1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
4 #define PROJECT_H_VERSION "$Id: project.h,v 1.199 2013/11/24 14:23:28 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...
113 typedef SOCKET jb_socket;
115 #define JB_INVALID_SOCKET INVALID_SOCKET
117 #else /* ndef _WIN32 */
120 * The type used by sockets. On UNIX it's an int. Microsoft decided to
121 * make it an unsigned.
123 typedef int jb_socket;
126 * The error value used for variables of type jb_socket. On UNIX this
127 * is -1, however Microsoft decided to make socket handles unsigned, so
128 * they use a different value.
131 #define JB_INVALID_SOCKET (-1)
133 #endif /* ndef _WIN32 */
137 * A standard error code. This should be JB_ERR_OK or one of the JB_ERR_xxx
142 #define JB_ERR_OK 0 /**< Success, no error */
143 #define JB_ERR_MEMORY 1 /**< Out of memory */
144 #define JB_ERR_CGI_PARAMS 2 /**< Missing or corrupt CGI parameters */
145 #define JB_ERR_FILE 3 /**< Error opening, reading or writing a file */
146 #define JB_ERR_PARSE 4 /**< Error parsing file */
147 #define JB_ERR_MODIFIED 5 /**< File has been modified outside of the
148 CGI actions editor. */
149 #define JB_ERR_COMPRESS 6 /**< Error on decompression */
152 * This macro is used to free a pointer that may be NULL.
153 * It also sets the variable to NULL after it's been freed.
154 * The paramater should be a simple variable without side effects.
156 #define freez(X) { if(X) { free((void*)X); X = NULL ; } }
160 * Macro definitions for platforms where isspace() and friends
161 * are macros that use their argument directly as an array index
162 * and thus better be positive. Supposedly that's the case on
163 * some unspecified Solaris versions.
164 * Note: Remember to #include <ctype.h> if you use these macros.
166 #define privoxy_isdigit(__X) isdigit((int)(unsigned char)(__X))
167 #define privoxy_isupper(__X) isupper((int)(unsigned char)(__X))
168 #define privoxy_toupper(__X) toupper((int)(unsigned char)(__X))
169 #define privoxy_tolower(__X) tolower((int)(unsigned char)(__X))
170 #define privoxy_isspace(__X) isspace((int)(unsigned char)(__X))
173 * Use for statically allocated buffers if you have no other choice.
174 * Remember to check the length of what you write into the buffer
175 * - we don't want any buffer overflows!
177 #define BUFFER_SIZE 5000
180 * Max length of CGI parameters (arbitrary limit).
182 #define CGI_PARAM_LEN_MAX 500U
185 * Buffer size for capturing struct hostent data in the
186 * gethostby(name|addr)_r library calls. Since we don't
187 * loop over gethostbyname_r, the buffer must be sufficient
188 * to accommodate multiple IN A RRs, as used in DNS round robin
189 * load balancing. W3C's wwwlib uses 1K, so that should be
190 * good enough for us, too.
193 * XXX: Temporary doubled, for some configurations
194 * 1K is still too small and we didn't get the
195 * real fix ready for inclusion.
197 #define HOSTENT_BUFFER_SIZE 2048
200 * Do not use. Originally this was so that you can
201 * say "while (FOREVER) { ...do something... }".
202 * However, this gives a warning with some compilers (e.g. VC++).
203 * Instead, use "for (;;) { ...do something... }".
208 * Default TCP/IP address to listen on, as a string.
209 * Set to "127.0.0.1:8118".
211 #define HADDR_DEFAULT "127.0.0.1:8118"
214 /* Forward def for struct client_state */
215 struct configuration_spec;
219 * Entry in a linked list of strings.
224 * The string pointer. It must point to a dynamically malloc()ed
225 * string or be NULL for the list functions to work. In the latter
226 * case, just be careful next time you iterate through the list in
231 /** Next entry in the linked list, or NULL if no more. */
232 struct list_entry *next;
236 * A header for a linked list of strings.
240 /** First entry in the list, or NULL if the list is empty. */
241 struct list_entry *first;
243 /** Last entry in the list, or NULL if the list is empty. */
244 struct list_entry *last;
249 * An entry in a map. This is a name=value pair.
253 /** The key for the map. */
255 /** The value associated with that key. */
257 /** The next map entry, or NULL if none. */
258 struct map_entry *next;
262 * A map from a string to another string.
263 * This is used for the paramaters passed in a HTTP GET request, and
264 * to store the exports when the CGI interface is filling in a template.
268 /** The first map entry, or NULL if the map is empty. */
269 struct map_entry *first;
270 /** The last map entry, or NULL if the map is empty. */
271 struct map_entry *last;
276 * A HTTP request. This includes the method (GET, POST) and
279 * This is also used whenever we want to match a URL against a
280 * URL pattern. This always contains the URL to match, and never
281 * a URL pattern. (See struct url_spec).
