER 0
862 /** Index into current_action_spec::multi[] for headers to add. */
863 #define ACTION_MULTI_WAFER 1
864 /** Index into current_action_spec::multi[] for filters to apply. */
865 #define ACTION_MULTI_FILTER 2
866 /** Number of multi-string actions. */
867 #define ACTION_MULTI_COUNT 3
871 * This structure contains a list of actions to apply to a URL.
872 * It only contains positive instructions - no "-" options.
873 * It is not used to store the actions list itself, only for
874 * url_actions() to return the current values.
876 struct current_action_spec
878 /** Actions to apply. A bit set to "1" means perform the action. */
882 * Paramaters for those actions that require them.
883 * Each entry is valid if & only if the corresponding entry in "flags" is
886 char * string[ACTION_STRING_COUNT];
888 /** Lists of strings for multi-string actions. */
889 struct list multi[ACTION_MULTI_COUNT][1];
894 * This structure contains a set of changes to actions.
895 * It can contain both positive and negative instructions.
896 * It is used to store an entry in the actions list.
900 unsigned long mask; /**< Actions to keep. A bit set to "0" means remove action. */
901 unsigned long add; /**< Actions to add. A bit set to "1" means add action. */
904 * Paramaters for those actions that require them.
905 * Each entry is valid if & only if the corresponding entry in "flags" is
908 char * string[ACTION_STRING_COUNT];
910 /** Lists of strings to remove, for multi-string actions. */
911 struct list multi_remove[ACTION_MULTI_COUNT][1];
913 /** If nonzero, remove *all* strings from the multi-string action. */
914 int multi_remove_all[ACTION_MULTI_COUNT];
916 /** Lists of strings to add, for multi-string actions. */
917 struct list multi_add[ACTION_MULTI_COUNT][1];
922 * This structure is used to store the actions list.
924 * It contains a URL pattern, and the chages to the actions.
925 * It is a linked list.
929 struct url_spec url[1]; /**< URL pattern. */
931 struct action_spec action[1]; /**< Actions. */
933 struct url_actions * next; /**< Next action in file, or NULL. */
938 * Flags for use in csp->flags
942 * Flag for csp->flags: Set if this client is processing data.
943 * Cleared when the thread associated with this structure dies.
945 #define CSP_FLAG_ACTIVE 0x01
948 * Flag for csp->flags: Set if the server's reply is in "chunked"
951 #define CSP_FLAG_CHUNKED 0x02
954 * Flag for csp->flags: Set if this request was enforced, although it would
955 * normally have been blocked.
957 #define CSP_FLAG_FORCED 0x04
960 * Flag for csp->flags: Set if any modification to the body was done.
962 #define CSP_FLAG_MODIFIED 0x08
965 * Flag for csp->flags: Set if request was blocked.
967 #define CSP_FLAG_REJECTED 0x10
970 * Flag for csp->flags: Set if we are toggled on (FEATURE_TOGGLE).
972 #define CSP_FLAG_TOGGLED_ON 0x20
976 * Maximum number of actions files. This limit is arbitrary - it's just used
979 #define MAX_ACTION_FILES 10
982 * The state of a Privoxy processing thread.
986 /** The proxy's configuration */
987 struct configuration_spec * config;
989 /** The actions to perform on the current request */
990 struct current_action_spec action[1];
992 /** socket to talk to client (web browser) */
995 /** socket to talk to server (web server or proxy) */
998 /** Multi-purpose flag container, see CSP_FLAG_* above */
999 unsigned short int flags;
1001 /** Client PC's IP address, as reported by the accept() function.
1004 /** Client PC's IP address, as reported by the accept() function.
1008 /** Our IP address. I.e. the IP address that the client used to reach us,
1010 char *my_ip_addr_str;
1012 /** Our hostname. I.e. the reverse DNS of the IP address that the client
1013 used to reach us, as a string. */
1016 /** The URL that was requested */
1017 struct http_request http[1];
1019 /** An I/O buffer used for buffering data read from the network */
1022 /** List of all headers for this request */
1023 struct list headers[1];
1025 /** List of all cookies for this request */
1026 struct list cookie_list[1];
1028 /** MIME-Type key, see CT_* above */
1029 unsigned short int content_type;
1031 /** The "X-Forwarded-For:" header sent by the client */
1034 /** Actions files associated with this client */
1035 struct file_list *actions_list[MAX_ACTION_FILES];
1037 /** pcrs job file. */
1038 struct file_list *rlist;
1040 /** Length after content modification. */
1041 size_t content_length;
1043 #ifdef FEATURE_TRUST
1046 struct file_list *tlist;
1048 #endif /* def FEATURE_TRUST */
1050 /** Next thread in linked list. Only read or modify from the main thread! */
1051 struct client_state *next;
1056 * A function to add a header
1058 typedef jb_err (*add_header_func_ptr)(struct client_state *);
1061 * A function to process a header
1063 typedef jb_err (*parser_func_ptr )(struct client_state *, char **);
1067 * List of functions to run on a list of headers
1071 /** The header prefix to match */
1074 /** The length of the prefix to match */
1077 /** The function to apply to this line */
1078 parser_func_ptr parser;
1083 * List of available CGI functions.
