1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
4 #define PROJECT_H_VERSION "$Id: project.h,v 1.176 2012/06/19 12:50:22 fabiankeil Exp $"
5 /*********************************************************************
7 * File : $Source: /cvsroot/ijbswa/current/project.h,v $
9 * Purpose : Defines data structures which are widely used in the
10 * project. Does not define any variables or functions
11 * (though it does declare some macros).
13 * Copyright : Written by and Copyright (C) 2001-2012 the
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.
38 *********************************************************************/
41 /* Declare struct FILE for vars and funcs. */
44 /* Need time_t for file_list */
46 /* Needed for pcre choice */
49 /* Need for struct sockaddr_storage */
53 # include <sys/socket.h>
56 # include <winsock2.h>
57 # include <ws2tcpip.h>
58 typedef unsigned short in_port_t;
64 * Include appropriate regular expression libraries.
65 * Note that pcrs and pcre (native) are needed for cgi
66 * and are included anyway.
72 # ifdef PCRE_H_IN_SUBDIR
73 # include <pcre/pcre.h>
86 # include "pcreposix.h"
88 # ifdef PCRE_H_IN_SUBDIR
89 # include <pcre/pcreposix.h>
91 # include <pcreposix.h>
97 #endif /* def AMIGA */
101 * I don't want to have to #include all this just for the declaration
102 * of SOCKET. However, it looks like we have to...
117 typedef SOCKET jb_socket;
119 #define JB_INVALID_SOCKET INVALID_SOCKET
121 #else /* ndef _WIN32 */
124 * The type used by sockets. On UNIX it's an int. Microsoft decided to
125 * make it an unsigned.
127 typedef int jb_socket;
130 * The error value used for variables of type jb_socket. On UNIX this
131 * is -1, however Microsoft decided to make socket handles unsigned, so
132 * they use a different value.
135 #define JB_INVALID_SOCKET (-1)
137 #endif /* ndef _WIN32 */
141 * A standard error code. This should be JB_ERR_OK or one of the JB_ERR_xxx
146 #define JB_ERR_OK 0 /**< Success, no error */
147 #define JB_ERR_MEMORY 1 /**< Out of memory */
148 #define JB_ERR_CGI_PARAMS 2 /**< Missing or corrupt CGI parameters */
149 #define JB_ERR_FILE 3 /**< Error opening, reading or writing a file */
150 #define JB_ERR_PARSE 4 /**< Error parsing file */
151 #define JB_ERR_MODIFIED 5 /**< File has been modified outside of the
152 CGI actions editor. */
153 #define JB_ERR_COMPRESS 6 /**< Error on decompression */
156 * This macro is used to free a pointer that may be NULL.
157 * It also sets the variable to NULL after it's been freed.
158 * The paramater should be a simple variable without side effects.
160 #define freez(X) { if(X) { free((void*)X); X = NULL ; } }
164 * Macro definitions for platforms where isspace() and friends
165 * are macros that use their argument directly as an array index
166 * and thus better be positive. Supposedly that's the case on
167 * some unspecified Solaris versions.
168 * Note: Remember to #include <ctype.h> if you use these macros.
170 #define privoxy_isdigit(__X) isdigit((int)(unsigned char)(__X))
171 #define privoxy_isupper(__X) isupper((int)(unsigned char)(__X))
172 #define privoxy_toupper(__X) toupper((int)(unsigned char)(__X))
173 #define privoxy_tolower(__X) tolower((int)(unsigned char)(__X))
174 #define privoxy_isspace(__X) isspace((int)(unsigned char)(__X))
177 * Use for statically allocated buffers if you have no other choice.
178 * Remember to check the length of what you write into the buffer
179 * - we don't want any buffer overflows!
181 #define BUFFER_SIZE 5000
184 * Max length of CGI parameters (arbitrary limit).
186 #define CGI_PARAM_LEN_MAX 500U
189 * Buffer size for capturing struct hostent data in the
190 * gethostby(name|addr)_r library calls. Since we don't
191 * loop over gethostbyname_r, the buffer must be sufficient
192 * to accommodate multiple IN A RRs, as used in DNS round robin
193 * load balancing. W3C's wwwlib uses 1K, so that should be
194 * good enough for us, too.
197 * XXX: Temporary doubled, for some configurations
198 * 1K is still too small and we didn't get the
199 * real fix ready for inclusion.
201 #define HOSTENT_BUFFER_SIZE 2048
204 * Do not use. Originally this was so that you can
205 * say "while (FOREVER) { ...do something... }".
206 * However, this gives a warning with some compilers (e.g. VC++).
207 * Instead, use "for (;;) { ...do something... }".
212 * Default TCP/IP address to listen on, as a string.
213 * Set to "127.0.0.1:8118".
215 #define HADDR_DEFAULT "127.0.0.1:8118"
218 /* Forward def for struct client_state */
219 struct configuration_spec;
223 * Entry in a linked list of strings.
228 * The string pointer. It must point to a dynamically malloc()ed
229 * string or be NULL for the list functions to work. In the latter
230 * case, just be careful next time you iterate through the list in
235 /** Next entry in the linked list, or NULL if no more. */
236 struct list_entry *next;
240 * A header for a linked list of strings.
