X-Git-Url: http://www.privoxy.org/gitweb/?p=privoxy.git;a=blobdiff_plain;f=parsers.c;h=6bf55ea00b21617bd2f718714062e5967292cd83;hp=cd03beba3f86a7f5f639d423c280e2b156da698c;hb=9285f53c6431ca054e4a970d97ee22db090639ab;hpb=ffc1ab733579543abf77003e7d4b1a373d81c7a3 diff --git a/parsers.c b/parsers.c index cd03beba..6bf55ea0 100644 --- a/parsers.c +++ b/parsers.c @@ -1,4 +1,4 @@ -const char parsers_rcs[] = "$Id: parsers.c,v 1.7 2001/05/27 13:19:06 oes Exp $"; +const char parsers_rcs[] = "$Id: parsers.c,v 1.42 2001/11/22 21:59:30 jongfoster Exp $"; /********************************************************************* * * File : $Source: /cvsroot/ijbswa/current/parsers.c,v $ @@ -18,10 +18,10 @@ const char parsers_rcs[] = "$Id: parsers.c,v 1.7 2001/05/27 13:19:06 oes Exp $"; * IJBSWA team. http://ijbswa.sourceforge.net * * Based on the Internet Junkbuster originally written - * by and Copyright (C) 1997 Anonymous Coders and + * by and Copyright (C) 1997 Anonymous Coders and * Junkbusters Corporation. http://www.junkbusters.com * - * This program is free software; you can redistribute it + * This program is free software; you can redistribute it * and/or modify it under the terms of the GNU General * Public License as published by the Free Software * Foundation; either version 2 of the License, or (at @@ -41,6 +41,201 @@ const char parsers_rcs[] = "$Id: parsers.c,v 1.7 2001/05/27 13:19:06 oes Exp $"; * * Revisions : * $Log: parsers.c,v $ + * Revision 1.42 2001/11/22 21:59:30 jongfoster + * Adding code to handle +no-cookies-keep + * + * Revision 1.41 2001/11/05 23:43:05 steudten + * Add time+date to log files. + * + * Revision 1.40 2001/10/26 20:13:09 jongfoster + * ctype.h is needed in Windows, too. + * + * Revision 1.39 2001/10/26 17:40:04 oes + * Introduced get_header_value() + * Removed http->user_agent, csp->referrer and csp->accept_types + * Removed client_accept() + * + * Revision 1.38 2001/10/25 03:40:48 david__schmidt + * Change in porting tactics: OS/2's EMX porting layer doesn't allow multiple + * threads to call select() simultaneously. So, it's time to do a real, live, + * native OS/2 port. See defines for __EMX__ (the porting layer) vs. __OS2__ + * (native). Both versions will work, but using __OS2__ offers multi-threading. + * + * Revision 1.37 2001/10/23 21:36:02 jongfoster + * Documenting sed()'s error behaviou (doc change only) + * + * Revision 1.36 2001/10/13 12:51:51 joergs + * Removed client_host, (was only required for the old 2.0.2-11 http://noijb. + * force-load), instead crumble Host: and add it (again) in client_host_adder + * (in case we get a HTTP/1.0 request without Host: header and forward it to + * a HTTP/1.1 server/proxy). + * + * Revision 1.35 2001/10/09 22:39:21 jongfoster + * assert.h is also required under Win32, so moving out of #ifndef _WIN32 + * block. + * + * Revision 1.34 2001/10/07 18:50:55 oes + * Added server_content_encoding, renamed server_transfer_encoding + * + * Revision 1.33 2001/10/07 18:04:49 oes + * Changed server_http11 to server_http and its pattern to "HTTP". + * Additional functionality: it now saves the HTTP status into + * csp->http->status and sets CT_TABOO for Status 206 (partial range) + * + * Revision 1.32 2001/10/07 15:43:28 oes + * Removed FEATURE_DENY_GZIP and replaced it with client_accept_encoding, + * client_te and client_accept_encoding_adder, triggered by the new + * +no-compression action. For HTTP/1.1 the Accept-Encoding header is + * changed to allow only identity and chunked, and the TE header is + * crunched. For HTTP/1.0, Accept-Encoding is crunched. + * + * parse_http_request no longer does anything than parsing. The rewriting + * of http->cmd and version mangling are gone. It now also recognizes + * the put and delete methods and saves the url in http->url. Removed + * unused variable. + * + * renamed content_type and content_length to have the server_ prefix + * + * server_content_type now only works if csp->content_type != CT_TABOO + * + * added server_transfer_encoding, which + * - Sets CT_TABOO to prohibit filtering if encoding compresses + * - Raises the CSP_FLAG_CHUNKED flag if Encoding is "chunked" + * - Change from "chunked" to "identity" if body was chunked + * but has been de-chunked for filtering. + * + * added server_content_md5 which crunches any Content-MD5 headers + * if the body was modified. + * + * made server_http11 conditional on +downgrade action + * + * Replaced 6 boolean members of csp with one bitmap (csp->flags) + * + * Revision 1.31 2001/10/05 14:25:02 oes + * Crumble Keep-Alive from Server + * + * Revision 1.30 2001/09/29 12:56:03 joergs + * IJB now changes HTTP/1.1 to HTTP/1.0 in requests and answers. + * + * Revision 1.29 2001/09/24 21:09:24 jongfoster + * Fixing 2 memory leaks that Guy spotted, where the paramater to + * enlist() was not being free()d. + * + * Revision 1.28 2001/09/22 16:32:28 jongfoster + * Removing unused #includes. + * + * Revision 1.27 2001/09/20 15:45:25 steudten + * + * add casting from size_t to int for printf() + * remove local variable shadow s2 + * + * Revision 1.26 2001/09/16 17:05:14 jongfoster + * Removing unused #include showarg.h + * + * Revision 1.25 2001/09/16 13:21:27 jongfoster + * Changes to use new list functions. + * + * Revision 1.24 2001/09/13 23:05:50 jongfoster + * Changing the string paramater to the header parsers a "const". + * + * Revision 1.23 2001/09/12 18:08:19 steudten + * + * In parse_http_request() header rewriting miss the host value, so + * from http://www.mydomain.com the result was just " / " not + * http://www.mydomain.com/ in case we forward. + * + * Revision 1.22 2001/09/10 10:58:53 oes + * Silenced compiler warnings + * + * Revision 1.21 2001/07/31 14:46:00 oes + * - Persistant connections now suppressed + * - sed() no longer appends empty header to csp->headers + * + * Revision 1.20 2001/07/30 