1085 struct cgi_dispatcher
1087 /** The URL of the CGI, relative to the CGI root. */
1088 const char * const name;
1090 /** The handler function for the CGI */
1091 jb_err (* const handler)(struct client_state *csp, struct http_response *rsp, const struct map *parameters);
1093 /** The description of the CGI, to appear on the main menu, or NULL to hide it. */
1094 const char * const description;
1099 * A data file used by Privoxy. Kept in a linked list.
1104 * This is a pointer to the data structures associated with the file.
1105 * Read-only once the structure has been created.
1110 * The unloader function.
1111 * Normally NULL. When we are finished with file (i.e. when we have
1112 * loaded a new one), set to a pointer to an unloader function.
1113 * Unloader will be called by sweep() (called from main loop) when
1114 * all clients using this file are done. This prevents threading
1117 void (*unloader)(void *);
1120 * Used internally by sweep(). Do not access from elsewhere.
1125 * File last-modified time, so we can check if file has been changed.
1126 * Read-only once the structure has been created.
1128 time_t lastmodified;
1131 * The full filename.
1136 * Pointer to next entry in the linked list of all "file_list"s.
1137 * This linked list is so that sweep() can navigate it.
1138 * Since sweep() can remove items from the list, we must be careful
1139 * to only access this value from main thread (when we know sweep
1140 * won't be running).
1142 struct file_list *next;
1146 #ifdef FEATURE_TRUST
1149 * The format of a trust file when loaded into memory.
1153 struct url_spec url[1]; /**< The URL pattern */
1154 int reject; /**< FIXME: Please document this! */
1155 struct block_spec *next; /**< Next entry in linked list */
1158 #endif /* def FEATURE_TRUST */
1161 #define SOCKS_NONE 0 /**< Don't use a SOCKS server */
1162 #define SOCKS_4 40 /**< original SOCKS 4 protocol */
1163 #define SOCKS_4A 41 /**< as modified for hosts w/o external DNS */
1167 * How to forward a connection to a parent proxy.
1171 /** URL pattern that this forward_spec is for. */
1172 struct url_spec url[1];
1174 /** Connection type. Must be SOCKS_NONE, SOCKS_4, or SOCKS_4A. */
1177 /** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
1180 /** SOCKS server port. */
1183 /** Parent HTTP proxy hostname, or NULL for none. */
1186 /** Parent HTTP proxy port. */
1189 /** Next entry in the linked list. */
1190 struct forward_spec *next;
1195 * Initializer for a static struct forward_spec.
1197 #define FORWARD_SPEC_INITIALIZER { { URL_SPEC_INITIALIZER }, 0, NULL, 0, NULL, 0, NULL }
1201 * This struct represents one filter (one block) from
1202 * the re_filterfile. If there is more than one filter
1203 * in the file, the file will be represented by a
1204 * chained list of re_filterfile specs.
1206 struct re_filterfile_spec
1208 char *name; /**< Name from FILTER: statement in re_filterfile. */
1209 char *description; /**< Description from FILTER: statement in re_filterfile. */
1210 struct list patterns[1]; /**< The patterns from the re_filterfile. */
1211 pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
1212 struct re_filterfile_spec *next; /**< The pointer for chaining. */
1218 #define ACL_PERMIT 1 /**< Accept connection request */
1219 #define ACL_DENY 2 /**< Reject connection request */
1222 * An IP address pattern. Used to specify networks in the ACL.
1224 struct access_control_addr
1226 unsigned long addr; /**< The IP address as an integer. */
1227 unsigned long mask; /**< The network mask as an integer. */
1228 unsigned long port; /**< The port number. */
1232 * An access control list (ACL) entry.
1234 * This is a linked list.
1236 struct access_control_list
1238 struct access_control_addr src[1]; /**< Client IP address */
1239 struct access_control_addr dst[1]; /**< Website or parent proxy IP address */
1241 short action; /**< ACL_PERMIT or ACL_DENY */
1242 struct access_control_list *next; /**< The next entry in the ACL. */
1245 #endif /* def FEATURE_ACL */
1248 /** Maximum number of loaders (actions, re_filter, ...) */
1252 /** configuration_spec::feature_flags: CGI actions editor. */
1253 #define RUNTIME_FEATURE_CGI_EDIT_ACTIONS 1
1255 /** configuration_spec::feature_flags: Web-based toggle. */
1256 #define RUNTIME_FEATURE_CGI_TOGGLE 2
1260 * Data loaded from the configuration file.
