/**
* Convert special chars (like german umlauts) to ASCII characters.
*
* @todo dh> IMHO this function should not be included in a file that gets used often/always.
* @param string
* @return string
*/
function replace_special_chars($str)
{
global $evo_charset;
if (can_convert_charsets('UTF-8', $evo_charset) && can_convert_charsets('UTF-8', 'ISO-8859-1')) {
$str = convert_charset($str, 'UTF-8', $evo_charset);
// TODO: add more...?!
$search = array('Ä', 'ä', 'Ö', 'ö', 'Ü', 'ü', 'ß', 'à', 'ç', 'è', 'é', 'ì', 'ò', 'ô', 'ù');
// iso-8859-1
$replace = array('Ae', 'ae', 'Oe', 'oe', 'Ue', 'ue', 'ss', 'a', 'c', 'e', 'e', 'i', 'o', 'o', 'u');
foreach ($search as $k => $v) {
// convert $search to UTF-8
$search[$k] = convert_charset($v, 'UTF-8', 'ISO-8859-1');
}
$str = str_replace($search, $replace, $str);
// Replace HTML entities
$str = htmlentities($str, ENT_NOQUOTES, 'UTF-8');
} else {
// Replace HTML entities only
$str = htmlentities($str, ENT_NOQUOTES, $evo_charset);
}
// Keep only one char in entities!
$str = preg_replace('/&(.).+?;/', '$1', $str);
// Replace non acceptable chars
$str = preg_replace('/[^A-Za-z0-9_]+/', '-', $str);
// Remove '-' at start and end:
$str = preg_replace('/^-+/', '', $str);
$str = preg_replace('/-+$/', '', $str);
return $str;
}
/**
* Test {@link bpost_count_words()}.
*/
function test_bpost_count_words()
{
global $evo_charset;
if (!can_convert_charsets('ISO-8859-1', 'UTF-8')) {
echo 'Skipping tests (cannot convert charsets)...<br />', "\n";
return;
}
$old_evo_charset = $evo_charset;
$evo_charset = 'ISO-8859-1';
$this->assertEqual(bpost_count_words(convert_charset('eine gleichung wie 1 + 2 = 9 /', 'ISO-8859-1', 'UTF-8')), 3);
$this->assertEqual(bpost_count_words(convert_charset('mixed with the 3 ümläuts: äää ööö üüü ÄÄÄ ÖÖÖ ÜÜÜ', 'ISO-8859-1', 'UTF-8')), 10);
$evo_charset = 'UTF-8';
$this->assertEqual(bpost_count_words('möre (again 3) ümläüts... öö üü ää ÄÄ ÖÖ ÜÜ'), 9);
$this->assertEqual(bpost_count_words('russian: Расширенные возможности - это удобный'), 5);
$this->assertEqual(bpost_count_words('A versão foi apelidade de Tilqi, porque era aniversário dele. numbers: 42'), 11);
$this->assertEqual(bpost_count_words('HTML tags -> <a href="http://b2evolution.net" target="_blank">visit b2evo!</a>. Some other chars: "\' \' " <<< < >>> > ``` -- versão удобный überladen'), 10);
$evo_charset = $old_evo_charset;
}
/**
* Convert special chars (like german umlauts) to ASCII characters.
*
* @param string Input string to operate on
* @param NULL|string The post locale or NULL if there is no specific locale.
* Gets passed to evo_iconv_transliterate().
* @return string The input string with replaced chars.
*/
function replace_special_chars($str, $post_locale = NULL)
{
global $evo_charset, $default_locale, $current_locale, $locales;
// Decode entities to be able to transliterate the associated chars:
// Tblue> TODO: Check if this could have side effects.