285 char *cmd; /**< Whole command line: method, URL, Version */
286 char *ocmd; /**< Backup of original cmd for CLF logging */
287 char *gpc; /**< HTTP method: GET, POST, ... */
288 char *url; /**< The URL */
289 char *ver; /**< Protocol version */
290 int status; /**< HTTP Status */
292 char *host; /**< Host part of URL */
293 int port; /**< Port of URL or 80 (default) */
294 char *path; /**< Path of URL */
295 char *hostport; /**< host[:port] */
296 int ssl; /**< Flag if protocol is https */
298 char *host_ip_addr_str; /**< String with dotted decimal representation
299 of host's IP. NULL before connect_to() */
301 #ifndef FEATURE_EXTENDED_HOST_PATTERNS
302 char *dbuffer; /**< Buffer with '\0'-delimited domain name. */
303 char **dvec; /**< List of pointers to the strings in dbuffer. */
304 int dcount; /**< How many parts to this domain? (length of dvec) */
305 #endif /* ndef FEATURE_EXTENDED_HOST_PATTERNS */
309 * Reasons for generating a http_response instead of delivering
310 * the requested resource. Mostly ordered the way they are checked
330 * Response generated by CGI, blocker, or error handler
334 char *status; /**< HTTP status (string). */
335 struct list headers[1]; /**< List of header lines. */
336 char *head; /**< Formatted http response head. */
337 size_t head_length; /**< Length of http response head. */
338 char *body; /**< HTTP document body. */
339 size_t content_length; /**< Length of body, REQUIRED if binary body. */
340 int is_static; /**< Nonzero if the content will never change and
341 should be cached by the browser (e.g. images). */
342 enum crunch_reason crunch_reason; /**< Why the response was generated in the first place. */
347 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
348 regex_t *host_regex;/**< Regex for host matching */
350 char *dbuffer; /**< Buffer with '\0'-delimited domain name, or NULL to match all hosts. */
351 char **dvec; /**< List of pointers to the strings in dbuffer. */
352 int dcount; /**< How many parts to this domain? (length of dvec) */
353 int unanchored; /**< Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT. */
354 #endif /* defined FEATURE_EXTENDED_HOST_PATTERNS */
356 char *port_list; /**< List of acceptable ports, or NULL to match all ports */
358 regex_t *preg; /**< Regex for matching path part */
362 * A URL or a tag pattern.
366 /** The string which was parsed to produce this pattern_spec.
367 Used for debugging or display only. */
372 struct url_spec url_spec;
376 unsigned int flags; /**< Bitmap with various pattern properties. */
380 * Constant for host part matching in URLs. If set, indicates that the start of
381 * the pattern must match the start of the URL. E.g. this is not set for the
382 * pattern ".example.com", so that it will match both "example.com" and
383 * "www.example.com". It is set for the pattern "example.com", which makes it
384 * match "example.com" only, not "www.example.com".
386 #define ANCHOR_LEFT 1
389 * Constant for host part matching in URLs. If set, indicates that the end of
390 * the pattern must match the end of the URL. E.g. this is not set for the
391 * pattern "ad.", so that it will match any host called "ad", irrespective
392 * of how many subdomains are in the fully-qualified domain name.
394 #define ANCHOR_RIGHT 2
396 /** Pattern spec bitmap: It's an URL pattern. */
397 #define PATTERN_SPEC_URL_PATTERN 0x00000001UL
399 /** Pattern spec bitmap: It's a TAG pattern. */
400 #define PATTERN_SPEC_TAG_PATTERN 0x00000002UL
402 /** Pattern spec bitmap: It's a NO-REQUEST-TAG pattern. */
403 #define PATTERN_SPEC_NO_REQUEST_TAG_PATTERN 0x00000004UL
405 /** Pattern spec bitmap: It's a NO-RESPONSE-TAG pattern. */
406 #define PATTERN_SPEC_NO_RESPONSE_TAG_PATTERN 0x00000008UL
409 * An I/O buffer. Holds a string which can be appended to, and can have data
410 * removed from the beginning.
414 char *buf; /**< Start of buffer */
415 char *cur; /**< Start of relevant data */
416 char *eod; /**< End of relevant data */
417 size_t size; /**< Size as malloc()ed */
422 * Return the number of bytes in the I/O buffer associated with the passed
423 * I/O buffer. May be zero.
425 #define IOB_PEEK(IOB) ((IOB->cur > IOB->eod) ? (IOB->eod - IOB->cur) : 0)
428 /* Bits for csp->content_type bitmask: */
429 #define CT_TEXT 0x0001U /**< Suitable for pcrs filtering. */
430 #define CT_GIF 0x0002U /**< Suitable for GIF filtering. */
431 #define CT_TABOO 0x0004U /**< DO NOT filter, irrespective of other flags. */
433 /* Although these are not, strictly speaking, content types
434 * (they are content encodings), it is simple to handle them
437 #define CT_GZIP 0x0010U /**< gzip-compressed data. */
438 #define CT_DEFLATE 0x0020U /**< zlib-compressed data. */
441 * Flag to signal that the server declared the content type,
442 * so we can differentiate between unknown and undeclared
445 #define CT_DECLARED 0x0040U
448 * The mask which includes all actions.
450 #define ACTION_MASK_ALL (~0UL)
453 * The most compatible set of actions - i.e. none.