244 /** First entry in the list, or NULL if the list is empty. */
245 struct list_entry *first;
247 /** Last entry in the list, or NULL if the list is empty. */
248 struct list_entry *last;
253 * An entry in a map. This is a name=value pair.
257 /** The key for the map. */
259 /** The value associated with that key. */
261 /** The next map entry, or NULL if none. */
262 struct map_entry *next;
266 * A map from a string to another string.
267 * This is used for the paramaters passed in a HTTP GET request, and
268 * to store the exports when the CGI interface is filling in a template.
272 /** The first map entry, or NULL if the map is empty. */
273 struct map_entry *first;
274 /** The last map entry, or NULL if the map is empty. */
275 struct map_entry *last;
280 * A HTTP request. This includes the method (GET, POST) and
283 * This is also used whenever we want to match a URL against a
284 * URL pattern. This always contains the URL to match, and never
285 * a URL pattern. (See struct url_spec).
289 char *cmd; /**< Whole command line: method, URL, Version */
290 char *ocmd; /**< Backup of original cmd for CLF logging */
291 char *gpc; /**< HTTP method: GET, POST, ... */
292 char *url; /**< The URL */
293 char *ver; /**< Protocol version */
294 int status; /**< HTTP Status */
296 char *host; /**< Host part of URL */
297 int port; /**< Port of URL or 80 (default) */
298 char *path; /**< Path of URL */
299 char *hostport; /**< host[:port] */
300 int ssl; /**< Flag if protocol is https */
302 char *host_ip_addr_str; /**< String with dotted decimal representation
303 of host's IP. NULL before connect_to() */
305 #ifndef FEATURE_EXTENDED_HOST_PATTERNS
306 char *dbuffer; /**< Buffer with '\0'-delimited domain name. */
307 char **dvec; /**< List of pointers to the strings in dbuffer. */
308 int dcount; /**< How many parts to this domain? (length of dvec) */
309 #endif /* ndef FEATURE_EXTENDED_HOST_PATTERNS */
313 * Reasons for generating a http_response instead of delivering
314 * the requested resource. Mostly ordered the way they are checked
334 * Response generated by CGI, blocker, or error handler
338 enum crunch_reason crunch_reason; /**< Why the response was generated in the first place. */
339 char *status; /**< HTTP status (string). */
340 struct list headers[1]; /**< List of header lines. */
341 char *head; /**< Formatted http response head. */
342 size_t head_length; /**< Length of http response head. */
343 char *body; /**< HTTP document body. */
344 size_t content_length; /**< Length of body, REQUIRED if binary body. */
345 int is_static; /**< Nonzero if the content will never change and
346 should be cached by the browser (e.g. images). */
350 * A URL or a tag pattern.
354 /** The string which was parsed to produce this url_spec.
355 Used for debugging or display only. */
358 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
359 regex_t *host_regex;/**< Regex for host matching */
361 char *dbuffer; /**< Buffer with '\0'-delimited domain name, or NULL to match all hosts. */
362 char **dvec; /**< List of pointers to the strings in dbuffer. */
363 int dcount; /**< How many parts to this domain? (length of dvec) */
364 int unanchored; /**< Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT. */
365 #endif /* defined FEATURE_EXTENDED_HOST_PATTERNS */
367 char *port_list; /**< List of acceptable ports, or NULL to match all ports */
369 regex_t *preg; /**< Regex for matching path part */
370 regex_t *tag_regex; /**< Regex for matching tags */
374 * If you declare a static url_spec, this is the value to initialize it to zero.
376 #ifndef FEATURE_EXTENDED_HOST_PATTERNS
377 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, 0, 0, NULL, NULL, NULL }
379 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, NULL, NULL }
380 #endif /* def FEATURE_EXTENDED_HOST_PATTERNS */
383 * Constant for host part matching in URLs. If set, indicates that the start of
384 * the pattern must match the start of the URL. E.g. this is not set for the
385 * pattern ".example.com", so that it will match both "example.com" and
386 * "www.example.com". It is set for the pattern "example.com", which makes it
387 * match "example.com" only, not "www.example.com".
389 #define ANCHOR_LEFT 1
392 * Constant for host part matching in URLs. If set, indicates that the end of
393 * the pattern must match the end of the URL. E.g. this is not set for the
394 * pattern "ad.", so that it will match any host called "ad", irrespective
395 * of how many subdomains are in the fully-qualified domain name.
397 #define ANCHOR_RIGHT 2
401 * An I/O buffer. Holds a string which can be appended to, and can have data
402 * removed from the beginning.
406 char *buf; /**< Start of buffer */
407 char *cur; /**< Start of relevant data */
408 char *eod; /**< End of relevant data */
409 size_t size; /**< Size as malloc()ed */
414 * Return the number of bytes in the I/O buffer associated with the passed
415 * client_state pointer.
418 #define IOB_PEEK(CSP) ((CSP->iob->cur > CSP->iob->eod) ? (CSP->iob->eod - CSP->iob->cur) : 0)
422 * Remove any data in the I/O buffer associated with the passed
423 * client_state pointer.
