1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
4 #define PROJECT_H_VERSION "$Id: project.h,v 1.72 2002/05/14 21:35:49 oes 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 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 2002/05/14 21:35:49 oes
41 * Split HELP_LINK_PREFIX into ACTIONS_HELP_PREFIX and CONFIG_HELP_PREFIX
42 * because of split in user-manual
44 * Revision 1.71 2002/05/12 21:39:36 jongfoster
45 * - Adding Doxygen-style comments to structures and #defines.
47 * Revision 1.70 2002/05/12 16:05:50 jongfoster
48 * Fixing ACTION_MASK_ALL to be unsigned long rather than
49 * just unsigned int. I don't know if anyone is porting
50 * Privoxy to 16-bit platforms, but if so, +limit-connect
51 * wouldn't have worked because of this bug.
53 * Revision 1.69 2002/05/08 16:00:16 oes
54 * Added size member to struct iob, so it can
55 * be alloced larger than needed.
57 * Revision 1.68 2002/04/26 12:56:00 oes
58 * Killed REDIRECT_URL, added USER_MANUAL_URL and HELP_LINK_PREFIX
60 * Revision 1.67 2002/04/24 02:12:43 oes
61 * - Jon's multiple AF patch:
62 * - Make csp->actions_list an array
63 * - #define MAX_ACTION_FILES
64 * - Moved CGI_PARAM_LEN_MAX (500) here
66 * Revision 1.66 2002/04/15 19:06:43 jongfoster
69 * Revision 1.65 2002/04/04 00:36:36 gliptak
70 * always use pcre for matching
72 * Revision 1.64 2002/04/03 22:28:03 gliptak
73 * Removed references to gnu_regex
75 * Revision 1.63 2002/03/31 17:19:00 jongfoster
76 * Win32 only: Enabling STRICT to fix a VC++ compile warning.
78 * Revision 1.62 2002/03/26 22:48:49 swa
81 * Revision 1.61 2002/03/26 22:29:55 swa
82 * we have a new homepage!
84 * Revision 1.60 2002/03/24 15:52:17 jongfoster
85 * Changing CGI URL prefixes for new name
87 * Revision 1.59 2002/03/24 15:23:33 jongfoster
90 * Revision 1.58 2002/03/24 13:25:43 swa
91 * name change related issues
93 * Revision 1.57 2002/03/16 20:28:34 oes
94 * Added descriptions to the filters so users will know what they select in the cgi editor
96 * Revision 1.56 2002/03/13 20:27:30 oes
97 * Fixing bug with CT_TABOO
99 * Revision 1.55 2002/03/12 01:42:50 oes
100 * Introduced modular filters
102 * Revision 1.54 2002/03/09 20:03:52 jongfoster
103 * - Making various functions return int rather than size_t.
104 * (Undoing a recent change). Since size_t is unsigned on
105 * Windows, functions like read_socket that return -1 on
106 * error cannot return a size_t.
108 * THIS WAS A MAJOR BUG - it caused frequent, unpredictable
109 * crashes, and also frequently caused JB to jump to 100%
110 * CPU and stay there. (Because it thought it had just
111 * read ((unsigned)-1) == 4Gb of data...)
113 * - The signature of write_socket has changed, it now simply
114 * returns success=0/failure=nonzero.
116 * - Trying to get rid of a few warnings --with-debug on
117 * Windows, I've introduced a new type "jb_socket". This is
118 * used for the socket file descriptors. On Windows, this
119 * is SOCKET (a typedef for unsigned). Everywhere else, it's
120 * an int. The error value can't be -1 any more, so it's
121 * now JB_INVALID_SOCKET (which is -1 on UNIX, and in
122 * Windows it maps to the #define INVALID_SOCKET.)
124 * - The signature of bind_port has changed.
126 * Revision 1.53 2002/03/08 16:48:55 oes
127 * Added FEATURE_NO_GIFS and BUILTIN_IMAGE_MIMETYPE
129 * Revision 1.52 2002/03/07 03:46:17 oes
130 * Fixed compiler warnings
132 * Revision 1.51 2002/03/05 04:52:42 oes
133 * Deleted non-errlog debugging code
135 * Revision 1.50 2002/03/04 19:32:07 oes
136 * Changed default port to 8118
138 * Revision 1.49 2002/03/04 18:28:55 oes
139 * Deleted PID_FILE_NAME
141 * Revision 1.48 2002/03/03 14:50:40 oes
142 * Fixed CLF logging: Added ocmd member for client's request to struct http_request
144 * Revision 1.47 2002/02/20 23:15:13 jongfoster
145 * Parsing functions now handle out-of-memory gracefully by returning
148 * Revision 1.46 2002/01/17 21:06:09 jongfoster
149 * Now #defining the URLs of the config interface
151 * Minor changes to struct http_request and struct url_spec due to
152 * standardizing that struct http_request is used to represent a URL, and
153 * struct url_spec is used to represent a URL pattern. (Before, URLs were
154 * represented as seperate variables and a partially-filled-in url_spec).
156 * Revision 1.45 2002/01/09 14:33:27 oes
157 * Added HOSTENT_BUFFER_SIZE
159 * Revision 1.44 2001/12/30 14:07:32 steudten
160 * - Add signal handling (unix)
161 * - Add SIGHUP handler (unix)
162 * - Add creation of pidfile (unix)
163 * - Add action 'top' in rc file (RH)
164 * - Add entry 'SIGNALS' to manpage
165 * - Add exit message to logfile (unix)
167 * Revision 1.43 2001/11/22 21:57:51 jongfoster
168 * Making action_spec->flags into an unsigned long rather than just an
170 * Adding ACTION_NO_COOKIE_KEEP
172 * Revision 1.42 2001/11/05 21:42:41 steudten
173 * Include DBG() macro.
