1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
4 #define PROJECT_H_VERSION "$Id: project.h,v 1.70 2002/05/12 16:05:50 jongfoster 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 1.70 2002/05/12 16:05:50 jongfoster
41 * Fixing ACTION_MASK_ALL to be unsigned long rather than
42 * just unsigned int. I don't know if anyone is porting
43 * Privoxy to 16-bit platforms, but if so, +limit-connect
44 * wouldn't have worked because of this bug.
46 * Revision 1.69 2002/05/08 16:00:16 oes
47 * Added size member to struct iob, so it can
48 * be alloced larger than needed.
50 * Revision 1.68 2002/04/26 12:56:00 oes
51 * Killed REDIRECT_URL, added USER_MANUAL_URL and HELP_LINK_PREFIX
53 * Revision 1.67 2002/04/24 02:12:43 oes
54 * - Jon's multiple AF patch:
55 * - Make csp->actions_list an array
56 * - #define MAX_ACTION_FILES
57 * - Moved CGI_PARAM_LEN_MAX (500) here
59 * Revision 1.66 2002/04/15 19:06:43 jongfoster
62 * Revision 1.65 2002/04/04 00:36:36 gliptak
63 * always use pcre for matching
65 * Revision 1.64 2002/04/03 22:28:03 gliptak
66 * Removed references to gnu_regex
68 * Revision 1.63 2002/03/31 17:19:00 jongfoster
69 * Win32 only: Enabling STRICT to fix a VC++ compile warning.
71 * Revision 1.62 2002/03/26 22:48:49 swa
74 * Revision 1.61 2002/03/26 22:29:55 swa
75 * we have a new homepage!
77 * Revision 1.60 2002/03/24 15:52:17 jongfoster
78 * Changing CGI URL prefixes for new name
80 * Revision 1.59 2002/03/24 15:23:33 jongfoster
83 * Revision 1.58 2002/03/24 13:25:43 swa
84 * name change related issues
86 * Revision 1.57 2002/03/16 20:28:34 oes
87 * Added descriptions to the filters so users will know what they select in the cgi editor
89 * Revision 1.56 2002/03/13 20:27:30 oes
90 * Fixing bug with CT_TABOO
92 * Revision 1.55 2002/03/12 01:42:50 oes
93 * Introduced modular filters
95 * Revision 1.54 2002/03/09 20:03:52 jongfoster
96 * - Making various functions return int rather than size_t.
97 * (Undoing a recent change). Since size_t is unsigned on
98 * Windows, functions like read_socket that return -1 on
99 * error cannot return a size_t.
101 * THIS WAS A MAJOR BUG - it caused frequent, unpredictable
102 * crashes, and also frequently caused JB to jump to 100%
103 * CPU and stay there. (Because it thought it had just
104 * read ((unsigned)-1) == 4Gb of data...)
106 * - The signature of write_socket has changed, it now simply
107 * returns success=0/failure=nonzero.
109 * - Trying to get rid of a few warnings --with-debug on
110 * Windows, I've introduced a new type "jb_socket". This is
111 * used for the socket file descriptors. On Windows, this
112 * is SOCKET (a typedef for unsigned). Everywhere else, it's
113 * an int. The error value can't be -1 any more, so it's
114 * now JB_INVALID_SOCKET (which is -1 on UNIX, and in
115 * Windows it maps to the #define INVALID_SOCKET.)
117 * - The signature of bind_port has changed.
119 * Revision 1.53 2002/03/08 16:48:55 oes
120 * Added FEATURE_NO_GIFS and BUILTIN_IMAGE_MIMETYPE
122 * Revision 1.52 2002/03/07 03:46:17 oes
123 * Fixed compiler warnings
125 * Revision 1.51 2002/03/05 04:52:42 oes
126 * Deleted non-errlog debugging code
128 * Revision 1.50 2002/03/04 19:32:07 oes
129 * Changed default port to 8118
131 * Revision 1.49 2002/03/04 18:28:55 oes
132 * Deleted PID_FILE_NAME
134 * Revision 1.48 2002/03/03 14:50:40 oes
135 * Fixed CLF logging: Added ocmd member for client's request to struct http_request
137 * Revision 1.47 2002/02/20 23:15:13 jongfoster
138 * Parsing functions now handle out-of-memory gracefully by returning
141 * Revision 1.46 2002/01/17 21:06:09 jongfoster
142 * Now #defining the URLs of the config interface
144 * Minor changes to struct http_request and struct url_spec due to
145 * standardizing that struct http_request is used to represent a URL, and
146 * struct url_spec is used to represent a URL pattern. (Before, URLs were
147 * represented as seperate variables and a partially-filled-in url_spec).
149 * Revision 1.45 2002/01/09 14:33:27 oes
150 * Added HOSTENT_BUFFER_SIZE
152 * Revision 1.44 2001/12/30 14:07:32 steudten
153 * - Add signal handling (unix)
154 * - Add SIGHUP handler (unix)
155 * - Add creation of pidfile (unix)
156 * - Add action 'top' in rc file (RH)
157 * - Add entry 'SIGNALS' to manpage
158 * - Add exit message to logfile (unix)
160 * Revision 1.43 2001/11/22 21:57:51 jongfoster
161 * Making action_spec->flags into an unsigned long rather than just an
163 * Adding ACTION_NO_COOKIE_KEEP
165 * Revision 1.42 2001/11/05 21:42:41 steudten
166 * Include DBG() macro.
