X-Git-Url: http://www.privoxy.org/gitweb/?p=privoxy.git;a=blobdiff_plain;f=urlmatch.c;h=c731ec3b1336f20e5065c908f96c4bcf1cebbb73;hp=d888e37dbf0ed86e3889218f647488e80e703fb2;hb=c88024b1f6a64ecc2d923ff875e10f15b64daff3;hpb=e3c7acd6fdb41505246948ee9cb09f9b4915e027 diff --git a/urlmatch.c b/urlmatch.c index d888e37d..c731ec3b 100644 --- a/urlmatch.c +++ b/urlmatch.c @@ -1,4 +1,4 @@ -const char urlmatch_rcs[] = "$Id: urlmatch.c,v 1.51 2009/05/16 13:27:20 fabiankeil Exp $"; +const char urlmatch_rcs[] = "$Id: urlmatch.c,v 1.74 2012/12/07 12:49:20 fabiankeil Exp $"; /********************************************************************* * * File : $Source: /cvsroot/ijbswa/current/urlmatch.c,v $ @@ -6,7 +6,7 @@ const char urlmatch_rcs[] = "$Id: urlmatch.c,v 1.51 2009/05/16 13:27:20 fabianke * Purpose : Declares functions to match URLs against URL * patterns. * - * Copyright : Written by and Copyright (C) 2001-2009 + * Copyright : Written by and Copyright (C) 2001-2011 * the Privoxy team. http://www.privoxy.org/ * * Based on the Internet Junkbuster originally written @@ -58,7 +58,13 @@ const char urlmatch_rcs[] = "$Id: urlmatch.c,v 1.51 2009/05/16 13:27:20 fabianke const char urlmatch_h_rcs[] = URLMATCH_H_VERSION; -enum regex_anchoring {NO_ANCHORING, LEFT_ANCHORED, RIGHT_ANCHORED}; +enum regex_anchoring +{ + NO_ANCHORING, + LEFT_ANCHORED, + RIGHT_ANCHORED, + RIGHT_ANCHORED_HOST +}; static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern); /********************************************************************* @@ -86,12 +92,15 @@ void free_http_request(struct http_request *http) freez(http->path); freez(http->ver); freez(http->host_ip_addr_str); +#ifndef FEATURE_EXTENDED_HOST_PATTERNS freez(http->dbuffer); freez(http->dvec); http->dcount = 0; +#endif } +#ifndef FEATURE_EXTENDED_HOST_PATTERNS /********************************************************************* * * Function : init_domain_components @@ -106,7 +115,6 @@ void free_http_request(struct http_request *http) * 1 : http = pointer to the http structure to hold elements. * * Returns : JB_ERR_OK on success - * JB_ERR_MEMORY on out of memory * JB_ERR_PARSE on malformed command/URL * or >100 domains deep. * @@ -117,20 +125,16 @@ jb_err init_domain_components(struct http_request *http) size_t size; char *p; - http->dbuffer = strdup(http->host); - if (NULL == http->dbuffer) - { - return JB_ERR_MEMORY; - } + http->dbuffer = strdup_or_die(http->host); /* map to lower case */ for (p = http->dbuffer; *p ; p++) { - *p = (char)tolower((int)(unsigned char)*p); + *p = (char)privoxy_tolower(*p); } /* split the domain name into components */ - http->dcount = ssplit(http->dbuffer, ".", vec, SZ(vec), 1, 1); + http->dcount = ssplit(http->dbuffer, ".", vec, SZ(vec)); if (http->dcount <= 0) { @@ -145,16 +149,60 @@ jb_err init_domain_components(struct http_request *http) /* save a copy of the pointers in dvec */ size = (size_t)http->dcount * sizeof(*http->dvec); - http->dvec = (char **)malloc(size); - if (NULL == http->dvec) - { - return JB_ERR_MEMORY; - } + http->dvec = malloc_or_die(size); memcpy(http->dvec, vec, size); return JB_ERR_OK; } +#endif /* ndef FEATURE_EXTENDED_HOST_PATTERNS */ + + +/********************************************************************* + * + * Function : url_requires_percent_encoding + * + * Description : Checks if an URL contains invalid characters + * according to RFC 3986 that should be percent-encoded. + * Does not