455 #define ACTION_MOST_COMPATIBLE 0x00000000UL
457 /** Action bitmap: Block the request. */
458 #define ACTION_BLOCK 0x00000001UL
459 /** Action bitmap: Deanimate if it's a GIF. */
460 #define ACTION_DEANIMATE 0x00000002UL
461 /** Action bitmap: Downgrade HTTP/1.1 to 1.0. */
462 #define ACTION_DOWNGRADE 0x00000004UL
463 /** Action bitmap: Fast redirects. */
464 #define ACTION_FAST_REDIRECTS 0x00000008UL
465 /** Action bitmap: Remove or add "X-Forwarded-For" header. */
466 #define ACTION_CHANGE_X_FORWARDED_FOR 0x00000010UL
467 /** Action bitmap: Hide "From" header. */
468 #define ACTION_HIDE_FROM 0x00000020UL
469 /** Action bitmap: Hide "Referer" header. (sic - follow HTTP, not English). */
470 #define ACTION_HIDE_REFERER 0x00000040UL
471 /** Action bitmap: Hide "User-Agent" and similar headers. */
472 #define ACTION_HIDE_USER_AGENT 0x00000080UL
473 /** Action bitmap: This is an image. */
474 #define ACTION_IMAGE 0x00000100UL
475 /** Action bitmap: Sets the image blocker. */
476 #define ACTION_IMAGE_BLOCKER 0x00000200UL
477 /** Action bitmap: Prevent compression. */
478 #define ACTION_NO_COMPRESSION 0x00000400UL
479 /** Action bitmap: Change cookies to session only cookies. */
480 #define ACTION_SESSION_COOKIES_ONLY 0x00000800UL
481 /** Action bitmap: Block cookies coming from the client. */
482 #define ACTION_CRUNCH_OUTGOING_COOKIES 0x00001000UL
483 /** Action bitmap: Block cookies coming from the server. */
484 #define ACTION_CRUNCH_INCOMING_COOKIES 0x00002000UL
485 /** Action bitmap: Override the forward settings in the config file */
486 #define ACTION_FORWARD_OVERRIDE 0x00004000UL
487 /** Action bitmap: Block as empty document */
488 #define ACTION_HANDLE_AS_EMPTY_DOCUMENT 0x00008000UL
489 /** Action bitmap: Limit CONNECT requests to safe ports. */
490 #define ACTION_LIMIT_CONNECT 0x00010000UL
491 /** Action bitmap: Redirect request. */
492 #define ACTION_REDIRECT 0x00020000UL
493 /** Action bitmap: Crunch or modify "if-modified-since" header. */
494 #define ACTION_HIDE_IF_MODIFIED_SINCE 0x00040000UL
495 /** Action bitmap: Overwrite Content-Type header. */
496 #define ACTION_CONTENT_TYPE_OVERWRITE 0x00080000UL
497 /** Action bitmap: Crunch specified server header. */
498 #define ACTION_CRUNCH_SERVER_HEADER 0x00100000UL
499 /** Action bitmap: Crunch specified client header */
500 #define ACTION_CRUNCH_CLIENT_HEADER 0x00200000UL
501 /** Action bitmap: Enable text mode by force */
502 #define ACTION_FORCE_TEXT_MODE 0x00400000UL
503 /** Action bitmap: Enable text mode by force */
504 #define ACTION_CRUNCH_IF_NONE_MATCH 0x00800000UL
505 /** Action bitmap: Enable content-disposition crunching */
506 #define ACTION_HIDE_CONTENT_DISPOSITION 0x01000000UL
507 /** Action bitmap: Replace or block Last-Modified header */
508 #define ACTION_OVERWRITE_LAST_MODIFIED 0x02000000UL
509 /** Action bitmap: Replace or block Accept-Language header */
510 #define ACTION_HIDE_ACCEPT_LANGUAGE 0x04000000UL
511 /** Action bitmap: Limit the cookie lifetime */
512 #define ACTION_LIMIT_COOKIE_LIFETIME 0x08000000UL
515 /** Action string index: How to deanimate GIFs */
516 #define ACTION_STRING_DEANIMATE 0
517 /** Action string index: Replacement for "From:" header */
518 #define ACTION_STRING_FROM 1
519 /** Action string index: How to block images */
520 #define ACTION_STRING_IMAGE_BLOCKER 2
521 /** Action string index: Replacement for "Referer:" header */
522 #define ACTION_STRING_REFERER 3
523 /** Action string index: Replacement for "User-Agent:" header */
524 #define ACTION_STRING_USER_AGENT 4
525 /** Action string index: Legal CONNECT ports. */
526 #define ACTION_STRING_LIMIT_CONNECT 5
527 /** Action string index: Server headers containing this pattern are crunched*/
528 #define ACTION_STRING_SERVER_HEADER 6
529 /** Action string index: Client headers containing this pattern are crunched*/
530 #define ACTION_STRING_CLIENT_HEADER 7
531 /** Action string index: Replacement for the "Accept-Language:" header*/
532 #define ACTION_STRING_LANGUAGE 8
533 /** Action string index: Replacement for the "Content-Type:" header*/
534 #define ACTION_STRING_CONTENT_TYPE 9
535 /** Action string index: Replacement for the "content-disposition:" header*/
536 #define ACTION_STRING_CONTENT_DISPOSITION 10
537 /** Action string index: Replacement for the "If-Modified-Since:" header*/
538 #define ACTION_STRING_IF_MODIFIED_SINCE 11
539 /** Action string index: Replacement for the "Last-Modified:" header. */
540 #define ACTION_STRING_LAST_MODIFIED 12
541 /** Action string index: Redirect URL */
542 #define ACTION_STRING_REDIRECT 13
543 /** Action string index: Decode before redirect? */
544 #define ACTION_STRING_FAST_REDIRECTS 14
545 /** Action string index: Overriding forward rule. */
546 #define ACTION_STRING_FORWARD_OVERRIDE 15