425 #define IOB_RESET(CSP) if(CSP->iob->buf) free(CSP->iob->buf); memset(CSP->iob, '\0', sizeof(CSP->iob));
427 /* Bits for csp->content_type bitmask: */
428 #define CT_TEXT 0x0001U /**< Suitable for pcrs filtering. */
429 #define CT_GIF 0x0002U /**< Suitable for GIF filtering. */
430 #define CT_TABOO 0x0004U /**< DO NOT filter, irrespective of other flags. */
432 /* Although these are not, strictly speaking, content types
433 * (they are content encodings), it is simple to handle them
436 #define CT_GZIP 0x0010U /**< gzip-compressed data. */
437 #define CT_DEFLATE 0x0020U /**< zlib-compressed data. */
440 * Flag to signal that the server declared the content type,
441 * so we can differentiate between unknown and undeclared
444 #define CT_DECLARED 0x0040U
447 * The mask which includes all actions.
449 #define ACTION_MASK_ALL (~0UL)
452 * The most compatible set of actions - i.e. none.
454 #define ACTION_MOST_COMPATIBLE 0x00000000UL
456 /** Action bitmap: Block the request. */
457 #define ACTION_BLOCK 0x00000001UL
458 /** Action bitmap: Deanimate if it's a GIF. */
459 #define ACTION_DEANIMATE 0x00000002UL
460 /** Action bitmap: Downgrade HTTP/1.1 to 1.0. */
461 #define ACTION_DOWNGRADE 0x00000004UL
462 /** Action bitmap: Fast redirects. */
463 #define ACTION_FAST_REDIRECTS 0x00000008UL
464 /** Action bitmap: Remove or add "X-Forwarded-For" header. */
465 #define ACTION_CHANGE_X_FORWARDED_FOR 0x00000010UL
466 /** Action bitmap: Hide "From" header. */
467 #define ACTION_HIDE_FROM 0x00000020UL
468 /** Action bitmap: Hide "Referer" header. (sic - follow HTTP, not English). */
469 #define ACTION_HIDE_REFERER 0x00000040UL
470 /** Action bitmap: Hide "User-Agent" and similar headers. */
471 #define ACTION_HIDE_USER_AGENT 0x00000080UL
472 /** Action bitmap: This is an image. */
473 #define ACTION_IMAGE 0x00000100UL
474 /** Action bitmap: Sets the image blocker. */
475 #define ACTION_IMAGE_BLOCKER 0x00000200UL
476 /** Action bitmap: Prevent compression. */
477 #define ACTION_NO_COMPRESSION 0x00000400UL
478 /** Action bitmap: Change cookies to session only cookies. */
479 #define ACTION_NO_COOKIE_KEEP 0x00000800UL
480 /** Action bitmap: Block rending cookies. */
481 #define ACTION_NO_COOKIE_READ 0x00001000UL
482 /** Action bitmap: Block setting cookies. */
483 #define ACTION_NO_COOKIE_SET 0x00002000UL
484 /** Action bitmap: Override the forward settings in the config file */
485 #define ACTION_FORWARD_OVERRIDE 0x00004000UL
486 /** Action bitmap: Block as empty document */
487 #define ACTION_HANDLE_AS_EMPTY_DOCUMENT 0x00008000UL
488 /** Action bitmap: Limit CONNECT requests to safe ports. */
489 #define ACTION_LIMIT_CONNECT 0x00010000UL
490 /** Action bitmap: Redirect request. */
491 #define ACTION_REDIRECT 0x00020000UL
492 /** Action bitmap: Crunch or modify "if-modified-since" header. */
493 #define ACTION_HIDE_IF_MODIFIED_SINCE 0x00040000UL
494 /** Action bitmap: Overwrite Content-Type header. */
495 #define ACTION_CONTENT_TYPE_OVERWRITE 0x00080000UL
496 /** Action bitmap: Crunch specified server header. */
497 #define ACTION_CRUNCH_SERVER_HEADER 0x00100000UL
498 /** Action bitmap: Crunch specified client header */
499 #define ACTION_CRUNCH_CLIENT_HEADER 0x00200000UL
500 /** Action bitmap: Enable text mode by force */
501 #define ACTION_FORCE_TEXT_MODE 0x00400000UL
502 /** Action bitmap: Enable text mode by force */
503 #define ACTION_CRUNCH_IF_NONE_MATCH 0x00800000UL
504 /** Action bitmap: Enable content-disposition crunching */
505 #define ACTION_HIDE_CONTENT_DISPOSITION 0x01000000UL
506 /** Action bitmap: Replace or block Last-Modified header */
507 #define ACTION_OVERWRITE_LAST_MODIFIED 0x02000000UL
508 /** Action bitmap: Replace or block Accept-Language header */
509 #define ACTION_HIDE_ACCEPT_LANGUAGE 0x04000000UL
512 /** Action string index: How to deanimate GIFs */
513 #define ACTION_STRING_DEANIMATE 0
514 /** Action string index: Replacement for "From:" header */
515 #define ACTION_STRING_FROM 1
516 /** Action string index: How to block images */
517 #define ACTION_STRING_IMAGE_BLOCKER 2
518 /** Action string index: Replacement for "Referer:" header */
519 #define ACTION_STRING_REFERER 3
520 /** Action string index: Replacement for "User-Agent:" header */
521 #define ACTION_STRING_USER_AGENT 4
522 /** Action string index: Legal CONNECT ports. */
523 #define ACTION_STRING_LIMIT_CONNECT 5
524 /** Action string index: Server headers containing this pattern are crunched*/
525 #define ACTION_STRING_SERVER_HEADER 6
526 /** Action string index: Client headers containing this pattern are crunched*/
