X-Git-Url: http://www.privoxy.org/gitweb/?p=privoxy.git;a=blobdiff_plain;f=cgi.c;h=ea63193e536747b468719477dc50a3e0e01f81b3;hp=84f8f9db67768e22cdc1d68372116f1ce060a2b5;hb=98a2b4811d4316ff8797495a803fa0d712d25176;hpb=57051e10b50bc3321e76c1999fd5a7e1177489d1 diff --git a/cgi.c b/cgi.c index 84f8f9db..ea63193e 100644 --- a/cgi.c +++ b/cgi.c @@ -1,4 +1,4 @@ -const char cgi_rcs[] = "$Id: cgi.c,v 1.19 2001/09/13 23:31:25 jongfoster Exp $"; +const char cgi_rcs[] = "$Id: cgi.c,v 1.24 2001/09/16 11:38:01 jongfoster Exp $"; /********************************************************************* * * File : $Source: /cvsroot/ijbswa/current/cgi.c,v $ @@ -36,6 +36,27 @@ const char cgi_rcs[] = "$Id: cgi.c,v 1.19 2001/09/13 23:31:25 jongfoster Exp $"; * * Revisions : * $Log: cgi.c,v $ + * Revision 1.24 2001/09/16 11:38:01 jongfoster + * Splitting fill_template() into 2 functions: + * template_load() loads the file + * template_fill() performs the PCRS regexps. + * This is because the CGI edit interface has a "table row" + * template which is used many times in the page - this + * change means it's only loaded from disk once. + * + * Revision 1.23 2001/09/16 11:16:05 jongfoster + * Better error handling in dispatch_cgi() and parse_cgi_parameters() + * + * Revision 1.22 2001/09/16 11:00:10 jongfoster + * New function alloc_http_response, for symmetry with free_http_response + * + * Revision 1.21 2001/09/13 23:53:03 jongfoster + * Support for both static and dynamically generated CGI pages. + * Correctly setting Last-Modified: and Expires: HTTP headers. + * + * Revision 1.20 2001/09/13 23:40:36 jongfoster + * (Cosmetic only) Indentation correction + * * Revision 1.19 2001/09/13 23:31:25 jongfoster * Moving image data to cgi.c rather than cgi.h. * @@ -161,6 +182,7 @@ const char cgi_rcs[] = "$Id: cgi.c,v 1.19 2001/09/13 23:31:25 jongfoster Exp $"; #include #include #include +#include #ifdef _WIN32 #define snprintf _snprintf @@ -182,6 +204,9 @@ const char cgi_rcs[] = "$Id: cgi.c,v 1.19 2001/09/13 23:31:25 jongfoster Exp $"; const char cgi_h_rcs[] = CGI_H_VERSION; const struct cgi_dispatcher cgi_dispatcher[] = { + { "robots.txt", + 10, cgi_robots_txt, + "HIDE Sends a robots.txt file to tell robots to go away." }, { "show-status", 11, cgi_show_status, "Show information about the current configuration" }, @@ -281,7 +306,7 @@ struct http_response *dispatch_cgi(struct client_state *csp) */ /* Get mem for response or fail*/ - if (NULL == ( rsp = zalloc(sizeof(*rsp)))) + if (NULL == (rsp = alloc_http_response())) { return NULL; } @@ -301,10 +326,18 @@ struct http_response *dispatch_cgi(struct client_state *csp) { if (strncmp(argstring, d->name, d->name_length) == 0) { - param_list = parse_cgi_parameters(argstring + d->name_length); + if (NULL == (param_list = + parse_cgi_parameters(argstring + d->name_length))) + { + free_map(param_list); + free_http_response(rsp); + return(NULL); + } if ((d->handler)(csp, rsp, param_list)) { - freez(rsp); + free_map(param_list); + free_http_response(rsp); + return(NULL); } free_map(param_list); @@ -313,9 +346,8 @@ struct http_response *dispatch_cgi(struct client_state *csp) } /* Can't get here, since cgi_default will match all requests */ - freez(rsp); + free_http_response(rsp); return(NULL); - } @@ -329,7 +361,7 @@ struct http_response *dispatch_cgi(struct client_state *csp) * Parameters : * 1 : string = string to be parsed * - * Returns : pointer to param list + * Returns : pointer to param list, or NULL if out of memory. * *********************************************************************/ struct map *parse_cgi_parameters(char *argstring) @@ -337,13 +369,22 @@ struct map *parse_cgi_parameters(char *argstring) char *tmp, *p; char *vector[BUFFER_SIZE]; int pairs, i; - struct map *cgi_params = new_map(); + struct map *cgi_params; + + if (NULL == (cgi_params = new_map())) + { + return