547 /** Action string index: Reason for the block. */
548 #define ACTION_STRING_BLOCK 16
549 /** Action string index: what to do with the "X-Forwarded-For" header. */
550 #define ACTION_STRING_CHANGE_X_FORWARDED_FOR 17
551 /** Action string index: how many minutes cookies should be valid. */
552 #define ACTION_STRING_LIMIT_COOKIE_LIFETIME 18
553 /** Number of string actions. */
554 #define ACTION_STRING_COUNT 19
557 /* To make the ugly hack in sed easier to understand */
558 #define CHECK_EVERY_HEADER_REMAINING 0
561 /** Index into current_action_spec::multi[] for headers to add. */
562 #define ACTION_MULTI_ADD_HEADER 0
563 /** Index into current_action_spec::multi[] for content filters to apply. */
564 #define ACTION_MULTI_FILTER 1
565 /** Index into current_action_spec::multi[] for server-header filters to apply. */
566 #define ACTION_MULTI_SERVER_HEADER_FILTER 2
567 /** Index into current_action_spec::multi[] for client-header filters to apply. */
568 #define ACTION_MULTI_CLIENT_HEADER_FILTER 3
569 /** Index into current_action_spec::multi[] for client-header tags to apply. */
570 #define ACTION_MULTI_CLIENT_HEADER_TAGGER 4
571 /** Index into current_action_spec::multi[] for server-header tags to apply. */
572 #define ACTION_MULTI_SERVER_HEADER_TAGGER 5
573 /** Number of multi-string actions. */
574 #define ACTION_MULTI_COUNT 6
578 * This structure contains a list of actions to apply to a URL.
579 * It only contains positive instructions - no "-" options.
580 * It is not used to store the actions list itself, only for
581 * url_actions() to return the current values.
583 struct current_action_spec
585 /** Actions to apply. A bit set to "1" means perform the action. */
589 * Paramaters for those actions that require them.
590 * Each entry is valid if & only if the corresponding entry in "flags" is
593 char * string[ACTION_STRING_COUNT];
595 /** Lists of strings for multi-string actions. */
596 struct list multi[ACTION_MULTI_COUNT][1];
601 * This structure contains a set of changes to actions.
602 * It can contain both positive and negative instructions.
603 * It is used to store an entry in the actions list.
607 unsigned long mask; /**< Actions to keep. A bit set to "0" means remove action. */
608 unsigned long add; /**< Actions to add. A bit set to "1" means add action. */
611 * Paramaters for those actions that require them.
612 * Each entry is valid if & only if the corresponding entry in "flags" is
615 char * string[ACTION_STRING_COUNT];
617 /** Lists of strings to remove, for multi-string actions. */
618 struct list multi_remove[ACTION_MULTI_COUNT][1];
620 /** If nonzero, remove *all* strings from the multi-string action. */
621 int multi_remove_all[ACTION_MULTI_COUNT];
623 /** Lists of strings to add, for multi-string actions. */
624 struct list multi_add[ACTION_MULTI_COUNT][1];
629 * This structure is used to store action files.
631 * It contains an URL or tag pattern, and the changes to
632 * the actions. It's a linked list and should only be
633 * free'd through unload_actions_file() unless there's
634 * only a single entry.
638 struct pattern_spec url[1]; /**< The URL or tag pattern. */
640 struct action_spec *action; /**< Action settings that might be shared with
641 the list entry before or after the current
642 one and can't be free'd willy nilly. */
644 struct url_actions *next; /**< Next action section in file, or NULL. */
647 enum forwarder_type {
648 /**< Don't use a SOCKS server */
650 /**< original SOCKS 4 protocol */
652 /**< SOCKS 4A, DNS resolution is done by the SOCKS server */
654 /**< SOCKS 5 with hostnames, DNS resolution is done by the SOCKS server */
656 /**< Like SOCKS5, but uses non-standard Tor extensions (currently only optimistic data) */
661 * Structure to hold the server socket and the information
662 * required to make sure we only reuse the connection if
663 * the host and forwarding settings are the same.
665 struct reusable_connection
669 time_t timestamp; /* XXX: rename? */
672 time_t response_received;
675 * Number of seconds after which this
676 * connection will no longer be reused.
678 unsigned int keep_alive_timeout;
680 * Number of requests that were sent to this connection.
681 * This is currently only for debugging purposes.
683 unsigned int requests_sent_total;
687 enum forwarder_type forwarder_type;
696 * Flags for use in csp->flags
700 * Flag for csp->flags: Set if this client is processing data.
701 * Cleared when the thread associated with this structure dies.
703 #define CSP_FLAG_ACTIVE 0x01U
706 * Flag for csp->flags: Set if the server's reply is in "chunked"
709 #define CSP_FLAG_CHUNKED 0x02U
712 * Flag for csp->flags: Set if this request was enforced, although it would
713 * normally have been blocked.