175 * Revision 1.41 2001/10/28 19:12:06 jongfoster
176 * Adding ijb_toupper()
178 * Revision 1.40 2001/10/26 17:40:47 oes
179 * Moved ijb_isspace and ijb_tolower to project.h
180 * Removed http->user_agent, csp->referrer and csp->accept_types
182 * Revision 1.39 2001/10/25 03:45:02 david__schmidt
183 * Adding a (void*) cast to freez() because Visual Age C++ won't expand the
184 * macro when called with a cast; so moving the cast to the macro def'n
185 * seems to both eliminate compiler warnings (on darwin and OS/2, anyway) and
186 * doesn't make macro expansion complain. Hope this works for everyone else
189 * Revision 1.38 2001/10/23 21:19:04 jongfoster
190 * New error-handling support: jb_err type and JB_ERR_xxx constants
191 * CGI functions now return a jb_err, and their parameters map is const.
192 * Support for RUNTIME_FEATUREs to enable/disable config editor
193 * Adding a few comments
195 * Revision 1.37 2001/10/14 22:14:01 jongfoster
196 * Removing name_length field from struct cgi_dispatcher, as this is
197 * now calculated at runtime from the "name" field.
199 * Revision 1.36 2001/10/10 16:45:15 oes
200 * Added LIMIT_CONNECT action and string
201 * Fixed HTTP message line termination
202 * Added CFORBIDDEN HTTP message
204 * Revision 1.35 2001/10/07 18:06:43 oes
205 * Added status member to struct http_request
207 * Revision 1.34 2001/10/07 15:45:25 oes
208 * Added url member to struct http_request and commented all
213 * Added ACTION_DOWNGRADE and ACTION_NO_COMPRESSION
215 * Replaced struct client_state members rejected,
216 * force, active and toggled_on with "flags" bitmap.
218 * Added CSP_FLAG_MODIFIED and CSP_FLAG_CHUNKED
220 * Added buffer_limit to struct configuration_spec
222 * Revision 1.33 2001/09/20 13:30:08 steudten
224 * Make freez() more secure in case of: if (exp) { free(z) ; a=*z }
225 * Last case will set z to NULL in free(z) and thats bad..
227 * Revision 1.32 2001/09/16 23:02:51 jongfoster
230 * Revision 1.31 2001/09/16 13:20:29 jongfoster
231 * Rewrite of list library. Now has seperate header and list_entry
232 * structures. Also added a large sprinking of assert()s to the list
235 * Revision 1.30 2001/09/13 23:52:00 jongfoster
236 * Support for both static and dynamically generated CGI pages
238 * Revision 1.29 2001/09/13 23:29:43 jongfoster
239 * Defining FORWARD_SPEC_INITIALIZER
241 * Revision 1.28 2001/09/13 23:05:50 jongfoster
242 * Changing the string paramater to the header parsers a "const".
244 * Revision 1.27 2001/08/05 16:06:20 jongfoster
245 * Modifiying "struct map" so that there are now separate header and
246 * "map_entry" structures. This means that functions which modify a
247 * map no longer need to return a pointer to the modified map.
248 * Also, it no longer reverses the order of the entries (which may be
249 * important with some advanced template substitutions).
251 * Revision 1.26 2001/07/30 22:08:36 jongfoster
252 * Tidying up #defines:
253 * - All feature #defines are now of the form FEATURE_xxx
254 * - Permanently turned off WIN_GUI_EDIT
255 * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
257 * Revision 1.25 2001/07/29 18:43:08 jongfoster
258 * Changing #ifdef _FILENAME_H to FILENAME_H_INCLUDED, to conform to
261 * Revision 1.24 2001/07/25 17:20:27 oes
262 * Introduced http->user_agent
264 * Revision 1.23 2001/07/18 12:32:23 oes
265 * - Added ACTION_STRING_DEANIMATE
266 * - moved #define freez from jcc.h to project.h
268 * Revision 1.22 2001/07/15 17:51:41 jongfoster
269 * Renaming #define STATIC to STATIC_PCRE
271 * Revision 1.21 2001/07/13 14:03:19 oes
272 * - Reorganized regex header inclusion and #defines to
273 * comply to the scheme in configure.in
274 * - Added csp->content_type and its CT_* keys
275 * - Added ACTION_DEANIMATE
276 * - Removed all #ifdef PCRS
278 * Revision 1.20 2001/06/29 21:45:41 oes
279 * Indentation, CRLF->LF, Tab-> Space
281 * Revision 1.19 2001/06/29 13:33:36 oes
282 * - Improved comments
283 * - Introduced http_request.host_ip_addr_str
284 * - Introduced http_response.head_length
285 * - Introduced config.my_ip_addr_str, config.my_hostname,
286 * config.admin_address and config.proxy_info_url
287 * - Removed config.proxy_args_header and config.proxy_args_trailer,
288 * renamed config.proxy_args_invocation to config.proxy_args
289 * - Removed HTML snipplets and GIFs
290 * - Removed logentry from cancelled commit
292 * Revision 1.18 2001/06/09 10:57:39 jongfoster
293 * Adding definition of BUFFER_SIZE.
294 * Changing struct cgi_dispatcher to use "const" strings.
296 * Revision 1.17 2001/06/07 23:15:09 jongfoster
297 * Merging ACL and forward files into config file.
