require_once('library/HTML5/Parser.php');
require_once('include/crypto.php');
+require_once('include/feed.php');
if(! function_exists('scrape_dfrn')) {
-function scrape_dfrn($url) {
+function scrape_dfrn($url, $dont_probe = false) {
$a = get_app();
logger('scrape_dfrn: url=' . $url);
+ // Try to fetch the data from noscrape. This is faster than parsing the HTML
+ $noscrape = str_replace("/hcard/", "/noscrape/", $url);
+ $noscrapejson = fetch_url($noscrape);
+ $noscrapedata = array();
+ if ($noscrapejson) {
+ $noscrapedata = json_decode($noscrapejson, true);
+
+ if (is_array($noscrapedata)) {
+ if ($noscrapedata["nick"] != "")
+ return($noscrapedata);
+ else
+ unset($noscrapedata["nick"]);
+ } else
+ $noscrapedata = array();
+ }
+
$s = fetch_url($url);
- if(! $s)
+ if (!$s)
return $ret;
+ if (!$dont_probe) {
+ $probe = probe_url($url);
+
+ if (isset($probe["addr"]))
+ $ret["addr"] = $probe["addr"];
+ }
+
$headers = $a->get_curl_headers();
logger('scrape_dfrn: headers=' . $headers, LOGGER_DEBUG);
}
}
}
-
- return $ret;
+ return array_merge($ret, $noscrapedata);
}}
$a = get_app();
$ret = array();
- $s = fetch_url($url);
+ $cookiejar = tempnam(get_temppath(), 'cookiejar-scrape-feed-');
+ $s = fetch_url($url, false, $redirects, 0, Null, $cookiejar);
+ unlink($cookiejar);
$headers = $a->get_curl_headers();
$code = $a->get_curl_code();
$ret['feed_atom'] = $url;
return $ret;
}
- if(stristr($line,'application/rss+xml') || stristr($s,'<rss')) {
+ if(stristr($line,'application/rss+xml') || stristr($s,'<rss')) {
$ret['feed_rss'] = $url;
return $ret;
}
}
}
- try {
- // Cleanup invalid HTML
- $doc = new DOMDocument();
- @$doc->loadHTML($s);
- $s = $doc->saveHTML();
+ $basename = implode('/', array_slice(explode('/',$url),0,3)) . '/';
- $dom = HTML5_Parser::parse($s);
- } catch (DOMException $e) {
- logger('scrape_feed: parse error: ' . $e);
- }
+ $doc = new DOMDocument();
+ @$doc->loadHTML($s);
+ $xpath = new DomXPath($doc);
- if(! $dom) {
- logger('scrape_feed: failed to parse.');
- return $ret;
- }
+ $base = $xpath->query("//base");
+ foreach ($base as $node) {
+ $attr = array();
+ if ($node->attributes->length)
+ foreach ($node->attributes as $attribute)
+ $attr[$attribute->name] = $attribute->value;
- $head = $dom->getElementsByTagName('base');
- if($head) {
- foreach($head as $head0) {
- $basename = $head0->getAttribute('href');
- break;
- }
+ if ($attr["href"] != "")
+ $basename = $attr["href"] ;
}
- if(! $basename)
- $basename = implode('/', array_slice(explode('/',$url),0,3)) . '/';
- $items = $dom->getElementsByTagName('link');
+ $list = $xpath->query("//link");
+ foreach ($list as $node) {
+ $attr = array();
- // get Atom/RSS link elements, take the first one of either.
+ if ($node->attributes->length)
+ foreach ($node->attributes as $attribute)
+ $attr[$attribute->name] = $attribute->value;
- if($items) {
- foreach($items as $item) {
- $x = $item->getAttribute('rel');
- if(($x === 'alternate') && ($item->getAttribute('type') === 'application/atom+xml')) {
- if(! x($ret,'feed_atom'))
- $ret['feed_atom'] = $item->getAttribute('href');
- }
- if(($x === 'alternate') && ($item->getAttribute('type') === 'application/rss+xml')) {
- if(! x($ret,'feed_rss'))
- $ret['feed_rss'] = $item->getAttribute('href');
- }
- }
+ if (($attr["rel"] == "alternate") AND ($attr["type"] == "application/atom+xml"))
+ $ret["feed_atom"] = $attr["href"];
+
+ if (($attr["rel"] == "alternate") AND ($attr["type"] == "application/rss+xml"))
+ $ret["feed_rss"] = $attr["href"];
}
- // Drupal and perhaps others only provide relative URL's. Turn them into absolute.
