1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
4 #define PROJECT_H_VERSION "$Id: project.h,v 1.191 2012/10/23 10:19:04 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 * I/O buffer. May be zero.
417 #define IOB_PEEK(IOB) ((IOB->cur > IOB->eod) ? (IOB->eod - IOB->cur) : 0)
420 /* Bits for csp->content_type bitmask: */
421 #define CT_TEXT 0x0001U /**< Suitable for pcrs filtering. */
422 #define CT_GIF 0x0002U /**< Suitable for GIF filtering. */
423 #define CT_TABOO 0x0004U /**< DO NOT filter, irrespective of other flags. */
425 /* Although these are not, strictly speaking, content types
426 * (they are content encodings), it is simple to handle them
429 #define CT_GZIP 0x0010U /**< gzip-compressed data. */
430 #define CT_DEFLATE 0x0020U /**< zlib-compressed data. */
433 * Flag to signal that the server declared the content type,
434 * so we can differentiate between unknown and undeclared
437 #define CT_DECLARED 0x0040U
440 * The mask which includes all actions.
442 #define ACTION_MASK_ALL (~0UL)
445 * The most compatible set of actions - i.e. none.
447 #define ACTION_MOST_COMPATIBLE 0x00000000UL
449 /** Action bitmap: Block the request. */
450 #define ACTION_BLOCK 0x00000001UL
451 /** Action bitmap: Deanimate if it's a GIF. */
452 #define ACTION_DEANIMATE 0x00000002UL
453 /** Action bitmap: Downgrade HTTP/1.1 to 1.0. */
454 #define ACTION_DOWNGRADE 0x00000004UL
455 /** Action bitmap: Fast redirects. */
456 #define ACTION_FAST_REDIRECTS 0x00000008UL
457 /** Action bitmap: Remove or add "X-Forwarded-For" header. */
458 #define ACTION_CHANGE_X_FORWARDED_FOR 0x00000010UL
459 /** Action bitmap: Hide "From" header. */
460 #define ACTION_HIDE_FROM 0x00000020UL
461 /** Action bitmap: Hide "Referer" header. (sic - follow HTTP, not English). */
462 #define ACTION_HIDE_REFERER 0x00000040UL
463 /** Action bitmap: Hide "User-Agent" and similar headers. */
464 #define ACTION_HIDE_USER_AGENT 0x00000080UL
465 /** Action bitmap: This is an image. */
466 #define ACTION_IMAGE 0x00000100UL
467 /** Action bitmap: Sets the image blocker. */
468 #define ACTION_IMAGE_BLOCKER 0x00000200UL
469 /** Action bitmap: Prevent compression. */
470 #define ACTION_NO_COMPRESSION 0x00000400UL
471 /** Action bitmap: Change cookies to session only cookies. */
472 #define ACTION_SESSION_COOKIES_ONLY 0x00000800UL
473 /** Action bitmap: Block cookies coming from the client. */
474 #define ACTION_CRUNCH_OUTGOING_COOKIES 0x00001000UL
475 /** Action bitmap: Block cookies coming from the server. */
476 #define ACTION_CRUNCH_INCOMING_COOKIES 0x00002000UL
477 /** Action bitmap: Override the forward settings in the config file */
478 #define ACTION_FORWARD_OVERRIDE 0x00004000UL
479 /** Action bitmap: Block as empty document */
480 #define ACTION_HANDLE_AS_EMPTY_DOCUMENT 0x00008000UL
481 /** Action bitmap: Limit CONNECT requests to safe ports. */
482 #define ACTION_LIMIT_CONNECT 0x00010000UL
483 /** Action bitmap: Redirect request. */
484 #define ACTION_REDIRECT 0x00020000UL
485 /** Action bitmap: Crunch or modify "if-modified-since" header. */
486 #define ACTION_HIDE_IF_MODIFIED_SINCE 0x00040000UL
487 /** Action bitmap: Overwrite Content-Type header. */
488 #define ACTION_CONTENT_TYPE_OVERWRITE 0x00080000UL
489 /** Action bitmap: Crunch specified server header. */
490 #define ACTION_CRUNCH_SERVER_HEADER 0x00100000UL
491 /** Action bitmap: Crunch specified client header */
492 #define ACTION_CRUNCH_CLIENT_HEADER 0x00200000UL
493 /** Action bitmap: Enable text mode by force */
494 #define ACTION_FORCE_TEXT_MODE 0x00400000UL
495 /** Action bitmap: Enable text mode by force */
496 #define ACTION_CRUNCH_IF_NONE_MATCH 0x00800000UL