527 #define ACTION_STRING_CLIENT_HEADER 7
528 /** Action string index: Replacement for the "Accept-Language:" header*/
529 #define ACTION_STRING_LANGUAGE 8
530 /** Action string index: Replacement for the "Content-Type:" header*/
531 #define ACTION_STRING_CONTENT_TYPE 9
532 /** Action string index: Replacement for the "content-disposition:" header*/
533 #define ACTION_STRING_CONTENT_DISPOSITION 10
534 /** Action string index: Replacement for the "If-Modified-Since:" header*/
535 #define ACTION_STRING_IF_MODIFIED_SINCE 11
536 /** Action string index: Replacement for the "Last-Modified:" header. */
537 #define ACTION_STRING_LAST_MODIFIED 12
538 /** Action string index: Redirect URL */
539 #define ACTION_STRING_REDIRECT 13
540 /** Action string index: Decode before redirect? */
541 #define ACTION_STRING_FAST_REDIRECTS 14
542 /** Action string index: Overriding forward rule. */
543 #define ACTION_STRING_FORWARD_OVERRIDE 15
544 /** Action string index: Reason for the block. */
545 #define ACTION_STRING_BLOCK 16
546 /** Action string index: what to do with the "X-Forwarded-For" header. */
547 #define ACTION_STRING_CHANGE_X_FORWARDED_FOR 17
548 /** Number of string actions. */
549 #define ACTION_STRING_COUNT 18
552 /* To make the ugly hack in sed easier to understand */
553 #define CHECK_EVERY_HEADER_REMAINING 0
556 /** Index into current_action_spec::multi[] for headers to add. */
557 #define ACTION_MULTI_ADD_HEADER 0
558 /** Index into current_action_spec::multi[] for content filters to apply. */
559 #define ACTION_MULTI_FILTER 1
560 /** Index into current_action_spec::multi[] for server-header filters to apply. */
561 #define ACTION_MULTI_SERVER_HEADER_FILTER 2
562 /** Index into current_action_spec::multi[] for client-header filters to apply. */
563 #define ACTION_MULTI_CLIENT_HEADER_FILTER 3
564 /** Index into current_action_spec::multi[] for client-header tags to apply. */
565 #define ACTION_MULTI_CLIENT_HEADER_TAGGER 4
566 /** Index into current_action_spec::multi[] for server-header tags to apply. */
567 #define ACTION_MULTI_SERVER_HEADER_TAGGER 5
568 /** Number of multi-string actions. */
569 #define ACTION_MULTI_COUNT 6
573 * This structure contains a list of actions to apply to a URL.
574 * It only contains positive instructions - no "-" options.
575 * It is not used to store the actions list itself, only for
576 * url_actions() to return the current values.
578 struct current_action_spec
580 /** Actions to apply. A bit set to "1" means perform the action. */
584 * Paramaters for those actions that require them.
585 * Each entry is valid if & only if the corresponding entry in "flags" is
588 char * string[ACTION_STRING_COUNT];
590 /** Lists of strings for multi-string actions. */
591 struct list multi[ACTION_MULTI_COUNT][1];
596 * This structure contains a set of changes to actions.
597 * It can contain both positive and negative instructions.
598 * It is used to store an entry in the actions list.
602 unsigned long mask; /**< Actions to keep. A bit set to "0" means remove action. */
603 unsigned long add; /**< Actions to add. A bit set to "1" means add action. */
606 * Paramaters for those actions that require them.
607 * Each entry is valid if & only if the corresponding entry in "flags" is
610 char * string[ACTION_STRING_COUNT];
612 /** Lists of strings to remove, for multi-string actions. */
613 struct list multi_remove[ACTION_MULTI_COUNT][1];
615 /** If nonzero, remove *all* strings from the multi-string action. */
616 int multi_remove_all[ACTION_MULTI_COUNT];
618 /** Lists of strings to add, for multi-string actions. */
619 struct list multi_add[ACTION_MULTI_COUNT][1];
624 * This structure is used to store action files.
626 * It contains an URL or tag pattern, and the changes to
627 * the actions. It's a linked list and should only be
628 * free'd through unload_actions_file() unless there's
629 * only a single entry.
633 struct url_spec url[1]; /**< The URL or tag pattern. */
635 struct action_spec *action; /**< Action settings that might be shared with
636 the list entry before or after the current
637 one and can't be free'd willy nilly. */
639 struct url_actions *next; /**< Next action section in file, or NULL. */
644 * Structure to hold the server socket and the information
645 * required to make sure we only reuse the connection if
646 * the host and forwarding settings are the same.
648 struct reusable_connection
652 time_t timestamp; /* XXX: rename? */
655 time_t response_received;
658 * Number of seconds after which this
659 * connection will no longer be reused.