1262 * (Anomaly: toggle is still handled through a global, not this structure)
1264 struct configuration_spec
1269 /** Nonzero to enable multithreading. */
1273 * Bitmask of features that can be enabled/disabled through the config
1274 * file. Currently defined bits:
1276 * - RUNTIME_FEATURE_CGI_EDIT_ACTIONS
1277 * - RUNTIME_FEATURE_CGI_TOGGLE
1279 unsigned feature_flags;
1281 /** The log file name. */
1282 const char *logfile;
1284 /** The config file directory. */
1285 const char *confdir;
1287 /** The log file directory. */
1290 /** The full paths to the actions files. */
1291 const char *actions_file[MAX_ACTION_FILES];
1293 /** The short names of the actions files. */
1294 const char *actions_file_short[MAX_ACTION_FILES];
1296 /** The administrator's email address */
1297 char *admin_address;
1299 /** A URL with info on this proxy */
1300 char *proxy_info_url;
1302 /** URL to the user manual (on our website or local copy) */
1305 /** The file name of the pcre filter file */
1306 const char *re_filterfile;
1308 #ifdef FEATURE_COOKIE_JAR
1310 /** The file name of the cookie jar file */
1311 const char * jarfile;
1313 /** The handle to the cookie jar file */
1316 #endif /* def FEATURE_COOKIE_JAR */
1318 /** IP address to bind to. Defaults to HADDR_DEFAULT == 127.0.0.1. */
1321 /** Port to bind to. Defaults to HADDR_PORT == 8118. */
1324 /** Size limit for IOB */
1325 size_t buffer_limit;
1327 #ifdef FEATURE_TRUST
1329 /** The file name of the trust file. */
1330 const char * trustfile;
1332 /** FIXME: DOCME: Document this. */
1333 struct list trust_info[1];
1335 /** FIXME: DOCME: Document this. */
1336 struct url_spec *trust_list[64];
1338 #endif /* def FEATURE_TRUST */
1342 /** The access control list (ACL). */
1343 struct access_control_list *acl;
1345 #endif /* def FEATURE_ACL */
1347 /** Information about parent proxies (forwarding). */
1348 struct forward_spec *forward;
1350 /** All options from the config file, HTML-formatted. */
1353 /** The configuration file object. */
1354 struct file_list *config_file_list;
1356 /** List of loaders */
1357 int (*loaders[NLOADERS])(struct client_state *);
1359 /** Nonzero if we need to bind() to the new port. */
1363 /** Calculates the number of elements in an array, using sizeof. */
1364 #define SZ(X) (sizeof(X) / sizeof(*X))
1366 #ifdef FEATURE_FORCE_LOAD
1367 /** The force load URL prefix. */
1368 #define FORCE_PREFIX "/PRIVOXY-FORCE"
1369 #endif /* def FEATURE_FORCE_LOAD */
1371 #ifdef FEATURE_NO_GIFS
1372 /** The MIME type for images ("image/png" or "image/gif"). */
1373 #define BUILTIN_IMAGE_MIMETYPE "image/png"
1375 #define BUILTIN_IMAGE_MIMETYPE "image/gif"
1376 #endif /* def FEATURE_NO_GIFS */
1383 /** URL for the Privoxy home page. */
1384 #define HOME_PAGE_URL "http://www.privoxy.org/"
1386 /** URL for the Privoxy user manual. */
1387 #define USER_MANUAL_URL HOME_PAGE_URL "/" VERSION "/user-manual/"
1389 /** Prefix for actions help links (append to USER_MANUAL_URL). */
1390 #define ACTIONS_HELP_PREFIX "actions-file.html#"
1392 /** Prefix for config option help links (append to USER_MANUAL_URL). */
1393 #define CONFIG_HELP_PREFIX "config.html#"
1396 * The "hosts" to intercept and display CGI pages.
1397 * First one is a hostname only, second one can specify host and path.
1400 * 1) Do not specify the http: prefix
1401 * 2) CGI_SITE_2_PATH must not end with /, one will be added automatically.
1402 * 3) CGI_SITE_2_PATH must start with /, unless it is the empty string.
1404 #define CGI_SITE_1_HOST "p.p"
1405 #define CGI_SITE_2_HOST "config.privoxy.org"
1406 #define CGI_SITE_2_PATH ""
1409 * The prefix for CGI pages. Written out in generated HTML.
1410 * INCLUDES the trailing slash.
1412 #define CGI_PREFIX "http://" CGI_SITE_2_HOST CGI_SITE_2_PATH "/"
1419 #endif /* ndef PROJECT_H_INCLUDED */