168 * Revision 1.41 2001/10/28 19:12:06 jongfoster
169 * Adding ijb_toupper()
171 * Revision 1.40 2001/10/26 17:40:47 oes
172 * Moved ijb_isspace and ijb_tolower to project.h
173 * Removed http->user_agent, csp->referrer and csp->accept_types
175 * Revision 1.39 2001/10/25 03:45:02 david__schmidt
176 * Adding a (void*) cast to freez() because Visual Age C++ won't expand the
177 * macro when called with a cast; so moving the cast to the macro def'n
178 * seems to both eliminate compiler warnings (on darwin and OS/2, anyway) and
179 * doesn't make macro expansion complain. Hope this works for everyone else
182 * Revision 1.38 2001/10/23 21:19:04 jongfoster
183 * New error-handling support: jb_err type and JB_ERR_xxx constants
184 * CGI functions now return a jb_err, and their parameters map is const.
185 * Support for RUNTIME_FEATUREs to enable/disable config editor
186 * Adding a few comments
188 * Revision 1.37 2001/10/14 22:14:01 jongfoster
189 * Removing name_length field from struct cgi_dispatcher, as this is
190 * now calculated at runtime from the "name" field.
192 * Revision 1.36 2001/10/10 16:45:15 oes
193 * Added LIMIT_CONNECT action and string
194 * Fixed HTTP message line termination
195 * Added CFORBIDDEN HTTP message
197 * Revision 1.35 2001/10/07 18:06:43 oes
198 * Added status member to struct http_request
200 * Revision 1.34 2001/10/07 15:45:25 oes
201 * Added url member to struct http_request and commented all
206 * Added ACTION_DOWNGRADE and ACTION_NO_COMPRESSION
208 * Replaced struct client_state members rejected,
209 * force, active and toggled_on with "flags" bitmap.
211 * Added CSP_FLAG_MODIFIED and CSP_FLAG_CHUNKED
213 * Added buffer_limit to struct configuration_spec
215 * Revision 1.33 2001/09/20 13:30:08 steudten
217 * Make freez() more secure in case of: if (exp) { free(z) ; a=*z }
218 * Last case will set z to NULL in free(z) and thats bad..
220 * Revision 1.32 2001/09/16 23:02:51 jongfoster
223 * Revision 1.31 2001/09/16 13:20:29 jongfoster
224 * Rewrite of list library. Now has seperate header and list_entry
225 * structures. Also added a large sprinking of assert()s to the list
228 * Revision 1.30 2001/09/13 23:52:00 jongfoster
229 * Support for both static and dynamically generated CGI pages
231 * Revision 1.29 2001/09/13 23:29:43 jongfoster
232 * Defining FORWARD_SPEC_INITIALIZER
234 * Revision 1.28 2001/09/13 23:05:50 jongfoster
235 * Changing the string paramater to the header parsers a "const".
237 * Revision 1.27 2001/08/05 16:06:20 jongfoster
238 * Modifiying "struct map" so that there are now separate header and
239 * "map_entry" structures. This means that functions which modify a
240 * map no longer need to return a pointer to the modified map.
241 * Also, it no longer reverses the order of the entries (which may be
242 * important with some advanced template substitutions).
244 * Revision 1.26 2001/07/30 22:08:36 jongfoster
245 * Tidying up #defines:
246 * - All feature #defines are now of the form FEATURE_xxx
247 * - Permanently turned off WIN_GUI_EDIT
248 * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
250 * Revision 1.25 2001/07/29 18:43:08 jongfoster
251 * Changing #ifdef _FILENAME_H to FILENAME_H_INCLUDED, to conform to
254 * Revision 1.24 2001/07/25 17:20:27 oes
255 * Introduced http->user_agent
257 * Revision 1.23 2001/07/18 12:32:23 oes
258 * - Added ACTION_STRING_DEANIMATE
259 * - moved #define freez from jcc.h to project.h
261 * Revision 1.22 2001/07/15 17:51:41 jongfoster
262 * Renaming #define STATIC to STATIC_PCRE
264 * Revision 1.21 2001/07/13 14:03:19 oes
265 * - Reorganized regex header inclusion and #defines to
266 * comply to the scheme in configure.in
267 * - Added csp->content_type and its CT_* keys
268 * - Added ACTION_DEANIMATE
269 * - Removed all #ifdef PCRS
271 * Revision 1.20 2001/06/29 21:45:41 oes
272 * Indentation, CRLF->LF, Tab-> Space
274 * Revision 1.19 2001/06/29 13:33:36 oes
275 * - Improved comments
276 * - Introduced http_request.host_ip_addr_str
277 * - Introduced http_response.head_length
278 * - Introduced config.my_ip_addr_str, config.my_hostname,
279 * config.admin_address and config.proxy_info_url
280 * - Removed config.proxy_args_header and config.proxy_args_trailer,
281 * renamed config.proxy_args_invocation to config.proxy_args
282 * - Removed HTML snipplets and GIFs
283 * - Removed logentry from cancelled commit
285 * Revision 1.18 2001/06/09 10:57:39 jongfoster
286 * Adding definition of BUFFER_SIZE.
287 * Changing struct cgi_dispatcher to use "const" strings.
289 * Revision 1.17 2001/06/07 23:15:09 jongfoster
290 * Merging ACL and forward files into config file.
291 * Moving struct gateway members into struct forward_spec
292 * Removing config->proxy_args_gateways
293 * Cosmetic: Adding a few comments
295 * Revision 1.16 2001/06/04 18:31:58 swa
296 * files are now prefixed with either `confdir' or `logdir'.