298 * Moving struct gateway members into struct forward_spec
299 * Removing config->proxy_args_gateways
300 * Cosmetic: Adding a few comments
302 * Revision 1.16 2001/06/04 18:31:58 swa
303 * files are now prefixed with either `confdir' or `logdir'.
304 * `make redhat-dist' replaces both entries confdir and logdir
307 * Revision 1.15 2001/06/04 11:28:53 swa
308 * redirect did not work due to missing /
310 * Revision 1.14 2001/06/03 11:03:48 oes
312 * added struct http_response,
313 * changed struct interceptors to struct cgi_dispatcher,
314 * moved HTML stuff to cgi.h
316 * Revision 1.13 2001/06/01 20:05:36 jongfoster
317 * Support for +image-blocker{}: added ACTION_IMAGE_BLOCKER
318 * constant, and removed csp->tinygif.
320 * Revision 1.12 2001/06/01 18:49:17 jongfoster
321 * Replaced "list_share" with "list" - the tiny memory gain was not
322 * worth the extra complexity.
324 * Revision 1.11 2001/06/01 10:32:47 oes
325 * Added constants for anchoring selection bitmap
327 * Revision 1.10 2001/05/31 21:33:53 jongfoster
328 * Changes for new actions file, replacing permissionsfile
329 * and parts of the config file. Also added support for
332 * Revision 1.9 2001/05/31 17:32:31 oes
334 * - Enhanced domain part globbing with infix and prefix asterisk
335 * matching and optional unanchored operation
337 * Revision 1.8 2001/05/29 20:09:15 joergs
338 * HTTP_REDIRECT_TEMPLATE fixed.
340 * Revision 1.7 2001/05/29 09:50:24 jongfoster
341 * Unified blocklist/imagelist/actionslist.
342 * File format is still under discussion, but the internal changes
345 * Also modified interceptor behaviour:
346 * - We now intercept all URLs beginning with one of the following
347 * prefixes (and *only* these prefixes):
349 * * http://ijbswa.sf.net/config/
350 * * http://ijbswa.sourceforge.net/config/
351 * - New interceptors "home page" - go to http://i.j.b/ to see it.
352 * - Internal changes so that intercepted and fast redirect pages
353 * are not replaced with an image.
354 * - Interceptors now have the option to send a binary page direct
355 * to the client. (i.e. ijb-send-banner uses this)
356 * - Implemented show-url-info interceptor. (Which is why I needed
357 * the above interceptors changes - a typical URL is
358 * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif".
359 * The previous mechanism would not have intercepted that, and
360 * if it had been intercepted then it then it would have replaced
363 * Revision 1.6 2001/05/27 22:17:04 oes
365 * - re_process_buffer no longer writes the modified buffer
366 * to the client, which was very ugly. It now returns the
367 * buffer, which it is then written by chat.
369 * - content_length now adjusts the Content-Length: header
370 * for modified documents rather than crunch()ing it.
371 * (Length info in csp->content_length, which is 0 for
372 * unmodified documents)
374 * - For this to work, sed() is called twice when filtering.
376 * Revision 1.5 2001/05/26 00:28:36 jongfoster
377 * Automatic reloading of config file.
378 * Removed obsolete SIGHUP support (Unix) and Reload menu option (Win32).
379 * Most of the global variables have been moved to a new
380 * struct configuration_spec, accessed through csp->config->globalname
381 * Most of the globals remaining are used by the Win32 GUI.
383 * Revision 1.4 2001/05/22 18:46:04 oes
385 * - Enabled filtering banners by size rather than URL
386 * by adding patterns that replace all standard banner
387 * sizes with the "Junkbuster" gif to the re_filterfile
389 * - Enabled filtering WebBugs by providing a pattern
390 * which kills all 1x1 images
392 * - Added support for PCRE_UNGREEDY behaviour to pcrs,
393 * which is selected by the (nonstandard and therefore
394 * capital) letter 'U' in the option string.
395 * It causes the quantifiers to be ungreedy by default.
396 * Appending a ? turns back to greedy (!).
398 * - Added a new interceptor ijb-send-banner, which
399 * sends back the "Junkbuster" gif. Without imagelist or
400 * MSIE detection support, or if tinygif = 1, or the
401 * URL isn't recognized as an imageurl, a lame HTML
402 * explanation is sent instead.
404 * - Added new feature, which permits blocking remote
405 * script redirects and firing back a local redirect
407 * The feature is conditionally compiled, i.e. it
408 * can be disabled with --disable-fast-redirects,
409 * plus it must be activated by a "fast-redirects"
410 * line in the config file, has its own log level
411 * and of course wants to be displayed by show-proxy-args
412 * Note: Boy, all the #ifdefs in 1001 locations and
413 * all the fumbling with configure.in and acconfig.h
414 * were *way* more work than the feature itself :-(
416 * - Because a generic redirect template was needed for
417 * this, tinygif = 3 now uses the same.
419 * - Moved GIFs, and other static HTTP response templates
424 * - Removed some >400 CRs again (Jon, you really worked
427 * Revision 1.3 2001/05/20 01:21:20 jongfoster
428 * Version 2.9.4 checkin.
429 * - Merged popupfile and cookiefile, and added control over PCRS
430 * filtering, in new "actionsfile".
431 * - Implemented LOG_LEVEL_FATAL, so that if there is a configuration
432 * file error you now get a message box (in the Win32 GUI) rather
433 * than the program exiting with no explanation.