NULL; + } if(*argstring == '?') { argstring++; } - tmp = strdup(argstring); + if (NULL == (tmp = strdup(argstring))) + { + free_map(cgi_params); + return NULL; + } pairs = ssplit(tmp, "&", vector, SZ(vector), 1, 1); @@ -399,7 +440,8 @@ int cgi_default(struct client_state *csp, struct http_response *rsp, map(exports, "cgi-parameters", 1, "", 1); } - rsp->body = fill_template(csp, "default", exports); + rsp->body = template_load(csp, "default"); + template_fill(&rsp->body, exports); free_map(exports); return(0); @@ -439,6 +481,7 @@ int cgi_send_banner(struct client_state *csp, struct http_response *rsp, } enlist(rsp->headers, "Content-Type: image/gif"); + rsp->is_static = 1; return(0); @@ -471,7 +514,8 @@ int cgi_show_version(struct client_state *csp, struct http_response *rsp, map(exports, "sourceversions", 1, show_rcs(), 0); - rsp->body = fill_template(csp, "show-version", exports); + rsp->body = template_load(csp, "show-version"); + template_fill(&rsp->body, exports); free_map(exports); return(0); @@ -508,6 +552,11 @@ int cgi_show_status(struct client_state *csp, struct http_response *rsp, char * p; const char * filename = NULL; char * file_description = NULL; +#ifdef FEATURE_STATISTICS + float perc_rej; /* Percentage of http requests rejected */ + int local_urls_read; + int local_urls_rejected; +#endif /* ndef FEATURE_STATISTICS */ struct map * exports = default_exports(csp, "show-status"); @@ -564,7 +613,8 @@ int cgi_show_status(struct client_state *csp, struct http_response *rsp, fclose(fp); map(exports, "contents", 1, s, 0); } - rsp->body = fill_template(csp, "show-status-file", exports); + rsp->body = template_load(csp, "show-status-file"); + template_fill(&rsp->body, exports); free_map(exports); return(0); @@ -584,7 +634,40 @@ int cgi_show_status(struct client_state *csp, struct http_response *rsp, show_defines(exports); #ifdef FEATURE_STATISTICS - add_stats(exports); + local_urls_read = urls_read; + local_urls_rejected = urls_rejected; + + /* + * Need to alter the stats not to include the fetch of this + * page. + * + * Can't do following thread safely! doh! + * + * urls_read--; + * urls_rejected--; * This will be incremented subsequently * + */ + + if (local_urls_read == 0) + { + map_block_killer(exports, "have-stats"); + } + else + { + map_block_killer(exports, "have-no-stats"); + + perc_rej = (float)local_urls_rejected * 100.0F / + (float)local_urls_read; + + sprintf(buf, "%d", local_urls_read); + map(exports, "requests-received", 1, buf, 1); + + sprintf(buf, "%d", local_urls_rejected); + map(exports, "requests-blocked", 1, buf, 1); + + sprintf(buf, "%6.2f", perc_rej); + map(exports, "percent-blocked", 1, buf, 1); + } + #else /* ndef FEATURE_STATISTICS */ map_block_killer(exports, "statistics"); #endif /* ndef FEATURE_STATISTICS */ @@ -620,14 +703,15 @@ int cgi_show_status(struct client_state *csp, struct http_response *rsp, map_block_killer(exports, "trust-support"); #endif /* ndef FEATURE_TRUST */ - rsp->body = fill_template(csp, "show-status", exports); + rsp->body = template_load(csp, "show-status"); + template_fill(&rsp->body, exports); free_map(exports); return(0); } - /********************************************************************* +/********************************************************************* * * Function : cgi_show_url_info * @@ -692,7 +776,8 @@ int cgi_show_url_info(struct client_state *csp, struct http_response *rsp, freez(url_param); free_current_action(action); - rsp->body = fill_template(csp, "show-url-info", exports); + rsp->body = template_load(csp, "show-url-info"); + template_fill(&rsp->body, exports); free_map(exports); return 0; @@ -728,7 +813,8 @@ int cgi_show_url_info(struct client_state *csp, struct http_response *rsp, freez(path); free_current_action(action); - rsp->body = fill_template(csp, "show-url-info", exports); + rsp->body = template_load(csp, "show-url-info"); + template_fill(&rsp->body, exports); free_map(exports); return 0; @@ -786,7 +872,8 @@ int cgi_show_url_info(struct client_state *csp, struct http_response *rsp, free_current_action(action); } - rsp->body = fill_template(csp, "show-url-info", exports); + rsp->body = template_load(csp, "show-url-info"); + template_fill(&rsp->body, exports); free_map(exports); return 0; @@ -813,10 +900,10 @@ struct http_response *error_response(struct client_state *csp, const char *templ struct http_response *rsp; struct map * exports = default_exports(csp, NULL); - if (NULL == ( rsp = (struct http_response *)zalloc(sizeof(*rsp)))) + if (NULL == (rsp = alloc_http_response())) { return NULL; - } + } map(exports, "host-html", 1, html_encode(csp->http->host), 0); map(exports, "hostport", 1, csp->http->hostport, 1); @@ -826,9 +913,10 @@ struct http_response *error_response(struct client_state *csp, const char *templ map(exports, "error", 1, safe_strerror(err), 0); map(exports, "host-ip", 1, csp->http->host_ip_addr_str, 1); - rsp->body = fill_template(csp, templatename, exports); + rsp->body = template_load(csp, templatename); + template_fill(&rsp->body, exports); free_map(exports); - + if (!strcmp(templatename, "no-such-domain")) { rsp->status = strdup("404 No such domain"); @@ -842,6 +930,65 @@ struct http_response *error_response(struct client_state *csp, const char *templ } +/********************************************************************* + * + * Function : get_http_time + * + * Description : Get the time in a format suitable for use in a + * HTTP header - e.g.: + * "Sun, 06 Nov 1994 08:49:37 GMT" + * + * Parameters : + * 1 : time_offset = Time returned will be current time + * plus this number of seconds. + * 2 : buf = Destination for result. Must be long enough + * to hold 29 characters plus a trailing zero. + * + * Returns : N/A + * + *********************************************************************/ +static void get_http_time(int time_offset, char * buf) +{ + static const char day_names[7][4] = + { "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat" }; + static const char month_names[12][4] = + { "Jan", "Feb", "Mar", "Apr", "May", "Jun", + "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" }; + + struct tm *t; + time_t current_time; + + assert(buf); + + time(¤t_time); /* get current time */ + +/* FIXME: is this needed? time() can't fail on Win32. What about Linux? + if(current_time <= 0) + { + return NULL; + } +*/ + + current_time += time_offset; + + /* get and save the gmt */ + t = gmtime(¤t_time); + + /* Format: "Sun, 06 Nov 1994 08:49:37 GMT" */ + snprintf(buf, 30, + "%s, %02d %s %4d %02d:%02d:%02d GMT", + day_names[t->tm_wday], + t->tm_mday, + month_names[t->tm_mon], + t->tm_year + 1900, + t->tm_hour, + t->tm_min, + t->tm_sec + ); + buf[32] = '\0'; +} + + /********************************************************************* * * Function : finish_http_response @@ -876,21 +1023,48 @@ struct http_response *finish_http_response(struct http_response *rsp) enlist(rsp->headers, buf); /* - * Fill in the default headers FIXME: Are these correct? sequence OK? check rfc! - * FIXME: Should have: - * "JunkBuster" GIF: Last-Modified: any *fixed* date in the past (as now). - * Expires: 5 minutes after the time when reply sent - * CGI, "blocked", & all other requests: - * Last-Modified: Time when reply sent - * Expires: Time when reply sent - * "Cache-Control: no-cache" - * + * Fill in the default headers: + * + * Content-Type: default to text/html if not already specified. + * Date: set to current date/time. + * Last-Modified: set to date/time the page was last changed. + * Expires: set to date/time page next needs reloading. + * Cache-Control: set to "no-cache" if applicable. + * * See http://www.w3.org/Protocols/rfc2068/rfc2068 */ - enlist_unique(rsp->headers, "Last-Modified: Thu Jul 31, 1997 07:42:22 pm