1 #ifndef PROJECT_H_INCLUDED
2 #define PROJECT_H_INCLUDED
3 #define PROJECT_H_VERSION "$Id: project.h,v 1.28 2001/09/13 23:05:50 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 1.28 2001/09/13 23:05:50 jongfoster
40 * Changing the string paramater to the header parsers a "const".
42 * Revision 1.27 2001/08/05 16:06:20 jongfoster
43 * Modifiying "struct map" so that there are now separate header and
44 * "map_entry" structures. This means that functions which modify a
45 * map no longer need to return a pointer to the modified map.
46 * Also, it no longer reverses the order of the entries (which may be
47 * important with some advanced template substitutions).
49 * Revision 1.26 2001/07/30 22:08:36 jongfoster
50 * Tidying up #defines:
51 * - All feature #defines are now of the form FEATURE_xxx
52 * - Permanently turned off WIN_GUI_EDIT
53 * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
55 * Revision 1.25 2001/07/29 18:43:08 jongfoster
56 * Changing #ifdef _FILENAME_H to FILENAME_H_INCLUDED, to conform to
59 * Revision 1.24 2001/07/25 17:20:27 oes
60 * Introduced http->user_agent
62 * Revision 1.23 2001/07/18 12:32:23 oes
63 * - Added ACTION_STRING_DEANIMATE
64 * - moved #define freez from jcc.h to project.h
66 * Revision 1.22 2001/07/15 17:51:41 jongfoster
67 * Renaming #define STATIC to STATIC_PCRE
69 * Revision 1.21 2001/07/13 14:03:19 oes
70 * - Reorganized regex header inclusion and #defines to
71 * comply to the scheme in configure.in
72 * - Added csp->content_type and its CT_* keys
73 * - Added ACTION_DEANIMATE
74 * - Removed all #ifdef PCRS
76 * Revision 1.20 2001/06/29 21:45:41 oes
77 * Indentation, CRLF->LF, Tab-> Space
79 * Revision 1.19 2001/06/29 13:33:36 oes
81 * - Introduced http_request.host_ip_addr_str
82 * - Introduced http_response.head_length
83 * - Introduced config.my_ip_addr_str, config.my_hostname,
84 * config.admin_address and config.proxy_info_url
85 * - Removed config.proxy_args_header and config.proxy_args_trailer,
86 * renamed config.proxy_args_invocation to config.proxy_args
87 * - Removed HTML snipplets and GIFs
88 * - Removed logentry from cancelled commit
90 * Revision 1.18 2001/06/09 10:57:39 jongfoster
91 * Adding definition of BUFFER_SIZE.
92 * Changing struct cgi_dispatcher to use "const" strings.
94 * Revision 1.17 2001/06/07 23:15:09 jongfoster
95 * Merging ACL and forward files into config file.
96 * Moving struct gateway members into struct forward_spec
97 * Removing config->proxy_args_gateways
98 * Cosmetic: Adding a few comments
100 * Revision 1.16 2001/06/04 18:31:58 swa
101 * files are now prefixed with either `confdir' or `logdir'.
102 * `make redhat-dist' replaces both entries confdir and logdir
105 * Revision 1.15 2001/06/04 11:28:53 swa
106 * redirect did not work due to missing /
108 * Revision 1.14 2001/06/03 11:03:48 oes
110 * added struct http_response,
111 * changed struct interceptors to struct cgi_dispatcher,
112 * moved HTML stuff to cgi.h
114 * Revision 1.13 2001/06/01 20:05:36 jongfoster
115 * Support for +image-blocker{}: added ACTION_IMAGE_BLOCKER
116 * constant, and removed csp->tinygif.
118 * Revision 1.12 2001/06/01 18:49:17 jongfoster
119 * Replaced "list_share" with "list" - the tiny memory gain was not
120 * worth the extra complexity.
122 * Revision 1.11 2001/06/01 10:32:47 oes
123 * Added constants for anchoring selection bitmap
125 * Revision 1.10 2001/05/31 21:33:53 jongfoster
126 * Changes for new actions file, replacing permissionsfile
127 * and parts of the config file. Also added support for
130 * Revision 1.9 2001/05/31 17:32:31 oes
132 * - Enhanced domain part globbing with infix and prefix asterisk
133 * matching and optional unanchored operation
135 * Revision 1.8 2001/05/29 20:09:15 joergs
136 * HTTP_REDIRECT_TEMPLATE fixed.