434 * - Made killpopup use the PCRS MIME-type checking and HTTP-header
436 * - Removed tabs from "config"
437 * - Moved duplicated url parsing code in "loaders.c" to a new funcition.
438 * - Bumped up version number.
440 * Revision 1.2 2001/05/17 23:01:01 oes
441 * - Cleaned CRLF's from the sources and related files
443 * Revision 1.1.1.1 2001/05/15 13:59:03 oes
444 * Initial import of version 2.9.3 source tree
447 *********************************************************************/
450 /* Declare struct FILE for vars and funcs. */
453 /* Need time_t for file_list */
457 * Include appropriate regular expression libraries.
458 * Note that pcrs and pcre (native) are needed for cgi
459 * and are included anyway.
475 # include "pcreposix.h"
477 # include <pcreposix.h>
482 #endif /* def AMIGA */
486 * I don't want to have to #include all this just for the declaration
487 * of SOCKET. However, it looks like we have to...
502 typedef SOCKET jb_socket;
504 #define JB_INVALID_SOCKET INVALID_SOCKET
506 #else /* ndef _WIN32 */
509 * The type used by sockets. On UNIX it's an int. Microsoft decided to
510 * make it an unsigned.
512 typedef int jb_socket;
515 * The error value used for variables of type jb_socket. On UNIX this
516 * is -1, however Microsoft decided to make socket handles unsigned, so
517 * they use a different value.
520 #define JB_INVALID_SOCKET (-1)
522 #endif /* ndef _WIN32 */
526 * A standard error code. This should be JB_ERR_OK or one of the JB_ERR_xxx
531 #define JB_ERR_OK 0 /**< Success, no error */
532 #define JB_ERR_MEMORY 1 /**< Out of memory */
533 #define JB_ERR_CGI_PARAMS 2 /**< Missing or corrupt CGI parameters */
534 #define JB_ERR_FILE 3 /**< Error opening, reading or writing a file */
535 #define JB_ERR_PARSE 4 /**< Error parsing file */
536 #define JB_ERR_MODIFIED 5 /**< File has been modified outside of the
537 CGI actions editor. */
541 * This macro is used to free a pointer that may be NULL.
542 * It also sets the variable to NULL after it's been freed.
543 * The paramater should be a simple variable without side effects.
545 #define freez(X) { if(X) { free((void*)X); X = NULL ; } }
549 * Fix a problem with Solaris. There should be no effect on other
552 * Solaris's isspace() is a macro which uses it's argument directly
553 * as an array index. Therefore we need to make sure that high-bit
554 * characters generate +ve values, and ideally we also want to make
555 * the argument match the declared parameter type of "int".
557 * Note: Remember to #include <ctype.h> if you use these macros.
559 #define ijb_toupper(__X) toupper((int)(unsigned char)(__X))
560 #define ijb_tolower(__X) tolower((int)(unsigned char)(__X))
561 #define ijb_isspace(__X) isspace((int)(unsigned char)(__X))
564 * Use for statically allocated buffers if you have no other choice.
565 * Remember to check the length of what you write into the buffer
566 * - we don't want any buffer overflows!
568 #define BUFFER_SIZE 5000
571 * Max length of CGI parameters (arbitrary limit).
573 #define CGI_PARAM_LEN_MAX 500
576 * Buffer size for capturing struct hostent data in the
577 * gethostby(name|addr)_r library calls. Since we don't
578 * loop over gethostbyname_r, the buffer must be sufficient
579 * to accomodate multiple IN A RRs, as used in DNS round robin
580 * load balancing. W3C's wwwlib uses 1K, so that should be
581 * good enough for us, too.
583 #define HOSTENT_BUFFER_SIZE 1024
586 * Do not use. Originally this was so that you can
587 * say "while (FOREVER) { ...do something... }".
588 * However, this gives a warning with some compilers (e.g. VC++).
589 * Instead, use "for (;;) { ...do something... }".
594 * Default IP address to listen on, as a string.
595 * Set to "127.0.0.1".
597 #define HADDR_DEFAULT "127.0.0.1"
600 * Default port to listen on, as a number.
603 #define HADDR_PORT 8118
606 /* Forward def for struct client_state */
607 struct configuration_spec;
611 * Entry in a linked list of strings.
616 * The string. The "const" is only to discourage modification,
617 * you can actually change it if you *really* want to.
618 * You can even freez() it and replace it with another
619 * malloc()d string. If you replace it with NULL, the list
620 * functions will work, just be careful next time you iterate
621 * through the list in your own code.
623 * FIXME: Should we remove the "const"?
627 /** Next entry in the linked list, or NULL if no more. */
628 struct list_entry *next;
632 * A header for a linked list of strings.
636 /** First entry in the list, or NULL if the list is empty. */
637 struct list_entry *first;
639 /** Last entry in the list, or NULL if the list is empty. */
640 struct list_entry *last;
645 * An entry in a map. This is a name=value pair.
649 /** The key for the map. */
651 /** The value associated with that key. */
653 /** The next map entry, or NULL if none. */
654 struct map_entry *next;
658 * A map from a string to another string.
659 * This is used for the paramaters passed in a HTTP GET request, and
660 * to store the exports when the CGI interface is filling in a template.
664 /** The first map entry, or NULL if the map is empty. */
665 struct map_entry *first;
666 /** The last map entry, or NULL if the map is empty. */
667 struct map_entry *last;
672 * A HTTP request. This includes the method (GET, POST) and
675 * This is also used whenever we want to match a URL against a
676 * URL pattern. This always contains the URL to match, and never
677 * a URL pattern. (See struct url_spec).