+ // Drupal and perhaps others only provide relative URLs. Turn them into absolute.
if(x($ret,'feed_atom') && (! strstr($ret['feed_atom'],'://')))
$ret['feed_atom'] = $basename . $ret['feed_atom'];
* PROBE_DIASPORA has a bias towards returning Diaspora information
* while PROBE_NORMAL has a bias towards dfrn/zot - in the case where
* an address (such as a Friendica address) supports more than one type
- * of network.
+ * of network.
*
*/
define ( 'PROBE_NORMAL', 0);
define ( 'PROBE_DIASPORA', 1);
-function probe_url($url, $mode = PROBE_NORMAL) {
+function probe_url($url, $mode = PROBE_NORMAL, $level = 1) {
require_once('include/email.php');
$result = array();
- if(! $url)
+ if (!$url)
return $result;
+ $result = Cache::get("probe_url:".$mode.":".$url);
+ if (!is_null($result)) {
+ $result = unserialize($result);
+ return $result;
+ }
+
+ $original_url = $url;
$network = null;
$diaspora = false;
$diaspora_base = '';
$diaspora_key = '';
$has_lrdd = false;
$email_conversant = false;
+ $connectornetworks = false;
+ $appnet = false;
+
+ if (strpos($url,'twitter.com')) {
+ $connectornetworks = true;
+ $network = NETWORK_TWITTER;
+ }
- // Twitter is deactivated since twitter closed its old API
- //$twitter = ((strpos($url,'twitter.com') !== false) ? true : false);
$lastfm = ((strpos($url,'last.fm/user') !== false) ? true : false);
$at_addr = ((strpos($url,'@') !== false) ? true : false);
- if((! $twitter) && (! $lastfm)) {
+ if((!$appnet) && (!$lastfm) && !$connectornetworks) {
if(strpos($url,'mailto:') !== false && $at_addr) {
$url = str_replace('mailto:','',$url);
else
$links = lrdd($url);
- if(count($links)) {
+ if ((count($links) == 0) AND strstr($url, "/index.php")) {
+ $url = str_replace("/index.php", "", $url);
+ $links = lrdd($url);
+ }
+
+ if (count($links)) {
$has_lrdd = true;
logger('probe_url: found lrdd links: ' . print_r($links,true), LOGGER_DATA);
$pubkey = $diaspora_key;
$diaspora = true;
}
+ if(($link['@attributes']['rel'] === 'http://ostatus.org/schema/1.0/subscribe') AND ($mode == PROBE_NORMAL)) {
+ $diaspora = false;
+ }
}
// Status.Net can have more than one profile URL. We need to match the profile URL
// to a contact on incoming messages to prevent spam, and we won't know which one
// to match. So in case of two, one of them is stored as an alias. Only store URL's
// and not webfinger user@host aliases. If they've got more than two non-email style
- // aliases, let's hope we're lucky and get one that matches the feed author-uri because
+ // aliases, let's hope we're lucky and get one that matches the feed author-uri because
// otherwise we're screwed.
+ $backup_alias = "";
+
foreach($links as $link) {
if($link['@attributes']['rel'] === 'alias') {
if(strpos($link['@attributes']['href'],'@') === false) {
if(isset($profile)) {
- if($link['@attributes']['href'] !== $profile)
- $alias = unamp($link['@attributes']['href']);
+ $alias_url = $link['@attributes']['href'];
+
+ if(($alias_url !== $profile) AND ($backup_alias == "") AND
+ ($alias_url !== str_replace("/index.php", "", $profile)))
+ $backup_alias = $alias_url;
+
+ if(($alias_url !== $profile) AND !strstr($alias_url, "index.php") AND
+ ($alias_url !== str_replace("/index.php", "", $profile)))
+ $alias = $alias_url;
}
else
$profile = unamp($link['@attributes']['href']);
}
}
}
+
+ if ($alias == "")
+ $alias = $backup_alias;
+
+ // If the profile is different from the url then the url is abviously an alias
+ if (($alias == "") AND ($profile != "") AND !$at_addr AND (normalise_link($profile) != normalise_link($url)))
+ $alias = $url;
}
elseif($mode == PROBE_NORMAL) {
}
if($mode == PROBE_NORMAL) {
+
if(strlen($zot)) {
$s = fetch_url($zot);
if($s) {
if($j) {
$network = NETWORK_ZOT;
$vcard = array(
- 'fn' => $j->fullname,
- 'nick' => $j->nickname,
+ 'fn' => $j->fullname,
+ 'nick' => $j->nickname,
'photo' => $j->photo
);
$profile = $j->url;
}
}
+
if(strlen($dfrn)) {
- $ret = scrape_dfrn(($hcard) ? $hcard : $dfrn);
+ $ret = scrape_dfrn(($hcard) ? $hcard : $dfrn, true);
if(is_array($ret) && x($ret,'dfrn-request')) {
$network = NETWORK_DFRN;
$request = $ret['dfrn-request'];
}
}
+ // Scrape the public key from the hcard.
