-const char urlmatch_rcs[] = "$Id: urlmatch.c,v 1.62 2011/09/04 11:10:56 fabiankeil Exp $";
+const char urlmatch_rcs[] = "$Id: urlmatch.c,v 1.76 2013/11/24 14:22:51 fabiankeil Exp $";
/*********************************************************************
*
* File : $Source: /cvsroot/ijbswa/current/urlmatch.c,v $
* 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
RIGHT_ANCHORED,
RIGHT_ANCHORED_HOST
};
-static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern);
+static jb_err compile_host_pattern(struct pattern_spec *url, const char *host_pattern);
/*********************************************************************
*
* 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.
*
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)
{
/* 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);
#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;
+
+}
+
+
/*********************************************************************
*
* Function : parse_http_url
* 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.
*
/*
* 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;
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;
* 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
{
* 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)
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, '@');
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
{
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;
- }
}
#ifdef FEATURE_EXTENDED_HOST_PATTERNS
* 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.
*
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);
/*
* 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;
}
*
*********************************************************************/
static jb_err compile_pattern(const char *pattern, enum regex_anchoring anchoring,
- struct url_spec *url, regex_t **regex)
+ struct pattern_spec *url, regex_t **regex)
{
int errcode;
char rebuf[BUFFER_SIZE];
* Description : Compiles the three parts of an URL pattern.
*
* Parameters :
- * 1 : url = Target url_spec to be filled in.
+ * 1 : url = Target pattern_spec to be filled in.
* 2 : buf = The url pattern to compile. Will be messed up.
*
* Returns : JB_ERR_OK - Success
* JB_ERR_PARSE - Cannot parse regex
*
*********************************************************************/
-static jb_err compile_url_pattern(struct url_spec *url, char *buf)
+static jb_err compile_url_pattern(struct pattern_spec *url, char *buf)
{
char *p;
/*
* XXX: does it make sense to compile the slash at the beginning?
*/
- jb_err err = compile_pattern(p, LEFT_ANCHORED, url, &url->preg);
+ jb_err err = compile_pattern(p, LEFT_ANCHORED, url, &url->pattern.url_spec.preg);
if (JB_ERR_OK != err)
{
if (NULL != p)
{
*p++ = '\0';
- url->port_list = strdup(p);
- if (NULL == url->port_list)
- {
- return JB_ERR_MEMORY;
- }
+ url->pattern.url_spec.port_list = strdup_or_die(p);
}
else
{
- url->port_list = NULL;
+ url->pattern.url_spec.port_list = NULL;
}
if (buf[0] != '\0')
*
* 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.
+ * 1 : url = Target pattern_spec to be filled in.
* 2 : host_pattern = Host pattern to compile.
*
* Returns : JB_ERR_OK - Success
* JB_ERR_PARSE - Cannot parse regex
*
*********************************************************************/
-static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern)
+static jb_err compile_host_pattern(struct pattern_spec *url, const char *host_pattern)
{
- return compile_pattern(host_pattern, RIGHT_ANCHORED_HOST, url, &url->host_regex);
+ return compile_pattern(host_pattern, RIGHT_ANCHORED_HOST, url, &url->pattern.url_spec.host_regex);
}
#else
* Description : Parses and "compiles" an old-school host pattern.
*
* Parameters :
- * 1 : url = Target url_spec to be filled in.
+ * 1 : url = Target pattern_spec to be filled in.
* 2 : host_pattern = Host pattern to parse.
*
* Returns : JB_ERR_OK - Success
- * JB_ERR_MEMORY - Out of memory
* JB_ERR_PARSE - Cannot parse regex
*
*********************************************************************/
-static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern)
+static jb_err compile_host_pattern(struct pattern_spec *url, const char *host_pattern)
{
char *v[150];
size_t size;
/*
* 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)
{
*/
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);
}
/*
* 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)))))
{
/*
* Success: Go ahead
}
/* Cut off extra '*'s */
- if(*pat == '*') pat++;
+ if (*pat == '*') pat++;
/* If this is the pattern's end, fine! */
return(*pat);
* Returns : 0 => domains are equivalent, else no match.