497 /** Action bitmap: Enable content-disposition crunching */
498 #define ACTION_HIDE_CONTENT_DISPOSITION 0x01000000UL
499 /** Action bitmap: Replace or block Last-Modified header */
500 #define ACTION_OVERWRITE_LAST_MODIFIED 0x02000000UL
501 /** Action bitmap: Replace or block Accept-Language header */
502 #define ACTION_HIDE_ACCEPT_LANGUAGE 0x04000000UL
503 /** Action bitmap: Limit the cookie lifetime */
504 #define ACTION_LIMIT_COOKIE_LIFETIME 0x08000000UL
507 /** Action string index: How to deanimate GIFs */
508 #define ACTION_STRING_DEANIMATE 0
509 /** Action string index: Replacement for "From:" header */
510 #define ACTION_STRING_FROM 1
511 /** Action string index: How to block images */
512 #define ACTION_STRING_IMAGE_BLOCKER 2
513 /** Action string index: Replacement for "Referer:" header */
514 #define ACTION_STRING_REFERER 3
515 /** Action string index: Replacement for "User-Agent:" header */
516 #define ACTION_STRING_USER_AGENT 4
517 /** Action string index: Legal CONNECT ports. */
518 #define ACTION_STRING_LIMIT_CONNECT 5
519 /** Action string index: Server headers containing this pattern are crunched*/
520 #define ACTION_STRING_SERVER_HEADER 6
521 /** Action string index: Client headers containing this pattern are crunched*/
522 #define ACTION_STRING_CLIENT_HEADER 7
523 /** Action string index: Replacement for the "Accept-Language:" header*/
524 #define ACTION_STRING_LANGUAGE 8
525 /** Action string index: Replacement for the "Content-Type:" header*/
526 #define ACTION_STRING_CONTENT_TYPE 9
527 /** Action string index: Replacement for the "content-disposition:" header*/
528 #define ACTION_STRING_CONTENT_DISPOSITION 10
529 /** Action string index: Replacement for the "If-Modified-Since:" header*/
530 #define ACTION_STRING_IF_MODIFIED_SINCE 11
531 /** Action string index: Replacement for the "Last-Modified:" header. */
532 #define ACTION_STRING_LAST_MODIFIED 12
533 /** Action string index: Redirect URL */
534 #define ACTION_STRING_REDIRECT 13
535 /** Action string index: Decode before redirect? */
536 #define ACTION_STRING_FAST_REDIRECTS 14
537 /** Action string index: Overriding forward rule. */
538 #define ACTION_STRING_FORWARD_OVERRIDE 15
539 /** Action string index: Reason for the block. */
540 #define ACTION_STRING_BLOCK 16
541 /** Action string index: what to do with the "X-Forwarded-For" header. */
542 #define ACTION_STRING_CHANGE_X_FORWARDED_FOR 17
543 /** Action string index: how many minutes cookies should be valid. */
544 #define ACTION_STRING_LIMIT_COOKIE_LIFETIME 18
545 /** Number of string actions. */
546 #define ACTION_STRING_COUNT 19
549 /* To make the ugly hack in sed easier to understand */
550 #define CHECK_EVERY_HEADER_REMAINING 0
553 /** Index into current_action_spec::multi[] for headers to add. */
554 #define ACTION_MULTI_ADD_HEADER 0
555 /** Index into current_action_spec::multi[] for content filters to apply. */
556 #define ACTION_MULTI_FILTER 1
557 /** Index into current_action_spec::multi[] for server-header filters to apply. */
558 #define ACTION_MULTI_SERVER_HEADER_FILTER 2
559 /** Index into current_action_spec::multi[] for client-header filters to apply. */
560 #define ACTION_MULTI_CLIENT_HEADER_FILTER 3
561 /** Index into current_action_spec::multi[] for client-header tags to apply. */
562 #define ACTION_MULTI_CLIENT_HEADER_TAGGER 4
563 /** Index into current_action_spec::multi[] for server-header tags to apply. */
564 #define ACTION_MULTI_SERVER_HEADER_TAGGER 5
565 /** Number of multi-string actions. */
566 #define ACTION_MULTI_COUNT 6
570 * This structure contains a list of actions to apply to a URL.