661 unsigned int keep_alive_timeout;
674 * Flags for use in csp->flags
678 * Flag for csp->flags: Set if this client is processing data.
679 * Cleared when the thread associated with this structure dies.
681 #define CSP_FLAG_ACTIVE 0x01U
684 * Flag for csp->flags: Set if the server's reply is in "chunked"
687 #define CSP_FLAG_CHUNKED 0x02U
690 * Flag for csp->flags: Set if this request was enforced, although it would
691 * normally have been blocked.
693 #define CSP_FLAG_FORCED 0x04U
696 * Flag for csp->flags: Set if any modification to the body was done.
698 #define CSP_FLAG_MODIFIED 0x08U
701 * Flag for csp->flags: Set if request was blocked.
703 #define CSP_FLAG_REJECTED 0x10U
706 * Flag for csp->flags: Set if we are toggled on (FEATURE_TOGGLE).
708 #define CSP_FLAG_TOGGLED_ON 0x20U
711 * Flag for csp->flags: Set if we answered the request ourselve.
713 #define CSP_FLAG_CRUNCHED 0x40U
716 * Flag for csp->flags: Set if an acceptable Connection header
717 * has already been set by the client.
719 #define CSP_FLAG_CLIENT_CONNECTION_HEADER_SET 0x00000040U
722 * Flag for csp->flags: Set if an acceptable Connection header
723 * has already been set by the server.
725 #define CSP_FLAG_SERVER_CONNECTION_HEADER_SET 0x00000080U
728 * Flag for csp->flags: Signals header parsers whether they
729 * are parsing server or client headers.
731 #define CSP_FLAG_CLIENT_HEADER_PARSING_DONE 0x00000100U
734 * Flag for csp->flags: Set if adding the Host: header
737 #define CSP_FLAG_HOST_HEADER_IS_SET 0x00000200U
740 * Flag for csp->flags: Set if filtering is disabled by X-Filter: No
741 * XXX: As we now have tags we might as well ditch this.
743 #define CSP_FLAG_NO_FILTERING 0x00000400U
746 * Flag for csp->flags: Set the client IP has appended to
747 * an already existing X-Forwarded-For header in which case
748 * no new header has to be generated.
750 #define CSP_FLAG_X_FORWARDED_FOR_APPENDED 0x00000800U
753 * Flag for csp->flags: Set if the server wants to keep
754 * the connection alive.
756 #define CSP_FLAG_SERVER_CONNECTION_KEEP_ALIVE 0x00001000U
758 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
760 * Flag for csp->flags: Set if the server specified the
763 #define CSP_FLAG_SERVER_CONTENT_LENGTH_SET 0x00002000U
766 * Flag for csp->flags: Set if we know the content length,
767 * either because the server set it, or we figured it out
770 #define CSP_FLAG_CONTENT_LENGTH_SET 0x00004000U
773 * Flag for csp->flags: Set if the client wants to keep
774 * the connection alive.
776 #define CSP_FLAG_CLIENT_CONNECTION_KEEP_ALIVE 0x00008000U
779 * Flag for csp->flags: Set if we think we got the whole
780 * client request and shouldn't read any additional data
781 * coming from the client until the current request has
784 #define CSP_FLAG_CLIENT_REQUEST_COMPLETELY_READ 0x00010000U
787 * Flag for csp->flags: Set if the server promised us to
788 * keep the connection open for a known number of seconds.
790 #define CSP_FLAG_SERVER_KEEP_ALIVE_TIMEOUT_SET 0x00020000U
792 #endif /* def FEATURE_CONNECTION_KEEP_ALIVE */
795 * Flag for csp->flags: Set if we think we can't reuse
798 #define CSP_FLAG_SERVER_SOCKET_TAINTED 0x00040000U
801 * Flag for csp->flags: Set if the Proxy-Connection header
802 * is among the server headers.
804 #define CSP_FLAG_SERVER_PROXY_CONNECTION_HEADER_SET 0x00080000U
807 * Flag for csp->flags: Set if the client reused its connection.
809 #define CSP_FLAG_REUSED_CLIENT_CONNECTION 0x00100000U
812 * Flag for csp->flags: Set if the supports deflate compression.
814 #define CSP_FLAG_CLIENT_SUPPORTS_DEFLATE 0x00200000U
817 * Flag for csp->flags: Set if the content has been deflated by Privoxy
819 #define CSP_FLAG_BUFFERED_CONTENT_DEFLATED 0x00400000U
823 * Flags for use in return codes of child processes
827 * Flag for process return code: Set if exiting porcess has been toggled
828 * during its lifetime.
830 #define RC_FLAG_TOGGLED 0x10
833 * Flag for process return code: Set if exiting porcess has blocked its
836 #define RC_FLAG_BLOCKED 0x20
839 * Maximum number of actions/filter files. This limit is arbitrary - it's just used
842 #define MAX_AF_FILES 10
845 * Maximum number of sockets to listen to. This limit is arbitrary - it's just used
848 #define MAX_LISTENING_SOCKETS 10
851 * The state of a Privoxy processing thread.