681 char *cmd; /**< Whole command line: method, URL, Version */
682 char *ocmd; /**< Backup of original cmd for CLF logging */
683 char *gpc; /**< HTTP method: GET, POST, ... */
684 char *url; /**< The URL */
685 char *ver; /**< Protocol version */
686 int status; /**< HTTP Status */
688 char *host; /**< Host part of URL */
689 int port; /**< Port of URL or 80 (default) */
690 char *path; /**< Path of URL */
691 char *hostport; /**< host[:port] */
692 int ssl; /**< Flag if protocol is https */
694 char *host_ip_addr_str; /**< String with dotted decimal representation
695 of host's IP. NULL before connect_to() */
697 char *dbuffer; /**< Buffer with '\0'-delimited domain name. */
698 char **dvec; /**< List of pointers to the strings in dbuffer. */
699 int dcount; /**< How many parts to this domain? (length of dvec) */
704 * Response generated by CGI, blocker, or error handler
708 char *status; /**< HTTP status (string). */
709 struct list headers[1]; /**< List of header lines. */
710 char *head; /**< Formatted http response head. */
711 size_t head_length; /**< Length of http response head. */
712 char *body; /**< HTTP document body. */
713 size_t content_length; /**< Length of body, REQUIRED if binary body. */
714 int is_static; /**< Nonzero if the content will never change and
715 should be cached by the browser (e.g. images). */
723 /** The string which was parsed to produce this url_spec.
724 Used for debugging or display only. */
727 char *dbuffer; /**< Buffer with '\0'-delimited domain name, or NULL to match all hosts. */
728 char **dvec; /**< List of pointers to the strings in dbuffer. */
729 int dcount; /**< How many parts to this domain? (length of dvec) */
730 int unanchored; /**< Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT. */
732 int port; /**< The port number, or 0 to match all ports. */
734 char *path; /**< The source for the regex. */
735 int pathlen; /**< ==strlen(path). Needed for prefix matching. FIXME: Now obsolete? */
736 regex_t *preg; /**< Regex for matching path part */
740 * If you declare a static url_spec, this is the value to initialize it to zero.
742 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, 0, 0, 0, NULL, 0, NULL }
745 * Constant for host part matching in URLs. If set, indicates that the start of
746 * the pattern must match the start of the URL. E.g. this is not set for the
747 * pattern ".example.com", so that it will match both "example.com" and
748 * "www.example.com". It is set for the pattern "example.com", which makes it
749 * match "example.com" only, not "www.example.com".
751 #define ANCHOR_LEFT 1
754 * Constant for host part matching in URLs. If set, indicates that the end of
755 * the pattern must match the end of the URL. E.g. this is not set for the
756 * pattern "ad.", so that it will match any host called "ad", irrespective
757 * of how many subdomains are in the fully-qualified domain name.
759 #define ANCHOR_RIGHT 2
763 * An I/O buffer. Holds a string which can be appended to, and can have data
764 * removed from the beginning.
768 char *buf; /**< Start of buffer */
769 char *cur; /**< Start of relevant data */
770 char *eod; /**< End of relevant data */
771 size_t size; /**< Size as malloc()ed */
776 * Return the number of bytes in the I/O buffer associated with the passed
777 * client_state pointer.
780 #define IOB_PEEK(CSP) ((CSP->iob->cur > CSP->iob->eod) ? (CSP->iob->eod - CSP->iob->cur) : 0)
784 * Remove any data in the I/O buffer associated with the passed
785 * client_state pointer.
787 #define IOB_RESET(CSP) if(CSP->iob->buf) free(CSP->iob->buf); memset(CSP->iob, '\0', sizeof(CSP->iob));
789 /* Bits for csp->content_type */
790 #define CT_TEXT 1 /**< csp->content_type bitmask:
791 Suitable for pcrs filtering. */
792 #define CT_GIF 2 /**< csp->content_type bitmask:
793 Suitable for GIF filtering. */
794 #define CT_TABOO 4 /**< csp->content_type bitmask:
795 DO NOT filter, irrespective of other flags. */
798 * The mask which includes all actions.
800 #define ACTION_MASK_ALL (~0UL)
803 * The most compatible set of actions - i.e. none.