571 * It only contains positive instructions - no "-" options.
572 * It is not used to store the actions list itself, only for
573 * url_actions() to return the current values.
575 struct current_action_spec
577 /** Actions to apply. A bit set to "1" means perform the action. */
581 * Paramaters for those actions that require them.
582 * Each entry is valid if & only if the corresponding entry in "flags" is
585 char * string[ACTION_STRING_COUNT];
587 /** Lists of strings for multi-string actions. */
588 struct list multi[ACTION_MULTI_COUNT][1];
593 * This structure contains a set of changes to actions.
594 * It can contain both positive and negative instructions.
595 * It is used to store an entry in the actions list.
599 unsigned long mask; /**< Actions to keep. A bit set to "0" means remove action. */
600 unsigned long add; /**< Actions to add. A bit set to "1" means add action. */
603 * Paramaters for those actions that require them.
604 * Each entry is valid if & only if the corresponding entry in "flags" is
607 char * string[ACTION_STRING_COUNT];
609 /** Lists of strings to remove, for multi-string actions. */
610 struct list multi_remove[ACTION_MULTI_COUNT][1];
612 /** If nonzero, remove *all* strings from the multi-string action. */
613 int multi_remove_all[ACTION_MULTI_COUNT];
615 /** Lists of strings to add, for multi-string actions. */
616 struct list multi_add[ACTION_MULTI_COUNT][1];
621 * This structure is used to store action files.
623 * It contains an URL or tag pattern, and the changes to
624 * the actions. It's a linked list and should only be
625 * free'd through unload_actions_file() unless there's
626 * only a single entry.
630 struct url_spec url[1]; /**< The URL or tag pattern. */
632 struct action_spec *action; /**< Action settings that might be shared with
633 the list entry before or after the current
634 one and can't be free'd willy nilly. */
636 struct url_actions *next; /**< Next action section in file, or NULL. */
639 enum forwarder_type {
640 /**< Don't use a SOCKS server */
642 /**< original SOCKS 4 protocol */
644 /**< SOCKS 4A, DNS resolution is done by the SOCKS server */
646 /**< SOCKS 5 with hostnames, DNS resolution is done by the SOCKS server */
651 * Structure to hold the server socket and the information
652 * required to make sure we only reuse the connection if
653 * the host and forwarding settings are the same.
655 struct reusable_connection
659 time_t timestamp; /* XXX: rename? */
662 time_t response_received;
665 * Number of seconds after which this
666 * connection will no longer be reused.
668 unsigned int keep_alive_timeout;
670 * Number of requests that were sent to this connection.
671 * This is currently only for debugging purposes.
673 unsigned int requests_sent_total;
677 enum forwarder_type forwarder_type;
686 * Flags for use in csp->flags
690 * Flag for csp->flags: Set if this client is processing data.
691 * Cleared when the thread associated with this structure dies.
693 #define CSP_FLAG_ACTIVE 0x01U
696 * Flag for csp->flags: Set if the server's reply is in "chunked"
699 #define CSP_FLAG_CHUNKED 0x02U
702 * Flag for csp->flags: Set if this request was enforced, although it would
703 * normally have been blocked.
705 #define CSP_FLAG_FORCED 0x04U
708 * Flag for csp->flags: Set if any modification to the body was done.
710 #define CSP_FLAG_MODIFIED 0x08U
713 * Flag for csp->flags: Set if request was blocked.