*
*********************************************************************/
-static int domain_match(const struct url_spec *pattern, const struct http_request *fqdn)
+static int domain_match(const struct pattern_spec *pattern, const struct http_request *fqdn)
{
char **pv, **fv; /* vectors */
int plen, flen;
* 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)
*
*********************************************************************/
-jb_err create_url_spec(struct url_spec *url, char *buf)
+jb_err create_url_spec(struct pattern_spec *url, char *buf)
{
assert(url);
assert(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 a tag pattern? */
+ /* Is it a positive 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);
+ url->flags |= PATTERN_SPEC_TAG_PATTERN;
+ return compile_pattern(tag_pattern, NO_ANCHORING, url, &url->pattern.tag_regex);
+ }
+ /* Is it a negative tag pattern? */
+ if (0 == strncmpic(url->spec, "NO-REQUEST-TAG:", 15))
+ {
+ /* The pattern starts with the first character after "NO-REQUEST-TAG:" */
+ const char *tag_pattern = buf + 15;
+ url->flags |= PATTERN_SPEC_NO_REQUEST_TAG_PATTERN;
+ return compile_pattern(tag_pattern, NO_ANCHORING, url, &url->pattern.tag_regex);
+ }
+ if (0 == strncmpic(url->spec, "NO-RESPONSE-TAG:", 16))
+ {
+ /* The pattern starts with the first character after "NO-RESPONSE-TAG:" */
+ const char *tag_pattern = buf + 16;
+ url->flags |= PATTERN_SPEC_NO_RESPONSE_TAG_PATTERN;
+ return compile_pattern(tag_pattern, NO_ANCHORING, url, &url->pattern.tag_regex);
}
+ url->flags |= PATTERN_SPEC_URL_PATTERN;
+
/* If it isn't a tag pattern it must be an URL pattern. */
return compile_url_pattern(url, buf);
}
* structure elements.
*
* Parameters :
- * 1 : url = pointer to a url_spec structure.
+ * 1 : url = pointer to a pattern_spec structure.
*
* Returns : N/A
*
*********************************************************************/
-void free_url_spec(struct url_spec *url)
+void free_url_spec(struct pattern_spec *url)
{
if (url == NULL) return;
freez(url->spec);
#ifdef FEATURE_EXTENDED_HOST_PATTERNS
- if (url->host_regex)
+ if (url->pattern.url_spec.host_regex)
{
- regfree(url->host_regex);
- freez(url->host_regex);
+ regfree(url->pattern.url_spec.host_regex);
+ freez(url->pattern.url_spec.host_regex);
}
#else
- freez(url->dbuffer);
- freez(url->dvec);
- url->dcount = 0;
+ freez(url->pattern.url_spec.dbuffer);
+ freez(url->pattern.url_spec.dvec);
+ url->pattern.url_spec.dcount = 0;
#endif /* ndef FEATURE_EXTENDED_HOST_PATTERNS */
- freez(url->port_list);
- if (url->preg)
+ freez(url->pattern.url_spec.port_list);
+ if (url->pattern.url_spec.preg)
{
- regfree(url->preg);
- freez(url->preg);
+ regfree(url->pattern.url_spec.preg);
+ freez(url->pattern.url_spec.preg);
}
- if (url->tag_regex)
+ if (url->pattern.tag_regex)
{
- regfree(url->tag_regex);
- freez(url->tag_regex);
+ regfree(url->pattern.tag_regex);
+ freez(url->pattern.tag_regex);
}
}
*
*********************************************************************/
static int host_matches(const struct http_request *http,
- const struct url_spec *pattern)
+ const struct pattern_spec *pattern)
{
#ifdef FEATURE_EXTENDED_HOST_PATTERNS
- return ((NULL == pattern->host_regex)
- || (0 == regexec(pattern->host_regex, http->host, 0, NULL, 0)));
+ return ((NULL == pattern->pattern.url_spec.host_regex)
+ || (0 == regexec(pattern->pattern.url_spec.host_regex, http->host, 0, NULL, 0)));
#else
- return ((NULL == pattern->dbuffer) || (0 == domain_match(pattern, http)));
+ return ((NULL == pattern->pattern.url_spec.dbuffer) || (0 == domain_match(pattern, http)));
#endif
}
* Returns : TRUE for yes, FALSE otherwise.
*
*********************************************************************/
-static int path_matches(const char *path, const struct url_spec *pattern)
+static int path_matches(const char *path, const struct pattern_spec *pattern)
{
- return ((NULL == pattern->preg)
- || (0 == regexec(pattern->preg, path, 0, NULL, 0)));
+ return ((NULL == pattern->pattern.url_spec.preg)
+ || (0 == regexec(pattern->pattern.url_spec.preg, path, 0, NULL, 0)));
}
* Returns : Nonzero if the URL matches the pattern, else 0.
*
*********************************************************************/
-int url_match(const struct url_spec *pattern,
+int url_match(const struct pattern_spec *pattern,
const struct http_request *http)
{
- if (pattern->tag_regex != NULL)
+ if (!(pattern->flags & PATTERN_SPEC_URL_PATTERN))
{
- /* It's a tag pattern and shouldn't be matched against URLs */
+ /* It's not an URL pattern and thus shouldn't be matched against URLs */
return 0;
}
- return (port_matches(http->port, pattern->port_list)
+ return (port_matches(http->port, pattern->pattern.url_spec.port_list)
&& host_matches(http, pattern) && path_matches(http->path, pattern));
}
{
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
* 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);
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, ']'))))
{