function makeAbsoluteStr($base, $url)
{
$base = new SimplePie_IRI($base);
// remove '//' in URL path (causes URLs not to resolve properly)
if (isset($base->path)) {
$base->path = preg_replace('!//+!', '/', $base->path);
}
if (preg_match('!^https?://!i', $url)) {
// already absolute
return $url;
} else {
if ($absolute = SimplePie_IRI::absolutize($base, $url)) {
return $absolute;
}
return false;
}
}
/**
* Create a new IRI object by resolving a relative IRI
*
* Returns false if $base is not absolute, otherwise an IRI.
*
* @param IRI|string $base (Absolute) Base IRI
* @param IRI|string $relative Relative IRI
* @return IRI|false
*/
public static function absolutize($base, $relative)
{
if (!$relative instanceof SimplePie_IRI) {
$relative = new SimplePie_IRI($relative);
}
if (!$relative->is_valid()) {
return false;
} elseif ($relative->scheme !== null) {
return clone $relative;
} else {
if (!$base instanceof SimplePie_IRI) {
$base = new SimplePie_IRI($base);
}
if ($base->scheme !== null && $base->is_valid()) {
if ($relative->get_iri() !== '') {
if ($relative->iuserinfo !== null || $relative->ihost !== null || $relative->port !== null) {
$target = clone $relative;
$target->scheme = $base->scheme;
} else {
$target = new SimplePie_IRI();
$target->scheme = $base->scheme;
$target->iuserinfo = $base->iuserinfo;
$target->ihost = $base->ihost;
$target->port = $base->port;
if ($relative->ipath !== '') {
if ($relative->ipath[0] === '/') {
$target->ipath = $relative->ipath;
} elseif (($base->iuserinfo !== null || $base->ihost !== null || $base->port !== null) && $base->ipath === '') {
$target->ipath = '/' . $relative->ipath;
} elseif (($last_segment = strrpos($base->ipath, '/')) !== false) {
$target->ipath = substr($base->ipath, 0, $last_segment + 1) . $relative->ipath;
} else {
$target->ipath = $relative->ipath;
}
$target->ipath = $target->remove_dot_segments($target->ipath);
$target->iquery = $relative->iquery;
} else {
$target->ipath = $base->ipath;
if ($relative->iquery !== null) {
$target->iquery = $relative->iquery;
} elseif ($base->iquery !== null) {
$target->iquery = $base->iquery;
}
}
$target->ifragment = $relative->ifragment;
}
} else {
$target = clone $base;
$target->ifragment = null;
}
$target->scheme_normalization();
return $target;
} else {
return false;
}
}
}
/**
* Try to find the refresh url from the meta.
*
* @param string $url Absolute url
* @param string $html First characters of the response (hopefully it'll be enough to find some meta)
*
* @return false|string
*/
private function getMetaRefreshURL($url, $html)
{
if ($html == '') {
return false;
}
// <meta HTTP-EQUIV="REFRESH" content="0; url=http://www.bernama.com/bernama/v6/newsindex.php?id=943513">
if (!preg_match('!<meta http-equiv=["\']?refresh["\']? content=["\']?[0-9];\\s*url=["\']?([^"\'>]+)["\']?!i', $html, $match)) {
return false;
}
$redirect_url = trim($match[1]);
if (preg_match('!^https?://!i', $redirect_url)) {
// already absolute
$this->logger->log('debug', 'Meta refresh redirect found (http-equiv="refresh"), new URL: ' . $redirect_url);
return $redirect_url;
}
// absolutize redirect URL
$base = new \SimplePie_IRI($url);
// remove '//' in URL path (causes URLs not to resolve properly)
if (isset($base->ipath)) {
$base->ipath = str_replace('//', '/', $base->ipath);
}
if ($absolute = \SimplePie_IRI::absolutize($base, $redirect_url)) {
$this->logger->log('debug', 'Meta refresh redirect found (http-equiv="refresh"), new URL: ' . $absolute);
return $absolute->get_iri();
}
return false;
}
public static function compress_parse_url($scheme = '', $authority = '', $path = '', $query = '', $fragment = '')
{
$iri = new SimplePie_IRI('');
$iri->scheme = $scheme;
$iri->authority = $authority;
$iri->path = $path;
$iri->query = $query;
$iri->fragment = $fragment;
return $iri->get_uri();
}
public static function normalize_url($url)
{
$iri = new SimplePie_IRI($url);
return $iri->get_uri();
}
public static function absolutize($base, $relative)
{
$relative = (string) $relative;
if ($relative !== '') {
$relative = new SimplePie_IRI($relative);
if ($relative->get_scheme() !== null) {
$target = $relative;
} elseif ($base->get_iri() !== null) {
if ($relative->get_authority() !== null) {
$target = $relative;
$target->set_scheme($base->get_scheme());
} else {
$target = new SimplePie_IRI('');
$target->set_scheme($base->get_scheme());
$target->set_userinfo($base->get_userinfo());
$target->set_host($base->get_host());
$target->set_port($base->get_port());
if ($relative->get_path() !== null) {
if (strpos($relative->get_path(), '/') === 0) {
$target->set_path($relative->get_path());
} elseif (($base->get_userinfo() !== null || $base->get_host() !== null || $base->get_port() !== null) && $base->get_path() === null) {
$target->set_path('/' . $relative->get_path());
} elseif (($last_segment = strrpos($base->get_path(), '/')) !== false) {
$target->set_path(substr($base->get_path(), 0, $last_segment + 1) . $relative->get_path());
} else {
$target->set_path($relative->get_path());
}
$target->set_query($relative->get_query());
} else {
$target->set_path($base->get_path());
if ($relative->get_query() !== null) {
$target->set_query($relative->get_query());
} elseif ($base->get_query() !== null) {
$target->set_query($base->get_query());
}
}
}
$target->set_fragment($relative->get_fragment());
} else {
$target = $relative;
}
} else {
$target = $base;
}
return $target;
}
private function makeAbsoluteAttr($base, $e, $attr)
{
if ($e->hasAttribute($attr)) {
// Trim leading and trailing white space. I don't really like this but
// unfortunately it does appear on some sites. e.g. <img src=" /path/to/image.jpg" />
$url = trim(str_replace('%20', ' ', $e->getAttribute($attr)));
$url = str_replace(' ', '%20', $url);
if (!preg_match('!https?://!i', $url)) {
if ($absolute = \SimplePie_IRI::absolutize($base, $url)) {
$e->setAttribute($attr, $absolute);
}
}
}
}
/**
* @dataProvider normalization_tests
*/
public function testStringNormalization($input, $output)
{
$input = new SimplePie_IRI($input);
$this->assertEquals($output, $input->get_iri());
}