$str = html_entity_decode($str, ENT_NOQUOTES, $evo_charset);
$our_locale = $post_locale;
if ($our_locale === NULL) {
// post locale is not set, try to guess current locale
if (!empty($default_locale)) {
$our_locale = $default_locale;
}
if (!empty($current_locale)) {
// Override with current locale if available
$our_locale = $current_locale;
}
}
if ($our_locale !== NULL && isset($locales[$our_locale]) && !empty($locales[$our_locale]['transliteration_map'])) {
// Use locale 'transliteration_map' if present
if (!array_key_exists('', $locales[$our_locale]['transliteration_map'])) {
// Make sure there's no empty string key, otherwise strtr() returns false
if ($tmp_str = strtr($str, $locales[$our_locale]['transliteration_map'])) {
}
// Use newly transliterated string
$str = $tmp_str;
}
}
if (($newstr = evo_iconv_transliterate($str, $post_locale)) !== false) {
// iconv allows us to get nice URL titles by transliterating non-ASCII chars.
// Tblue> htmlentities() does not know anything about ASCII?! ISO-8859-1 will work too, though.
$newstr_charset = 'ISO-8859-1';
} else {
if (can_convert_charsets('UTF-8', $evo_charset) && can_convert_charsets('UTF-8', 'ISO-8859-1')) {
// Fallback to the limited old method: Transliterate only a few known chars.
$newstr = convert_charset($str, 'UTF-8', $evo_charset);
$newstr_charset = 'UTF-8';
$search = array('Ä', 'ä', 'Ö', 'ö', 'Ü', 'ü', 'ß', 'à', 'ç', 'è', 'é', 'ì', 'ò', 'ô', 'ù');
// iso-8859-1
$replace = array('Ae', 'ae', 'Oe', 'oe', 'Ue', 'ue', 'ss', 'a', 'c', 'e', 'e', 'i', 'o', 'o', 'u');
foreach ($search as $k => $v) {
// convert $search to UTF-8
$search[$k] = convert_charset($v, 'UTF-8', 'ISO-8859-1');
}
$newstr = str_replace($search, $replace, $newstr);
} else {
// Replace HTML entities only.
$newstr = $str;
$newstr_charset = $evo_charset;
}
}
// Replace HTML entities
$newstr = htmlentities($newstr, ENT_NOQUOTES, $newstr_charset);
// Handle special entities (e.g., use "-" instead of "a" for "&"):
$newstr = str_replace(array('&', '«', '»'), '-', $newstr);
// Keep only one char in entities!
$newstr = preg_replace('/&(.).+?;/', '$1', $newstr);
// Replace non acceptable chars
$newstr = preg_replace('/[^A-Za-z0-9_]+/', '-', $newstr);
// Remove '-' at start and end:
$newstr = preg_replace('/^-+/', '', $newstr);
$newstr = preg_replace('/-+$/', '', $newstr);
//pre_dump( $str, $newstr );
return $newstr;
}
/**
* Check the content of a given URL (referer), if the requested URI (with different hostname variations)
* is present.
*
* @todo Use DB cache to avoid checking the same page again and again! (Plugin DB table)
*
* @param string
* @param string URI to append to matching pattern for hostnames
* @return boolean
*/
function is_referer_linking_us($referer, $uri)
{
global $misc_inc_path, $lib_subdir, $ReqHost;
if (empty($referer)) {
return false;
}
// Load page content (max. 500kb), using fsockopen:
$url_parsed = @parse_url($referer);
if (!$url_parsed) {
return false;
}
if (empty($url_parsed['scheme'])) {
$url_parsed = parse_url('http://' . $referer);
}
$host = $url_parsed['host'];
$port = empty($url_parsed['port']) ? 80 : $url_parsed['port'];
$path = empty($url_parsed['path']) ? '/' : $url_parsed['path'];
if (!empty($url_parsed['query'])) {
$path .= '?' . $url_parsed['query'];
}
$fp = @fsockopen($host, $port, $errno, $errstr, 30);
if (!$fp) {
// could not access referring page
$this->debug_log('is_referer_linking_us(): could not access «' . $referer . '» (host: ' . $host . '): ' . $errstr . ' (#' . $errno . ')');
return false;
}
// Set timeout for data:
if (function_exists('stream_set_timeout')) {
stream_set_timeout($fp, 20);
} else {
socket_set_timeout($fp, 20);
}
// PHP 4
// Send request:
$out = "GET {$path} HTTP/1.0\r\n";
$out .= "Host: {$host}:{$port}\r\n";
$out .= "Connection: Close\r\n\r\n";
fwrite($fp, $out);
// Skip headers:
$i = 0;
$source_charset = 'iso-8859-1';
// default
while (($s = fgets($fp, 4096)) !== false) {
$i++;
if ($s == "\r\n" || $i > 100) {
break;
}
if (preg_match('~^Content-Type:.*?charset=([\\w-]+)~i', $s, $match)) {
$source_charset = $match[1];
}
}
// Get the refering page's content
$content_ref_page = '';
$bytes_read = 0;
while (($s = fgets($fp, 4096)) !== false) {
$content_ref_page .= $s;
$bytes_read += strlen($s);
if ($bytes_read > 512000) {
// do not pull more than 500kb of data!