715 #define CSP_FLAG_REJECTED 0x10U
718 * Flag for csp->flags: Set if we are toggled on (FEATURE_TOGGLE).
720 #define CSP_FLAG_TOGGLED_ON 0x20U
723 * Flag for csp->flags: Set if we answered the request ourselve.
725 #define CSP_FLAG_CRUNCHED 0x40U
728 * Flag for csp->flags: Set if an acceptable Connection header
729 * has already been set by the client.
731 #define CSP_FLAG_CLIENT_CONNECTION_HEADER_SET 0x00000040U
734 * Flag for csp->flags: Set if an acceptable Connection header
735 * has already been set by the server.
737 #define CSP_FLAG_SERVER_CONNECTION_HEADER_SET 0x00000080U
740 * Flag for csp->flags: Signals header parsers whether they
741 * are parsing server or client headers.
743 #define CSP_FLAG_CLIENT_HEADER_PARSING_DONE 0x00000100U
746 * Flag for csp->flags: Set if adding the Host: header
749 #define CSP_FLAG_HOST_HEADER_IS_SET 0x00000200U
752 * Flag for csp->flags: Set if filtering is disabled by X-Filter: No
753 * XXX: As we now have tags we might as well ditch this.
755 #define CSP_FLAG_NO_FILTERING 0x00000400U
758 * Flag for csp->flags: Set the client IP has appended to
759 * an already existing X-Forwarded-For header in which case
760 * no new header has to be generated.
762 #define CSP_FLAG_X_FORWARDED_FOR_APPENDED 0x00000800U
765 * Flag for csp->flags: Set if the server wants to keep
766 * the connection alive.
768 #define CSP_FLAG_SERVER_CONNECTION_KEEP_ALIVE 0x00001000U
771 * Flag for csp->flags: Set if the server specified the
774 #define CSP_FLAG_SERVER_CONTENT_LENGTH_SET 0x00002000U
777 * Flag for csp->flags: Set if we know the content length,
778 * either because the server set it, or we figured it out
781 #define CSP_FLAG_CONTENT_LENGTH_SET 0x00004000U
784 * Flag for csp->flags: Set if the client wants to keep
785 * the connection alive.
787 #define CSP_FLAG_CLIENT_CONNECTION_KEEP_ALIVE 0x00008000U
790 * Flag for csp->flags: Set if we think we got the whole
791 * client request and shouldn't read any additional data
792 * coming from the client until the current request has
795 #define CSP_FLAG_CLIENT_REQUEST_COMPLETELY_READ 0x00010000U
798 * Flag for csp->flags: Set if the server promised us to
799 * keep the connection open for a known number of seconds.
801 #define CSP_FLAG_SERVER_KEEP_ALIVE_TIMEOUT_SET 0x00020000U
804 * Flag for csp->flags: Set if we think we can't reuse
805 * the server socket. XXX: It's also set after sabotaging
806 * pipelining attempts which is somewhat inconsistent with
809 #define CSP_FLAG_SERVER_SOCKET_TAINTED 0x00040000U
812 * Flag for csp->flags: Set if the Proxy-Connection header
813 * is among the server headers.
815 #define CSP_FLAG_SERVER_PROXY_CONNECTION_HEADER_SET 0x00080000U
818 * Flag for csp->flags: Set if the client reused its connection.
820 #define CSP_FLAG_REUSED_CLIENT_CONNECTION 0x00100000U
823 * Flag for csp->flags: Set if the supports deflate compression.
825 #define CSP_FLAG_CLIENT_SUPPORTS_DEFLATE 0x00200000U
828 * Flag for csp->flags: Set if the content has been deflated by Privoxy
830 #define CSP_FLAG_BUFFERED_CONTENT_DEFLATED 0x00400000U
833 * Flag for csp->flags: Set if we already read (parts of)
834 * a pipelined request in which case the client obviously
835 * isn't done talking.
837 #define CSP_FLAG_PIPELINED_REQUEST_WAITING 0x00800000U
840 * Flags for use in return codes of child processes
844 * Flag for process return code: Set if exiting porcess has been toggled
845 * during its lifetime.