138 * Revision 1.7 2001/05/29 09:50:24 jongfoster
139 * Unified blocklist/imagelist/actionslist.
140 * File format is still under discussion, but the internal changes
143 * Also modified interceptor behaviour:
144 * - We now intercept all URLs beginning with one of the following
145 * prefixes (and *only* these prefixes):
147 * * http://ijbswa.sf.net/config/
148 * * http://ijbswa.sourceforge.net/config/
149 * - New interceptors "home page" - go to http://i.j.b/ to see it.
150 * - Internal changes so that intercepted and fast redirect pages
151 * are not replaced with an image.
152 * - Interceptors now have the option to send a binary page direct
153 * to the client. (i.e. ijb-send-banner uses this)
154 * - Implemented show-url-info interceptor. (Which is why I needed
155 * the above interceptors changes - a typical URL is
156 * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif".
157 * The previous mechanism would not have intercepted that, and
158 * if it had been intercepted then it then it would have replaced
161 * Revision 1.6 2001/05/27 22:17:04 oes
163 * - re_process_buffer no longer writes the modified buffer
164 * to the client, which was very ugly. It now returns the
165 * buffer, which it is then written by chat.
167 * - content_length now adjusts the Content-Length: header
168 * for modified documents rather than crunch()ing it.
169 * (Length info in csp->content_length, which is 0 for
170 * unmodified documents)
172 * - For this to work, sed() is called twice when filtering.
174 * Revision 1.5 2001/05/26 00:28:36 jongfoster
175 * Automatic reloading of config file.
176 * Removed obsolete SIGHUP support (Unix) and Reload menu option (Win32).
177 * Most of the global variables have been moved to a new
178 * struct configuration_spec, accessed through csp->config->globalname
179 * Most of the globals remaining are used by the Win32 GUI.
181 * Revision 1.4 2001/05/22 18:46:04 oes
183 * - Enabled filtering banners by size rather than URL
184 * by adding patterns that replace all standard banner
185 * sizes with the "Junkbuster" gif to the re_filterfile
187 * - Enabled filtering WebBugs by providing a pattern
188 * which kills all 1x1 images
190 * - Added support for PCRE_UNGREEDY behaviour to pcrs,
191 * which is selected by the (nonstandard and therefore
192 * capital) letter 'U' in the option string.
193 * It causes the quantifiers to be ungreedy by default.
194 * Appending a ? turns back to greedy (!).
196 * - Added a new interceptor ijb-send-banner, which
197 * sends back the "Junkbuster" gif. Without imagelist or
198 * MSIE detection support, or if tinygif = 1, or the
199 * URL isn't recognized as an imageurl, a lame HTML
200 * explanation is sent instead.
202 * - Added new feature, which permits blocking remote
203 * script redirects and firing back a local redirect
205 * The feature is conditionally compiled, i.e. it
206 * can be disabled with --disable-fast-redirects,
207 * plus it must be activated by a "fast-redirects"
208 * line in the config file, has its own log level
209 * and of course wants to be displayed by show-proxy-args
210 * Note: Boy, all the #ifdefs in 1001 locations and
211 * all the fumbling with configure.in and acconfig.h
212 * were *way* more work than the feature itself :-(
214 * - Because a generic redirect template was needed for
215 * this, tinygif = 3 now uses the same.
217 * - Moved GIFs, and other static HTTP response templates
222 * - Removed some >400 CRs again (Jon, you really worked
225 * Revision 1.3 2001/05/20 01:21:20 jongfoster
226 * Version 2.9.4 checkin.
227 * - Merged popupfile and cookiefile, and added control over PCRS
228 * filtering, in new "actionsfile".
229 * - Implemented LOG_LEVEL_FATAL, so that if there is a configuration
230 * file error you now get a message box (in the Win32 GUI) rather
231 * than the program exiting with no explanation.
232 * - Made killpopup use the PCRS MIME-type checking and HTTP-header
234 * - Removed tabs from "config"
235 * - Moved duplicated url parsing code in "loaders.c" to a new funcition.
236 * - Bumped up version number.
238 * Revision 1.2 2001/05/17 23:01:01 oes
239 * - Cleaned CRLF's from the sources and related files
241 * Revision 1.1.1.1 2001/05/15 13:59:03 oes
242 * Initial import of version 2.9.3 source tree
245 *********************************************************************/
248 /* Declare struct FILE for vars and funcs. */
251 /* Need time_t for file_list */
255 * Include appropriate regular expression libraries.