break;
}
}
fclose($fp);
if (!strlen($content_ref_page)) {
$this->debug_log('is_referer_linking_us(): empty $content_ref_page (' . bytesreadable($bytes_read) . ' read)');
return false;
}
$have_idn_name = false;
// Build the search pattern:
// We match for basically for 'href="[SERVER][URI]', where [SERVER] is a list of possible hosts (especially IDNA)
$search_pattern = '~\\shref=["\']?https?://(';
$possible_hosts = array($_SERVER['HTTP_HOST']);
if ($_SERVER['SERVER_NAME'] != $_SERVER['HTTP_HOST']) {
$possible_hosts[] = $_SERVER['SERVER_NAME'];
}
$search_pattern_hosts = array();
foreach ($possible_hosts as $l_host) {
if (preg_match('~^([^.]+\\.)(.*?)([^.]+\\.[^.]+)$~', $l_host, $match)) {
// we have subdomains in this hostname
if (stristr($match[1], 'www')) {
// search also for hostname without 'www.'
$search_pattern_hosts[] = $match[2] . $match[3];
}
}
$search_pattern_hosts[] = $l_host;
}
$search_pattern_hosts = array_unique($search_pattern_hosts);
foreach ($search_pattern_hosts as $l_host) {
// add IDN, because this could be linked:
$l_idn_host = idna_decode($l_host);
// the decoded puny-code ("xn--..") name (utf8)
if ($l_idn_host != $l_host) {
$have_idn_name = true;
$search_pattern_hosts[] = $l_idn_host;
}
}
// add hosts to pattern, preg_quoted
for ($i = 0, $n = count($search_pattern_hosts); $i < $n; $i++) {
$search_pattern_hosts[$i] = preg_quote($search_pattern_hosts[$i], '~');
}
$search_pattern .= implode('|', $search_pattern_hosts) . ')';
if (empty($uri)) {
// host(s) should end with "/", "'", '"', "?" or whitespace
$search_pattern .= '[/"\'\\s?]';
} else {
$search_pattern .= preg_quote($uri, '~');
// URI should end with "'", '"' or whitespace
$search_pattern .= '["\'\\s]';
}
$search_pattern .= '~i';
if ($have_idn_name) {
// Convert charset to UTF-8, because the decoded domain name is UTF-8, too:
if (can_convert_charsets('utf-8', $source_charset)) {
$content_ref_page = convert_charset($content_ref_page, 'utf-8', $source_charset);
} else {
$this->debug_log('is_referer_linking_us(): warning: cannot convert charset of referring page');
}
}
if (preg_match($search_pattern, $content_ref_page)) {
$this->debug_log('is_referer_linking_us(): found current URL in page (' . bytesreadable($bytes_read) . ' read)');
return true;
} else {
if (strpos($referer, $ReqHost) === 0 && !empty($uri)) {
// Referer is the same host.. just search for $uri
if (strpos($content_ref_page, $uri) !== false) {
$this->debug_log('is_referer_linking_us(): found current URI in page (' . bytesreadable($bytes_read) . ' read)');
return true;
}
}
$this->debug_log('is_referer_linking_us(): ' . sprintf('did not find «%s» in «%s» (%s bytes read).', $search_pattern, $referer, bytesreadable($bytes_read)));
return false;
}
}