+ // Diaspora will remove it from the webfinger somewhere in the future.
+ if (($hcard != "") AND ($pubkey == "")) {
+ $ret = scrape_dfrn(($hcard) ? $hcard : $dfrn, true);
+ if (isset($ret["key"])) {
+ $hcard_key = $ret["key"];
+ if(strstr($hcard_key,'RSA '))
+ $pubkey = rsatopem($hcard_key);
+ else
+ $pubkey = $hcard_key;
+ }
+ }
if($diaspora && $diaspora_base && $diaspora_guid) {
- if($mode == PROBE_DIASPORA || ! $notify) {
- $notify = $diaspora_base . 'receive/users/' . $diaspora_guid;
+ $diaspora_notify = $diaspora_base.'receive/users/'.$diaspora_guid;
+
+ if($mode == PROBE_DIASPORA || !$notify || ($notify == $diaspora_notify)) {
+ $notify = $diaspora_notify;
$batch = $diaspora_base . 'receive/public' ;
}
if(strpos($url,'@'))
if($network !== NETWORK_ZOT && $network !== NETWORK_DFRN && $network !== NETWORK_MAIL) {
if($diaspora)
$network = NETWORK_DIASPORA;
- elseif($has_lrdd)
+ elseif($has_lrdd AND ($notify))
$network = NETWORK_OSTATUS;
+
+ if(strpos($url,'@'))
+ $addr = str_replace('acct:', '', $url);
+
$priority = 0;
if($hcard && ! $vcard) {
$vcard['nick'] = $addr_parts[0];
}
- /* if($twitter) {
- logger('twitter: setup');
- $tid = basename($url);
- $tapi = 'https://api.twitter.com/1/statuses/user_timeline.rss';
- if(intval($tid))
- $poll = $tapi . '?user_id=' . $tid;
- else
- $poll = $tapi . '?screen_name=' . $tid;
- $profile = 'http://twitter.com/#!/' . $tid;
- //$vcard['photo'] = 'https://api.twitter.com/1/users/profile_image/' . $tid;
- $vcard['photo'] = 'https://api.twitter.com/1/users/profile_image?screen_name=' . $tid . '&size=bigger';
- $vcard['nick'] = $tid;
- $vcard['fn'] = $tid;
- } */
-
if($lastfm) {
$profile = $url;
$poll = str_replace(array('www.','last.fm/'),array('','ws.audioscrobbler.com/1.0/'),$url) . '/recenttracks.rss';
// Will leave it to others to figure out how to grab the avatar, which is on the $url page in the open graph meta links
}
- if($twitter || ! $poll)
+ if($appnet || ! $poll)
$check_feed = true;
if((! isset($vcard)) || (! x($vcard,'fn')) || (! $profile))
$check_feed = true;
if(($at_addr) && (! count($links)))
$check_feed = false;
+ if ($connectornetworks)
+ $check_feed = false;
+
if($check_feed) {
$feedret = scrape_feed(($poll) ? $poll : $url);
+
logger('probe_url: scrape_feed ' . (($poll)? $poll : $url) . ' returns: ' . print_r($feedret,true), LOGGER_DATA);
if(count($feedret) && ($feedret['feed_atom'] || $feedret['feed_rss'])) {
$poll = ((x($feedret,'feed_atom')) ? unamp($feedret['feed_atom']) : unamp($feedret['feed_rss']));
if(x($feedret,'photo') && (! x($vcard,'photo')))
$vcard['photo'] = $feedret['photo'];
- require_once('library/simplepie/simplepie.inc');
- $feed = new SimplePie();
- $xml = fetch_url($poll);
+
+ $cookiejar = tempnam(get_temppath(), 'cookiejar-scrape-feed-');
+ $xml = fetch_url($poll, false, $redirects, 0, Null, $cookiejar);
+ unlink($cookiejar);
logger('probe_url: fetch feed: ' . $poll . ' returns: ' . $xml, LOGGER_DATA);
- $a = get_app();
-
- logger('probe_url: scrape_feed: headers: ' . $a->get_curl_headers(), LOGGER_DATA);
-
- // Don't try and parse an empty string
- $feed->set_raw_data(($xml) ? $xml : '<?xml version="1.0" encoding="utf-8" ?><xml></xml>');
-
- $feed->init();