715 #define CSP_FLAG_FORCED 0x04U
718 * Flag for csp->flags: Set if any modification to the body was done.
720 #define CSP_FLAG_MODIFIED 0x08U
723 * Flag for csp->flags: Set if request was blocked.
725 #define CSP_FLAG_REJECTED 0x10U
728 * Flag for csp->flags: Set if we are toggled on (FEATURE_TOGGLE).
730 #define CSP_FLAG_TOGGLED_ON 0x20U
733 * Flag for csp->flags: Set if we answered the request ourselve.
735 #define CSP_FLAG_CRUNCHED 0x40U
738 * Flag for csp->flags: Set if an acceptable Connection header
739 * has already been set by the client.
741 #define CSP_FLAG_CLIENT_CONNECTION_HEADER_SET 0x00000040U
744 * Flag for csp->flags: Set if an acceptable Connection header
745 * has already been set by the server.
747 #define CSP_FLAG_SERVER_CONNECTION_HEADER_SET 0x00000080U
750 * Flag for csp->flags: Signals header parsers whether they
751 * are parsing server or client headers.
753 #define CSP_FLAG_CLIENT_HEADER_PARSING_DONE 0x00000100U
756 * Flag for csp->flags: Set if adding the Host: header
759 #define CSP_FLAG_HOST_HEADER_IS_SET 0x00000200U
762 * Flag for csp->flags: Set if filtering is disabled by X-Filter: No
763 * XXX: As we now have tags we might as well ditch this.
765 #define CSP_FLAG_NO_FILTERING 0x00000400U
768 * Flag for csp->flags: Set the client IP has appended to
769 * an already existing X-Forwarded-For header in which case
770 * no new header has to be generated.
772 #define CSP_FLAG_X_FORWARDED_FOR_APPENDED 0x00000800U
775 * Flag for csp->flags: Set if the server wants to keep
776 * the connection alive.
778 #define CSP_FLAG_SERVER_CONNECTION_KEEP_ALIVE 0x00001000U
781 * Flag for csp->flags: Set if the server specified the
784 #define CSP_FLAG_SERVER_CONTENT_LENGTH_SET 0x00002000U
787 * Flag for csp->flags: Set if we know the content length,
788 * either because the server set it, or we figured it out
791 #define CSP_FLAG_CONTENT_LENGTH_SET 0x00004000U
794 * Flag for csp->flags: Set if the client wants to keep
795 * the connection alive.
797 #define CSP_FLAG_CLIENT_CONNECTION_KEEP_ALIVE 0x00008000U
800 * Flag for csp->flags: Set if we think we got the whole
801 * client request and shouldn't read any additional data
802 * coming from the client until the current request has
805 #define CSP_FLAG_CLIENT_REQUEST_COMPLETELY_READ 0x00010000U
808 * Flag for csp->flags: Set if the server promised us to
809 * keep the connection open for a known number of seconds.
811 #define CSP_FLAG_SERVER_KEEP_ALIVE_TIMEOUT_SET 0x00020000U
814 * Flag for csp->flags: Set if we think we can't reuse
815 * the server socket. XXX: It's also set after sabotaging
816 * pipelining attempts which is somewhat inconsistent with
819 #define CSP_FLAG_SERVER_SOCKET_TAINTED 0x00040000U
822 * Flag for csp->flags: Set if the Proxy-Connection header
823 * is among the server headers.
825 #define CSP_FLAG_SERVER_PROXY_CONNECTION_HEADER_SET 0x00080000U
828 * Flag for csp->flags: Set if the client reused its connection.
830 #define CSP_FLAG_REUSED_CLIENT_CONNECTION 0x00100000U
833 * Flag for csp->flags: Set if the supports deflate compression.
835 #define CSP_FLAG_CLIENT_SUPPORTS_DEFLATE 0x00200000U
838 * Flag for csp->flags: Set if the content has been deflated by Privoxy
840 #define CSP_FLAG_BUFFERED_CONTENT_DEFLATED 0x00400000U
843 * Flag for csp->flags: Set if we already read (parts of)
844 * a pipelined request in which case the client obviously
845 * isn't done talking.
847 #define CSP_FLAG_PIPELINED_REQUEST_WAITING 0x00800000U
850 * Flag for csp->flags: Set if the client body is chunk-encoded
852 #define CSP_FLAG_CHUNKED_CLIENT_BODY 0x01000000U
856 * Flags for use in return codes of child processes
860 * Flag for process return code: Set if exiting porcess has been toggled
861 * during its lifetime.
863 #define RC_FLAG_TOGGLED 0x10
866 * Flag for process return code: Set if exiting porcess has blocked its
869 #define RC_FLAG_BLOCKED 0x20
872 * Maximum number of actions/filter files. This limit is arbitrary - it's just used
875 #define MAX_AF_FILES 30
878 * Maximum number of sockets to listen to. This limit is arbitrary - it's just used
881 #define MAX_LISTENING_SOCKETS 10
884 * The state of a Privoxy processing thread.