256 * Note that pcrs and pcre (native) are needed for cgi
257 * and are included anyway.
260 #if defined(REGEX_PCRE) || defined (REGEX_GNU)
262 #endif /* defined(REGEX_PCRE) || defined (REGEX_GNU) */
276 #if defined(REGEX_PCRE)
278 # include "pcreposix.h"
280 # include <pcreposix.h>
282 #endif /* defined(REGEX_PCRE) */
284 #if defined(REGEX_GNU)
285 # include "gnu_regex.h"
290 #endif /* def AMIGA */
296 #define freez(X) if(X) free(X); X = NULL
298 #define BUFFER_SIZE 5000
302 /* Default IP and port to listen on */
303 #define HADDR_DEFAULT "127.0.0.1"
304 #define HADDR_PORT 8000
307 /* Need this for struct client_state */
308 struct configuration_spec;
310 /* Generic linked list of strings */
311 struct list /* FIXME: Why not separate entries and header? */
313 char * str; /* valid in an entry */
314 struct list *last; /* valid in header */
322 struct map_entry *next;
327 struct map_entry *first;
328 struct map_entry *last;
336 char *host_ip_addr_str; /* NULL before connect_to() */
340 char *hostport; /* "host[:port]" */
342 char *user_agent; /* Client's User-Agent: header value */
345 /* Response generated by CGI, blocker, or error handler */
348 char *status; /* HTTP status (string)*/
349 struct list headers[1]; /* List of header lines */
350 char *head; /* Formatted http response head */
351 int head_length; /* Length of http response head */
352 char *body; /* HTTP document body */
353 int content_length; /* Length of body, REQUIRED if binary body*/
359 char *spec; /* The string which was parsed to produce this */
360 /* url_spec. Used for debugging or display only. */
362 /* Hostname matching: */
363 char *domain; /* Fully qalified domain name (FQDN) pattern. */
364 /* May contain "*". */
365 char *dbuf; /* Buffer with '\0'-delimited fqdn */
366 char **dvec; /* Domain ptr vector into dbuf */
367 int dcnt; /* How many domains in fqdn? */
368 int unanchored; /* Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT */
371 int port; /* The port number, or 0 to match all ports. */
374 char *path; /* The path prefix (if not using regex), or source */
376 int pathlen; /* ==strlen(path). Needed for prefix matching. */
378 regex_t *preg; /* Regex for matching path part */
382 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, NULL, 0, 0, 0, NULL, 0, NULL }
383 #else /* ifndef REGEX */
384 #define URL_SPEC_INITIALIZER { NULL, NULL, NULL, NULL, 0, 0, 0, NULL, 0 }
385 #endif /* ndef REGEX */
387 /* Constants for host part matching in URLs */
388 #define ANCHOR_LEFT 1
389 #define ANCHOR_RIGHT 2
401 #define IOB_PEEK(CSP) ((CSP->iob->cur > CSP->iob->eod) ? (CSP->iob->eod - CSP->iob->cur) : 0)
402 #define IOB_RESET(CSP) if(CSP->iob->buf) free(CSP->iob->buf); memset(CSP->iob, '\0', sizeof(CSP->iob));
404 /* Keys for csp->content_type */
405 #define CT_TEXT 0x01U
408 #define ACTION_MASK_ALL (~0U)
410 #define ACTION_MOST_COMPATIBLE 0x0000U
412 #define ACTION_BLOCK 0x0001U
413 #define ACTION_DEANIMATE 0x2000U
414 #define ACTION_FAST_REDIRECTS 0x0002U
415 #define ACTION_FILTER 0x0004U
416 #define ACTION_HIDE_FORWARDED 0x0008U
417 #define ACTION_HIDE_FROM 0x0010U
418 #define ACTION_HIDE_REFERER 0x0020U /* sic - follow HTTP, not English */
419 #define ACTION_HIDE_USER_AGENT 0x0040U
420 #define ACTION_IMAGE 0x0080U
421 #define ACTION_IMAGE_BLOCKER 0x0100U
422 #define ACTION_NO_COOKIE_READ 0x0200U
423 #define ACTION_NO_COOKIE_SET 0x0400U
424 #define ACTION_NO_POPUPS 0x0800U
425 #define ACTION_VANILLA_WAFER 0x1000U
427 #define ACTION_STRING_DEANIMATE 0
428 #define ACTION_STRING_FROM 1
429 #define ACTION_STRING_IMAGE_BLOCKER 2
430 #define ACTION_STRING_REFERER 3
431 #define ACTION_STRING_USER_AGENT 4
432 #define ACTION_STRING_COUNT 5
435 #define ACTION_MULTI_ADD_HEADER 0
436 #define ACTION_MULTI_WAFER 1
437 #define ACTION_MULTI_COUNT 2
440 * This structure contains a list of actions to apply to a URL.