- if($feed->error())
- logger('probe_url: scrape_feed: Error parsing XML: ' . $feed->error());
-
-
- if(! x($vcard,'photo'))
- $vcard['photo'] = $feed->get_image_url();
- $author = $feed->get_author();
-
- if($author) {
- $vcard['fn'] = unxmlify(trim($author->get_name()));
- if(! $vcard['fn'])
- $vcard['fn'] = trim(unxmlify($author->get_email()));
- if(strpos($vcard['fn'],'@') !== false)
- $vcard['fn'] = substr($vcard['fn'],0,strpos($vcard['fn'],'@'));
- $email = unxmlify($author->get_email());
- if(! $profile && $author->get_link())
- $profile = trim(unxmlify($author->get_link()));
- if(! $vcard['photo']) {
- $rawtags = $feed->get_feed_tags( SIMPLEPIE_NAMESPACE_ATOM_10, 'author');
- if($rawtags) {
- $elems = $rawtags[0]['child'][SIMPLEPIE_NAMESPACE_ATOM_10];
- if((x($elems,'link')) && ($elems['link'][0]['attribs']['']['rel'] === 'photo'))
- $vcard['photo'] = $elems['link'][0]['attribs']['']['href'];
- }
+
+ if ($xml == "") {
+ logger("scrape_feed: XML is empty for feed ".$poll);
+ $network = NETWORK_PHANTOM;
+ } else {
+ $data = feed_import($xml,$dummy1,$dummy2, $dummy3, true);
+
+ if (!is_array($data)) {
+ logger("scrape_feed: This doesn't seem to be a feed: ".$poll);
+ $network = NETWORK_PHANTOM;
+ } else {
+ if (($vcard["photo"] == "") AND ($data["header"]["author-avatar"] != ""))
+ $vcard["photo"] = $data["header"]["author-avatar"];
+
+ if (($vcard["fn"] == "") AND ($data["header"]["author-name"] != ""))
+ $vcard["fn"] = $data["header"]["author-name"];
+
+ if (($vcard["nick"] == "") AND ($data["header"]["author-nick"] != ""))
+ $vcard["nick"] = $data["header"]["author-nick"];
+
+ if ($network == NETWORK_OSTATUS) {
+ if ($data["header"]["author-id"] != "")
+ $alias = $data["header"]["author-id"];
+
+ if ($data["header"]["author-link"] != "")
+ $profile = $data["header"]["author-link"];
+
+ } elseif(!$profile AND ($data["header"]["author-link"] != "") AND !in_array($network, array("", NETWORK_FEED)))
+ $profile = $data["header"]["author-link"];
}
}
- else {
- $item = $feed->get_item(0);
- if($item) {
- $author = $item->get_author();
- if($author) {
- $vcard['fn'] = trim(unxmlify($author->get_name()));
- if(! $vcard['fn'])
- $vcard['fn'] = trim(unxmlify($author->get_email()));
- if(strpos($vcard['fn'],'@') !== false)
- $vcard['fn'] = substr($vcard['fn'],0,strpos($vcard['fn'],'@'));
- $email = unxmlify($author->get_email());
- if(! $profile && $author->get_link())
- $profile = trim(unxmlify($author->get_link()));
- }
- if(! $vcard['photo']) {
- $rawmedia = $item->get_item_tags('http://search.yahoo.com/mrss/','thumbnail');
- if($rawmedia && $rawmedia[0]['attribs']['']['url'])
- $vcard['photo'] = unxmlify($rawmedia[0]['attribs']['']['url']);
- }
- if(! $vcard['photo']) {
- $rawtags = $item->get_item_tags( SIMPLEPIE_NAMESPACE_ATOM_10, 'author');
- if($rawtags) {
- $elems = $rawtags[0]['child'][SIMPLEPIE_NAMESPACE_ATOM_10];
- if((x($elems,'link')) && ($elems['link'][0]['attribs']['']['rel'] === 'photo'))
- $vcard['photo'] = $elems['link'][0]['attribs']['']['href'];
- }
- }
+
+ // Workaround for misconfigured Friendica servers
+ if (($network == "") AND (strstr($url, "/profile/"))) {
+ $noscrape = str_replace("/profile/", "/noscrape/", $url);