855 /** The proxy's configuration */
856 struct configuration_spec * config;
858 /** The actions to perform on the current request */
859 struct current_action_spec action[1];
861 /** socket to talk to client (web browser) */
864 /** current connection to the server (may go through a proxy) */
865 struct reusable_connection server_connection;
867 /** Multi-purpose flag container, see CSP_FLAG_* above */
870 /** Client PC's IP address, as reported by the accept() function.
874 /** Client PC's TCP address, as reported by the accept() function.
876 struct sockaddr_storage tcp_addr;
878 /** Client PC's IP address, as reported by the accept() function.
880 unsigned long ip_addr_long;
881 #endif /* def HAVE_RFC2553 */
883 /** The URL that was requested */
884 struct http_request http[1];
887 * The final forwarding settings.
888 * XXX: Currently this is only used for forward-override,
889 * so we can free the space in sweep.
891 struct forward_spec * fwd;
893 /** An I/O buffer used for buffering data read from the network */
896 /** List of all headers for this request */
897 struct list headers[1];
899 /** List of all tags that apply to this request */
902 /** MIME-Type key, see CT_* above */
903 unsigned int content_type;
905 /** Actions files associated with this client */
906 struct file_list *actions_list[MAX_AF_FILES];
908 /** pcrs job files. */
909 struct file_list *rlist[MAX_AF_FILES];
911 /** Length after content modification. */
912 unsigned long long content_length;
914 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
915 /* XXX: is this the right location? */
917 /** Expected length of content after which we
918 * should stop reading from the server socket.
920 unsigned long long expected_content_length;
922 /** Expected length of content after which we
923 * should stop reading from the client socket.
925 unsigned long long expected_client_content_length;
926 #endif /* def FEATURE_CONNECTION_KEEP_ALIVE */
931 struct file_list *tlist;
933 #endif /* def FEATURE_TRUST */
936 * Failure reason to embedded in the CGI error page,
937 * or NULL. Currently only used for socks errors.
943 * List of client states so the main thread can keep
944 * track of them and garbage collect their resources.
948 struct client_states *next;
949 struct client_state csp;
953 * A function to add a header
955 typedef jb_err (*add_header_func_ptr)(struct client_state *);
958 * A function to process a header
960 typedef jb_err (*parser_func_ptr )(struct client_state *, char **);
964 * List of available CGI functions.
966 struct cgi_dispatcher
968 /** The URL of the CGI, relative to the CGI root. */
969 const char * const name;
971 /** The handler function for the CGI */
972 jb_err (* const handler)(struct client_state *csp, struct http_response *rsp, const struct map *parameters);
974 /** The description of the CGI, to appear on the main menu, or NULL to hide it. */
975 const char * const description;
977 /** A flag that indicates whether unintentional calls to this CGI can cause damage */
983 * A data file used by Privoxy. Kept in a linked list.
988 * This is a pointer to the data structures associated with the file.
989 * Read-only once the structure has been created.
994 * The unloader function.
995 * Normally NULL. When we are finished with file (i.e. when we have
996 * loaded a new one), set to a pointer to an unloader function.
997 * Unloader will be called by sweep() (called from main loop) when
998 * all clients using this file are done. This prevents threading
1001 void (*unloader)(void *);
1004 * Used internally by sweep(). Do not access from elsewhere.
1009 * File last-modified time, so we can check if file has been changed.
1010 * Read-only once the structure has been created.
1012 time_t lastmodified;
1015 * The full filename.
1020 * Pointer to next entry in the linked list of all "file_list"s.
1021 * This linked list is so that sweep() can navigate it.
1022 * Since sweep() can remove items from the list, we must be careful
1023 * to only access this value from main thread (when we know sweep
1024 * won't be running).
1026 struct file_list *next;
1030 #ifdef FEATURE_TRUST
1033 * The format of a trust file when loaded into memory.
1037 struct url_spec url[1]; /**< The URL pattern */
1038 int reject; /**< FIXME: Please document this! */
1039 struct block_spec *next; /**< Next entry in linked list */
1043 * Arbitrary limit for the number of trusted referrers.
1045 #define MAX_TRUSTED_REFERRERS 512
1047 #endif /* def FEATURE_TRUST */
1049 enum forwarder_type {
1050 /**< Don't use a SOCKS server */
1052 /**< original SOCKS 4 protocol */
1054 /**< SOCKS 4A, DNS resolution is done by the SOCKS server */
1056 /**< SOCKS 5 with hostnames, DNS resolution is done by the SOCKS server */
1061 * How to forward a connection to a parent proxy.
1065 /** URL pattern that this forward_spec is for. */
1066 struct url_spec url[1];
1068 /** Connection type. Must be SOCKS_NONE, SOCKS_4, SOCKS_4A or SOCKS_5. */
1069 enum forwarder_type type;
1071 /** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
1074 /** SOCKS server port. */
1077 /** Parent HTTP proxy hostname, or NULL for none. */
1080 /** Parent HTTP proxy port. */
1083 /** Next entry in the linked list. */
1084 struct forward_spec *next;
1089 * Initializer for a static struct forward_spec.