verify whether or not the passed string + * actually is a valid URL. + * + * Parameters : + * 1 : url = URL to check + * + * Returns : True in case of valid URLs, false otherwise + * + *********************************************************************/ +int url_requires_percent_encoding(const char *url) +{ + static const char allowed_characters[128] = { + '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', + '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', + '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', + '\0', '\0', '\0', '!', '\0', '#', '$', '%', '&', '\'', + '(', ')', '*', '+', ',', '-', '.', '/', '0', '1', + '2', '3', '4', '5', '6', '7', '8', '9', ':', ';', + '\0', '=', '\0', '?', '@', 'A', 'B', 'C', 'D', 'E', + 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', + 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', + 'Z', '[', '\0', ']', '\0', '_', '\0', 'a', 'b', 'c', + 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', + 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', + 'x', 'y', 'z', '\0', '\0', '\0', '~', '\0' + }; + + while (*url != '\0') + { + const unsigned int i = (unsigned char)*url++; + if (i >= sizeof(allowed_characters) || '\0' == allowed_characters[i]) + { + return TRUE; + } + } + + return FALSE; + +} /********************************************************************* @@ -172,7 +220,6 @@ jb_err init_domain_components(struct http_request *http) * protocol are acceptable. * * Returns : JB_ERR_OK on success - * JB_ERR_MEMORY on out of memory * JB_ERR_PARSE on malformed command/URL * or >100 domains deep. * @@ -184,23 +231,15 @@ jb_err parse_http_url(const char *url, struct http_request *http, int require_pr /* * Save our initial URL */ - http->url = strdup(url); - if (http->url == NULL) - { - return JB_ERR_MEMORY; - } - + http->url = strdup_or_die(url); /* * Check for * URI. If found, we're done. - */ + */ if (*http->url == '*') { - if ( NULL == (http->path = strdup("*")) - || NULL == (http->hostport = strdup("")) ) - { - return JB_ERR_MEMORY; - } + http->path = strdup_or_die("*"); + http->hostport = strdup_or_die(""); if (http->url[1] != '\0') { return JB_ERR_PARSE; @@ -217,11 +256,7 @@ jb_err parse_http_url(const char *url, struct http_request *http, int require_pr char *url_noproto; char *url_path; - buf = strdup(url); - if (buf == NULL) - { - return JB_ERR_MEMORY; - } + buf = strdup_or_die(url); /* Find the start of the URL in our scratch space */ url_noproto = buf; @@ -264,9 +299,9 @@ jb_err parse_http_url(const char *url, struct http_request *http, int require_pr * https URL in and it's parsed by the function. (When the * URL is actually retrieved, SSL hides the path part). */ - http->path = strdup(http->ssl ? "/" : url_path); + http->path = strdup_or_die(http->ssl ? "/" : url_path); *url_path = '\0'; - http->hostport = strdup(url_noproto); + http->hostport = strdup_or_die(url_noproto); } else { @@ -274,17 +309,11 @@ jb_err parse_http_url(const char *url, struct http_request *http, int require_pr * Repair broken HTTP requests that don't contain a path, * or CONNECT requests */ - http->path = strdup("/"); - http->hostport = strdup(url_noproto); + http->path = strdup_or_die("/"); + http->hostport = strdup_or_die(url_noproto); } freez(buf); - - if ( (http->path == NULL) - || (http->hostport == NULL)) - { - return JB_ERR_MEMORY; - } } if (!host_available) @@ -301,11 +330,7 @@ jb_err parse_http_url(const char *url, struct http_request *http, int require_pr char *host; char *port; - buf = strdup(http->hostport); - if (buf == NULL) - { - return JB_ERR_MEMORY; - } + buf = strdup_or_die(http->hostport); /* check if url contains username and/or password */ host = strchr(buf, '@'); @@ -357,9 +382,18 @@ jb_err parse_http_url(const char *url, struct http_request *http, int require_pr if (port != NULL) { /* Contains port */ + char *endptr; + long parsed_port; /* Terminate hostname and point to start of port string */ *port++ = '\0'; - http->port = atoi(port); + parsed_port = strtol(port, &endptr, 10); + if ((parsed_port <= 0) || (parsed_port > 65535) || (*endptr != '\0')) + { + log_error(LOG_LEVEL_ERROR, "Invalid port in URL: %s.", url); + freez(buf); + return JB_ERR_PARSE; + } + http->port = (int)parsed_port; } else { @@ -367,20 +401,17 @@ jb_err parse_http_url(const char *url, struct http_request *http, int require_pr http->port = (http->ssl ? 443 : 80); } - http->host = strdup(host); + http->host = strdup_or_die(host); freez(buf); - - if (http->host == NULL) - { - return JB_ERR_MEMORY; - } } - /* - * Split domain name so we can compare it against wildcards - */ +#ifdef FEATURE_EXTENDED_HOST_PATTERNS + return JB_ERR_OK; +#else + /* Split domain name so we can compare it against wildcards */ return init_domain_components(http); +#endif /* def FEATURE_EXTENDED_HOST_PATTERNS */ } @@ -399,7 +430,7 @@ jb_err parse_http_url(const char *url, struct http_request *http, int require_pr *********************************************************************/ static int unknown_method(const char *method) { - static const char *known_http_methods[] = { + static const char * const known_http_methods[] = { /* Basic HTTP request type */ "GET", "HEAD", "POST", "PUT", "DELETE", "OPTIONS", "TRACE", "CONNECT", /* webDAV extensions (RFC2518) */ @@ -408,14 +439,14 @@ static int unknown_method(const char *method) * Microsoft webDAV extension for Exchange 2000. See: * http://lists.w3.org/Archives/Public/w3c-dist-auth/2002JanMar/0001.html * http://msdn.microsoft.com/library/en-us/wss/wss/_webdav_methods.asp - */ + */ "BCOPY", "BMOVE", "BDELETE", "BPROPFIND", "BPROPPATCH", /* * Another Microsoft webDAV extension for Exchange 2000. See: * http://systems.cs.colorado.edu/grunwald/MobileComputing/Papers/draft-cohen-gena-p-base-00.txt * http://lists.w3.org/Archives/Public/w3c-dist-auth/2002JanMar/0001.html * http://msdn.microsoft.com/library/en-us/wss/wss/_webdav_methods.asp - */ + */ "SUBSCRIBE", "UNSUBSCRIBE", "NOTIFY", "POLL", /* * Yet another WebDAV extension, this time for @@ -451,7 +482,6 @@ static int unknown_method(const char *method) * 2 : http = pointer to the http structure to hold elements * * Returns : JB_ERR_OK on success - * JB_ERR_MEMORY on out of memory * JB_ERR_CGI_PARAMS on malformed command/URL * or >100 domains deep. * @@ -465,13 +495,9 @@ jb_err parse_http_request(const char *req, struct http_request *http) memset(http, '\0', sizeof(*http)); - buf = strdup(req); - if (buf == NULL) - { - return JB_ERR_MEMORY; - } + buf = strdup_or_die(req); - n = ssplit(buf, " \r\n", v, SZ(v), 1, 1); + n = ssplit(buf, " \r\n", v, SZ(v)); if (n != 3) { freez(buf); @@ -514,19 +540,12 @@ jb_err parse_http_request(const char *req, struct http_request *http) /* * Copy the details into the structure */ - http->cmd = strdup(req); - http->gpc = strdup(v[0]); - http->ver = strdup(v[2]); + http->cmd = strdup_or_die(req); + http->gpc = strdup_or_die(v[0]); + http->ver = strdup_or_die(v[2]); freez(buf); - if ( (http->cmd == NULL) - || (http->gpc == NULL) - || (http->ver == NULL) ) - { - return JB_ERR_MEMORY; - } - return JB_ERR_OK; } @@ -540,9 +559,10 @@ jb_err parse_http_request(const char *req, struct http_request *http) * * Parameters : * 1 : pattern = The pattern to compile. - * 2 : anchoring = How the regex should be anchored. - * Can be either one of NO_ANCHORING, - * LEFT_ANCHORED or RIGHT_ANCHORED. + * 2 : anchoring = How the regex should be modified + * before compilation. Can be either + * one of NO_ANCHORING, LEFT_ANCHORED, + * RIGHT_ANCHORED or RIGHT_ANCHORED_HOST. * 3 : url = In case of failures, the spec member is * logged and the structure freed. * 4 : regex = Where the compiled regex should be stored. @@ -576,6 +596,9 @@ static jb_err compile_pattern(const char *pattern, enum regex_anchoring anchorin case RIGHT_ANCHORED: fmt = "%s$"; break; + case RIGHT_ANCHORED_HOST: + fmt = "%s\\.?$"; + break; case LEFT_ANCHORED: fmt = "^%s"; break; @@ -686,11 +709,7 @@ static jb_err compile_url_pattern(struct url_spec *url, char *buf) if (NULL != p) { *p++ = '\0'; - url->port_list = strdup(p); - if (NULL == url->port_list) - { - return JB_ERR_MEMORY; - } + url->port_list = strdup_or_die(p); } else { @@ -712,7 +731,7 @@ static jb_err compile_url_pattern(struct url_spec *url, char *buf) * * Function : compile_host_pattern * - * Description : Parses and compiles a host pattern.. + * Description : Parses and compiles a host pattern. * * Parameters : * 1 : url = Target url_spec to be filled in. @@ -725,7 +744,7 @@ static jb_err compile_url_pattern(struct url_spec *url, char *buf) *********************************************************************/ static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern) { - return compile_pattern(host_pattern, RIGHT_ANCHORED, url, &url->host_regex); + return compile_pattern(host_pattern, RIGHT_ANCHORED_HOST, url, &url->host_regex); } #else @@ -741,7 +760,6 @@ static jb_err compile_host_pattern(struct url_spec *url, const char *host_patter * 2 : host_pattern = Host pattern to parse. * * Returns : JB_ERR_OK - Success - * JB_ERR_MEMORY - Out of memory * JB_ERR_PARSE - Cannot parse regex * *********************************************************************/ @@ -763,47 +781,37 @@ static jb_err compile_host_pattern(struct url_spec *url, const char *host_patter url->unanchored |= ANCHOR_LEFT; } - /* + /* * Split domain into components */ - url->dbuffer = strdup(host_pattern); - if (NULL == url->dbuffer) - { - free_url_spec(url); - return JB_ERR_MEMORY; - } + url->dbuffer = strdup_or_die(host_pattern); - /* + /* * Map to lower case */ for (p = url->dbuffer; *p ; p++) { - *p = (char)tolower((int)(unsigned char)*p); + *p = (char)privoxy_tolower(*p); } - /* + /* * Split the domain name into components */ - url->dcount = ssplit(url->dbuffer, ".", v, SZ(v), 1, 1); + url->dcount = ssplit(url->dbuffer, ".", v, SZ(v)); if (url->dcount < 0) { free_url_spec(url); - return JB_ERR_MEMORY; + return JB_ERR_PARSE; } else if (url->dcount != 0) { - /* + /* * Save a copy of the pointers in dvec */ size = (size_t)url->dcount * sizeof(*url->dvec); - - url->dvec = (char **)malloc(size); - if (NULL == url->dvec) - { - free_url_spec(url); - return JB_ERR_MEMORY; - } + + url->dvec = malloc_or_die(size); memcpy(url->dvec, v, size); } @@ -836,13 +844,13 @@ static int simplematch(const char *pattern, const