297 * `make redhat-dist' replaces both entries confdir and logdir
300 * Revision 1.15 2001/06/04 11:28:53 swa
301 * redirect did not work due to missing /
303 * Revision 1.14 2001/06/03 11:03:48 oes
305 * added struct http_response,
306 * changed struct interceptors to struct cgi_dispatcher,
307 * moved HTML stuff to cgi.h
309 * Revision 1.13 2001/06/01 20:05:36 jongfoster
310 * Support for +image-blocker{}: added ACTION_IMAGE_BLOCKER
311 * constant, and removed csp->tinygif.
313 * Revision 1.12 2001/06/01 18:49:17 jongfoster
314 * Replaced "list_share" with "list" - the tiny memory gain was not
315 * worth the extra complexity.
317 * Revision 1.11 2001/06/01 10:32:47 oes
318 * Added constants for anchoring selection bitmap
320 * Revision 1.10 2001/05/31 21:33:53 jongfoster
321 * Changes for new actions file, replacing permissionsfile
322 * and parts of the config file. Also added support for
325 * Revision 1.9 2001/05/31 17:32:31 oes
327 * - Enhanced domain part globbing with infix and prefix asterisk
328 * matching and optional unanchored operation
330 * Revision 1.8 2001/05/29 20:09:15 joergs
331 * HTTP_REDIRECT_TEMPLATE fixed.
333 * Revision 1.7 2001/05/29 09:50:24 jongfoster
334 * Unified blocklist/imagelist/actionslist.
335 * File format is still under discussion, but the internal changes
338 * Also modified interceptor behaviour:
339 * - We now intercept all URLs beginning with one of the following
340 * prefixes (and *only* these prefixes):
342 * * http://ijbswa.sf.net/config/
343 * * http://ijbswa.sourceforge.net/config/
344 * - New interceptors "home page" - go to http://i.j.b/ to see it.
345 * - Internal changes so that intercepted and fast redirect pages
346 * are not replaced with an image.
347 * - Interceptors now have the option to send a binary page direct
348 * to the client. (i.e. ijb-send-banner uses this)
349 * - Implemented show-url-info interceptor. (Which is why I needed
350 * the above interceptors changes - a typical URL is
351 * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif".
352 * The previous mechanism would not have intercepted that, and
353 * if it had been intercepted then it then it would have replaced
356 * Revision 1.6 2001/05/27 22:17:04 oes
358 * - re_process_buffer no longer writes the modified buffer
359 * to the client, which was very ugly. It now returns the
360 * buffer, which it is then written by chat.
362 * - content_length now adjusts the Content-Length: header
363 * for modified documents rather than crunch()ing it.
364 * (Length info in csp->content_length, which is 0 for
365 * unmodified documents)
367 * - For this to work, sed() is called twice when filtering.
369 * Revision 1.5 2001/05/26 00:28:36 jongfoster
370 * Automatic reloading of config file.
371 * Removed obsolete SIGHUP support (Unix) and Reload menu option (Win32).
372 * Most of the global variables have been moved to a new
373 * struct configuration_spec, accessed through csp->config->globalname
374 * Most of the globals remaining are used by the Win32 GUI.
376 * Revision 1.4 2001/05/22 18:46:04 oes
378 * - Enabled filtering banners by size rather than URL
379 * by adding patterns that replace all standard banner
380 * sizes with the "Junkbuster" gif to the re_filterfile
382 * - Enabled filtering WebBugs by providing a pattern
383 * which kills all 1x1 images
385 * - Added support for PCRE_UNGREEDY behaviour to pcrs,
386 * which is selected by the (nonstandard and therefore
387 * capital) letter 'U' in the option string.
388 * It causes the quantifiers to be ungreedy by default.
389 * Appending a ? turns back to greedy (!).
391 * - Added a new interceptor ijb-send-banner, which
392 * sends back the "Junkbuster" gif. Without imagelist or
393 * MSIE detection support, or if tinygif = 1, or the
394 * URL isn't recognized as an imageurl, a lame HTML
395 * explanation is sent instead.
397 * - Added new feature, which permits blocking remote
398 * script redirects and firing back a local redirect
400 * The feature is conditionally compiled, i.e. it
401 * can be disabled with --disable-fast-redirects,
402 * plus it must be activated by a "fast-redirects"
403 * line in the config file, has its own log level
404 * and of course wants to be displayed by show-proxy-args
405 * Note: Boy, all the #ifdefs in 1001 locations and
406 * all the fumbling with configure.in and acconfig.h
407 * were *way* more work than the feature itself :-(
409 * - Because a generic redirect template was needed for
410 * this, tinygif = 3 now uses the same.
412 * - Moved GIFs, and other static HTTP response templates
417 * - Removed some >400 CRs again (Jon, you really worked
420 * Revision 1.3 2001/05/20 01:21:20 jongfoster
421 * Version 2.9.4 checkin.
422 * - Merged popupfile and cookiefile, and added control over PCRS
423 * filtering, in new "actionsfile".
424 * - Implemented LOG_LEVEL_FATAL, so that if there is a configuration
425 * file error you now get a message box (in the Win32 GUI) rather
426 * than the program exiting with no explanation.
427 * - Made killpopup use the PCRS MIME-type checking and HTTP-header
429 * - Removed tabs from "config"
430 * - Moved duplicated url parsing code in "loaders.c" to a new funcition.