847 #define RC_FLAG_TOGGLED 0x10
850 * Flag for process return code: Set if exiting porcess has blocked its
853 #define RC_FLAG_BLOCKED 0x20
856 * Maximum number of actions/filter files. This limit is arbitrary - it's just used
859 #define MAX_AF_FILES 10
862 * Maximum number of sockets to listen to. This limit is arbitrary - it's just used
865 #define MAX_LISTENING_SOCKETS 10
868 * The state of a Privoxy processing thread.
872 /** The proxy's configuration */
873 struct configuration_spec * config;
875 /** The actions to perform on the current request */
876 struct current_action_spec action[1];
878 /** socket to talk to client (web browser) */
881 /** Number of requests received on the client socket. */
882 unsigned int requests_received_total;
884 /** current connection to the server (may go through a proxy) */
885 struct reusable_connection server_connection;
887 /** Multi-purpose flag container, see CSP_FLAG_* above */
890 /** Client PC's IP address, as reported by the accept() function.
894 /** Client PC's TCP address, as reported by the accept() function.
896 struct sockaddr_storage tcp_addr;
898 /** Client PC's IP address, as reported by the accept() function.
900 unsigned long ip_addr_long;
901 #endif /* def HAVE_RFC2553 */
903 /** The URL that was requested */
904 struct http_request http[1];
907 * The final forwarding settings.
908 * XXX: Currently this is only used for forward-override,
909 * so we can free the space in sweep.
911 struct forward_spec * fwd;
913 /** An I/O buffer used for buffering data read from the server */
914 /* XXX: should be renamed to server_iob */
917 /** An I/O buffer used for buffering data read from the client */
918 struct iob client_iob[1];
920 /** List of all headers for this request */
921 struct list headers[1];
923 /** List of all tags that apply to this request */
926 /** MIME-Type key, see CT_* above */
927 unsigned int content_type;
929 /** Actions files associated with this client */
930 struct file_list *actions_list[MAX_AF_FILES];
932 /** pcrs job files. */
933 struct file_list *rlist[MAX_AF_FILES];
935 /** Length after content modification. */
936 unsigned long long content_length;
938 /* XXX: is this the right location? */
940 /** Expected length of content after which we
941 * should stop reading from the server socket.
943 unsigned long long expected_content_length;
945 /** Expected length of content after which we
946 * should stop reading from the client socket.
948 unsigned long long expected_client_content_length;
953 struct file_list *tlist;
955 #endif /* def FEATURE_TRUST */
958 * Failure reason to embedded in the CGI error page,
959 * or NULL. Currently only used for socks errors.
965 * List of client states so the main thread can keep
966 * track of them and garbage collect their resources.
970 struct client_states *next;
971 struct client_state csp;
975 * A function to add a header
977 typedef jb_err (*add_header_func_ptr)(struct client_state *);
980 * A function to process a header
982 typedef jb_err (*parser_func_ptr )(struct client_state *, char **);
986 * List of available CGI functions.
988 struct cgi_dispatcher
990 /** The URL of the CGI, relative to the CGI root. */
991 const char * const name;
993 /** The handler function for the CGI */
994 jb_err (* const handler)(struct client_state *csp, struct http_response *rsp, const struct map *parameters);
996 /** The description of the CGI, to appear on the main menu, or NULL to hide it. */
997 const char * const description;
999 /** A flag that indicates whether unintentional calls to this CGI can cause damage */
1005 * A data file used by Privoxy. Kept in a linked list.
1010 * This is a pointer to the data structures associated with the file.
1011 * Read-only once the structure has been created.
1016 * The unloader function.
1017 * Normally NULL. When we are finished with file (i.e. when we have
1018 * loaded a new one), set to a pointer to an unloader function.
1019 * Unloader will be called by sweep() (called from main loop) when
1020 * all clients using this file are done. This prevents threading
1023 void (*unloader)(void *);
1026 * Used internally by sweep(). Do not access from elsewhere.
1031 * File last-modified time, so we can check if file has been changed.
1032 * Read-only once the structure has been created.
1034 time_t lastmodified;
1037 * The full filename.