char *text) { const unsigned char *pat = (const unsigned char *)pattern; const unsigned char *txt = (const unsigned char *)text; - const unsigned char *fallback = pat; + const unsigned char *fallback = pat; int wildcard = 0; - + unsigned char lastchar = 'a'; unsigned i; unsigned char charmap[32]; - + while (*txt) { @@ -860,15 +868,15 @@ static int simplematch(const char *pattern, const char *text) } /* '*' in the pattern? */ - if (*pat == '*') + if (*pat == '*') { - + /* The pattern ends afterwards? Speed up the return. */ if (*++pat == '\0') { return 0; } - + /* Else, set wildcard mode and remember position after '*' */ wildcard = 1; fallback = pat; @@ -882,7 +890,7 @@ static int simplematch(const char *pattern, const char *text) while (*++pat != ']') { if (!*pat) - { + { return 1; } else if (*pat == '-') @@ -894,7 +902,7 @@ static int simplematch(const char *pattern, const char *text) for (i = lastchar; i <= *pat; i++) { charmap[i / 8] |= (unsigned char)(1 << (i % 8)); - } + } } else { @@ -905,21 +913,21 @@ static int simplematch(const char *pattern, const char *text) } /* -END- if Character range specification */ - /* - * Char match, or char range match? + /* + * Char match, or char range match? */ - if ( (*pat == *txt) - || (*pat == '?') - || ((*pat == ']') && (charmap[*txt / 8] & (1 << (*txt % 8)))) ) + if ((*pat == *txt) + || (*pat == '?') + || ((*pat == ']') && (charmap[*txt / 8] & (1 << (*txt % 8))))) { - /* - * Sucess: Go ahead + /* + * Success: Go ahead */ pat++; } else if (!wildcard) { - /* + /* * No match && no wildcard: No luck */ return 1; @@ -946,7 +954,7 @@ static int simplematch(const char *pattern, const char *text) } /* Cut off extra '*'s */ - if(*pat == '*') pat++; + if (*pat == '*') pat++; /* If this is the pattern's end, fine! */ return(*pat); @@ -1092,7 +1100,6 @@ static int domain_match(const struct url_spec *pattern, const struct http_reques * are lost forever. * * Returns : JB_ERR_OK - Success - * JB_ERR_MEMORY - Out of memory * JB_ERR_PARSE - Cannot parse regex (Detailed message * written to system log) * @@ -1105,21 +1112,17 @@ jb_err create_url_spec(struct url_spec *url, char *buf) memset(url, '\0', sizeof(*url)); /* Remember the original specification for the CGI pages. */ - url->spec = strdup(buf); - if (NULL == url->spec) - { - return JB_ERR_MEMORY; - } + url->spec = strdup_or_die(buf); - /* Is it tag pattern? */ - if (0 == strncmpic("TAG:", url->spec, 4)) + /* Is it a tag pattern? */ + if (0 == strncmpic(url->spec, "TAG:", 4)) { /* The pattern starts with the first character after "TAG:" */ const char *tag_pattern = buf + 4; return compile_pattern(tag_pattern, NO_ANCHORING, url, &url->tag_regex); } - /* If it isn't a tag pattern it must be a URL pattern. */ + /* If it isn't a tag pattern it must be an URL pattern. */ return compile_url_pattern(url, buf); } @@ -1167,6 +1170,70 @@ void free_url_spec(struct url_spec *url) } +/********************************************************************* + * + * Function : port_matches + * + * Description : Compares a port against a port list. + * + * Parameters : + * 1 : port = The port to check. + * 2 : port_list = The list of port to compare with. + * + * Returns : TRUE for yes, FALSE otherwise. + * + *********************************************************************/ +static int port_matches(const int port, const char *port_list) +{ + return ((NULL == port_list) || match_portlist(port_list, port)); +} + + +/********************************************************************* + * + * Function : host_matches + * + * Description : Compares a host against a host pattern. + * + * Parameters : + * 1 : url = The URL to match + * 2 : pattern = The URL pattern + * + * Returns : TRUE for yes, FALSE otherwise. + * + *********************************************************************/ +static int host_matches(const struct http_request *http, + const struct url_spec *pattern) +{ +#ifdef FEATURE_EXTENDED_HOST_PATTERNS + return ((NULL == pattern->host_regex) + || (0 == regexec(pattern->host_regex, http->host, 0, NULL, 0))); +#else + return ((NULL == pattern->dbuffer) || (0 == domain_match(pattern, http))); +#endif +} + + +/********************************************************************* + * + * Function : path_matches + * + * Description : Compares a path against a path pattern. + * + * Parameters : + * 1 : path = The path to match + * 2 : pattern = The URL pattern + * + * Returns : TRUE for yes, FALSE otherwise. + * + *********************************************************************/ +static int path_matches(const char *path, const struct url_spec *pattern) +{ + return ((NULL == pattern->preg) + || (0 == regexec(pattern->preg, path, 0, NULL, 0))); +} + + /********************************************************************* * * Function : url_match @@ -1183,22 +1250,14 @@ void free_url_spec(struct url_spec *url) int url_match(const struct url_spec *pattern, const struct http_request *http) { - /* XXX: these should probably be functions. */ -#define PORT_MATCHES ((NULL == pattern->port_list) || match_portlist(pattern->port_list, http->port)) -#ifdef FEATURE_EXTENDED_HOST_PATTERNS -#define DOMAIN_MATCHES ((NULL == pattern->host_regex) || (0 == regexec(pattern->host_regex, http->host, 0, NULL, 0))) -#else -#define DOMAIN_MATCHES ((NULL == pattern->dbuffer) || (0 == domain_match(pattern, http))) -#endif -#define PATH_MATCHES ((NULL == pattern->preg) || (0 == regexec(pattern->preg, http->path, 0, NULL, 0))) - if (pattern->tag_regex != NULL) { /* It's a tag pattern and shouldn't be matched against URLs */ return 0; - } + } - return (PORT_MATCHES && DOMAIN_MATCHES && PATH_MATCHES); + return (port_matches(http->port, pattern->port_list) + && host_matches(http, pattern) && path_matches(http->path, pattern)); } @@ -1222,7 +1281,7 @@ int match_portlist(const char *portlist, int port) { char *min, *max, *next, *portlist_copy; - min = portlist_copy = strdup(portlist); + min = portlist_copy = strdup_or_die(portlist); /* * Zero-terminate first item and remember offset for next @@ -1235,7 +1294,7 @@ int match_portlist(const char *portlist, int port) /* * Loop through all items, checking for match */ - while(min) + while (NULL != min) { if (NULL == (max = strchr(min, (int) '-'))) { @@ -1255,7 +1314,7 @@ int match_portlist(const char *portlist, int port) * or, if max was omitted, between min and 65K */ *max++ = '\0'; - if(port >= atoi(min) && port <= (atoi(max) ? atoi(max) : 65535)) + if (port >= atoi(min) && port <= (atoi(max) ? atoi(max) : 65535)) { freez(portlist_copy); return(1); @@ -1310,11 +1369,7 @@ jb_err parse_forwarder_address(char *address, char **hostname, int *port) return JB_ERR_PARSE; } - *hostname = strdup(address); - if (NULL == *hostname) - { - return JB_ERR_MEMORY; - } + *hostname = strdup_or_die(address); if ((**hostname == '[') && (NULL != (p = strchr(*hostname, ']')))) {