805 #define ACTION_MOST_COMPATIBLE 0x00000000UL
807 /** Action bitmap: Block the request. */
808 #define ACTION_BLOCK 0x00000001UL
809 /** Action bitmap: Deanimate if it's a GIF. */
810 #define ACTION_DEANIMATE 0x00000002UL
811 /** Action bitmap: Downgrade HTTP/1.1 to 1.0. */
812 #define ACTION_DOWNGRADE 0x00000004UL
813 /** Action bitmap: Fast redirects. */
814 #define ACTION_FAST_REDIRECTS 0x00000008UL
815 /** Action bitmap: Remove existing "Forwarded" header, and do not add another. */
816 #define ACTION_HIDE_FORWARDED 0x00000010UL
817 /** Action bitmap: Hide "From" header. */
818 #define ACTION_HIDE_FROM 0x00000020UL
819 /** Action bitmap: Hide "Referer" header. (sic - follow HTTP, not English). */
820 #define ACTION_HIDE_REFERER 0x00000040UL
821 /** Action bitmap: Hide "User-Agent" and similar headers. */
822 #define ACTION_HIDE_USER_AGENT 0x00000080UL
823 /** Action bitmap: This is an image. */
824 #define ACTION_IMAGE 0x00000100UL
825 /** Action bitmap: Sets the image blocker. */
826 #define ACTION_IMAGE_BLOCKER 0x00000200UL
827 /** Action bitmap: Prevent compression. */
828 #define ACTION_NO_COMPRESSION 0x00000400UL
829 /** Action bitmap: Change cookies to session only cookies. */
830 #define ACTION_NO_COOKIE_KEEP 0x00000800UL
831 /** Action bitmap: Block rending cookies. */
832 #define ACTION_NO_COOKIE_READ 0x00001000UL
833 /** Action bitmap: Block setting cookies. */
834 #define ACTION_NO_COOKIE_SET 0x00002000UL
835 /** Action bitmap: Filter out popups. */
836 #define ACTION_NO_POPUPS 0x00004000UL
837 /** Action bitmap: Send a vanilla wafer. */
838 #define ACTION_VANILLA_WAFER 0x00008000UL
839 /** Action bitmap: Limit CONNECT requests to safe ports. */
840 #define ACTION_LIMIT_CONNECT 0x00010000UL
842 /** Action string index: How to deanimate GIFs */
843 #define ACTION_STRING_DEANIMATE 0
844 /** Action string index: Replacement for "From:" header */
845 #define ACTION_STRING_FROM 1
846 /** Action string index: How to block images */
847 #define ACTION_STRING_IMAGE_BLOCKER 2
848 /** Action string index: Replacement for "Referer:" header */
849 #define ACTION_STRING_REFERER 3
850 /** Action string index: Replacement for "User-Agent:" header */
851 #define ACTION_STRING_USER_AGENT 4
852 /** Action string index: Legal CONNECT ports. */
853 #define ACTION_STRING_LIMIT_CONNECT 5
854 /** Number of string actions. */
855 #define ACTION_STRING_COUNT 6
857 /** Index into current_action_spec::multi[] for headers to add. */
858 #define ACTION_MULTI_ADD_HEADER 0
859 /** Index into current_action_spec::multi[] for headers to add. */
860 #define ACTION_MULTI_WAFER 1
861 /** Index into current_action_spec::multi[] for filters to apply. */
862 #define ACTION_MULTI_FILTER 2
863 /** Number of multi-string actions. */
864 #define ACTION_MULTI_COUNT 3
868 * This structure contains a list of actions to apply to a URL.
869 * It only contains positive instructions - no "-" options.
870 * It is not used to store the actions list itself, only for
871 * url_actions() to return the current values.
873 struct current_action_spec
875 /** Actions to apply. A bit set to "1" means perform the action. */
879 * Paramaters for those actions that require them.
880 * Each entry is valid if & only if the corresponding entry in "flags" is
883 char * string[ACTION_STRING_COUNT];
885 /** Lists of strings for multi-string actions. */
886 struct list multi[ACTION_MULTI_COUNT][1];
891 * This structure contains a set of changes to actions.
892 * It can contain both positive and negative instructions.
893 * It is used to store an entry in the actions list.
897 unsigned long mask; /**< Actions to keep. A bit set to "0" means remove action. */
898 unsigned long add; /**< Actions to add. A bit set to "1" means add action. */
901 * Paramaters for those actions that require them.
902 * Each entry is valid if & only if the corresponding entry in "flags" is
905 char * string[ACTION_STRING_COUNT];
907 /** Lists of strings to remove, for multi-string actions. */
908 struct list multi_remove[ACTION_MULTI_COUNT][1];
910 /** If nonzero, remove *all* strings from the multi-string action. */
911 int multi_remove_all[ACTION_MULTI_COUNT];
913 /** Lists of strings to add, for multi-string actions. */
914 struct list multi_add[ACTION_MULTI_COUNT][1];
919 * This structure is used to store the actions list.
921 * It contains a URL pattern, and the chages to the actions.
922 * It is a linked list.
926 struct url_spec url[1]; /**< URL pattern. */
928 struct action_spec action[1]; /**< Actions. */
930 struct url_actions * next; /**< Next action in file, or NULL. */
935 * Flags for use in csp->flags
939 * Flag for csp->flags: Set if this client is processing data.
940 * Cleared when the thread associated with this structure dies.
942 #define CSP_FLAG_ACTIVE 0x01
945 * Flag for csp->flags: Set if the server's reply is in "chunked"
948 #define CSP_FLAG_CHUNKED 0x02
951 * Flag for csp->flags: Set if this request was enforced, although it would
952 * normally have been blocked.
954 #define CSP_FLAG_FORCED 0x04
957 * Flag for csp->flags: Set if any modification to the body was done.
959 #define CSP_FLAG_MODIFIED 0x08
962 * Flag for csp->flags: Set if request was blocked.
964 #define CSP_FLAG_REJECTED 0x10
967 * Flag for csp->flags: Set if we are toggled on (FEATURE_TOGGLE).
969 #define CSP_FLAG_TOGGLED_ON 0x20
973 * Maximum number of actions files. This limit is arbitrary - it's just used
976 #define MAX_ACTION_FILES 10
979 * The state of a Privoxy processing thread.
983 /** The proxy's configuration */
984 struct configuration_spec * config;
986 /** The actions to perform on the current request */
987 struct current_action_spec action[1];
989 /** socket to talk to client (web browser) */
992 /** socket to talk to server (web server or proxy) */
995 /** Multi-purpose flag container, see CSP_FLAG_* above */
996 unsigned short int flags;
998 /** Client PC's IP address, as reported by the accept() function.
1001 /** Client PC's IP address, as reported by the accept() function.