441 * It only contains positive instructions - no "-" options.
442 * It is not used to store the actions list itself, only for
443 * url_actions() to return the current values.
445 struct current_action_spec
447 unsigned flags; /* a bit set to "1" = add action */
449 /* For those actions that require parameters: */
451 /* each entry is valid if & only if corresponding entry in "add" set. */
452 char * string[ACTION_STRING_COUNT];
455 struct list multi[ACTION_MULTI_COUNT][1];
460 * This structure contains a set of changes to actions.
461 * It can contain both positive and negative instructions.
462 * It is used to store an entry in the actions list.
466 unsigned mask; /* a bit set to "0" = remove action */
467 unsigned add; /* a bit set to "1" = add action */
469 /* For those actions that require parameters: */
471 /* each entry is valid if & only if corresponding entry in "add" set. */
472 char * string[ACTION_STRING_COUNT];
474 /* Strings to remove. */
475 struct list multi_remove[ACTION_MULTI_COUNT][1];
477 /* If nonzero, remove *all* strings. */
478 int multi_remove_all[ACTION_MULTI_COUNT];
481 struct list multi_add[ACTION_MULTI_COUNT][1];
485 * This structure is used to store the actions list.
487 * It contains a URL pattern, and the chages to the actions.
488 * It is a linked list.
492 struct url_spec url[1];
494 struct action_spec action[1];
496 struct url_actions * next;
500 /* Constants defining bitmask for csp->accept_types */
502 #ifdef FEATURE_IMAGE_DETECT_MSIE
504 /* MSIE detected by user-agent string */
505 #define ACCEPT_TYPE_IS_MSIE 0x0001
508 * *If* this is MSIE, it wants an image. (Or this is a shift-reload, or
509 * it's got an image from this URL before... yuck!)
510 * Only meaningful if ACCEPT_TYPE_IS_MSIE set
512 #define ACCEPT_TYPE_MSIE_IMAGE 0x0002
515 * *If* this is MSIE, it wants a HTML document.
516 * Only meaningful if ACCEPT_TYPE_IS_MSIE set
518 #define ACCEPT_TYPE_MSIE_HTML 0x0004
520 #endif /* def FEATURE_IMAGE_DETECT_MSIE */
525 /* The proxy's configuration */
526 struct configuration_spec * config;
528 /* The actions to perform on the current request */
529 struct current_action_spec action[1];
531 /* socket to talk to client (web browser) */
534 /* socket to talk to server (web server or proxy) */
538 #ifdef FEATURE_STATISTICS
539 /* 1 if this URL was rejected, 0 otherwise. Allows actual stats inc to
540 * occur in main thread only for thread-safety.
543 #endif /* def FEATURE_STATISTICS */
545 #ifdef FEATURE_FORCE_LOAD
547 #endif /* def FEATURE_FORCE_LOAD */
549 #ifdef FEATURE_TOGGLE
551 #endif /* def FEATURE_TOGGLE */
554 * Client PC's IP address, as reported by the accept()_ function.
555 * Both as string and number
561 /* Our IP address and hostname, i.e. the IP address that
562 the client used to reach us, and the associated hostname,
565 char *my_ip_addr_str;
569 /* The referer in this request, if one was specified. */
571 #endif /* def FEATURE_TRUST */
573 #if defined(FEATURE_IMAGE_DETECT_MSIE)
574 /* Types the client will accept.
575 * Bitmask - see ACCEPT_TYPE_XXX constants.
578 #endif /* defined(FEATURE_IMAGE_DETECT_MSIE) */
580 /* The URL that was requested */
581 struct http_request http[1];
583 /* An I/O buffer used for buffering data read from the client */
586 /* List of all headers for this request */
587 struct list headers[1];
589 /* List of all cookies for this request */
590 struct list cookie_list[1];
592 /* MIME-Type bitmap, see CT_* above */
593 unsigned char content_type;
595 /* The "X-Forwarded-For:" header sent by the client */
599 * Nonzero if this client is processing data.
600 * Set to zero when the thread associated with this structure dies.