1042 * Pointer to next entry in the linked list of all "file_list"s.
1043 * This linked list is so that sweep() can navigate it.
1044 * Since sweep() can remove items from the list, we must be careful
1045 * to only access this value from main thread (when we know sweep
1046 * won't be running).
1048 struct file_list *next;
1052 #ifdef FEATURE_TRUST
1055 * The format of a trust file when loaded into memory.
1059 struct url_spec url[1]; /**< The URL pattern */
1060 int reject; /**< FIXME: Please document this! */
1061 struct block_spec *next; /**< Next entry in linked list */
1065 * Arbitrary limit for the number of trusted referrers.
1067 #define MAX_TRUSTED_REFERRERS 512
1069 #endif /* def FEATURE_TRUST */
1072 * How to forward a connection to a parent proxy.
1076 /** URL pattern that this forward_spec is for. */
1077 struct url_spec url[1];
1079 /** Connection type. Must be SOCKS_NONE, SOCKS_4, SOCKS_4A or SOCKS_5. */
1080 enum forwarder_type type;
1082 /** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
1085 /** SOCKS server port. */
1088 /** Parent HTTP proxy hostname, or NULL for none. */
1091 /** Parent HTTP proxy port. */
1094 /** Next entry in the linked list. */
1095 struct forward_spec *next;
1100 * Initializer for a static struct forward_spec.
1102 #define FORWARD_SPEC_INITIALIZER { { URL_SPEC_INITIALIZER }, 0, NULL, 0, NULL, 0, NULL }
1104 /* Supported filter types */
1107 FT_CONTENT_FILTER = 0,
1108 FT_CLIENT_HEADER_FILTER = 1,
1109 FT_SERVER_HEADER_FILTER = 2,
1110 FT_CLIENT_HEADER_TAGGER = 3,
1111 FT_SERVER_HEADER_TAGGER = 4,
1112 FT_INVALID_FILTER = 42,
1114 #define MAX_FILTER_TYPES 5
1117 * This struct represents one filter (one block) from
1118 * the re_filterfile. If there is more than one filter
1119 * in the file, the file will be represented by a
1120 * chained list of re_filterfile specs.
1122 struct re_filterfile_spec
1124 char *name; /**< Name from FILTER: statement in re_filterfile. */
1125 char *description; /**< Description from FILTER: statement in re_filterfile. */
1126 struct list patterns[1]; /**< The patterns from the re_filterfile. */
1127 pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
1128 enum filter_type type; /**< Filter type (content, client-header, server-header). */
1129 int dynamic; /**< Set to one if the pattern might contain variables
1130 and has to be recompiled for every request. */
1131 struct re_filterfile_spec *next; /**< The pointer for chaining. */
1137 #define ACL_PERMIT 1 /**< Accept connection request */
1138 #define ACL_DENY 2 /**< Reject connection request */
1141 * An IP address pattern. Used to specify networks in the ACL.
1143 struct access_control_addr
1146 struct sockaddr_storage addr; /* <The TCP address in network order. */
1147 struct sockaddr_storage mask; /* <The TCP mask in network order. */
1149 unsigned long addr; /**< The IP address as an integer. */
1150 unsigned long mask; /**< The network mask as an integer. */
1151 unsigned long port; /**< The port number. */
1152 #endif /* HAVE_RFC2553 */
1156 * An access control list (ACL) entry.
1158 * This is a linked list.