888 /** The proxy's configuration */
889 struct configuration_spec * config;
891 /** The actions to perform on the current request */
892 struct current_action_spec action[1];
894 /** socket to talk to client (web browser) */
897 /** Number of requests received on the client socket. */
898 unsigned int requests_received_total;
900 /** current connection to the server (may go through a proxy) */
901 struct reusable_connection server_connection;
903 /** Multi-purpose flag container, see CSP_FLAG_* above */
906 /** Client PC's IP address, as reported by the accept() function.
910 /** Client PC's TCP address, as reported by the accept() function.
912 struct sockaddr_storage tcp_addr;
914 /** Client PC's IP address, as reported by the accept() function.
916 unsigned long ip_addr_long;
917 #endif /* def HAVE_RFC2553 */
919 /** The URL that was requested */
920 struct http_request http[1];
923 * The final forwarding settings.
924 * XXX: Currently this is only used for forward-override,
925 * so we can free the space in sweep.
927 struct forward_spec * fwd;
929 /** An I/O buffer used for buffering data read from the server */
930 /* XXX: should be renamed to server_iob */
933 /** An I/O buffer used for buffering data read from the client */
934 struct iob client_iob[1];
936 /** List of all headers for this request */
937 struct list headers[1];
939 /** List of all tags that apply to this request */
942 /** MIME-Type key, see CT_* above */
943 unsigned int content_type;
945 /** Actions files associated with this client */
946 struct file_list *actions_list[MAX_AF_FILES];
948 /** pcrs job files. */
949 struct file_list *rlist[MAX_AF_FILES];
951 /** Length after content modification. */
952 unsigned long long content_length;
954 /* XXX: is this the right location? */
956 /** Expected length of content after which we
957 * should stop reading from the server socket.
959 unsigned long long expected_content_length;
961 /** Expected length of content after which we
962 * should stop reading from the client socket.
964 unsigned long long expected_client_content_length;
969 struct file_list *tlist;
971 #endif /* def FEATURE_TRUST */
974 * Failure reason to embedded in the CGI error page,
975 * or NULL. Currently only used for socks errors.
981 * List of client states so the main thread can keep
982 * track of them and garbage collect their resources.
986 struct client_states *next;
987 struct client_state csp;
991 * A function to add a header
993 typedef jb_err (*add_header_func_ptr)(struct client_state *);
996 * A function to process a header
998 typedef jb_err (*parser_func_ptr )(struct client_state *, char **);
1002 * List of available CGI functions.
1004 struct cgi_dispatcher
1006 /** The URL of the CGI, relative to the CGI root. */
1007 const char * const name;
1009 /** The handler function for the CGI */
1010 jb_err (* const handler)(struct client_state *csp, struct http_response *rsp, const struct map *parameters);
1012 /** The description of the CGI, to appear on the main menu, or NULL to hide it. */
1013 const char * const description;
1015 /** A flag that indicates whether unintentional calls to this CGI can cause damage */
1021 * A data file used by Privoxy. Kept in a linked list.
1026 * This is a pointer to the data structures associated with the file.
1027 * Read-only once the structure has been created.
1032 * The unloader function.
1033 * Normally NULL. When we are finished with file (i.e. when we have
1034 * loaded a new one), set to a pointer to an unloader function.
1035 * Unloader will be called by sweep() (called from main loop) when
1036 * all clients using this file are done. This prevents threading
1039 void (*unloader)(void *);
1042 * Used internally by sweep(). Do not access from elsewhere.
1047 * File last-modified time, so we can check if file has been changed.
1048 * Read-only once the structure has been created.
1050 time_t lastmodified;
1053 * The full filename.
1058 * Pointer to next entry in the linked list of all "file_list"s.
1059 * This linked list is so that sweep() can navigate it.
1060 * Since sweep() can remove items from the list, we must be careful
1061 * to only access this value from main thread (when we know sweep
1062 * won't be running).
1064 struct file_list *next;
1068 #ifdef FEATURE_TRUST
1071 * The format of a trust file when loaded into memory.
1075 struct pattern_spec url[1]; /**< The URL pattern */
1076 int reject; /**< FIXME: Please document this! */
1077 struct block_spec *next; /**< Next entry in linked list */
1081 * Arbitrary limit for the number of trusted referrers.
1083 #define MAX_TRUSTED_REFERRERS 512
1085 #endif /* def FEATURE_TRUST */
1088 * How to forward a connection to a parent proxy.
1092 /** URL pattern that this forward_spec is for. */
1093 struct pattern_spec url[1];
1095 /** Connection type. Must be SOCKS_NONE, SOCKS_4, SOCKS_4A or SOCKS_5. */
1096 enum forwarder_type type;
1098 /** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
1101 /** SOCKS server port. */
1104 /** Parent HTTP proxy hostname, or NULL for none. */
1107 /** Parent HTTP proxy port. */
1110 /** Next entry in the linked list. */
1111 struct forward_spec *next;
1115 /* Supported filter types */
1118 FT_CONTENT_FILTER = 0,
1119 FT_CLIENT_HEADER_FILTER = 1,
1120 FT_SERVER_HEADER_FILTER = 2,
1121 FT_CLIENT_HEADER_TAGGER = 3,
1122 FT_SERVER_HEADER_TAGGER = 4,
1123 FT_INVALID_FILTER = 42,
1125 #define MAX_FILTER_TYPES 5
1128 * This struct represents one filter (one block) from
1129 * the re_filterfile. If there is more than one filter
1130 * in the file, the file will be represented by a
1131 * chained list of re_filterfile specs.