GMT", 14); - enlist_unique(rsp->headers, "Expires: Thu Jul 31, 1997 07:42:22 pm GMT", 8); enlist_unique(rsp->headers, "Content-Type: text/html", 13); + if (rsp->is_static) + { + /* + * Set Expires to about 10 min into the future so it'll get reloaded + * occasionally, e.g. if IJB gets upgraded. + */ + + get_http_time(0, buf); + enlist_unique_header(rsp->headers, "Date", buf); + + /* Some date in the past. */ + enlist_unique_header(rsp->headers, "Last-Modified", "Sat, 17 Jun 2000 12:00:00 GMT"); + + get_http_time(10 * 60, buf); /* 10 * 60sec = 10 minutes */ + enlist_unique_header(rsp->headers, "Expires", buf); + } + else + { + /* + * Compliant browsers should not cache this due to the "Cache-Control" + * setting. However, to be certain, we also set both "Last-Modified" + * and "Expires" to the current time. + */ + enlist_unique_header(rsp->headers, "Cache-Control", "no-cache"); + get_http_time(0, buf); + enlist_unique_header(rsp->headers, "Date", buf); + enlist_unique_header(rsp->headers, "Last-Modified", buf); + enlist_unique_header(rsp->headers, "Expires", buf); + } + /* * Write the head @@ -907,6 +1081,23 @@ struct http_response *finish_http_response(struct http_response *rsp) } +/********************************************************************* + * + * Function : alloc_http_response + * + * Description : Allocates a new http_response structure. + * + * Parameters : N/A + * + * Returns : pointer to a new http_response, or NULL. + * + *********************************************************************/ +struct http_response * alloc_http_response(void) +{ + return (struct http_response *) zalloc(sizeof(struct http_response)); +} + + /********************************************************************* * * Function : free_http_response @@ -922,13 +1113,13 @@ struct http_response *finish_http_response(struct http_response *rsp) *********************************************************************/ void free_http_response(struct http_response *rsp) { - if(rsp) + if (rsp) { freez(rsp->status); freez(rsp->head); freez(rsp->body); destroy_list(rsp->headers); - freez(rsp); + free(rsp); } } @@ -946,23 +1137,16 @@ void free_http_response(struct http_response *rsp) * Parameters : * 1 : csp = Current client state (buffers, headers, etc...) * 3 : template = name of the HTML template to be used - * 2 : exports = map with fill in symbol -> name pairs * - * Returns : char * with filled out form, or NULL if failiure + * Returns : char * with loaded template, or NULL if failure * *********************************************************************/ -char *fill_template(struct client_state *csp, const char *templatename, struct map *exports) +char *template_load(struct client_state *csp, const char *templatename) { - struct map_entry *m; - pcrs_job *job; char buf[BUFFER_SIZE]; - char *tmp_out_buffer; char *file_buffer = NULL; - int size; - int error; FILE *fp; - /* * Open template file or fail */ @@ -981,16 +1165,54 @@ char *fill_template(struct client_state *csp, const char *templatename, struct m while (fgets(buf, BUFFER_SIZE, fp)) { /* skip lines starting with '#' */ - if(*buf == '#') continue; + if(*buf == '#') + { + continue; + } file_buffer = strsav(file_buffer, buf); } fclose(fp); + return(file_buffer); +} - /* - * Execute the jobs - */ + +/********************************************************************* + * + * Function : fill_template + * + * Description : CGI support function that loads a given HTML + * template from the confdir, and fills it in + * by replacing @name@ with value using pcrs, + * for each item in the output map. + * + * Parameters : + * 1 : template_ptr = IN: Template to be filled out. + * Will be free()d. + * OUT: Filled out template. + * Caller must free(). + * 2 : exports = map with fill in symbol -> name pairs + * + * Returns : N/A + * + *********************************************************************/ +void template_fill(char ** template_ptr, struct map *exports) +{ + struct map_entry *m; + pcrs_job *job; + char buf[BUFFER_SIZE]; + char *tmp_out_buffer; + char *file_buffer; + int size; + int error; + const char * flags; + + assert(template_ptr); + assert(*template_ptr); + assert(exports); + + file_buffer = *template_ptr; size = strlen(file_buffer) + 1; /* @@ -998,11 +1220,34 @@ char *fill_template(struct client_state *csp, const char *templatename, struct m */ for (m = exports->first; m != NULL; m = m->next) { - /* Enclose name in @@ */ - snprintf(buf, BUFFER_SIZE, "@%s@", m->name); + if (*m->name == '$') + { + /* + * First character of name is '$', so remove this flag + * character and allow backreferences ($1 etc) in the + * "replace with" text. + */ + snprintf(buf, BUFFER_SIZE, "%s", m->name + 1); + flags = "sigU"; + } + else + { + /* + * Treat the "replace with" text as a literal string - + * no quoting needed, no backreferences allowed. + * ("Trivial" ['T'] flag). + */ + flags = "sigTU"; + + /* Enclose name in @@ */ + snprintf(buf, BUFFER_SIZE, "@%s@", m->name); + } + + + log_error(LOG_LEVEL_CGI, "Substituting: s/%s/%s/%s", buf, m->value, flags); /* Make and run job. */ - job = pcrs_compile(buf, m->value, "sigTU", &error); + job = pcrs_compile(buf, m->value, flags, &error); if (job == NULL) { log_error(LOG_LEVEL_ERROR, "Error compiling template fill job %s: %d", m->name, error); @@ -1019,12 +1264,10 @@ char *fill_template(struct client_state *csp, const char *templatename, struct m } } - /* * Return */ - return(file_buffer); - + *template_ptr = file_buffer; } @@ -1218,61 +1461,46 @@ char *dump_map(const struct map *the_map) } -#ifdef FEATURE_STATISTICS /********************************************************************* * - * Function : add_stats + * Function : cgi_robots_txt * - * Description : Add the blocking statistics to a given map. + * Description : CGI function to return "/robots.txt". * * Parameters : - * 1 : exports = map to write to. + * 1 : csp = Current client state (buffers, headers, etc...) + * 2 : rsp = http_response data structure for output + * 3 : parameters = map of cgi parameters * - * Returns : pointer to extended map + * CGI Parameters : None + * + * Returns : 0 * *********************************************************************/ -struct map *add_stats(struct map *exports) +int cgi_robots_txt(struct client_state *csp, struct http_response *rsp, + struct map *parameters) { - float perc_rej; /* Percentage of http requests rejected */ - char buf[1000]; - int local_urls_read = urls_read; - int local_urls_rejected = urls_rejected; + char buf[100]; - /* - * Need to alter the stats not to include the fetch of this - * page. - * - * Can't do following thread safely! doh! - * - * urls_read--; - * urls_rejected--; * This will be incremented subsequently * - */ - - if (local_urls_read == 0) - { - map_block_killer(exports, "have-stats"); - } - else - { - map_block_killer(exports, "have-no-stats"); - - perc_rej = (float)local_urls_rejected * 100.0F / - (float)local_urls_read; + rsp->body = strdup( + "# This is the Internet Junkbuster control interface.\n" + "# It isn't very useful to index it, and you're likely to break stuff.\n" + "# So go away!\n" + "\n" + "User-agent: *\n" + "Disallow: /\n" + "\n"); - sprintf(buf, "%d", local_urls_read); - map(exports, "requests-received", 1, buf, 1); + enlist_unique(rsp->headers, "Content-Type: text/plain", 13); - sprintf(buf, "%d", local_urls_rejected); - map(exports, "requests-blocked", 1, buf, 1); + rsp->is_static = 1; - sprintf(buf, "%6.2f", perc_rej); - map(exports, "percent-blocked", 1, buf, 1); - } - - return(exports); + get_http_time(7 * 24 * 60 * 60, buf); /* 7 days into future */ + enlist_unique_header(rsp->headers, "Expires", buf); + return 0; } -#endif /* def FEATURE_STATISTICS */ + /* Local Variables: