1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
3 #define PROJECT_H_VERSION "$Id: project.h,v 1.27 2001/08/05 16:06:20 jongfoster Exp $"
4 /*********************************************************************
6 * File : $Source: /cvsroot/ijbswa/current/project.h,v $
8 * Purpose : Defines data structures which are widely used in the
9 * project. Does not define any variables or functions
10 * (though it does declare some macros).
12 * Copyright : Written by and Copyright (C) 2001 the SourceForge
13 * IJBSWA team. http://ijbswa.sourceforge.net
15 * Based on the Internet Junkbuster originally written
16 * by and Copyright (C) 1997 Anonymous Coders and
17 * Junkbusters Corporation. http://www.junkbusters.com
19 * This program is free software; you can redistribute it
20 * and/or modify it under the terms of the GNU General
21 * Public License as published by the Free Software
22 * Foundation; either version 2 of the License, or (at
23 * your option) any later version.
25 * This program is distributed in the hope that it will
26 * be useful, but WITHOUT ANY WARRANTY; without even the
27 * implied warranty of MERCHANTABILITY or FITNESS FOR A
28 * PARTICULAR PURPOSE. See the GNU General Public
29 * License for more details.
31 * The GNU General Public License should be included with
32 * this file. If not, you can view it at
33 * http://www.gnu.org/copyleft/gpl.html
34 * or write to the Free Software Foundation, Inc., 59
35 * Temple Place - Suite 330, Boston, MA 02111-1307, USA.
39 * Revision 1.27 2001/08/05 16:06:20 jongfoster
40 * Modifiying "struct map" so that there are now separate header and
41 * "map_entry" structures. This means that functions which modify a
42 * map no longer need to return a pointer to the modified map.
43 * Also, it no longer reverses the order of the entries (which may be
44 * important with some advanced template substitutions).
46 * Revision 1.26 2001/07/30 22:08:36 jongfoster
47 * Tidying up #defines:
48 * - All feature #defines are now of the form FEATURE_xxx
49 * - Permanently turned off WIN_GUI_EDIT
50 * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
52 * Revision 1.25 2001/07/29 18:43:08 jongfoster
53 * Changing #ifdef _FILENAME_H to FILENAME_H_INCLUDED, to conform to
56 * Revision 1.24 2001/07/25 17:20:27 oes
57 * Introduced http->user_agent
59 * Revision 1.23 2001/07/18 12:32:23 oes
60 * - Added ACTION_STRING_DEANIMATE
61 * - moved #define freez from jcc.h to project.h
63 * Revision 1.22 2001/07/15 17:51:41 jongfoster
64 * Renaming #define STATIC to STATIC_PCRE
66 * Revision 1.21 2001/07/13 14:03:19 oes
67 * - Reorganized regex header inclusion and #defines to
68 * comply to the scheme in configure.in
69 * - Added csp->content_type and its CT_* keys
70 * - Added ACTION_DEANIMATE
71 * - Removed all #ifdef PCRS
73 * Revision 1.20 2001/06/29 21:45:41 oes
74 * Indentation, CRLF->LF, Tab-> Space
76 * Revision 1.19 2001/06/29 13:33:36 oes
78 * - Introduced http_request.host_ip_addr_str
79 * - Introduced http_response.head_length
80 * - Introduced config.my_ip_addr_str, config.my_hostname,
81 * config.admin_address and config.proxy_info_url
82 * - Removed config.proxy_args_header and config.proxy_args_trailer,
83 * renamed config.proxy_args_invocation to config.proxy_args
84 * - Removed HTML snipplets and GIFs
85 * - Removed logentry from cancelled commit
87 * Revision 1.18 2001/06/09 10:57:39 jongfoster
88 * Adding definition of BUFFER_SIZE.
89 * Changing struct cgi_dispatcher to use "const" strings.
91 * Revision 1.17 2001/06/07 23:15:09 jongfoster
92 * Merging ACL and forward files into config file.
93 * Moving struct gateway members into struct forward_spec
94 * Removing config->proxy_args_gateways
95 * Cosmetic: Adding a few comments
97 * Revision 1.16 2001/06/04 18:31:58 swa
98 * files are now prefixed with either `confdir' or `logdir'.
99 * `make redhat-dist' replaces both entries confdir and logdir
102 * Revision 1.15 2001/06/04 11:28:53 swa
103 * redirect did not work due to missing /
105 * Revision 1.14 2001/06/03 11:03:48 oes
107 * added struct http_response,
108 * changed struct interceptors to struct cgi_dispatcher,
109 * moved HTML stuff to cgi.h
111 * Revision 1.13 2001/06/01 20:05:36 jongfoster
112 * Support for +image-blocker{}: added ACTION_IMAGE_BLOCKER
113 * constant, and removed csp->tinygif.
115 * Revision 1.12 2001/06/01 18:49:17 jongfoster
116 * Replaced "list_share" with "list" - the tiny memory gain was not
117 * worth the extra complexity.
119 * Revision 1.11 2001/06/01 10:32:47 oes
120 * Added constants for anchoring selection bitmap
122 * Revision 1.10 2001/05/31 21:33:53 jongfoster
123 * Changes for new actions file, replacing permissionsfile
124 * and parts of the config file. Also added support for
127 * Revision 1.9 2001/05/31 17:32:31 oes
129 * - Enhanced domain part globbing with infix and prefix asterisk
130 * matching and optional unanchored operation
132 * Revision 1.8 2001/05/29 20:09:15 joergs
133 * HTTP_REDIRECT_TEMPLATE fixed.
135 * Revision 1.7 2001/05/29 09:50:24 jongfoster
136 * Unified blocklist/imagelist/actionslist.
137 * File format is still under discussion, but the internal changes
140 * Also modified interceptor behaviour:
141 * - We now intercept all URLs beginning with one of the following
142 * prefixes (and *only* these prefixes):
144 * * http://ijbswa.sf.net/config/
145 * * http://ijbswa.sourceforge.net/config/
146 * - New interceptors "home page" - go to http://i.j.b/ to see it.
147 * - Internal changes so that intercepted and fast redirect pages
148 * are not replaced with an image.
149 * - Interceptors now have the option to send a binary page direct
150 * to the client. (i.e. ijb-send-banner uses this)
151 * - Implemented show-url-info interceptor. (Which is why I needed
152 * the above interceptors changes - a typical URL is
153 * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif".
154 * The previous mechanism would not have intercepted that, and
155 * if it had been intercepted then it then it would have replaced
158 * Revision 1.6 2001/05/27 22:17:04 oes
160 * - re_process_buffer no longer writes the modified buffer
161 * to the client, which was very ugly. It now returns the
162 * buffer, which it is then written by chat.
164 * - content_length now adjusts the Content-Length: header
165 * for modified documents rather than crunch()ing it.
166 * (Length info in csp->content_length, which is 0 for
167 * unmodified documents)
169 * - For this to work, sed() is called twice when filtering.
171 * Revision 1.5 2001/05/26 00:28:36 jongfoster
172 * Automatic reloading of config file.
173 * Removed obsolete SIGHUP support (Unix) and Reload menu option (Win32).
174 * Most of the global variables have been moved to a new
175 * struct configuration_spec, accessed through csp->config->globalname
176 * Most of the globals remaining are used by the Win32 GUI.
178 * Revision 1.4 2001/05/22 18:46:04 oes
180 * - Enabled filtering banners by size rather than URL
181 * by adding patterns that replace all standard banner
182 * sizes with the "Junkbuster" gif to the re_filterfile
184 * - Enabled filtering WebBugs by providing a pattern
185 * which kills all 1x1 images
187 * - Added support for PCRE_UNGREEDY behaviour to pcrs,
188 * which is selected by the (nonstandard and therefore
189 * capital) letter 'U' in the option string.
190 * It causes the quantifiers to be ungreedy by default.
191 * Appending a ? turns back to greedy (!).
193 * - Added a new interceptor ijb-send-banner, which
194 * sends back the "Junkbuster" gif. Without imagelist or
195 * MSIE detection support, or if tinygif = 1, or the
196 * URL isn't recognized as an imageurl, a lame HTML
197 * explanation is sent instead.
199 * - Added new feature, which permits blocking remote
200 * script redirects and firing back a local redirect
202 * The feature is conditionally compiled, i.e. it
203 * can be disabled with --disable-fast-redirects,
204 * plus it must be activated by a "fast-redirects"
205 * line in the config file, has its own log level
206 * and of course wants to be displayed by show-proxy-args
207 * Note: Boy, all the #ifdefs in 1001 locations and
208 * all the fumbling with configure.in and acconfig.h
209 * were *way* more work than the feature itself :-(
211 * - Because a generic redirect template was needed for
212 * this, tinygif = 3 now uses the same.
214 * - Moved GIFs, and other static HTTP response templates
219 * - Removed some >400 CRs again (Jon, you really worked
222 * Revision 1.3 2001/05/20 01:21:20 jongfoster
223 * Version 2.9.4 checkin.
224 * - Merged popupfile and cookiefile, and added control over PCRS
225 * filtering, in new "actionsfile".
226 * - Implemented LOG_LEVEL_FATAL, so that if there is a configuration
227 * file error you now get a message box (in the Win32 GUI) rather
228 * than the program exiting with no explanation.
229 * - Made killpopup use the PCRS MIME-type checking and HTTP-header
231 * - Removed tabs from "config"
232 * - Moved duplicated url parsing code in "loaders.c" to a new funcition.
233 * - Bumped up version number.
235 * Revision 1.2 2001/05/17 23:01:01 oes
236 * - Cleaned CRLF's from the sources and related files
238 * Revision 1.1.1.1 2001/05/15 13:59:03 oes
239 * Initial import of version 2.9.3 source tree
242 *********************************************************************/
245 /* Declare struct FILE for vars and funcs. */
248 /* Need time_t for file_list */
252 * Include appropriate regular expression libraries.
253 * Note that pcrs and pcre (native) are needed for cgi
254 * and are included anyway.
257 #if defined(REGEX_PCRE) || defined (REGEX_GNU)
259 #endif /* defined(REGEX_PCRE) || defined (REGEX_GNU) */
273 #if defined(REGEX_PCRE)
275 # include "pcreposix.h"
277 # include <pcreposix.h>
279 #endif /* defined(REGEX_PCRE) */
281 #if defined(REGEX_GNU)
282 # include "gnu_regex.h"
287 #endif /* def AMIGA */
293 #define freez(X) if(X) free(X); X = NULL
295 #define BUFFER_SIZE 5000
299 /* Default IP and port to listen on */
300 #define HADDR_DEFAULT "127.0.0.1"
301 #define HADDR_PORT 8000
304 /* Need this for struct client_state */
305 struct configuration_spec;
307 /* Generic linked list of strings */
308 struct list /* FIXME: Why not separate entries and header? */
310 char * str; /* valid in an entry */
311 struct list *last; /* valid in header */
319 struct map_entry *next;
324 struct map_entry *first;
325 struct map_entry *last;
333 char *host_ip_addr_str; /* NULL before connect_to() */
337 char *hostport; /* "host[:port]" */
339 char *user_agent; /* Client's User-Agent: header value */
342 /* Response generated by CGI, blocker, or error handler */
345 char *status; /* HTTP status (string)*/
346 struct list headers[1]; /* List of header lines */
347 char *head; /* Formatted http response head */
348 int head_length; /* Length of http response head */
349 char *body; /* HTTP document body */
350 int content_length; /* Length of body, REQUIRED if binary body*/
356 char *spec; /* The string which was parsed to produce this */
357 /* url_spec. Used for debugging or display only. */
359 /* Hostname matching: */
360 char *domain; /* Fully qalified domain name (FQDN) pattern. */
361 /* May contain "*". */
362 char *dbuf; /* Buffer with '\0'-delimited fqdn */
363 char **dvec; /* Domain ptr vector into dbuf */
364 int dcnt; /* How many domains in fqdn? */
365 int unanchored; /* Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT */
368 int port; /* The port number, or 0 to match all ports. */
371 char *path; /* The path prefix (if not using regex), or source */
373 int pathlen; /* ==strlen(path). Needed for prefix matching. */
375 regex_t *preg; /* Regex for matching path part */
379 /* Constants for host part matching in URLs */
380 #define ANCHOR_LEFT 1
381 #define ANCHOR_RIGHT 2
393 #define IOB_PEEK(CSP) ((CSP->iob->cur > CSP->iob->eod) ? (CSP->iob->eod - CSP->iob->cur) : 0)
394 #define IOB_RESET(CSP) if(CSP->iob->buf) free(CSP->iob->buf); memset(CSP->iob, '\0', sizeof(CSP->iob));
396 /* Keys for csp->content_type */
397 #define CT_TEXT 0x01U
400 #define ACTION_MASK_ALL (~0U)
402 #define ACTION_MOST_COMPATIBLE 0x0000U
404 #define ACTION_BLOCK 0x0001U
405 #define ACTION_DEANIMATE 0x2000U
406 #define ACTION_FAST_REDIRECTS 0x0002U
407 #define ACTION_FILTER 0x0004U
408 #define ACTION_HIDE_FORWARDED 0x0008U
409 #define ACTION_HIDE_FROM 0x0010U
410 #define ACTION_HIDE_REFERER 0x0020U /* sic - follow HTTP, not English */
411 #define ACTION_HIDE_USER_AGENT 0x0040U
412 #define ACTION_IMAGE 0x0080U
413 #define ACTION_IMAGE_BLOCKER 0x0100U
414 #define ACTION_NO_COOKIE_READ 0x0200U
415 #define ACTION_NO_COOKIE_SET 0x0400U
416 #define ACTION_NO_POPUPS 0x0800U
417 #define ACTION_VANILLA_WAFER 0x1000U
419 #define ACTION_STRING_DEANIMATE 0
420 #define ACTION_STRING_FROM 1
421 #define ACTION_STRING_IMAGE_BLOCKER 2
422 #define ACTION_STRING_REFERER 3
423 #define ACTION_STRING_USER_AGENT 4
424 #define ACTION_STRING_COUNT 5
427 #define ACTION_MULTI_ADD_HEADER 0
428 #define ACTION_MULTI_WAFER 1
429 #define ACTION_MULTI_COUNT 2
432 * This structure contains a list of actions to apply to a URL.
433 * It only contains positive instructions - no "-" options.
434 * It is not used to store the actions list itself, only for
435 * url_actions() to return the current values.
437 struct current_action_spec
439 unsigned flags; /* a bit set to "1" = add action */
441 /* For those actions that require parameters: */
443 /* each entry is valid if & only if corresponding entry in "add" set. */
444 char * string[ACTION_STRING_COUNT];
447 struct list multi[ACTION_MULTI_COUNT][1];
452 * This structure contains a set of changes to actions.
453 * It can contain both positive and negative instructions.
454 * It is used to store an entry in the actions list.
458 unsigned mask; /* a bit set to "0" = remove action */
459 unsigned add; /* a bit set to "1" = add action */
461 /* For those actions that require parameters: */
463 /* each entry is valid if & only if corresponding entry in "add" set. */
464 char * string[ACTION_STRING_COUNT];
466 /* Strings to remove. */
467 struct list multi_remove[ACTION_MULTI_COUNT][1];
469 /* If nonzero, remove *all* strings. */
470 int multi_remove_all[ACTION_MULTI_COUNT];
473 struct list multi_add[ACTION_MULTI_COUNT][1];
477 * This structure is used to store the actions list.
479 * It contains a URL pattern, and the chages to the actions.
480 * It is a linked list.
484 struct url_spec url[1];
486 struct action_spec action[1];
488 struct url_actions * next;
492 /* Constants defining bitmask for csp->accept_types */
494 #ifdef FEATURE_IMAGE_DETECT_MSIE
496 /* MSIE detected by user-agent string */
497 #define ACCEPT_TYPE_IS_MSIE 0x0001
500 * *If* this is MSIE, it wants an image. (Or this is a shift-reload, or
501 * it's got an image from this URL before... yuck!)
502 * Only meaningful if ACCEPT_TYPE_IS_MSIE set
504 #define ACCEPT_TYPE_MSIE_IMAGE 0x0002
507 * *If* this is MSIE, it wants a HTML document.
508 * Only meaningful if ACCEPT_TYPE_IS_MSIE set
510 #define ACCEPT_TYPE_MSIE_HTML 0x0004
512 #endif /* def FEATURE_IMAGE_DETECT_MSIE */
517 /* The proxy's configuration */
518 struct configuration_spec * config;
520 /* The actions to perform on the current request */
521 struct current_action_spec action[1];
523 /* socket to talk to client (web browser) */
526 /* socket to talk to server (web server or proxy) */
530 #ifdef FEATURE_STATISTICS
531 /* 1 if this URL was rejected, 0 otherwise. Allows actual stats inc to
532 * occur in main thread only for thread-safety.
535 #endif /* def FEATURE_STATISTICS */
537 #ifdef FEATURE_FORCE_LOAD
539 #endif /* def FEATURE_FORCE_LOAD */
541 #ifdef FEATURE_TOGGLE
543 #endif /* def FEATURE_TOGGLE */
546 * Client PC's IP address, as reported by the accept()_ function.
547 * Both as string and number
553 /* Our IP address and hostname, i.e. the IP address that
554 the client used to reach us, and the associated hostname,
557 char *my_ip_addr_str;
561 /* The referer in this request, if one was specified. */
563 #endif /* def FEATURE_TRUST */
565 #if defined(FEATURE_IMAGE_DETECT_MSIE)
566 /* Types the client will accept.
567 * Bitmask - see ACCEPT_TYPE_XXX constants.
570 #endif /* defined(FEATURE_IMAGE_DETECT_MSIE) */
572 /* The URL that was requested */
573 struct http_request http[1];
575 /* An I/O buffer used for buffering data read from the client */
578 /* List of all headers for this request */
579 struct list headers[1];
581 /* List of all cookies for this request */
582 struct list cookie_list[1];
584 /* MIME-Type bitmap, see CT_* above */
585 unsigned char content_type;
587 /* The "X-Forwarded-For:" header sent by the client */
591 * Nonzero if this client is processing data.