1133 struct re_filterfile_spec
1135 char *name; /**< Name from FILTER: statement in re_filterfile. */
1136 char *description; /**< Description from FILTER: statement in re_filterfile. */
1137 struct list patterns[1]; /**< The patterns from the re_filterfile. */
1138 pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
1139 enum filter_type type; /**< Filter type (content, client-header, server-header). */
1140 int dynamic; /**< Set to one if the pattern might contain variables
1141 and has to be recompiled for every request. */
1142 struct re_filterfile_spec *next; /**< The pointer for chaining. */
1148 #define ACL_PERMIT 1 /**< Accept connection request */
1149 #define ACL_DENY 2 /**< Reject connection request */
1152 * An IP address pattern. Used to specify networks in the ACL.
1154 struct access_control_addr
1157 struct sockaddr_storage addr; /* <The TCP address in network order. */
1158 struct sockaddr_storage mask; /* <The TCP mask in network order. */
1160 unsigned long addr; /**< The IP address as an integer. */
1161 unsigned long mask; /**< The network mask as an integer. */
1162 unsigned long port; /**< The port number. */
1163 #endif /* HAVE_RFC2553 */
1167 * An access control list (ACL) entry.
1169 * This is a linked list.
1171 struct access_control_list
1173 struct access_control_addr src[1]; /**< Client IP address */
1174 struct access_control_addr dst[1]; /**< Website or parent proxy IP address */
1176 int wildcard_dst; /** < dst address is wildcard */
1179 short action; /**< ACL_PERMIT or ACL_DENY */
1180 struct access_control_list *next; /**< The next entry in the ACL. */
1183 #endif /* def FEATURE_ACL */
1186 /** Maximum number of loaders (actions, re_filter, ...) */
1190 /** configuration_spec::feature_flags: CGI actions editor. */
1191 #define RUNTIME_FEATURE_CGI_EDIT_ACTIONS 1U
1193 /** configuration_spec::feature_flags: Web-based toggle. */
1194 #define RUNTIME_FEATURE_CGI_TOGGLE 2U
1196 /** configuration_spec::feature_flags: HTTP-header-based toggle. */
1197 #define RUNTIME_FEATURE_HTTP_TOGGLE 4U
1199 /** configuration_spec::feature_flags: Split large forms to limit the number of GET arguments. */
1200 #define RUNTIME_FEATURE_SPLIT_LARGE_FORMS 8U
1202 /** configuration_spec::feature_flags: Check the host header for requests with host-less request lines. */
1203 #define RUNTIME_FEATURE_ACCEPT_INTERCEPTED_REQUESTS 16U
1205 /** configuration_spec::feature_flags: Don't allow to circumvent blocks with the force prefix. */
1206 #define RUNTIME_FEATURE_ENFORCE_BLOCKS 32U
1208 /** configuration_spec::feature_flags: Allow to block or redirect CGI requests. */
1209 #define RUNTIME_FEATURE_CGI_CRUNCHING 64U
1211 /** configuration_spec::feature_flags: Try to keep the connection to the server alive. */
1212 #define RUNTIME_FEATURE_CONNECTION_KEEP_ALIVE 128U
1214 /** configuration_spec::feature_flags: Share outgoing connections between different client connections. */
1215 #define RUNTIME_FEATURE_CONNECTION_SHARING 256U
1217 /** configuration_spec::feature_flags: Pages blocked with +handle-as-empty-doc get a return status of 200 OK. */
1218 #define RUNTIME_FEATURE_EMPTY_DOC_RETURNS_OK 512U
1220 /** configuration_spec::feature_flags: Buffered content is sent compressed if the client supports it. */
1221 #define RUNTIME_FEATURE_COMPRESSION 1024U
1223 /** configuration_spec::feature_flags: Pipelined requests are served instead of being discarded. */
1224 #define RUNTIME_FEATURE_TOLERATE_PIPELINING 2048U
1226 /** configuration_spec::feature_flags: Proxy authentication headers are forwarded instead of removed. */
1227 #define RUNTIME_FEATURE_FORWARD_PROXY_AUTHENTICATION_HEADERS 4096U
1230 * Data loaded from the configuration file.