+ $noscrapejson = fetch_url($noscrape);
+ if ($noscrapejson) {
+
+ $network = NETWORK_DFRN;
+
+ $poco = str_replace("/profile/", "/poco/", $url);
+
+ $noscrapedata = json_decode($noscrapejson, true);
+
+ if (isset($noscrapedata["addr"]))
+ $addr = $noscrapedata["addr"];
+
+ if (isset($noscrapedata["fn"]))
+ $vcard["fn"] = $noscrapedata["fn"];
+
+ if (isset($noscrapedata["key"]))
+ $pubkey = $noscrapedata["key"];
+
+ if (isset($noscrapedata["photo"]))
+ $vcard["photo"] = $noscrapedata["photo"];
+
+ if (isset($noscrapedata["dfrn-request"]))
+ $request = $noscrapedata["dfrn-request"];
+
+ if (isset($noscrapedata["dfrn-confirm"]))
+ $confirm = $noscrapedata["dfrn-confirm"];
+
+ if (isset($noscrapedata["dfrn-notify"]))
+ $notify = $noscrapedata["dfrn-notify"];
+
+ if (isset($noscrapedata["dfrn-poll"]))
+ $poll = $noscrapedata["dfrn-poll"];
+
}
}
- if((! $vcard['photo']) && strlen($email))
- $vcard['photo'] = avatar_img($email);
- if($poll === $profile)
- $lnk = $feed->get_permalink();
- if(isset($lnk) && strlen($lnk))
- $profile = $lnk;
-
- if(! $network) {
+ if(! $network)
$network = NETWORK_FEED;
- // If it is a feed, don't take the author name as feed name
- unset($vcard['fn']);
- }
- if(! (x($vcard,'fn')))
- $vcard['fn'] = notags($feed->get_title());
- if(! (x($vcard,'fn')))
- $vcard['fn'] = notags($feed->get_description());
-
- if(strpos($vcard['fn'],'Twitter / ') !== false) {
- $vcard['fn'] = substr($vcard['fn'],strpos($vcard['fn'],'/')+1);
- $vcard['fn'] = trim($vcard['fn']);
- }
+
if(! x($vcard,'nick')) {
$vcard['nick'] = strtolower(notags(unxmlify($vcard['fn'])));
if(strpos($vcard['nick'],' '))
if(! x($vcard,'photo')) {
$a = get_app();
- $vcard['photo'] = $a->get_baseurl() . '/images/person-175.jpg' ;
+ $vcard['photo'] = App::get_baseurl() . '/images/person-175.jpg' ;
}
if(! $profile)
if(($network === NETWORK_FEED) && ($poll) && (! x($vcard,'fn')))
$vcard['fn'] = $url;
+ if (($notify != "") AND ($poll != "")) {
+ $baseurl = matching_url(normalise_link($notify), normalise_link($poll));
+
+ $baseurl2 = matching_url($baseurl, normalise_link($profile));
+ if ($baseurl2 != "")
+ $baseurl = $baseurl2;
+ }
+
+ if (($baseurl == "") AND ($notify != ""))
+ $baseurl = matching_url(normalise_link($profile), normalise_link($notify));
+
+ if (($baseurl == "") AND ($poll != ""))
+ $baseurl = matching_url(normalise_link($profile), normalise_link($poll));
+
+ if (substr($baseurl, -10) == "/index.php")
+ $baseurl = str_replace("/index.php", "", $baseurl);
+
+ if ($network == "")
+ $network = NETWORK_PHANTOM;
+
+ $baseurl = rtrim($baseurl, "/");
+
+ if(strpos($url,'@') AND ($addr == "") AND ($network == NETWORK_DFRN))
+ $addr = str_replace('acct:', '', $url);
+
$vcard['fn'] = notags($vcard['fn']);
$vcard['nick'] = str_replace(' ','',notags($vcard['nick']));
$result['network'] = $network;
$result['alias'] = $alias;
$result['pubkey'] = $pubkey;
+ $result['baseurl'] = $baseurl;
logger('probe_url: ' . print_r($result,true), LOGGER_DEBUG);
+ if ($level == 1) {
+ // Trying if it maybe a diaspora account
+ if (($result['network'] == NETWORK_FEED) OR ($result['addr'] == "")) {
+ require_once('include/bbcode.php');
+ $address = GetProfileUsername($url, "", true);