1160 struct access_control_list
1162 struct access_control_addr src[1]; /**< Client IP address */
1163 struct access_control_addr dst[1]; /**< Website or parent proxy IP address */
1165 int wildcard_dst; /** < dst address is wildcard */
1168 short action; /**< ACL_PERMIT or ACL_DENY */
1169 struct access_control_list *next; /**< The next entry in the ACL. */
1172 #endif /* def FEATURE_ACL */
1175 /** Maximum number of loaders (actions, re_filter, ...) */
1179 /** configuration_spec::feature_flags: CGI actions editor. */
1180 #define RUNTIME_FEATURE_CGI_EDIT_ACTIONS 1U
1182 /** configuration_spec::feature_flags: Web-based toggle. */
1183 #define RUNTIME_FEATURE_CGI_TOGGLE 2U
1185 /** configuration_spec::feature_flags: HTTP-header-based toggle. */
1186 #define RUNTIME_FEATURE_HTTP_TOGGLE 4U
1188 /** configuration_spec::feature_flags: Split large forms to limit the number of GET arguments. */
1189 #define RUNTIME_FEATURE_SPLIT_LARGE_FORMS 8U
1191 /** configuration_spec::feature_flags: Check the host header for requests with host-less request lines. */
1192 #define RUNTIME_FEATURE_ACCEPT_INTERCEPTED_REQUESTS 16U
1194 /** configuration_spec::feature_flags: Don't allow to circumvent blocks with the force prefix. */
1195 #define RUNTIME_FEATURE_ENFORCE_BLOCKS 32U
1197 /** configuration_spec::feature_flags: Allow to block or redirect CGI requests. */
1198 #define RUNTIME_FEATURE_CGI_CRUNCHING 64U
1200 /** configuration_spec::feature_flags: Try to keep the connection to the server alive. */
1201 #define RUNTIME_FEATURE_CONNECTION_KEEP_ALIVE 128U
1203 /** configuration_spec::feature_flags: Share outgoing connections between different client connections. */
1204 #define RUNTIME_FEATURE_CONNECTION_SHARING 256U
1206 /** configuration_spec::feature_flags: Pages blocked with +handle-as-empty-doc get a return status of 200 OK. */
1207 #define RUNTIME_FEATURE_EMPTY_DOC_RETURNS_OK 512U
1209 /** configuration_spec::feature_flags: Buffered content is sent compressed if the client supports it. */
1210 #define RUNTIME_FEATURE_COMPRESSION 1024U
1212 /** configuration_spec::feature_flags: Pipelined requests are served instead of being discarded. */
1213 #define RUNTIME_FEATURE_TOLERATE_PIPELINING 2048U
1216 * Data loaded from the configuration file.
1218 * (Anomaly: toggle is still handled through a global, not this structure)
1220 struct configuration_spec
1225 /** Nonzero to enable multithreading. */
1229 * Bitmask of features that can be enabled/disabled through the config
1230 * file. Currently defined bits:
1232 * - RUNTIME_FEATURE_CGI_EDIT_ACTIONS
1233 * - RUNTIME_FEATURE_CGI_TOGGLE
1234 * - RUNTIME_FEATURE_HTTP_TOGGLE
1235 * - RUNTIME_FEATURE_SPLIT_LARGE_FORMS
1236 * - RUNTIME_FEATURE_ACCEPT_INTERCEPTED_REQUESTS
1237 * - RUNTIME_FEATURE_ENFORCE_BLOCKS
1238 * - RUNTIME_FEATURE_CGI_CRUNCHING
1239 * - RUNTIME_FEATURE_CONNECTION_KEEP_ALIVE
1240 * - RUNTIME_FEATURE_CONNECTION_SHARING
1241 * - RUNTIME_FEATURE_EMPTY_DOC_RETURNS_OK
1243 unsigned feature_flags;
1245 /** The log file name. */
1246 const char *logfile;
1248 /** The config file directory. */
1249 const char *confdir;
1251 /** The directory for customized CGI templates. */
1252 const char *templdir;
1254 /** The log file directory. */
1257 /** The full paths to the actions files. */
1258 const char *actions_file[MAX_AF_FILES];
1260 /** The short names of the actions files. */
1261 const char *actions_file_short[MAX_AF_FILES];
1263 /** The administrator's email address */
1264 char *admin_address;
1266 /** A URL with info on this proxy */
1267 char *proxy_info_url;
1269 /** URL to the user manual (on our website or local copy) */
1272 /** The file names of the pcre filter files. */