592 * Set to zero when the thread associated with this structure dies.
596 /* files associated with this client */
597 struct file_list *actions_list;
599 struct file_list *rlist; /* pcrs job file */
600 size_t content_length; /* Length after content modification */
603 struct file_list *tlist; /* trustfile */
604 #endif /* def FEATURE_TRUST */
606 struct client_state *next;
614 char *(*parser)(const struct parsers *, const char *, struct client_state *);
617 struct cgi_dispatcher
621 int (*handler)(struct client_state *csp, struct http_response *rsp, struct map *parameters);
622 const char *description;
628 * this is a pointer to the data structures associated with the file.
629 * Read-only once the structure has been created.
633 /* Normally NULL. When we are finished with file (i.e. when we have
634 * loaded a new one), set to a pointer to an unloader function.
635 * Unloader will be called by sweep() (called from main loop) when
636 * all clients using this file are done. This prevents threading
639 void (*unloader)(void *);
641 /* Used internally by sweep(). Do not access from elsewhere. */
644 /* Following variables allow us to check if file has been changed.
645 * Read-only once the structure has been created.
650 /* Pointer to next entry in the linked list of all "file_list"s.
651 * This linked list is so that sweep() can navigate it.
652 * Since sweep() can remove items from the list, we must be careful
653 * to only access this value from main thread (when we know sweep
656 struct file_list *next;
663 struct url_spec url[1];
665 struct block_spec *next;
667 #endif /* def FEATURE_TRUST */
670 #define SOCKS_NONE 0 /* Don't use a SOCKS server */
671 #define SOCKS_4 40 /* original SOCKS 4 protocol */
672 #define SOCKS_4A 41 /* as modified for hosts w/o external DNS */
676 struct url_spec url[1];
678 /* Connection type - must be a SOCKS_xxx constant */
685 /* Parent HTTP proxy */
689 /* For the linked list */
690 struct forward_spec *next;
693 struct re_filterfile_spec
697 struct list patterns[1];
702 #define ACL_PERMIT 1 /* accept connection request */
703 #define ACL_DENY 2 /* reject connection request */
705 struct access_control_addr
712 struct access_control_list
714 struct access_control_addr src[1];
715 struct access_control_addr dst[1];
718 struct access_control_list *next;
720 #endif /* def FEATURE_ACL */
723 /* Maximum number of loaders (actions, re_filter, ...) */
727 * Data loaded from the configuration file.
729 * (Anomaly: toggle is still handled through a global, not this structure)
731 struct configuration_spec
740 const char *actions_file;
742 /* The administrator's email address */
745 /* A URL with info on this proxy */
746 char *proxy_info_url;
748 const char *re_filterfile;
750 #ifdef FEATURE_COOKIE_JAR
751 const char * jarfile;
753 #endif /* def FEATURE_COOKIE_JAR */
756 * Port and IP to bind to.
757 * Defaults to HADDR_DEFAULT:HADDR_PORT == 127.0.0.1:8000
763 const char * trustfile;
765 struct list trust_info[1];
766 struct url_spec *trust_list[64];
767 #endif /* def FEATURE_TRUST */
770 struct access_control_list *acl;
771 #endif /* def FEATURE_ACL */
773 struct forward_spec *forward;
775 /* All options from the config file, HTML-formatted */
778 /* the configuration file object. */
779 struct file_list *config_file_list;
781 /* List of loaders */
782 int (*loaders[NLOADERS])(struct client_state *);
784 /* bool, nonzero if we need to bind() to the new port */
789 #define SZ(X) (sizeof(X) / sizeof(*X))
791 #ifdef FEATURE_FORCE_LOAD
792 #define FORCE_PREFIX "/IJB-FORCE-LOAD"
793 #endif /* def FEATURE_FORCE_LOAD */
796 #define HOME_PAGE_URL "http://ijbswa.sourceforge.net"
797 #define REDIRECT_URL HOME_PAGE_URL "/redirect.php?v=" VERSION "&to="
798 #define CGI_PREFIX_HOST "i.j.b"
801 static const char CSUCCEED[] =
802 "HTTP/1.0 200 Connection established\n"
803 "Proxy-Agent: IJ/" VERSION "\n\n";
805 static const char CHEADER[] =
806 "HTTP/1.0 400 Invalid header received from browser\n\n";
812 #endif /* ndef PROJECT_H_INCLUDED */