1005 /** Our IP address. I.e. the IP address that the client used to reach us,
1007 char *my_ip_addr_str;
1009 /** Our hostname. I.e. the reverse DNS of the IP address that the client
1010 used to reach us, as a string. */
1013 /** The URL that was requested */
1014 struct http_request http[1];
1016 /** An I/O buffer used for buffering data read from the network */
1019 /** List of all headers for this request */
1020 struct list headers[1];
1022 /** List of all cookies for this request */
1023 struct list cookie_list[1];
1025 /** MIME-Type key, see CT_* above */
1026 unsigned short int content_type;
1028 /** The "X-Forwarded-For:" header sent by the client */
1031 /** Actions files associated with this client */
1032 struct file_list *actions_list[MAX_ACTION_FILES];
1034 /** pcrs job file. */
1035 struct file_list *rlist;
1037 /** Length after content modification. */
1038 size_t content_length;
1040 #ifdef FEATURE_TRUST
1043 struct file_list *tlist;
1045 #endif /* def FEATURE_TRUST */
1047 /** Next thread in linked list. Only read or modify from the main thread! */
1048 struct client_state *next;
1053 * A function to add a header
1055 typedef jb_err (*add_header_func_ptr)(struct client_state *);
1058 * A function to process a header
1060 typedef jb_err (*parser_func_ptr )(struct client_state *, char **);
1064 * List of functions to run on a list of headers
1068 /** The header prefix to match */
1071 /** The length of the prefix to match */
1074 /** The function to apply to this line */
1075 parser_func_ptr parser;
1080 * List of available CGI functions.
1082 struct cgi_dispatcher
1084 /** The URL of the CGI, relative to the CGI root. */
1085 const char * const name;
1087 /** The handler function for the CGI */
1088 jb_err (* const handler)(struct client_state *csp, struct http_response *rsp, const struct map *parameters);
1090 /** The description of the CGI, to appear on the main menu, or NULL to hide it. */
1091 const char * const description;
1096 * A data file used by Privoxy. Kept in a linked list.
1101 * This is a pointer to the data structures associated with the file.
1102 * Read-only once the structure has been created.
1107 * The unloader function.
1108 * Normally NULL. When we are finished with file (i.e. when we have
1109 * loaded a new one), set to a pointer to an unloader function.
1110 * Unloader will be called by sweep() (called from main loop) when
1111 * all clients using this file are done. This prevents threading
1114 void (*unloader)(void *);
1117 * Used internally by sweep(). Do not access from elsewhere.
1122 * File last-modified time, so we can check if file has been changed.
1123 * Read-only once the structure has been created.
1125 time_t lastmodified;
1128 * The full filename.
1133 * Pointer to next entry in the linked list of all "file_list"s.
1134 * This linked list is so that sweep() can navigate it.
1135 * Since sweep() can remove items from the list, we must be careful
1136 * to only access this value from main thread (when we know sweep
1137 * won't be running).
1139 struct file_list *next;
1143 #ifdef FEATURE_TRUST
1146 * The format of a trust file when loaded into memory.
1150 struct url_spec url[1]; /**< The URL pattern */
1151 int reject; /**< FIXME: Please document this! */
1152 struct block_spec *next; /**< Next entry in linked list */
1155 #endif /* def FEATURE_TRUST */
1158 #define SOCKS_NONE 0 /**< Don't use a SOCKS server */
1159 #define SOCKS_4 40 /**< original SOCKS 4 protocol */
1160 #define SOCKS_4A 41 /**< as modified for hosts w/o external DNS */
1164 * How to forward a connection to a parent proxy.
1168 /** URL pattern that this forward_spec is for. */
1169 struct url_spec url[1];
1171 /** Connection type. Must be SOCKS_NONE, SOCKS_4, or SOCKS_4A. */
1174 /** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
1177 /** SOCKS server port. */
1180 /** Parent HTTP proxy hostname, or NULL for none. */
1183 /** Parent HTTP proxy port. */
1186 /** Next entry in the linked list. */
1187 struct forward_spec *next;
1192 * Initializer for a static struct forward_spec.
1194 #define FORWARD_SPEC_INITIALIZER { { URL_SPEC_INITIALIZER }, 0, NULL, 0, NULL, 0, NULL }
1198 * This struct represents one filter (one block) from
1199 * the re_filterfile. If there is more than one filter
1200 * in the file, the file will be represented by a
1201 * chained list of re_filterfile specs.
1203 struct re_filterfile_spec
1205 char *name; /**< Name from FILTER: statement in re_filterfile. */
1206 char *description; /**< Description from FILTER: statement in re_filterfile. */
1207 struct list patterns[1]; /**< The patterns from the re_filterfile. */
1208 pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
1209 struct re_filterfile_spec *next; /**< The pointer for chaining. */
1215 #define ACL_PERMIT 1 /**< Accept connection request */
1216 #define ACL_DENY 2 /**< Reject connection request */
1219 * An IP address pattern. Used to specify networks in the ACL.
1221 struct access_control_addr
1223 unsigned long addr; /**< The IP address as an integer. */
1224 unsigned long mask; /**< The network mask as an integer. */
1225 unsigned long port; /**< The port number. */
1229 * An access control list (ACL) entry.
1231 * This is a linked list.