1232 * (Anomaly: toggle is still handled through a global, not this structure)
1234 struct configuration_spec
1239 /** Nonzero to enable multithreading. */
1243 * Bitmask of features that can be enabled/disabled through the config
1244 * file. Currently defined bits:
1246 * - RUNTIME_FEATURE_CGI_EDIT_ACTIONS
1247 * - RUNTIME_FEATURE_CGI_TOGGLE
1248 * - RUNTIME_FEATURE_HTTP_TOGGLE
1249 * - RUNTIME_FEATURE_SPLIT_LARGE_FORMS
1250 * - RUNTIME_FEATURE_ACCEPT_INTERCEPTED_REQUESTS
1251 * - RUNTIME_FEATURE_ENFORCE_BLOCKS
1252 * - RUNTIME_FEATURE_CGI_CRUNCHING
1253 * - RUNTIME_FEATURE_CONNECTION_KEEP_ALIVE
1254 * - RUNTIME_FEATURE_CONNECTION_SHARING
1255 * - RUNTIME_FEATURE_EMPTY_DOC_RETURNS_OK
1257 unsigned feature_flags;
1259 /** The log file name. */
1260 const char *logfile;
1262 /** The config file directory. */
1263 const char *confdir;
1265 /** The directory for customized CGI templates. */
1266 const char *templdir;
1268 /** The log file directory. */
1271 /** The full paths to the actions files. */
1272 const char *actions_file[MAX_AF_FILES];
1274 /** The short names of the actions files. */
1275 const char *actions_file_short[MAX_AF_FILES];
1277 /** The administrator's email address */
1278 char *admin_address;
1280 /** A URL with info on this proxy */
1281 char *proxy_info_url;
1283 /** URL to the user manual (on our website or local copy) */
1286 /** The file names of the pcre filter files. */
1287 const char *re_filterfile[MAX_AF_FILES];
1289 /** The short names of the pcre filter files. */
1290 const char *re_filterfile_short[MAX_AF_FILES];
1292 /**< List of ordered client header names. */
1293 struct list ordered_client_headers[1];
1295 /** The hostname to show on CGI pages, or NULL to use the real one. */
1296 const char *hostname;
1298 /** IP addresses to bind to. Defaults to HADDR_DEFAULT == 127.0.0.1. */
1299 const char *haddr[MAX_LISTENING_SOCKETS];
1301 /** Ports to bind to. Defaults to HADDR_PORT == 8118. */
1302 int hport[MAX_LISTENING_SOCKETS];
1304 /** Size limit for IOB */
1305 size_t buffer_limit;
1307 #ifdef FEATURE_TRUST
1309 /** The file name of the trust file. */
1310 const char * trustfile;
1312 /** FIXME: DOCME: Document this. */
1313 struct list trust_info[1];
1315 /** FIXME: DOCME: Document this. */
1316 struct pattern_spec *trust_list[MAX_TRUSTED_REFERRERS];
1318 #endif /* def FEATURE_TRUST */
1322 /** The access control list (ACL). */
1323 struct access_control_list *acl;
1325 #endif /* def FEATURE_ACL */
1327 /** Information about parent proxies (forwarding). */
1328 struct forward_spec *forward;
1330 /** Number of retries in case a forwarded connection attempt fails */
1331 int forwarded_connect_retries;
1333 /** Maximum number of client connections. */
1334 int max_client_connections;
1336 /* Timeout when waiting on sockets for data to become available. */
1339 #ifdef FEATURE_CONNECTION_KEEP_ALIVE
1340 /* Maximum number of seconds after which an open connection will no longer be reused. */
1341 unsigned int keep_alive_timeout;
1343 /* Assumed server-side keep alive timeout if none is specified. */
1344 unsigned int default_server_timeout;
1347 #ifdef FEATURE_COMPRESSION
1348 int compression_level;
1351 /** All options from the config file, HTML-formatted. */
1354 /** The configuration file object. */
1355 struct file_list *config_file_list;
1357 /** List of loaders */
1358 int (*loaders[NLOADERS])(struct client_state *);
1360 /** Nonzero if we need to bind() to the new port. */
1364 /** Calculates the number of elements in an array, using sizeof. */
1365 #define SZ(X) (sizeof(X) / sizeof(*X))
1367 #ifdef FEATURE_FORCE_LOAD
1368 /** The force load URL prefix. */
1369 #define FORCE_PREFIX "/PRIVOXY-FORCE"
1370 #endif /* def FEATURE_FORCE_LOAD */
1372 #ifdef FEATURE_NO_GIFS
1373 /** The MIME type for images ("image/png" or "image/gif"). */
1374 #define BUILTIN_IMAGE_MIMETYPE "image/png"
1376 #define BUILTIN_IMAGE_MIMETYPE "image/gif"
1377 #endif /* def FEATURE_NO_GIFS */
1384 /** URL for the Privoxy home page. */
1385 #define HOME_PAGE_URL "http://www.privoxy.org/"
1387 /** URL for the Privoxy user manual. */
1388 #define USER_MANUAL_URL HOME_PAGE_URL VERSION "/user-manual/"
1390 /** Prefix for actions help links (append to USER_MANUAL_URL). */
1391 #define ACTIONS_HELP_PREFIX "actions-file.html#"
1393 /** Prefix for config option help links (append to USER_MANUAL_URL). */
1394 #define CONFIG_HELP_PREFIX "config.html#"
1397 * The "hosts" to intercept and display CGI pages.
1398 * First one is a hostname only, second one can specify host and path.
1401 * 1) Do not specify the http: prefix
1402 * 2) CGI_SITE_2_PATH must not end with /, one will be added automatically.
1403 * 3) CGI_SITE_2_PATH must start with /, unless it is the empty string.
1405 #define CGI_SITE_1_HOST "p.p"
1406 #define CGI_SITE_2_HOST "config.privoxy.org"
1407 #define CGI_SITE_2_PATH ""
1410 * The prefix for CGI pages. Written out in generated HTML.
1411 * INCLUDES the trailing slash.
1413 #define CGI_PREFIX "http://" CGI_SITE_2_HOST CGI_SITE_2_PATH "/"
1415 #endif /* ndef PROJECT_H_INCLUDED */