1091 #define FORWARD_SPEC_INITIALIZER { { URL_SPEC_INITIALIZER }, 0, NULL, 0, NULL, 0, NULL }
1093 /* Supported filter types */
1096 FT_CONTENT_FILTER = 0,
1097 FT_CLIENT_HEADER_FILTER = 1,
1098 FT_SERVER_HEADER_FILTER = 2,
1099 FT_CLIENT_HEADER_TAGGER = 3,
1100 FT_SERVER_HEADER_TAGGER = 4,
1101 FT_INVALID_FILTER = 42,
1103 #define MAX_FILTER_TYPES 5
1106 * This struct represents one filter (one block) from
1107 * the re_filterfile. If there is more than one filter
1108 * in the file, the file will be represented by a
1109 * chained list of re_filterfile specs.
1111 struct re_filterfile_spec
1113 char *name; /**< Name from FILTER: statement in re_filterfile. */
1114 char *description; /**< Description from FILTER: statement in re_filterfile. */
1115 struct list patterns[1]; /**< The patterns from the re_filterfile. */
1116 pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
1117 enum filter_type type; /**< Filter type (content, client-header, server-header). */
1118 int dynamic; /**< Set to one if the pattern might contain variables
1119 and has to be recompiled for every request. */
1120 struct re_filterfile_spec *next; /**< The pointer for chaining. */
1126 #define ACL_PERMIT 1 /**< Accept connection request */
1127 #define ACL_DENY 2 /**< Reject connection request */
1130 * An IP address pattern. Used to specify networks in the ACL.
1132 struct access_control_addr
1135 struct sockaddr_storage addr; /* <The TCP address in network order. */
1136 struct sockaddr_storage mask; /* <The TCP mask in network order. */
1138 unsigned long addr; /**< The IP address as an integer. */
1139 unsigned long mask; /**< The network mask as an integer. */
1140 unsigned long port; /**< The port number. */
1141 #endif /* HAVE_RFC2553 */
1145 * An access control list (ACL) entry.
1147 * This is a linked list.
1149 struct access_control_list
1151 struct access_control_addr src[1]; /**< Client IP address */
1152 struct access_control_addr dst[1]; /**< Website or parent proxy IP address */
1154 int wildcard_dst; /** < dst address is wildcard */
1157 short action; /**< ACL_PERMIT or ACL_DENY */
1158 struct access_control_list *next; /**< The next entry in the ACL. */
1161 #endif /* def FEATURE_ACL */
1164 /** Maximum number of loaders (actions, re_filter, ...) */
1168 /** configuration_spec::feature_flags: CGI actions editor. */
1169 #define RUNTIME_FEATURE_CGI_EDIT_ACTIONS 1U
1171 /** configuration_spec::feature_flags: Web-based toggle. */
1172 #define RUNTIME_FEATURE_CGI_TOGGLE 2U
1174 /** configuration_spec::feature_flags: HTTP-header-based toggle. */
1175 #define RUNTIME_FEATURE_HTTP_TOGGLE 4U
1177 /** configuration_spec::feature_flags: Split large forms to limit the number of GET arguments. */
1178 #define RUNTIME_FEATURE_SPLIT_LARGE_FORMS 8U
1180 /** configuration_spec::feature_flags: Check the host header for requests with host-less request lines. */
1181 #define RUNTIME_FEATURE_ACCEPT_INTERCEPTED_REQUESTS 16U
1183 /** configuration_spec::feature_flags: Don't allow to circumvent blocks with the force prefix. */
1184 #define RUNTIME_FEATURE_ENFORCE_BLOCKS 32U
1186 /** configuration_spec::feature_flags: Allow to block or redirect CGI requests. */
1187 #define RUNTIME_FEATURE_CGI_CRUNCHING 64U
1189 /** configuration_spec::feature_flags: Try to keep the connection to the server alive. */
1190 #define RUNTIME_FEATURE_CONNECTION_KEEP_ALIVE 128U
1192 /** configuration_spec::feature_flags: Share outgoing connections between different client connections. */
1193 #define RUNTIME_FEATURE_CONNECTION_SHARING 256U
1195 /** configuration_spec::feature_flags: Pages blocked with +handle-as-empty-doc get a return status of 200 OK. */
1196 #define RUNTIME_FEATURE_EMPTY_DOC_RETURNS_OK 512U
1198 /** configuration_spec::feature_flags: Buffered content is sent compressed if the client supports it. */
1199 #define RUNTIME_FEATURE_COMPRESSION 1024U
1202 * Data loaded from the configuration file.