431 * - Bumped up version number.
433 * Revision 1.2 2001/05/17 23:01:01 oes
434 * - Cleaned CRLF's from the sources and related files
436 * Revision 1.1.1.1 2001/05/15 13:59:03 oes
437 * Initial import of version 2.9.3 source tree
440 *********************************************************************/
443 /* Declare struct FILE for vars and funcs. */
446 /* Need time_t for file_list */
450 * Include appropriate regular expression libraries.
451 * Note that pcrs and pcre (native) are needed for cgi
452 * and are included anyway.
468 # include "pcreposix.h"
470 # include <pcreposix.h>
475 #endif /* def AMIGA */
479 * I don't want to have to #include all this just for the declaration
480 * of SOCKET. However, it looks like we have to...
495 typedef SOCKET jb_socket;
497 #define JB_INVALID_SOCKET INVALID_SOCKET
499 #else /* ndef _WIN32 */
502 * The type used by sockets. On UNIX it's an int. Microsoft decided to
503 * make it an unsigned.
505 typedef int jb_socket;
508 * The error value used for variables of type jb_socket. On UNIX this
509 * is -1, however Microsoft decided to make socket handles unsigned, so
510 * they use a different value.
513 #define JB_INVALID_SOCKET (-1)
515 #endif /* ndef _WIN32 */
519 * A standard error code. This should be JB_ERR_OK or one of the JB_ERR_xxx
524 #define JB_ERR_OK 0 /**< Success, no error */
525 #define JB_ERR_MEMORY 1 /**< Out of memory */
526 #define JB_ERR_CGI_PARAMS 2 /**< Missing or corrupt CGI parameters */
527 #define JB_ERR_FILE 3 /**< Error opening, reading or writing a file */
528 #define JB_ERR_PARSE 4 /**< Error parsing file */
529 #define JB_ERR_MODIFIED 5 /**< File has been modified outside of the
530 CGI actions editor. */
534 * This macro is used to free a pointer that may be NULL.
535 * It also sets the variable to NULL after it's been freed.
536 * The paramater should be a simple variable without side effects.
538 #define freez(X) { if(X) { free((void*)X); X = NULL ; } }
542 * Fix a problem with Solaris. There should be no effect on other
545 * Solaris's isspace() is a macro which uses it's argument directly
546 * as an array index. Therefore we need to make sure that high-bit
547 * characters generate +ve values, and ideally we also want to make
548 * the argument match the declared parameter type of "int".
550 * Note: Remember to #include <ctype.h> if you use these macros.
552 #define ijb_toupper(__X) toupper((int)(unsigned char)(__X))
553 #define ijb_tolower(__X) tolower((int)(unsigned char)(__X))
554 #define ijb_isspace(__X) isspace((int)(unsigned char)(__X))
557 * Use for statically allocated buffers if you have no other choice.
558 * Remember to check the length of what you write into the buffer
559 * - we don't want any buffer overflows!
561 #define BUFFER_SIZE 5000
564 * Max length of CGI parameters (arbitrary limit).
566 #define CGI_PARAM_LEN_MAX 500
569 * Buffer size for capturing struct hostent data in the
570 * gethostby(name|addr)_r library calls. Since we don't
571 * loop over gethostbyname_r, the buffer must be sufficient
572 * to accomodate multiple IN A RRs, as used in DNS round robin
573 * load balancing. W3C's wwwlib uses 1K, so that should be
574 * good enough for us, too.
576 #define HOSTENT_BUFFER_SIZE 1024
579 * Do not use. Originally this was so that you can
580 * say "while (FOREVER) { ...do something... }".
581 * However, this gives a warning with some compilers (e.g. VC++).
582 * Instead, use "for (;;) { ...do something... }".
587 * Default IP address to listen on, as a string.
588 * Set to "127.0.0.1".
590 #define HADDR_DEFAULT "127.0.0.1"
593 * Default port to listen on, as a number.
596 #define HADDR_PORT 8118
599 /* Forward def for struct client_state */
600 struct configuration_spec;
604 * Entry in a linked list of strings.
609 * The string. The "const" is only to discourage modification,
610 * you can actually change it if you *really* want to.
611 * You can even freez() it and replace it with another
612 * malloc()d string. If you replace it with NULL, the list
613 * functions will work, just be careful next time you iterate
614 * through the list in your own code.
616 * FIXME: Should we remove the "const"?
620 /** Next entry in the linked list, or NULL if no more. */
621 struct list_entry *next;
625 * A header for a linked list of strings.
629 /** First entry in the list, or NULL if the list is empty. */
630 struct list_entry *first;
632 /** Last entry in the list, or NULL if the list is empty. */
633 struct list_entry *last;
638 * An entry in a map. This is a name=value pair.
642 /** The key for the map. */
644 /** The value associated with that key. */
646 /** The next map entry, or NULL if none. */
647 struct map_entry *next;
651 * A map from a string to another string.
652 * This is used for the paramaters passed in a HTTP GET request, and
653 * to store the exports when the CGI interface is filling in a template.
657 /** The first map entry, or NULL if the map is empty. */
658 struct map_entry *first;
659 /** The last map entry, or NULL if the map is empty. */
660 struct map_entry *last;
665 * A HTTP request. This includes the method (GET, POST) and
668 * This is also used whenever we want to match a URL against a
669 * URL pattern. This always contains the URL to match, and never
670 * a URL pattern. (See struct url_spec).