1273 const char *re_filterfile[MAX_AF_FILES];
1275 /** The short names of the pcre filter files. */
1276 const char *re_filterfile_short[MAX_AF_FILES];
1278 /**< List of ordered client header names. */
1279 struct list ordered_client_headers[1];
1281 /** The hostname to show on CGI pages, or NULL to use the real one. */
1282 const char *hostname;
1284 /** IP addresses to bind to. Defaults to HADDR_DEFAULT == 127.0.0.1. */
1285 const char *haddr[MAX_LISTENING_SOCKETS];
1287 /** Ports to bind to. Defaults to HADDR_PORT == 8118. */
1288 int hport[MAX_LISTENING_SOCKETS];
1290 /** Size limit for IOB */
1291 size_t buffer_limit;
1293 #ifdef FEATURE_TRUST
1295 /** The file name of the trust file. */
1296 const char * trustfile;
1298 /** FIXME: DOCME: Document this. */
1299 struct list trust_info[1];
1301 /** FIXME: DOCME: Document this. */
1302 struct url_spec *trust_list[MAX_TRUSTED_REFERRERS];
1304 #endif /* def FEATURE_TRUST */
1308 /** The access control list (ACL). */
1309 struct access_control_list *acl;
1311 #endif /* def FEATURE_ACL */
1313 /** Information about parent proxies (forwarding). */
1314 struct forward_spec *forward;
1316 /** Number of retries in case a forwarded connection attempt fails */
1317 int forwarded_connect_retries;
1319 /** Maximum number of client connections. */
1320 int max_client_connections;
1322 /* Timeout when waiting on sockets for data to become available. */
1325 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
1326 /* Maximum number of seconds after which an open connection will no longer be reused. */
1327 unsigned int keep_alive_timeout;
1329 /* Assumed server-side keep alive timeout if none is specified. */
1330 unsigned int default_server_timeout;
1333 #ifdef FEATURE_COMPRESSION
1334 int compression_level;
1337 /** All options from the config file, HTML-formatted. */
1340 /** The configuration file object. */
1341 struct file_list *config_file_list;
1343 /** List of loaders */
1344 int (*loaders[NLOADERS])(struct client_state *);
1346 /** Nonzero if we need to bind() to the new port. */
1350 /** Calculates the number of elements in an array, using sizeof. */
1351 #define SZ(X) (sizeof(X) / sizeof(*X))
1353 #ifdef FEATURE_FORCE_LOAD
1354 /** The force load URL prefix. */
1355 #define FORCE_PREFIX "/PRIVOXY-FORCE"
1356 #endif /* def FEATURE_FORCE_LOAD */
1358 #ifdef FEATURE_NO_GIFS
1359 /** The MIME type for images ("image/png" or "image/gif"). */
1360 #define BUILTIN_IMAGE_MIMETYPE "image/png"
1362 #define BUILTIN_IMAGE_MIMETYPE "image/gif"
1363 #endif /* def FEATURE_NO_GIFS */
1370 /** URL for the Privoxy home page. */
1371 #define HOME_PAGE_URL "http://www.privoxy.org/"
1373 /** URL for the Privoxy user manual. */
1374 #define USER_MANUAL_URL HOME_PAGE_URL VERSION "/user-manual/"
1376 /** Prefix for actions help links (append to USER_MANUAL_URL). */
1377 #define ACTIONS_HELP_PREFIX "actions-file.html#"
1379 /** Prefix for config option help links (append to USER_MANUAL_URL). */
1380 #define CONFIG_HELP_PREFIX "config.html#"
1383 * The "hosts" to intercept and display CGI pages.
1384 * First one is a hostname only, second one can specify host and path.
1387 * 1) Do not specify the http: prefix
1388 * 2) CGI_SITE_2_PATH must not end with /, one will be added automatically.
1389 * 3) CGI_SITE_2_PATH must start with /, unless it is the empty string.
1391 #define CGI_SITE_1_HOST "p.p"
1392 #define CGI_SITE_2_HOST "config.privoxy.org"
1393 #define CGI_SITE_2_PATH ""
1396 * The prefix for CGI pages. Written out in generated HTML.
1397 * INCLUDES the trailing slash.
1399 #define CGI_PREFIX "http://" CGI_SITE_2_HOST CGI_SITE_2_PATH "/"
1405 #endif /* ndef PROJECT_H_INCLUDED */