1233 struct access_control_list
1235 struct access_control_addr src[1]; /**< Client IP address */
1236 struct access_control_addr dst[1]; /**< Website or parent proxy IP address */
1238 short action; /**< ACL_PERMIT or ACL_DENY */
1239 struct access_control_list *next; /**< The next entry in the ACL. */
1242 #endif /* def FEATURE_ACL */
1245 /** Maximum number of loaders (actions, re_filter, ...) */
1249 /** configuration_spec::feature_flags: CGI actions editor. */
1250 #define RUNTIME_FEATURE_CGI_EDIT_ACTIONS 1
1252 /** configuration_spec::feature_flags: Web-based toggle. */
1253 #define RUNTIME_FEATURE_CGI_TOGGLE 2
1257 * Data loaded from the configuration file.
1259 * (Anomaly: toggle is still handled through a global, not this structure)
1261 struct configuration_spec
1266 /** Nonzero to enable multithreading. */
1270 * Bitmask of features that can be enabled/disabled through the config
1271 * file. Currently defined bits:
1273 * - RUNTIME_FEATURE_CGI_EDIT_ACTIONS
1274 * - RUNTIME_FEATURE_CGI_TOGGLE
1276 unsigned feature_flags;
1278 /** The log file name. */
1279 const char *logfile;
1281 /** The config file directory. */
1282 const char *confdir;
1284 /** The log file directory. */
1287 /** The full paths to the actions files. */
1288 const char *actions_file[MAX_ACTION_FILES];
1290 /** The short names of the actions files. */
1291 const char *actions_file_short[MAX_ACTION_FILES];
1293 /** The administrator's email address */
1294 char *admin_address;
1296 /** A URL with info on this proxy */
1297 char *proxy_info_url;
1299 /** URL to the user manual (on our website or local copy) */
1302 /** The file name of the pcre filter file */
1303 const char *re_filterfile;
1305 #ifdef FEATURE_COOKIE_JAR
1307 /** The file name of the cookie jar file */
1308 const char * jarfile;
1310 /** The handle to the cookie jar file */
1313 #endif /* def FEATURE_COOKIE_JAR */
1315 /** IP address to bind to. Defaults to HADDR_DEFAULT == 127.0.0.1. */
1318 /** Port to bind to. Defaults to HADDR_PORT == 8118. */
1321 /** Size limit for IOB */
1322 size_t buffer_limit;
1324 #ifdef FEATURE_TRUST
1326 /** The file name of the trust file. */
1327 const char * trustfile;
1329 /** FIXME: DOCME: Document this. */
1330 struct list trust_info[1];
1332 /** FIXME: DOCME: Document this. */
1333 struct url_spec *trust_list[64];
1335 #endif /* def FEATURE_TRUST */
1339 /** The access control list (ACL). */
1340 struct access_control_list *acl;
1342 #endif /* def FEATURE_ACL */
1344 /** Information about parent proxies (forwarding). */
1345 struct forward_spec *forward;
1347 /** All options from the config file, HTML-formatted. */
1350 /** The configuration file object. */
1351 struct file_list *config_file_list;
1353 /** List of loaders */
1354 int (*loaders[NLOADERS])(struct client_state *);
1356 /** Nonzero if we need to bind() to the new port. */
1360 /** Calculates the number of elements in an array, using sizeof. */
1361 #define SZ(X) (sizeof(X) / sizeof(*X))
1363 #ifdef FEATURE_FORCE_LOAD
1364 /** The force load URL prefix. */
1365 #define FORCE_PREFIX "/PRIVOXY-FORCE"
1366 #endif /* def FEATURE_FORCE_LOAD */
1368 #ifdef FEATURE_NO_GIFS
1369 /** The MIME type for images ("image/png" or "image/gif"). */
1370 #define BUILTIN_IMAGE_MIMETYPE "image/png"
1372 #define BUILTIN_IMAGE_MIMETYPE "image/gif"
1373 #endif /* def FEATURE_NO_GIFS */
1380 /** URL for the Privoxy home page. */
1381 #define HOME_PAGE_URL "http://www.privoxy.org/"
1383 /** URL for the Privoxy user manual. */
1384 #define USER_MANUAL_URL HOME_PAGE_URL "/" VERSION "/user-manual/"
1386 /** Prefix for actions help links (append to USER_MANUAL_URL). */
1387 #define ACTIONS_HELP_PREFIX "actions-file.html#"
1389 /** Prefix for config option help links (append to USER_MANUAL_URL). */
1390 #define CONFIG_HELP_PREFIX "config.html#"
1393 * The "hosts" to intercept and display CGI pages.
1394 * First one is a hostname only, second one can specify host and path.
1397 * 1) Do not specify the http: prefix
1398 * 2) CGI_SITE_2_PATH must not end with /, one will be added automatically.
1399 * 3) CGI_SITE_2_PATH must start with /, unless it is the empty string.
1401 #define CGI_SITE_1_HOST "p.p"
1402 #define CGI_SITE_2_HOST "config.privoxy.org"
1403 #define CGI_SITE_2_PATH ""
1406 * The prefix for CGI pages. Written out in generated HTML.
1407 * INCLUDES the trailing slash.
1409 #define CGI_PREFIX "http://" CGI_SITE_2_HOST CGI_SITE_2_PATH "/"
1414 * FIXME: This is very inefficient. There could be one copy of these strings
1415 * for each .c file!! They should be "extern", not "static".
1417 static const char CSUCCEED[] =
1418 "HTTP/1.0 200 Connection established\n"
1419 "Proxy-Agent: Privoxy/" VERSION "\r\n\r\n";
1421 static const char CHEADER[] =
1422 "HTTP/1.0 400 Invalid header received from browser\r\n\r\n";
1424 static const char CFORBIDDEN[] =
1425 "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";
1431 #endif /* ndef PROJECT_H_INCLUDED */