674 char *cmd; /**< Whole command line: method, URL, Version */
675 char *ocmd; /**< Backup of original cmd for CLF logging */
676 char *gpc; /**< HTTP method: GET, POST, ... */
677 char *url; /**< The URL */
678 char *ver; /**< Protocol version */
679 int status; /**< HTTP Status */
681 char *host; /**< Host part of URL */
682 int port; /**< Port of URL or 80 (default) */
683 char *path; /**< Path of URL */
684 char *hostport; /**< host[:port] */
685 int ssl; /**< Flag if protocol is https */
687 char *host_ip_addr_str; /**< String with dotted decimal representation
688 of host's IP. NULL before connect_to() */
690 char *dbuffer; /**< Buffer with '\0'-delimited domain name. */
691 char **dvec; /**< List of pointers to the strings in dbuffer. */
692 int dcount; /**< How many parts to this domain? (length of dvec) */
697 * Response generated by CGI, blocker, or error handler
701 char *status; /**< HTTP status (string). */
702 struct list headers[1]; /**< List of header lines. */
703 char *head; /**< Formatted http response head. */
704 size_t head_length; /**< Length of http response head. */
705 char *body; /**< HTTP document body. */
706 size_t content_length; /**< Length of body, REQUIRED if binary body. */
707 int is_static; /**< Nonzero if the content will never change and
708 should be cached by the browser (e.g. images). */
716 /** The string which was parsed to produce this url_spec.
717 Used for debugging or display only. */
720 char *dbuffer; /**< Buffer with '\0'-delimited domain name, or NULL to match all hosts. */
721 char **dvec; /**< List of pointers to the strings in dbuffer. */
722 int dcount; /**< How many parts to this domain? (length of dvec) */
723 int unanchored; /**< Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT. */
725 int port; /**< The port number, or 0 to match all ports. */
727 char *path; /**< The source for the regex. */
728 int pathlen; /**< ==strlen(path). Needed for prefix matching. FIXME: Now obsolete? */
729 regex_t *preg; /**< Regex for matching path part */
733 * If you declare a static url_spec, this is the value to initialize it to zero.
735 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, 0, 0, 0, NULL, 0, NULL }
738 * Constant for host part matching in URLs. If set, indicates that the start of
739 * the pattern must match the start of the URL. E.g. this is not set for the
740 * pattern ".example.com", so that it will match both "example.com" and
741 * "www.example.com". It is set for the pattern "example.com", which makes it
742 * match "example.com" only, not "www.example.com".
744 #define ANCHOR_LEFT 1
747 * Constant for host part matching in URLs. If set, indicates that the end of
748 * the pattern must match the end of the URL. E.g. this is not set for the
749 * pattern "ad.", so that it will match any host called "ad", irrespective
750 * of how many subdomains are in the fully-qualified domain name.
752 #define ANCHOR_RIGHT 2
756 * An I/O buffer. Holds a string which can be appended to, and can have data
757 * removed from the beginning.
761 char *buf; /**< Start of buffer */
762 char *cur; /**< Start of relevant data */
763 char *eod; /**< End of relevant data */
764 size_t size; /**< Size as malloc()ed */
769 * Return the number of bytes in the I/O buffer associated with the passed
770 * client_state pointer.
773 #define IOB_PEEK(CSP) ((CSP->iob->cur > CSP->iob->eod) ? (CSP->iob->eod - CSP->iob->cur) : 0)
777 * Remove any data in the I/O buffer associated with the passed
778 * client_state pointer.
780 #define IOB_RESET(CSP) if(CSP->iob->buf) free(CSP->iob->buf); memset(CSP->iob, '\0', sizeof(CSP->iob));
782 /* Bits for csp->content_type */
783 #define CT_TEXT 1 /**< csp->content_type bitmask:
784 Suitable for pcrs filtering. */
785 #define CT_GIF 2 /**< csp->content_type bitmask:
786 Suitable for GIF filtering. */
787 #define CT_TABOO 4 /**< csp->content_type bitmask:
788 DO NOT filter, irrespective of other flags. */
791 * The mask which includes all actions.
793 #define ACTION_MASK_ALL (~0UL)
796 * The most compatible set of actions - i.e. none.