1204 * (Anomaly: toggle is still handled through a global, not this structure)
1206 struct configuration_spec
1211 /** Nonzero to enable multithreading. */
1215 * Bitmask of features that can be enabled/disabled through the config
1216 * file. Currently defined bits:
1218 * - RUNTIME_FEATURE_CGI_EDIT_ACTIONS
1219 * - RUNTIME_FEATURE_CGI_TOGGLE
1220 * - RUNTIME_FEATURE_HTTP_TOGGLE
1221 * - RUNTIME_FEATURE_SPLIT_LARGE_FORMS
1222 * - RUNTIME_FEATURE_ACCEPT_INTERCEPTED_REQUESTS
1223 * - RUNTIME_FEATURE_ENFORCE_BLOCKS
1224 * - RUNTIME_FEATURE_CGI_CRUNCHING
1225 * - RUNTIME_FEATURE_CONNECTION_KEEP_ALIVE
1226 * - RUNTIME_FEATURE_CONNECTION_SHARING
1227 * - RUNTIME_FEATURE_EMPTY_DOC_RETURNS_OK
1229 unsigned feature_flags;
1231 /** The log file name. */
1232 const char *logfile;
1234 /** The config file directory. */
1235 const char *confdir;
1237 /** The directory for customized CGI templates. */
1238 const char *templdir;
1240 /** The log file directory. */
1243 /** The full paths to the actions files. */
1244 const char *actions_file[MAX_AF_FILES];
1246 /** The short names of the actions files. */
1247 const char *actions_file_short[MAX_AF_FILES];
1249 /** The administrator's email address */
1250 char *admin_address;
1252 /** A URL with info on this proxy */
1253 char *proxy_info_url;
1255 /** URL to the user manual (on our website or local copy) */
1258 /** The file names of the pcre filter files. */
1259 const char *re_filterfile[MAX_AF_FILES];
1261 /** The short names of the pcre filter files. */
1262 const char *re_filterfile_short[MAX_AF_FILES];
1264 /** The hostname to show on CGI pages, or NULL to use the real one. */
1265 const char *hostname;
1267 /** IP addresses to bind to. Defaults to HADDR_DEFAULT == 127.0.0.1. */
1268 const char *haddr[MAX_LISTENING_SOCKETS];
1270 /** Ports to bind to. Defaults to HADDR_PORT == 8118. */
1271 int hport[MAX_LISTENING_SOCKETS];
1273 /** Size limit for IOB */
1274 size_t buffer_limit;
1276 #ifdef FEATURE_TRUST
1278 /** The file name of the trust file. */
1279 const char * trustfile;
1281 /** FIXME: DOCME: Document this. */
1282 struct list trust_info[1];
1284 /** FIXME: DOCME: Document this. */
1285 struct url_spec *trust_list[MAX_TRUSTED_REFERRERS];
1287 #endif /* def FEATURE_TRUST */
1291 /** The access control list (ACL). */
1292 struct access_control_list *acl;
1294 #endif /* def FEATURE_ACL */
1296 /** Information about parent proxies (forwarding). */
1297 struct forward_spec *forward;
1299 /** Number of retries in case a forwarded connection attempt fails */
1300 int forwarded_connect_retries;
1302 /** Maximum number of client connections. */
1303 int max_client_connections;
1305 /* Timeout when waiting on sockets for data to become available. */
1308 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
1309 /* Maximum number of seconds after which an open connection will no longer be reused. */
1310 unsigned int keep_alive_timeout;
1312 /* Assumed server-side keep alive timeout if none is specified. */
1313 unsigned int default_server_timeout;
1316 #ifdef FEATURE_COMPRESSION
1317 int compression_level;
1320 /** All options from the config file, HTML-formatted. */
1323 /** The configuration file object. */
1324 struct file_list *config_file_list;
1326 /** List of loaders */
1327 int (*loaders[NLOADERS])(struct client_state *);
1329 /** Nonzero if we need to bind() to the new port. */
1333 /** Calculates the number of elements in an array, using sizeof. */
1334 #define SZ(X) (sizeof(X) / sizeof(*X))
1336 #ifdef FEATURE_FORCE_LOAD
1337 /** The force load URL prefix. */
1338 #define FORCE_PREFIX "/PRIVOXY-FORCE"
1339 #endif /* def FEATURE_FORCE_LOAD */
1341 #ifdef FEATURE_NO_GIFS
1342 /** The MIME type for images ("image/png" or "image/gif"). */
1343 #define BUILTIN_IMAGE_MIMETYPE "image/png"
1345 #define BUILTIN_IMAGE_MIMETYPE "image/gif"
1346 #endif /* def FEATURE_NO_GIFS */
1353 /** URL for the Privoxy home page. */
1354 #define HOME_PAGE_URL "http://www.privoxy.org/"
1356 /** URL for the Privoxy user manual. */
1357 #define USER_MANUAL_URL HOME_PAGE_URL VERSION "/user-manual/"
1359 /** Prefix for actions help links (append to USER_MANUAL_URL). */
1360 #define ACTIONS_HELP_PREFIX "actions-file.html#"
1362 /** Prefix for config option help links (append to USER_MANUAL_URL). */
1363 #define CONFIG_HELP_PREFIX "config.html#"
1366 * The "hosts" to intercept and display CGI pages.
1367 * First one is a hostname only, second one can specify host and path.
1370 * 1) Do not specify the http: prefix
1371 * 2) CGI_SITE_2_PATH must not end with /, one will be added automatically.
1372 * 3) CGI_SITE_2_PATH must start with /, unless it is the empty string.
1374 #define CGI_SITE_1_HOST "p.p"
1375 #define CGI_SITE_2_HOST "config.privoxy.org"
1376 #define CGI_SITE_2_PATH ""
1379 * The prefix for CGI pages. Written out in generated HTML.
1380 * INCLUDES the trailing slash.
1382 #define CGI_PREFIX "http://" CGI_SITE_2_HOST CGI_SITE_2_PATH "/"
1388 #endif /* ndef PROJECT_H_INCLUDED */