604 /* files associated with this client */
605 struct file_list *actions_list;
607 struct file_list *rlist; /* pcrs job file */
608 size_t content_length; /* Length after content modification */
611 struct file_list *tlist; /* trustfile */
612 #endif /* def FEATURE_TRUST */
614 struct client_state *next;
622 char *(*parser)(const struct parsers *, const char *, struct client_state *);
625 struct cgi_dispatcher
629 int (*handler)(struct client_state *csp, struct http_response *rsp, struct map *parameters);
630 const char *description;
636 * this is a pointer to the data structures associated with the file.
637 * Read-only once the structure has been created.
641 /* Normally NULL. When we are finished with file (i.e. when we have
642 * loaded a new one), set to a pointer to an unloader function.
643 * Unloader will be called by sweep() (called from main loop) when
644 * all clients using this file are done. This prevents threading
647 void (*unloader)(void *);
649 /* Used internally by sweep(). Do not access from elsewhere. */
652 /* Following variables allow us to check if file has been changed.
653 * Read-only once the structure has been created.
658 /* Pointer to next entry in the linked list of all "file_list"s.
659 * This linked list is so that sweep() can navigate it.
660 * Since sweep() can remove items from the list, we must be careful
661 * to only access this value from main thread (when we know sweep
664 struct file_list *next;
671 struct url_spec url[1];
673 struct block_spec *next;
675 #endif /* def FEATURE_TRUST */
678 #define SOCKS_NONE 0 /* Don't use a SOCKS server */
679 #define SOCKS_4 40 /* original SOCKS 4 protocol */
680 #define SOCKS_4A 41 /* as modified for hosts w/o external DNS */
684 struct url_spec url[1];
686 /* Connection type - must be a SOCKS_xxx constant */
693 /* Parent HTTP proxy */
697 /* For the linked list */
698 struct forward_spec *next;
700 #define FORWARD_SPEC_INITIALIZER { URL_SPEC_INITIALIZER, 0, NULL, 0, NULL, 0, NULL }
703 struct re_filterfile_spec
707 struct list patterns[1];
712 #define ACL_PERMIT 1 /* accept connection request */
713 #define ACL_DENY 2 /* reject connection request */
715 struct access_control_addr
722 struct access_control_list
724 struct access_control_addr src[1];
725 struct access_control_addr dst[1];
728 struct access_control_list *next;
730 #endif /* def FEATURE_ACL */
733 /* Maximum number of loaders (actions, re_filter, ...) */
737 * Data loaded from the configuration file.
739 * (Anomaly: toggle is still handled through a global, not this structure)
741 struct configuration_spec
750 const char *actions_file;
752 /* The administrator's email address */
755 /* A URL with info on this proxy */
756 char *proxy_info_url;
758 const char *re_filterfile;
760 #ifdef FEATURE_COOKIE_JAR
761 const char * jarfile;
763 #endif /* def FEATURE_COOKIE_JAR */
766 * Port and IP to bind to.
767 * Defaults to HADDR_DEFAULT:HADDR_PORT == 127.0.0.1:8000
773 const char * trustfile;
775 struct list trust_info[1];
776 struct url_spec *trust_list[64];
777 #endif /* def FEATURE_TRUST */
780 struct access_control_list *acl;
781 #endif /* def FEATURE_ACL */
783 struct forward_spec *forward;
785 /* All options from the config file, HTML-formatted */
788 /* the configuration file object. */
789 struct file_list *config_file_list;
791 /* List of loaders */
792 int (*loaders[NLOADERS])(struct client_state *);
794 /* bool, nonzero if we need to bind() to the new port */
799 #define SZ(X) (sizeof(X) / sizeof(*X))
801 #ifdef FEATURE_FORCE_LOAD
802 #define FORCE_PREFIX "/IJB-FORCE-LOAD"
803 #endif /* def FEATURE_FORCE_LOAD */
806 #define HOME_PAGE_URL "http://ijbswa.sourceforge.net"
807 #define REDIRECT_URL HOME_PAGE_URL "/redirect.php?v=" VERSION "&to="
808 #define CGI_PREFIX_HOST "i.j.b"
811 static const char CSUCCEED[] =
812 "HTTP/1.0 200 Connection established\n"
813 "Proxy-Agent: IJ/" VERSION "\n\n";
815 static const char CHEADER[] =
816 "HTTP/1.0 400 Invalid header received from browser\n\n";
822 #endif /* ndef PROJECT_H_INCLUDED */