798 #define ACTION_MOST_COMPATIBLE 0x00000000UL
800 /** Action bitmap: Block the request. */
801 #define ACTION_BLOCK 0x00000001UL
802 /** Action bitmap: Deanimate if it's a GIF. */
803 #define ACTION_DEANIMATE 0x00000002UL
804 /** Action bitmap: Downgrade HTTP/1.1 to 1.0. */
805 #define ACTION_DOWNGRADE 0x00000004UL
806 /** Action bitmap: Fast redirects. */
807 #define ACTION_FAST_REDIRECTS 0x00000008UL
808 /** Action bitmap: Remove existing "Forwarded" header, and do not add another. */
809 #define ACTION_HIDE_FORWARDED 0x00000010UL
810 /** Action bitmap: Hide "From" header. */
811 #define ACTION_HIDE_FROM 0x00000020UL
812 /** Action bitmap: Hide "Referer" header. (sic - follow HTTP, not English). */
813 #define ACTION_HIDE_REFERER 0x00000040UL
814 /** Action bitmap: Hide "User-Agent" and similar headers. */
815 #define ACTION_HIDE_USER_AGENT 0x00000080UL
816 /** Action bitmap: This is an image. */
817 #define ACTION_IMAGE 0x00000100UL
818 /** Action bitmap: Sets the image blocker. */
819 #define ACTION_IMAGE_BLOCKER 0x00000200UL
820 /** Action bitmap: Prevent compression. */
821 #define ACTION_NO_COMPRESSION 0x00000400UL
822 /** Action bitmap: Change cookies to session only cookies. */
823 #define ACTION_NO_COOKIE_KEEP 0x00000800UL
824 /** Action bitmap: Block rending cookies. */
825 #define ACTION_NO_COOKIE_READ 0x00001000UL
826 /** Action bitmap: Block setting cookies. */
827 #define ACTION_NO_COOKIE_SET 0x00002000UL
828 /** Action bitmap: Filter out popups. */
829 #define ACTION_NO_POPUPS 0x00004000UL
830 /** Action bitmap: Send a vanilla wafer. */
831 #define ACTION_VANILLA_WAFER 0x00008000UL
832 /** Action bitmap: Limit CONNECT requests to safe ports. */
833 #define ACTION_LIMIT_CONNECT 0x00010000UL
835 /** Action string index: How to deanimate GIFs */
836 #define ACTION_STRING_DEANIMATE 0
837 /** Action string index: Replacement for "From:" header */
838 #define ACTION_STRING_FROM 1
839 /** Action string index: How to block images */
840 #define ACTION_STRING_IMAGE_BLOCKER 2
841 /** Action string index: Replacement for "Referer:" header */
842 #define ACTION_STRING_REFERER 3
843 /** Action string index: Replacement for "User-Agent:" header */
844 #define ACTION_STRING_USER_AGENT 4
845 /** Action string index: Legal CONNECT ports. */
846 #define ACTION_STRING_LIMIT_CONNECT 5
847 /** Number of string actions. */
848 #define ACTION_STRING_COUNT 6
850 /** Index into current_action_spec::multi[] for headers to add. */
851 #define ACTION_MULTI_ADD_HEADER 0
852 /** Index into current_action_spec::multi[] for headers to add. */
853 #define ACTION_MULTI_WAFER 1
854 /** Index into current_action_spec::multi[] for filters to apply. */
855 #define ACTION_MULTI_FILTER 2
856 /** Number of multi-string actions. */
857 #define ACTION_MULTI_COUNT 3
861 * This structure contains a list of actions to apply to a URL.
862 * It only contains positive instructions - no "-" options.
863 * It is not used to store the actions list itself, only for
864 * url_actions() to return the current values.
866 struct current_action_spec
868 /** Actions to apply. A bit set to "1" means perform the action. */
872 * Paramaters for those actions that require them.
873 * Each entry is valid if & only if the corresponding entry in "flags" is
876 char * string[ACTION_STRING_COUNT];
878 /** Lists of strings for multi-string actions. */
879 struct list multi[ACTION_MULTI_COUNT][1];
884 * This structure contains a set of changes to actions.
885 * It can contain both positive and negative instructions.
886 * It is used to store an entry in the actions list.
890 unsigned long mask; /**< Actions to keep. A bit set to "0" means remove action. */
891 unsigned long add; /**< Actions to add. A bit set to "1" means add 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 to remove, for multi-string actions. */
901 struct list multi_remove[ACTION_MULTI_COUNT][1];
903 /** If nonzero, remove *all* strings from the multi-string action. */
904 int multi_remove_all[ACTION_MULTI_COUNT];
906 /** Lists of strings to add, for multi-string actions. */
907 struct list multi_add[ACTION_MULTI_COUNT][1];
912 * This structure is used to store the actions list.
914 * It contains a URL pattern, and the chages to the actions.
915 * It is a linked list.
919 struct url_spec url[1]; /**< URL pattern. */
921 struct action_spec action[1]; /**< Actions. */
923 struct url_actions * next; /**< Next action in file, or NULL. */
928 * Flags for use in csp->flags
932 * Flag for csp->flags: Set if this client is processing data.
933 * Cleared when the thread associated with this structure dies.
935 #define CSP_FLAG_ACTIVE 0x01
938 * Flag for csp->flags: Set if the server's reply is in "chunked"
941 #define CSP_FLAG_CHUNKED 0x02
944 * Flag for csp->flags: Set if this request was enforced, although it would
945 * normally have been blocked.
947 #define CSP_FLAG_FORCED 0x04
950 * Flag for csp->flags: Set if any modification to the body was done.
952 #define CSP_FLAG_MODIFIED 0x08
955 * Flag for csp->flags: Set if request was blocked.
957 #define CSP_FLAG_REJECTED 0x10
960 * Flag for csp->flags: Set if we are toggled on (FEATURE_TOGGLE).
962 #define CSP_FLAG_TOGGLED_ON 0x20
966 * Maximum number of actions files. This limit is arbitrary - it's just used
969 #define MAX_ACTION_FILES 10
972 * The state of a Privoxy processing thread.
976 /** The proxy's configuration */
977 struct configuration_spec * config;
979 /** The actions to perform on the current request */
980 struct current_action_spec action[1];
982 /** socket to talk to client (web browser) */
985 /** socket to talk to server (web server or proxy) */
988 /** Multi-purpose flag container, see CSP_FLAG_* above */
989 unsigned short int flags;
991 /** Client PC's IP address, as reported by the accept() function.
994 /** Client PC's IP address, as reported by the accept() function.
998 /** Our IP address. I.e. the IP address that the client used to reach us,
1000 char *my_ip_addr_str;
1002 /** Our hostname. I.e. the reverse DNS of the IP address that the client
1003 used to reach us, as a string. */
1006 /** The URL that was requested */
1007 struct http_request http[1];
1009 /** An I/O buffer used for buffering data read from the network */
1012 /** List of all headers for this request */
1013 struct list headers[1];
1015 /** List of all cookies for this request */
1016 struct list cookie_list[1];
1018 /** MIME-Type key, see CT_* above */
1019 unsigned short int content_type;
1021 /** The "X-Forwarded-For:" header sent by the client */
1024 /** Actions files associated with this client */
1025 struct file_list *actions_list[MAX_ACTION_FILES];
1027 /** pcrs job file. */
1028 struct file_list *rlist;
1030 /** Length after content modification. */
1031 size_t content_length;
1033 #ifdef FEATURE_TRUST
1036 struct file_list *tlist;
1038 #endif /* def FEATURE_TRUST */
1040 /** Next thread in linked list. Only read or modify from the main thread! */
1041 struct client_state *next;
1046 * A function to add a header
1048 typedef jb_err (*add_header_func_ptr)(struct client_state *);
1051 * A function to process a header
1053 typedef jb_err (*parser_func_ptr )(struct client_state *, char **);
1057 * List of functions to run on a list of headers
1061 /** The header prefix to match */
1064 /** The length of the prefix to match */
1067 /** The function to apply to this line */
1068 parser_func_ptr parser;
1073 * List of available CGI functions.
1075 struct cgi_dispatcher
1077 /** The URL of the CGI, relative to the CGI root. */
1078 const char * const name;
1080 /** The handler function for the CGI */
1081 jb_err (* const handler)(struct client_state *csp, struct http_response *rsp, const struct map *parameters);
1083 /** The description of the CGI, to appear on the main menu, or NULL to hide it. */
1084 const char * const description;
1089 * A data file used by Privoxy. Kept in a linked list.
1094 * This is a pointer to the data structures associated with the file.
1095 * Read-only once the structure has been created.
1100 * The unloader function.
1101 * Normally NULL. When we are finished with file (i.e. when we have
1102 * loaded a new one), set to a pointer to an unloader function.
1103 * Unloader will be called by sweep() (called from main loop) when
1104 * all clients using this file are done. This prevents threading
1107 void (*unloader)(void *);
1110 * Used internally by sweep(). Do not access from elsewhere.
1115 * File last-modified time, so we can check if file has been changed.
1116 * Read-only once the structure has been created.
1118 time_t lastmodified;
1121 * The full filename.
1126 * Pointer to next entry in the linked list of all "file_list"s.
1127 * This linked list is so that sweep() can navigate it.
1128 * Since sweep() can remove items from the list, we must be careful
1129 * to only access this value from main thread (when we know sweep
1130 * won't be running).
1132 struct file_list *next;
1136 #ifdef FEATURE_TRUST
1139 * The format of a trust file when loaded into memory.
1143 struct url_spec url[1]; /**< The URL pattern */
1144 int reject; /**< FIXME: Please document this! */
1145 struct block_spec *next; /**< Next entry in linked list */
1148 #endif /* def FEATURE_TRUST */
1151 #define SOCKS_NONE 0 /**< Don't use a SOCKS server */
1152 #define SOCKS_4 40 /**< original SOCKS 4 protocol */
1153 #define SOCKS_4A 41 /**< as modified for hosts w/o external DNS */
1157 * How to forward a connection to a parent proxy.
1161 /** URL pattern that this forward_spec is for. */
1162 struct url_spec url[1];
1164 /** Connection type. Must be SOCKS_NONE, SOCKS_4, or SOCKS_4A. */
1167 /** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
1170 /** SOCKS server port. */
1173 /** Parent HTTP proxy hostname, or NULL for none. */
1176 /** Parent HTTP proxy port. */
1179 /** Next entry in the linked list. */
1180 struct forward_spec *next;
1185 * Initializer for a static struct forward_spec.
1187 #define FORWARD_SPEC_INITIALIZER { { URL_SPEC_INITIALIZER }, 0, NULL, 0, NULL, 0, NULL }
1191 * This struct represents one filter (one block) from
1192 * the re_filterfile. If there is more than one filter
1193 * in the file, the file will be represented by a
1194 * chained list of re_filterfile specs.
1196 struct re_filterfile_spec
1198 char *name; /**< Name from FILTER: statement in re_filterfile. */
1199 char *description; /**< Description from FILTER: statement in re_filterfile. */
1200 struct list patterns[1]; /**< The patterns from the re_filterfile. */
1201 pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
1202 struct re_filterfile_spec *next; /**< The pointer for chaining. */
1208 #define ACL_PERMIT 1 /**< Accept connection request */
1209 #define ACL_DENY 2 /**< Reject connection request */
1212 * An IP address pattern. Used to specify networks in the ACL.
1214 struct access_control_addr
1216 unsigned long addr; /**< The IP address as an integer. */
1217 unsigned long mask; /**< The network mask as an integer. */
1218 unsigned long port; /**< The port number. */
1222 * An access control list (ACL) entry.
1224 * This is a linked list.