22:08:36 jongfoster + * Tidying up #defines: + * - All feature #defines are now of the form FEATURE_xxx + * - Permanently turned off WIN_GUI_EDIT + * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS + * + * Revision 1.19 2001/07/25 17:21:54 oes + * client_uagent now saves copy of User-Agent: header value + * + * Revision 1.18 2001/07/13 14:02:46 oes + * - Included fix to repair broken HTTP requests that + * don't contain a path, not even '/'. + * - Removed all #ifdef PCRS + * - content_type now always inspected and classified as + * text, gif or other. + * - formatting / comments + * + * Revision 1.17 2001/06/29 21:45:41 oes + * Indentation, CRLF->LF, Tab-> Space + * + * Revision 1.16 2001/06/29 13:32:42 oes + * - Fixed a comment + * - Adapted free_http_request + * - Removed logentry from cancelled commit + * + * Revision 1.15 2001/06/03 19:12:38 oes + * deleted const struct interceptors + * + * Revision 1.14 2001/06/01 18:49:17 jongfoster + * Replaced "list_share" with "list" - the tiny memory gain was not + * worth the extra complexity. + * + * Revision 1.13 2001/05/31 21:30:33 jongfoster + * Removed list code - it's now in list.[ch] + * Renamed "permission" to "action", and changed many features + * to use the actions file rather than the global config. + * + * Revision 1.12 2001/05/31 17:33:13 oes + * + * CRLF -> LF + * + * Revision 1.11 2001/05/29 20:11:19 joergs + * '/ * inside comment' warning removed. + * + * Revision 1.10 2001/05/29 09:50:24 jongfoster + * Unified blocklist/imagelist/permissionslist. + * File format is still under discussion, but the internal changes + * are (mostly) done. + * + * Also modified interceptor behaviour: + * - We now intercept all URLs beginning with one of the following + * prefixes (and *only* these prefixes): + * * http://i.j.b/ + * * http://ijbswa.sf.net/config/ + * * http://ijbswa.sourceforge.net/config/ + * - New interceptors "home page" - go to http://i.j.b/ to see it. + * - Internal changes so that intercepted and fast redirect pages + * are not replaced with an image. + * - Interceptors now have the option to send a binary page direct + * to the client. (i.e. ijb-send-banner uses this) + * - Implemented show-url-info interceptor. (Which is why I needed + * the above interceptors changes - a typical URL is + * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif". + * The previous mechanism would not have intercepted that, and + * if it had been intercepted then it then it would have replaced + * it with an image.) + * + * Revision 1.9 2001/05/28 17:26:33 jongfoster + * Fixing segfault if last header was crunched. + * Fixing Windows build (snprintf() is _snprintf() under Win32, but we + * can use the cross-platform sprintf() instead.) + * + * Revision 1.8 2001/05/27 22:17:04 oes + * + * - re_process_buffer no longer writes the modified buffer + * to the client, which was very ugly. It now returns the + * buffer, which it is then written by chat. + * + * - content_length now adjusts the Content-Length: header + * for modified documents rather than crunch()ing it. + * (Length info in csp->content_length, which is 0 for + * unmodified documents) + * + * - For this to work, sed() is called twice when filtering. + * * Revision 1.7 2001/05/27 13:19:06 oes * Patched Joergs solution for the content-length in. * @@ -124,27 +319,29 @@ const char parsers_rcs[] = "$Id: parsers.c,v 1.7 2001/05/27 13:19:06 oes Exp $"; #include "config.h" +#ifndef _WIN32 #include #include +#endif + #include #include +#include #include -#ifndef _WIN32 +#if !defined(_WIN32) && !defined(__OS2__) #include #endif #include "project.h" +#include "list.h" #include "parsers.h" #include "encode.h" -#include "filters.h" -#include "loaders.h" -#include "showargs.h" -#include "jcc.h" #include "ssplit.h" #include "errlog.h" #include "jbsockets.h" #include "miscutil.h" +#include "list.h" const char parsers_h_rcs[] = PARSERS_H_VERSION; @@ -154,8 +351,8 @@ const char parsers_h_rcs[] = PARSERS_H_VERSION; * as an array index. Therefore we need to make sure that high-bit * characters generate +ve values, and ideally we also want to make * the argument match the declared parameter type of "int". - * - * Why did they write a character function that can't take a simple + * + * Why did they write a character function that can't take a simple * "char" argument? Doh! */ #define ijb_isupper(__X) isupper((int)(unsigned char)(__X)) @@ -169,51 +366,44 @@ const struct parsers client_patterns[] = { { "from:", 5, client_from }, { "cookie:", 7, client_send_cookie }, { "x-forwarded-for:", 16, client_x_forwarded }, - { "proxy-connection:", 17, crumble }, -#ifdef DENY_GZIP - { "Accept-Encoding: gzip", 21, crumble }, -#endif /* def DENY_GZIP */ -#if defined(DETECT_MSIE_IMAGES) - { "Accept:", 7, client_accept }, -#endif /* defined(DETECT_MSIE_IMAGES) */ -#ifdef FORCE_LOAD - { "Host:", 5, client_host }, -#endif /* def FORCE_LOAD */ + { "Accept-Encoding:", 16, client_accept_encoding }, + { "TE:", 3, client_te }, + { "Host:", 5, crumble }, /* { "if-modified-since:", 18, crumble }, */ + { "Keep-Alive:", 11, crumble }, + { "connection:", 11, crumble }, + { "proxy-connection:", 17, crumble }, { NULL, 0, NULL } }; -const struct interceptors intercept_patterns[] = { - { "show-proxy-args", 14, show_proxy_args }, - { "ijb-send-banner", 14, ijb_send_banner }, -#ifdef TRUST_FILES - { "ij-untrusted-url", 15, ij_untrusted_url }, -#endif /* def TRUST_FILES */ - { NULL, 0, NULL } -}; const struct parsers server_patterns[] = { + { "HTTP", 4, server_http }, { "set-cookie:", 11, server_set_cookie }, { "connection:", 11, crumble }, -#if defined(PCRS) || defined(KILLPOPUPS) - { "Content-Type:", 13, content_type }, -#endif /* defined(PCRS) || defined(KILLPOPUPS) */ -#ifdef PCRS - { "Content-Length:", 15, content_length }, -#endif /* def PCRS */ + { "Content-Type:", 13, server_content_type }, + { "Content-Length:", 15, server_content_length }, + { "Content-MD5:", 12, server_content_md5 }, + { "Content-Encoding:", 17, server_content_encoding }, + { "Transfer-Encoding:", 18, server_transfer_coding }, + { "Keep-Alive:", 11, crumble }, { NULL, 0, NULL } }; void (* const add_client_headers[])(struct client_state *) = { + client_host_adder, client_cookie_adder, client_x_forwarded_adder, client_xtra_adder, + client_accept_encoding_adder, + connection_close_adder, NULL }; void (* const add_server_headers[])(struct client_state *) = { + connection_close_adder, NULL }; @@ -376,123 +566,54 @@ char *get_header(struct client_state *csp) /********************************************************************* * - * Function : enlist - * - * Description : Append a string into a specified string list. - * - * Parameters : - * 1 : h = pointer to list 'dummy' header - * 2 : s = string to add to the list - * - * Returns : N/A - * - *********************************************************************/ -void enlist(struct list *h, const char *s) -{ - struct list *n = (struct list *)malloc(sizeof(*n)); - struct list *l; - - if (n) - { - n->str = strdup(s); - n->next = NULL; - - if ((l = h->last)) - { - l->next = n; - } - else - { - h->next = n; - } - - h->last = n; - } - -} - - -/********************************************************************* - * - * Function : destroy_list + * Function : get_header_value * - * Description : Destroy a string list (opposite of enlist) + * Description : Get the value of a given header from a chained list + * of header lines or return NULL if no such header is + * present in the list. * * Parameters : - * 1 : h = pointer to list 'dummy' header + * 1 : header_list = pointer to list + * 2 : header_name = string with name of header to look for. + * Trailing colon required, capitalization + * doesn't matter. * - * Returns : N/A + * Returns : NULL if not found, else value of header * *********************************************************************/ -void destroy_list(struct list *h) +char *get_header_value(const struct list *header_list, const char *header_name) { - struct list *p, *n; - - for (p = h->next; p ; p = n) - { - n = p->next; - freez(p->str); - freez(p); - } - - memset(h, '\0', sizeof(*h)); - -} - - -/********************************************************************* - * - * Function : list_to_text - * - * Description : "Flaten" a string list into 1 long \r\n delimited string. - * - * Parameters : - * 1 : h = pointer to list 'dummy' header - * - * Returns : NULL on malloc error, else new long string. - * - *********************************************************************/ -static char *list_to_text(struct list *h) -{ - struct list *p; + struct list_entry *cur_entry; char *ret = NULL; - char *s; - int size; - - size = 0; - - for (p = h->next; p ; p = p->next) - { - if (p->str) - { - size += strlen(p->str) + 2; - } - } + size_t length = 0; - if ((ret = (char *)malloc(size + 1)) == NULL) - { - return(NULL); - } + assert(header_list); + assert(header_name); + length = strlen(header_name); - ret[size] = '\0'; - - s = ret; - - for (p = h->next; p ; p = p->next) + for (cur_entry = header_list->first; cur_entry ; cur_entry = cur_entry->next) { - if (p->str) + if (cur_entry->str) { - strcpy(s, p->str); - s += strlen(s); - *s++ = '\r'; *s++ = '\n'; + if (!strncmpic(cur_entry->str, header_name, length)) + { + /* + * Found: return pointer to start of value + */ + ret = (char *) (cur_entry->str + length); + while (*ret && ijb_isspace(*ret)) ret++; + return(ret); + } } } - return(ret); + /* + * Not found + */ + return NULL; } - /********************************************************************* * * Function : sed @@ -511,19 +632,20 @@ static char *list_to_text(struct list *h) * headers (client or server) * 3 : csp = Current client state (buffers, headers, etc...) * - * Returns : Single pointer to a fully formed header. + * Returns : Single pointer to a fully formed header, or NULL + * on out-of-memory error. * *********************************************************************/ char *sed(const struct parsers pats[], void (* const more_headers[])(struct client_state *), struct client_state *csp) { - struct list *p; + struct list_entry *p; const struct parsers *v; char *hdr; void (* const *f)(); for (v = pats; v->str ; v++) { - for (p = csp->headers->next; p ; p = p->next) + for (p = csp->headers->first; p ; p = p->next) { /* Header crunch()ed in previous run? -> ignore */ if (p->str == NULL) continue; @@ -533,7 +655,7 @@ char *sed(const struct parsers pats[], void (* const more_headers[])(struct clie if (strncmpic(p->str, v->str, v->len) == 0) { hdr = v->parser(v, p->str, csp); - freez(p->str); + freez(p->str); /* FIXME: Yuck! patching a list...*/ p->str = hdr; } } @@ -545,15 +667,9 @@ char *sed(const struct parsers pats[], void (* const more_headers[])(struct clie (*f)(csp); } - /* add the blank line at the end of the header, if necessary */ - if(strlen(csp->headers->last->str) != 0) - { - enlist(csp->headers, ""); - } - hdr = list_to_text(csp->headers); - return(hdr); + return hdr; } @@ -572,12 +688,16 @@ char *sed(const struct parsers pats[], void (* const more_headers[])(struct clie *********************************************************************/ void free_http_request(struct http_request *http) { + assert(http); + freez(http->cmd); freez(http->gpc); freez(http->host); + freez(http->url); freez(http->hostport); freez(http->path); freez(http->ver); + freez(http->host_ip_addr_str); } @@ -603,11 +723,9 @@ void parse_http_request(char *req, struct http_request *http, struct client_stat int n; memset(http, '\0', sizeof(*http)); - http->cmd = strdup(req); buf = strdup(req); - n = ssplit(buf, " \r\n", v, SZ(v), 1, 1); if (n == 3) @@ -621,41 +739,29 @@ void parse_http_request(char *req, struct http_request *http, struct