+ $result2 = probe_url($address, $mode, ++$level);
+ if (!in_array($result2['network'], array("", NETWORK_PHANTOM, NETWORK_FEED)))
+ $result = $result2;
+ }
+
+ // Maybe it's some non standard GNU Social installation (Single user, subfolder or no uri rewrite)
+ if (($result['network'] == NETWORK_FEED) AND ($result['baseurl'] != "") AND ($result['nick'] != "")) {
+ $addr = $result['nick'].'@'.str_replace("http://", "", $result['baseurl']);
+ $result2 = probe_url($addr, $mode, ++$level);
+ if (!in_array($result2['network'], array("", NETWORK_PHANTOM, NETWORK_FEED)))
+ $result = $result2;
+ }
+
+ // Quickfix for Hubzilla systems with enabled OStatus plugin
+ if (($result['network'] == NETWORK_DIASPORA) AND ($result["batch"] == "")) {
+ $result2 = probe_url($url, PROBE_DIASPORA, ++$level);
+ if ($result2['network'] == NETWORK_DIASPORA) {
+ $addr = $result["addr"];
+ $result = $result2;
+
+ if (($result["addr"] == "") AND ($addr != ""))
+ $result["addr"] = $addr;
+ }
+ }
+ }
+
+ // Only store into the cache if the value seems to be valid
+ if ($result['network'] != NETWORK_PHANTOM) {
+ Cache::set("probe_url:".$mode.":".$original_url,serialize($result), CACHE_DAY);
+
+ /// @todo temporary fix - we need a real contact update function that updates only changing fields
+ /// The biggest problem is the avatar picture that could have a reduced image size.
+ /// It should only be updated if the existing picture isn't existing anymore.
+ if (($result['network'] != NETWORK_FEED) AND ($mode == PROBE_NORMAL) AND
+ $result["name"] AND $result["nick"] AND $result["url"] AND $result["addr"] AND $result["poll"])
+ q("UPDATE `contact` SET `name` = '%s', `nick` = '%s', `url` = '%s', `addr` = '%s',
+ `notify` = '%s', `poll` = '%s', `alias` = '%s', `success_update` = '%s'
+ WHERE `nurl` = '%s' AND NOT `self` AND `uid` = 0",
+ dbesc($result["name"]),
+ dbesc($result["nick"]),
+ dbesc($result["url"]),
+ dbesc($result["addr"]),
+ dbesc($result["notify"]),
+ dbesc($result["poll"]),
+ dbesc($result["alias"]),
+ dbesc(datetime_convert()),
+ dbesc(normalise_link($result['url']))
+ );
+ }
+
return $result;
}
+
+/**
+ * @brief Find the matching part between two url
+ *
+ * @param string $url1
+ * @param string $url2
+ * @return string The matching part
+ */
+function matching_url($url1, $url2) {
+
+ if (($url1 == "") OR ($url2 == ""))
+ return "";
+
+ $url1 = normalise_link($url1);
+ $url2 = normalise_link($url2);
+
+ $parts1 = parse_url($url1);
+ $parts2 = parse_url($url2);
+
+ if (!isset($parts1["host"]) OR !isset($parts2["host"]))
+ return "";
+
+ if ($parts1["scheme"] != $parts2["scheme"])
+ return "";
+
+ if ($parts1["host"] != $parts2["host"])
+ return "";
+
+ if ($parts1["port"] != $parts2["port"])
+ return "";
+
+ $match = $parts1["scheme"]."://".$parts1["host"];
+
+ if ($parts1["port"])
+ $match .= ":".$parts1["port"];
+
+ $pathparts1 = explode("/", $parts1["path"]);
+ $pathparts2 = explode("/", $parts2["path"]);
+
+ $i = 0;
+ $path = "";
+ do {
+ $path1 = $pathparts1[$i];
+ $path2 = $pathparts2[$i];
+
+ if ($path1 == $path2)
+ $path .= $path1."/";
+
+ } while (($path1 == $path2) AND ($i++ <= count($pathparts1)));
+
+ $match .= $path;
+
+ return normalise_link($match);
+}