1226 struct access_control_list
1228 struct access_control_addr src[1]; /**< Client IP address */
1229 struct access_control_addr dst[1]; /**< Website or parent proxy IP address */
1231 short action; /**< ACL_PERMIT or ACL_DENY */
1232 struct access_control_list *next; /**< The next entry in the ACL. */
1235 #endif /* def FEATURE_ACL */
1238 /** Maximum number of loaders (actions, re_filter, ...) */
1242 /** configuration_spec::feature_flags: CGI actions editor. */
1243 #define RUNTIME_FEATURE_CGI_EDIT_ACTIONS 1
1245 /** configuration_spec::feature_flags: Web-based toggle. */
1246 #define RUNTIME_FEATURE_CGI_TOGGLE 2
1250 * Data loaded from the configuration file.
1252 * (Anomaly: toggle is still handled through a global, not this structure)
1254 struct configuration_spec
1259 /** Nonzero to enable multithreading. */
1263 * Bitmask of features that can be enabled/disabled through the config
1264 * file. Currently defined bits:
1266 * - RUNTIME_FEATURE_CGI_EDIT_ACTIONS
1267 * - RUNTIME_FEATURE_CGI_TOGGLE
1269 unsigned feature_flags;
1271 /** The log file name. */
1272 const char *logfile;
1274 /** The config file directory. */
1275 const char *confdir;
1277 /** The log file directory. */
1280 /** The full paths to the actions files. */
1281 const char *actions_file[MAX_ACTION_FILES];
1283 /** The short names of the actions files. */
1284 const char *actions_file_short[MAX_ACTION_FILES];
1286 /** The administrator's email address */
1287 char *admin_address;
1289 /** A URL with info on this proxy */
1290 char *proxy_info_url;
1292 /** URL to the user manual (on our website or local copy) */
1295 /** The file name of the pcre filter file */
1296 const char *re_filterfile;
1298 #ifdef FEATURE_COOKIE_JAR
1300 /** The file name of the cookie jar file */
1301 const char * jarfile;
1303 /** The handle to the cookie jar file */
1306 #endif /* def FEATURE_COOKIE_JAR */
1308 /** IP address to bind to. Defaults to HADDR_DEFAULT == 127.0.0.1. */
1311 /** Port to bind to. Defaults to HADDR_PORT == 8118. */
1314 /** Size limit for IOB */
1315 size_t buffer_limit;
1317 #ifdef FEATURE_TRUST
1319 /** The file name of the trust file. */
1320 const char * trustfile;
1322 /** FIXME: DOCME: Document this. */
1323 struct list trust_info[1];
1325 /** FIXME: DOCME: Document this. */
1326 struct url_spec *trust_list[64];
1328 #endif /* def FEATURE_TRUST */
1332 /** The access control list (ACL). */
1333 struct access_control_list *acl;
1335 #endif /* def FEATURE_ACL */
1337 /** Information about parent proxies (forwarding). */
1338 struct forward_spec *forward;
1340 /** All options from the config file, HTML-formatted. */
1343 /** The configuration file object. */
1344 struct file_list *config_file_list;
1346 /** List of loaders */
1347 int (*loaders[NLOADERS])(struct client_state *);
1349 /** Nonzero if we need to bind() to the new port. */
1353 /** Calculates the number of elements in an array, using sizeof. */
1354 #define SZ(X) (sizeof(X) / sizeof(*X))
1356 #ifdef FEATURE_FORCE_LOAD
1357 /** The force load URL prefix. */
1358 #define FORCE_PREFIX "/PRIVOXY-FORCE"
1359 #endif /* def FEATURE_FORCE_LOAD */
1361 #ifdef FEATURE_NO_GIFS
1362 /** The MIME type for images ("image/png" or "image/gif"). */
1363 #define BUILTIN_IMAGE_MIMETYPE "image/png"
1365 #define BUILTIN_IMAGE_MIMETYPE "image/gif"
1366 #endif /* def FEATURE_NO_GIFS */
1369 /* Hardwired URLs */
1371 /** URL for the Privoxy home page. No trailing "/". */
1372 #define HOME_PAGE_URL "http://www.privoxy.org"
1374 /** URL for the Privoxy user manual. With a trailing "/". */
1375 #define USER_MANUAL_URL HOME_PAGE_URL "/" VERSION "/user-manual/"
1377 /** FIXME: DOCME: Document this. */
1378 #define HELP_LINK_PREFIX "configuration.html#"
1381 * The "hosts" to intercept and display CGI pages.
1382 * First one is a hostname only, second one can specify host and path.
1385 * 1) Do not specify the http: prefix
1386 * 2) CGI_SITE_2_PATH must not end with /, one will be added automatically.
1387 * 3) CGI_SITE_2_PATH must start with /, unless it is the empty string.
1389 #define CGI_SITE_1_HOST "p.p"
1390 #define CGI_SITE_2_HOST "config.privoxy.org"
1391 #define CGI_SITE_2_PATH ""
1394 * The prefix for CGI pages. Written out in generated HTML.
1395 * INCLUDES the trailing slash.
1397 #define CGI_PREFIX "http://" CGI_SITE_2_HOST CGI_SITE_2_PATH "/"
1402 * FIXME: This is very inefficient. There could be one copy of these strings
1403 * for each .c file!! They should be "extern", not "static".
1405 static const char CSUCCEED[] =
1406 "HTTP/1.0 200 Connection established\n"
1407 "Proxy-Agent: Privoxy/" VERSION "\r\n\r\n";
1409 static const char CHEADER[] =
1410 "HTTP/1.0 400 Invalid header received from browser\r\n\r\n";
1412 static const char CFORBIDDEN[] =
1413 "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";
1419 #endif /* ndef PROJECT_H_INCLUDED */