client_stat http->ver = strdup(v[2]); } -#ifdef WEBDAV - -/* This next line is a little ugly, but it simplifies the if statement below. */ -/* Basically if using webDAV, we want the OR condition to use these too. */ - -/* - * by haroon - * These are the headers as defined in RFC2518 to add webDAV support - */ - -#define OR_WEBDAV || \ - (0 == strcmpic(v[0], "propfind")) || \ - (0 == strcmpic(v[0], "proppatch")) || \ - (0 == strcmpic(v[0], "move")) || \ - (0 == strcmpic(v[0], "copy")) || \ - (0 == strcmpic(v[0], "mkcol")) || \ - (0 == strcmpic(v[0], "lock")) || \ - (0 == strcmpic(v[0], "unlock")) - -#else /* No webDAV support is enabled. Provide an empty OR_WEBDAV macro. */ - -#define OR_WEBDAV - -#endif - - /* or it could be a GET or a POST (possibly webDAV too) */ - if ((strcmpic(v[0], "get") == 0) || - (strcmpic(v[0], "head") == 0) OR_WEBDAV || - (strcmpic(v[0], "post") == 0)) + /* or it could be any other basic HTTP request type */ + if ((0 == strcmpic(v[0], "get")) + || (0 == strcmpic(v[0], "head")) + || (0 == strcmpic(v[0], "post")) + || (0 == strcmpic(v[0], "put")) + || (0 == strcmpic(v[0], "delete")) + + /* or a webDAV extension (RFC2518) */ + || (0 == strcmpic(v[0], "propfind")) + || (0 == strcmpic(v[0], "proppatch")) + || (0 == strcmpic(v[0], "move")) + || (0 == strcmpic(v[0], "copy")) + || (0 == strcmpic(v[0], "mkcol")) + || (0 == strcmpic(v[0], "lock")) + || (0 == strcmpic(v[0], "unlock")) + ) { - http->ssl = 0; - http->gpc = strdup(v[0]); - url = v[1]; - http->ver = strdup(v[2]); + http->ssl = 0; + http->gpc = strdup(v[0]); + http->url = strdup(v[1]); + http->ver = strdup(v[2]); + url = v[1]; if (strncmpic(url, "http://", 7) == 0) { url += 7; @@ -669,11 +775,22 @@ void parse_http_request(char *req, struct http_request *http, struct client_stat url = NULL; } - if (url && (p = strchr(url, '/'))) + if (url) { - http->path = strdup(p); - *p = '\0'; - http->hostport = strdup(url); + if ((p = strchr(url, '/'))) + { + http->path = strdup(p); + *p = '\0'; + http->hostport = strdup(url); + } + /* + * Repair broken HTTP requests that don't contain a path + */ + else + { + http->path = strdup("/"); + http->hostport = strdup(url); + } } } } @@ -723,7 +840,7 @@ void parse_http_request(char *req, struct http_request *http, struct client_stat if (http->path == NULL) { - http->path = strdup(""); + http->path = strdup("/"); } } @@ -747,7 +864,7 @@ void parse_http_request(char *req, struct http_request *http, struct client_stat * Returns : Always NULL. * *********************************************************************/ -char *crumble(const struct parsers *v, char *s, struct client_state *csp) +char *crumble(const struct parsers *v, const char *s, struct client_state *csp) { log_error(LOG_LEVEL_HEADER, "crunch!"); return(NULL); @@ -755,13 +872,47 @@ char *crumble(const struct parsers *v, char *s, struct client_state *csp) } -#if defined(PCRS) || defined(KILLPOPUPS) +/********************************************************************* + * + * Function : server_content_type + * + * Description : Set the content-type for filterable types (text/.*, + * javascript and image/gif) unless filtering has been + * forbidden (CT_TABOO) while parsing earlier headers. + * + * Parameters : + * 1 : v = ignored + * 2 : s = header string we are "considering" + * 3 : csp = Current client state (buffers, headers, etc...) + * + * Returns : A duplicate string pointer to this header (ie. pass thru) + * + *********************************************************************/ +char *server_content_type(const struct parsers *v, const char *s, struct client_state *csp) +{ + if (csp->content_type != CT_TABOO) + { + if (strstr(s, " text/") || strstr(s, "application/x-javascript")) + csp->content_type = CT_TEXT; + else if (strstr(s, " image/gif")) + csp->content_type = CT_GIF; + else + csp->content_type = 0; + } + + return(strdup(s)); + +} + /********************************************************************* * - * Function : content_type + * Function : server_transfer_coding * - * Description : Is this a text/* or javascript MIME Type? + * Description : - Prohibit filtering (CT_TABOO) if transfer coding compresses + * - Raise the CSP_FLAG_CHUNKED flag if coding is "chunked" + * - Change from "chunked" to "identity" if body was chunked + * but has been de-chunked for filtering. * * Parameters : * 1 : v = ignored @@ -771,26 +922,43 @@ char *crumble(const struct parsers *v, char *s, struct client_state *csp) * Returns : A duplicate string pointer to this header (ie. pass thru) * *********************************************************************/ -char *content_type(const struct parsers *v, char *s, struct client_state *csp) +char *server_transfer_coding(const struct parsers *v, const char *s, struct client_state *csp) { - if (strstr (s, " text/") || strstr (s, "application/x-javascript")) - csp->is_text = 1; - else - csp->is_text = 0; + /* + * Turn off pcrs and gif filtering if body compressed + */ + if (strstr(s, "gzip") || strstr(s, "compress") || strstr(s, "deflate")) + { + csp->content_type = CT_TABOO; + } + + /* + * Raise flag if body chunked + */ + if (strstr(s, "chunked")) + { + csp->flags |= CSP_FLAG_CHUNKED; + + /* + * If the body was modified, it has been + * de-chunked first, so adjust the header: + */ + if (csp->flags & CSP_FLAG_MODIFIED) + { + return(strdup("Transfer-Encoding: identity")); + } + } return(strdup(s)); } -#endif /* defined(PCRS) || defined(KILLPOPUPS) */ -#ifdef PCRS /********************************************************************* * - * Function : content_length + * Function : server_content_encoding * - * Description : Crunch Content-Length header if & only if we are - * filtering this page through PCRS. + * Description : Prohibit filtering (CT_TABOO) if content encoding compresses * * Parameters : * 1 : v = ignored @@ -800,24 +968,152 @@ char *content_type(const struct parsers *v, char *s, struct client_state *csp) * Returns : A duplicate string pointer to this header (ie. pass thru) * *********************************************************************/ -char *content_length(const struct parsers *v, char *s, struct client_state *csp) +char *server_content_encoding(const struct parsers *v, const char *s, struct client_state *csp) { - if (csp->content_length != 0) /* Content has been modified */ - { - s = (char *) zalloc(100); - snprintf(s, 100, "Content-Length: %d", csp->content_length); - log_error(LOG_LEVEL_HEADER, "Adjust Content-Length to %d", csp->content_length); - return(s); - } + /* + * Turn off pcrs and gif filtering if body compressed + */ + if (strstr(s, "gzip") || strstr(s, "compress") || strstr(s, "deflate")) + { + csp->content_type = CT_TABOO; + } + + return(strdup(s)); + +} + + +/********************************************************************* + * + * Function : server_content_length + * + * Description : Adjust Content-Length header if we modified + * the body. + * + * Parameters : + * 1 : v = ignored + * 2 : s = header string we are "considering" + * 3 : csp = Current client state (buffers, headers, etc...) + * + * Returns : A duplicate string pointer to this header (ie. pass thru) + * + *********************************************************************/ +char *server_content_length(const struct parsers *v, const char *s, struct client_state *csp) +{ + if (csp->content_length != 0) /* Content length has been modified */ + { + char * s2 = (char *) zalloc(100); + sprintf(s2, "Content-Length: %d", (int) csp->content_length); + + log_error(LOG_LEVEL_HEADER, "Adjust Content-Length to %d", (int) csp->content_length); + return(s2); + } else { return(strdup(s)); } + } -#endif /* def PCRS */ + +/********************************************************************* + * + * Function : server_content_md5 + * + * Description : Crumble any Content-MD5 headers if the document was + * modified. FIXME: Should we re-compute instead? + * + * Parameters : + * 1 : v = ignored + * 2 : s = header string we are "considering" + * 3 : csp = Current client state (buffers, headers, etc...) + * + * Returns : A duplicate string pointer to this header (ie. pass thru) + * + *********************************************************************/ +char *server_content_md5(const struct parsers *v, const char *s, struct client_state *csp) +{ + if (csp->flags & CSP_FLAG_MODIFIED) + { + log_error(LOG_LEVEL_HEADER, "Crunching Content-MD5"); + return(NULL); + } + else + { + return(strdup(s)); + } + +} +/********************************************************************* + * + * Function : client_accept_encoding + * + * Description : Rewrite the client's Accept-Encoding header so that + * if doesn't allow compression, if the action applies. + * Note: For HTTP/1.0 the absence of the header is enough. + * + * Parameters : + * 1 : v = ignored + * 2 : s = header string we are "considering" + * 3 : csp = Current client state (buffers, headers, etc...) + * + * Returns : A copy of the client's original or the modified header. + * + *********************************************************************/ +char *client_accept_encoding(const struct parsers *v, const char *s, struct client_state *csp) +{ + if ((csp->action->flags & ACTION_NO_COMPRESSION) == 0) + { + return(strdup(s)); + } + else + { + log_error(LOG_LEVEL_HEADER, "Supressed offer to compress content"); + + if (!strcmpic(csp->http->ver, "HTTP/1.1")) + { + return(strdup("Accept-Encoding: identity;q=1.0, *;q=0")); + } + else + { + return(NULL); + } + } + +} + + +/********************************************************************* + * + * Function : client_te + * + * Description : Rewrite the client's TE header so that + * if doesn't allow compression, if the action applies. + * + * Parameters : + * 1 : v = ignored + * 2 : s = header string we are "considering" + * 3 : csp = Current client state (buffers, headers, etc...) + * + * Returns : A copy of the client's original or the modified header. + * + *********************************************************************/ +char *client_te(const struct parsers *v, const char *s, struct client_state *csp) +{ + if ((csp->action->flags & ACTION_NO_COMPRESSION) == 0) + { + return(strdup(s)); + } + else + { + log_error(LOG_LEVEL_HEADER, "Supressed offer to compress transfer"); + return(NULL); + } + +} + /********************************************************************* * * Function : client_referrer @@ -834,70 +1130,86 @@ char *content_length(const struct parsers *v, char *s, struct client_state *csp) * or modified header * *********************************************************************/ -char *client_referrer(const struct parsers *v, char *s, struct client_state *csp) +char *client_referrer(const struct parsers *v, const char *s, struct client_state *csp) { -#ifdef FORCE_LOAD + const char * newval; + char * s2; +#ifdef FEATURE_FORCE_LOAD /* Since the referrer can include the prefix even * even if the request itself is non-forced, we must - * clean it unconditionally + * clean it unconditionally */ strclean(s, FORCE_PREFIX); -#endif /* def FORCE_LOAD */ +#endif /* def FEATURE_FORCE_LOAD */ - csp->referrer = strdup(s); - - if (csp->config->referrer == NULL) + /* + * Are we sending referer? + */ + if ((csp->action->flags & ACTION_HIDE_REFERER) == 0) { - log_error(LOG_LEVEL_HEADER, "crunch!"); - return(NULL); + return(strdup(s)); } - if (*csp->config->referrer == '.') + newval = csp->action->string[ACTION_STRING_REFERER]; + + /* + * Are we blocking referer? + */ + if ((newval == NULL) || (0 == strcmpic(newval, "block")) ) { - return(strdup(s)); + log_error(LOG_LEVEL_HEADER, "crunch!"); + return(NULL); } - if (*csp->config->referrer == '@') + /* + * Are we forging referer? + */ + if (0 == strcmpic(newval, "forge")) { - if (csp->permissions & PERMIT_COOKIE_READ) - { - return(strdup(s)); - } - else - { - log_error(LOG_LEVEL_HEADER, "crunch!"); - return(NULL); - } + /* + * Forge a referer as http://[hostname:port of REQUEST]/ + * to fool stupid checks for in-site links + */ + log_error(LOG_LEVEL_HEADER, "crunch+forge!"); + s2 = strsav(NULL, "Referer: "); + s2 = strsav(s2, "http://"); + s2 = strsav(s2, csp->http->hostport); + s2 = strsav(s2, "/"); + return(s2); } /* - * New option § or L: Forge a referer as http://[hostname:port of REQUEST]/ - * to fool stupid checks for in-site links + * Have we got a fixed referer? */ - - if (*csp->config->referrer == '§' || *csp->config->referrer == 'L') + if (0 == strncmpic(newval, "http://", 7)) { - if (csp->permissions & PERMIT_COOKIE_READ) - { - return(strdup(s)); - } - else - { - log_error(LOG_LEVEL_HEADER, "crunch+forge!"); - s = strsav(NULL, "Referer: "); - s = strsav(s, "http://"); - s = strsav(s, csp->http->hostport); - s = strsav(s, "/"); - return(s); - } - } + /* + * We have a specific (fixed) referer we want to send. + */ + char * s3; - log_error(LOG_LEVEL_HEADER, "modified"); + log_error(LOG_LEVEL_HEADER, "modified"); + + s3 = strsav( NULL, "Referer: " ); + s3 = strsav( s3, newval ); + return(s3); + } - s = strsav( NULL, "Referer: " ); - s = strsav( s, csp->config->referrer ); - return(s); + /* Should never get here! */ + log_error(LOG_LEVEL_ERROR, "Bad parameter: +referer{%s}", newval); + /* + * Forge is probably the best default. + * + * Forge a referer as http://[hostname:port of REQUEST]/ + * to fool stupid checks for in-site links + */ + log_error(LOG_LEVEL_HEADER, "crunch+forge!"); + s2 = strsav(NULL, "Referer: "); + s2 = strsav(s2, "http://"); + s2 = strsav(s2, csp->http->hostport); + s2 = strsav(s2, "/"); + return(s2); } @@ -905,8 +1217,9 @@ char *client_referrer(const struct parsers *v, char *s, struct client_state *csp * * Function : client_uagent * - * Description : Handle the "user-agent" config setting properly. - * Called from `sed'. + * Description : Handle the "user-agent" config setting properly + * and remember its original value to enable browser + * bug workarounds. Called from `sed'. * * Parameters : * 1 : v = ignored @@ -917,47 +1230,27 @@ char *client_referrer(const struct parsers *v, char *s, struct client_state *csp * a malloc'ed string pointer to this header (ie. pass thru). * *********************************************************************/ -char *client_uagent(const struct parsers *v, char *s, struct client_state *csp) +char *client_uagent(const struct parsers *v, const char *s, struct client_state *csp) { -#ifdef DETECT_MSIE_IMAGES - if (strstr (s, "MSIE ")) - { - /* This is Microsoft Internet Explorer. - * Enable auto-detect. - */ - csp->accept_types |= ACCEPT_TYPE_IS_MSIE; - } -#endif /* def DETECT_MSIE_IMAGES */ + const char * newval; + char * s2; - if (csp->config->uagent == NULL) - { - log_error(LOG_LEVEL_HEADER, "default"); - return(strdup(DEFAULT_USER_AGENT)); - } - - if (*csp->config->uagent == '.') + if ((csp->action->flags & ACTION_HIDE_USER_AGENT) == 0) { return(strdup(s)); } - if (*csp->config->uagent == '@') + newval = csp->action->string[ACTION_STRING_USER_AGENT]; + if (newval == NULL) { - if (csp->permissions & PERMIT_COOKIE_READ) - { - return(strdup(s)); - } - else - { - log_error(LOG_LEVEL_HEADER, "default"); - return(strdup(DEFAULT_USER_AGENT)); - } + return(strdup(s)); } log_error(LOG_LEVEL_HEADER, "modified"); - s = strsav( NULL, "User-Agent: " ); - s = strsav( s, csp->config->uagent ); - return(s); + s2 = strsav( NULL, "User-Agent: " ); + s2 = strsav( s2, newval ); + return(s2); } @@ -976,35 +1269,17 @@ char *client_uagent(const struct parsers *v, char *s, struct client_state *csp) * Returns : NULL if crunched, or a malloc'ed string to original header * *********************************************************************/ -char *client_ua(const struct parsers *v, char *s, struct client_state *csp) +char *client_ua(const struct parsers *v, const char *s, struct client_state *csp) { - if (csp->config->uagent == NULL) - { - log_error(LOG_LEVEL_HEADER, "crunch!"); - return(NULL); - } - - if (*csp->config->uagent == '.') + if ((csp->action->flags & ACTION_HIDE_USER_AGENT) == 0) { return(strdup(s)); } - - if (*csp->config->uagent == '@') + else { - if (csp->permissions & PERMIT_COOKIE_READ) - { - return(strdup(s)); - } - else - { - log_error(LOG_LEVEL_HEADER, "crunch!"); - return(NULL); - } + log_error(LOG_LEVEL_HEADER, "crunch!"); + return(NULL); } - - log_error(LOG_LEVEL_HEADER, "crunch!"); - return(NULL); - } @@ -1024,25 +1299,32 @@ char *client_ua(const struct parsers *v, char *s, struct client_state *csp) * modified/original header. * *********************************************************************/ -char *client_from(const struct parsers *v, char *s, struct client_state *csp) +char *client_from(const struct parsers *v, const char *s, struct client_state *csp) { - /* if not set, zap it */ - if (csp->config->from == NULL) + const char * newval; + char * s2; + + if ((csp->action->flags & ACTION_HIDE_FROM) == 0) { - log_error(LOG_LEVEL_HEADER, "crunch!"); - return(NULL); + return(strdup(s)); } - if (*csp->config->from == '.') + newval = csp->action->string[ACTION_STRING_FROM]; + + /* + * Are we blocking referer? + */ + if ((newval == NULL) || (0 == strcmpic(newval, "block")) ) { - return(strdup(s)); + log_error(LOG_LEVEL_HEADER, "crunch!"); + return(NULL); } log_error(LOG_LEVEL_HEADER, " modified"); - s = strsav( NULL, "From: " ); - s = strsav( s, csp->config->from ); - return(s); + s2 = strsav( NULL, "From: " ); + s2 = strsav( s2, newval ); + return(s2); } @@ -1063,9 +1345,9 @@ char *client_from(const struct parsers *v, char *s, struct client_state *csp) * Returns : Always NULL. * *********************************************************************/ -char *client_send_cookie(const struct parsers *v, char *s, struct client_state *csp) +char *client_send_cookie(const struct parsers *v, const char *s, struct client_state *csp) { - if (csp->permissions & PERMIT_COOKIE_READ) + if ((csp->action->flags & ACTION_NO_COOKIE_READ) == 0) { enlist(csp->cookie_list, s + v->len + 1); } @@ -1098,10 +1380,11 @@ char *client_send_cookie(const struct parsers *v, char *s, struct client_state * * Returns : Always NULL. * *********************************************************************/ -char *client_x_forwarded(const struct parsers *v, char *s, struct client_state *csp) +char *client_x_forwarded(const struct parsers *v, const char *s, struct client_state *csp) { - if (csp->config->add_forwarded) + if ((csp->action->flags & ACTION_HIDE_FORWARDED) == 0) { + /* Save it so we can re-add it later */ csp->x_forwarded = strdup(s); } @@ -1114,48 +1397,32 @@ char *client_x_forwarded(const struct parsers *v, char *s, struct client_state * } -#if defined(DETECT_MSIE_IMAGES) +/* the following functions add headers directly to the header list */ + /********************************************************************* * - * Function : client_accept + * Function : client_host_adder * - * Description : Detect whether the client wants HTML or an image. - * Clients do not always make this information available - * in a sane way. Always passes the header through - * the proxy unchanged. + * Description : (re)adds the host header. Called from `sed'. * * Parameters : - * 1 : v = Ignored. - * 2 : s = Header string. Null terminated. - * 3 : csp = Current client state (buffers, headers, etc...) + * 1 : csp = Current client state (buffers, headers, etc...) * - * Returns : Duplicate of argument s. + * Returns : N/A * *********************************************************************/ -char *client_accept(const struct parsers *v, char *s, struct client_state *csp) +void client_host_adder(struct client_state *csp) { -#ifdef DETECT_MSIE_IMAGES - if (strstr (s, "image/gif")) - { - /* Client will accept HTML. If this seems counterintuitive, - * blame Microsoft. - */ - csp->accept_types |= ACCEPT_TYPE_MSIE_HTML; - } - else - { - csp->accept_types |= ACCEPT_TYPE_MSIE_IMAGE; - } -#endif /* def DETECT_MSIE_IMAGES */ - - return(strdup(s)); - -} -#endif /* defined(DETECT_MSIE_IMAGES) */ + char *p = NULL; + p = strsav(p, "Host: "); + p = strsav(p, csp->http->hostport); + log_error(LOG_LEVEL_HEADER, "addh: %s", p); + enlist(csp->headers, p); -/* the following functions add headers directly to the header list */ + freez(p); +} /********************************************************************* @@ -1172,27 +1439,27 @@ char *client_accept(const struct parsers *v, char *s, struct client_state *csp) *********************************************************************/ void client_cookie_adder(struct client_state *csp) { - struct list *l; + struct list_entry *lst; char *tmp = NULL; char *e; - for (l = csp->cookie_list->next; l ; l = l->next) + for (lst = csp->cookie_list->first; lst ; lst = lst->next) { if (tmp) { tmp = strsav(tmp, "; "); } - tmp = strsav(tmp, l->str); + tmp = strsav(tmp, lst->str); } - for (l = csp->config->wafer_list->next; l ; l = l->next) + for (lst = csp->action->multi[ACTION_MULTI_WAFER]->first; lst ; lst = lst->next) { if (tmp) { tmp = strsav(tmp, "; "); } - if ((e = cookie_encode(l->str))) + if ((e = cookie_encode(lst->str))) { tmp = strsav(tmp, e); freez(e); @@ -1214,6 +1481,32 @@ void client_cookie_adder(struct client_state *csp) } +/********************************************************************* + * + * Function : client_accept_encoding_adder + * + * Description : Add an Accept-Encoding header to the client's request + * that disables compression if the action applies, and + * the header is not already there. Called from `sed'. + * Note: For HTTP/1.0, the absence of the header is enough. + * + * Parameters : + * 1 : csp = Current client state (buffers, headers, etc...) + * + * Returns : N/A + * + *********************************************************************/ +void client_accept_encoding_adder(struct client_state *csp) +{ + if ( ((csp->action->flags & ACTION_NO_COMPRESSION) != 0) + && (!strcmpic(csp->http->ver, "HTTP/1.1")) ) + { + enlist_unique(csp->headers, "Accept-Encoding: identity;q=1.0, *;q=0", 16); + } + +} + + /********************************************************************* * * Function : client_xtra_adder @@ -1228,12 +1521,13 @@ void client_cookie_adder(struct client_state *csp) *********************************************************************/ void client_xtra_adder(struct client_state *csp) { - struct list *l; + struct list_entry *lst; - for (l = csp->config->xtra_list->next; l ; l = l->next) + for (lst = csp->action->multi[ACTION_MULTI_ADD_HEADER]->first; + lst ; lst = lst->next) { - log_error(LOG_LEVEL_HEADER, "addh: %s", l->str); - enlist(csp->headers, l->str); + log_error(LOG_LEVEL_HEADER, "addh: %s", lst->str); + enlist(csp->headers, lst->str); } } @@ -1255,7 +1549,7 @@ void client_x_forwarded_adder(struct client_state *csp) { char *p = NULL; - if (csp->config->add_forwarded == 0) + if ((csp->action->flags & ACTION_HIDE_FORWARDED) != 0) { return; } @@ -1275,78 +1569,200 @@ void client_x_forwarded_adder(struct client_state *csp) log_error(LOG_LEVEL_HEADER, "addh: %s", p); enlist(csp->headers, p); + freez(p); } /********************************************************************* * - * Function : server_set_cookie + * Function : connection_close_adder * - * Description : Handle the server "cookie" header properly. - * Log cookie to the jar file. Then "crunch" it, - * or accept it. Called from `sed'. + * Description : Adds a "Connection: close" header to csp->headers + * as a temporary fix for the needed but missing HTTP/1.1 + * support. Called from `sed'. + * FIXME: This whole function shouldn't be neccessary! + * + * Parameters : + * 1 : csp = Current client state (buffers, headers, etc...) + * + * Returns : N/A + * + *********************************************************************/ +void connection_close_adder(struct client_state *csp) +{ + enlist(csp->headers, "Connection: close"); +} + + +/********************************************************************* + * + * Function : server_http + * + * Description : - Save the HTTP Status into csp->http->status + * - Set CT_TABOO to prevent filtering if the answer + * is a partial range (HTTP status 206) + * - Rewrite HTTP/1.1 answers to HTTP/1.0 if +downgrade + * action applies. * * Parameters : * 1 : v = parser pattern that matched this header * 2 : s = header that matched this pattern * 3 : csp = Current client state (buffers, headers, etc...) * - * Returns : `crumble' or a newly malloc'ed string. + * Returns : Copy of changed or original answer. * *********************************************************************/ -char *server_set_cookie(const struct parsers *v, char *s, struct client_state *csp) +char *server_http(const struct parsers *v, const char *s, struct client_state *csp) { -#ifdef JAR_FILES - if (csp->config->jar) + char *ret = strdup(s); + + sscanf(ret, "HTTP/%*d.%*d %d", &(csp->http->status)); + if (csp->http->status == 206) { - fprintf(csp->config->jar, "%s\t%s\n", csp->http->host, (s + v->len + 1)); + csp->content_type = CT_TABOO; } -#endif /* def JAR_FILES */ - if (!(csp->permissions & PERMIT_COOKIE_SET)) + if ((csp->action->flags & ACTION_DOWNGRADE) != 0) { - return(crumble(v, s, csp)); + ret[7] = '0'; + log_error(LOG_LEVEL_HEADER, "Downgraded answer to HTTP/1.0"); } - - return(strdup(s)); + return(ret); } -#ifdef FORCE_LOAD /********************************************************************* * - * Function : client_host + * Function : server_set_cookie * - * Description : Clean the FORCE_PREFIX out of the 'host' http - * header, if we use force + * Description : Handle the server "cookie" header properly. + * Log cookie to the jar file. Then "crunch" it, + * or accept it. Called from `sed'. * * Parameters : - * 1 : v = ignored - * 2 : s = header (from sed) to clean + * 1 : v = parser pattern that matched this header + * 2 : s = header that matched this pattern * 3 : csp = Current client state (buffers, headers, etc...) * - * Returns : A malloc'ed pointer to the cleaned host header + * Returns : `crumble' or a newly malloc'ed string. * *********************************************************************/ -char *client_host(const struct parsers *v, char *s, struct client_state *csp) +char *server_set_cookie(const struct parsers *v, const char *s, struct client_state *csp) { - char *cleanhost = strdup(s); - - if(csp->force) - strclean(cleanhost, FORCE_PREFIX); - - return(cleanhost); +#ifdef FEATURE_COOKIE_JAR + if (csp->config->jar) + { + /* + * Write timestamp into outbuf. + * + * Complex because not all OSs have tm_gmtoff or + * the %z field in strftime() + */ + char tempbuf[ BUFFER_SIZE ]; + time_t now; + struct tm *tm_now; + time (&now); + tm_now = localtime (&now); + strftime (tempbuf, BUFFER_SIZE-6, "%b %d %H:%M:%S ", tm_now); + + fprintf(csp->config->jar, "%s %s\t%s\n", tempbuf, csp->http->host, (s + v->len + 1)); + } +#endif /* def FEATURE_COOKIE_JAR */ + + if ((csp->action->flags & ACTION_NO_COOKIE_SET) != 0) + { + return(crumble(v, s, csp)); + } + else if ((csp->action->flags & ACTION_NO_COOKIE_KEEP) != 0) + { + /* Flag whether or not to log a message */ + int changed = 0; + + /* A variable to store the tag we're working on */ + char * cur_tag; + + /* Make a copy of the header we can write to */ + char * result = strdup(s); + if (result == NULL) + { + /* FIXME: This error handling is incorrect */ + return NULL; + } + + /* Skip "Set-Cookie:" (11 characters) in header */ + cur_tag = result + 11; + + /* skip whitespace between "Set-Cookie:" and value */ + while (*cur_tag && ijb_isspace(*cur_tag)) + { + cur_tag++; + } + + /* Loop through each tag in the cookie */ + while (*cur_tag) + { + /* Find next tag */ + char * next_tag = strchr(cur_tag, ';'); + if (next_tag != NULL) + { + /* Skip the ';' character itself */ + next_tag++; + + /* skip whitespace ";" and start of tag */ + while (*next_tag && ijb_isspace(*next_tag)) + { + next_tag++; + } + } + else + { + /* "Next tag" is the end of the string */ + next_tag = cur_tag + strlen(cur_tag); + } + + /* Is this the "Expires" tag? */ + if (strncmpic(cur_tag, "expires=", 8) == 0) + { + /* Delete the tag by copying the rest of the string over it. + * (Note that we cannot just use "strcpy(cur_tag, next_tag)", + * since the behaviour of strcpy is undefined for overlapping + * strings.) + */ + memmove(cur_tag, next_tag, strlen(next_tag) + 1); + + /* That changed the header, need to issue a log message */ + changed = 1; + + /* Note that the next tag has now been moved to *cur_tag, + * so we do not need to update the cur_tag pointer. + */ + } + else + { + /* Move on to next cookie tag */ + cur_tag = next_tag; + } + } + + if (changed) + { + log_error(LOG_LEVEL_HEADER, "Changed cookie to a temporary one."); + } + + return result; + } + + return(strdup(s)); } -#endif /* def FORCE_LOAD */ - - -#ifdef FORCE_LOAD + + +#ifdef FEATURE_FORCE_LOAD /********************************************************************* * * Function : strclean * - * Description : In-Situ-Eliminate all occurances of substring in + * Description : In-Situ-Eliminate all occurances of substring in * string * * Parameters : @@ -1366,7 +1782,7 @@ int strclean(const char *string, const char *substring) p = pos + len; do { - *(p - len) = *p; + *(p - len) = *p; } while (*p++ != '\0'); @@ -1375,7 +1791,7 @@ int strclean(const char *string, const char *substring) return(hits); } -#endif /* def FORCE_LOAD */ +#endif /* def FEATURE_FORCE_LOAD */ /*