/**
* ping back links referenced in some text
*
* This is the client implementation of
* [link=trackback]http://www.movabletype.org/docs/mttrackback.html[/link]
* and [link=pingback]http://www.hixie.ch/specs/pingback/pingback[/link] specifications.
*
* This function is triggered by publishing scripts, either [script]articles/publish.php[/script],
* [script]services/blog.php[/script], [script]agents/messages.php[/script] or [script]agents/uploads.php[/script].
*
* @see articles/publish.php
* @see services/blog.php
* @see agents/messages.php
* @see agents/uploads.php
*
* It is used to efficiently link pages across a set of web sites according to the following mechanism:
* - The list of external links is built for this page
* - Only the 7 first links are kept from the list; others are stripped
* - If links do not exist, create additional records in the table used for links
* - Each link (actually, only the 7 most recent) is checked, to see if it's trackback- or pingback-enabled or not
* - Each trackback-/pingback-enabled link is activated, providing the full URL of the anchor page
*
* We are claiming to support most of the trackback client interface here, as described in the [link=trackback]http://www.movabletype.org/docs/mttrackback.html[/link] specification.
* A foreign page is considered as being trackback-enabled if it has a special RDF section
* linking its reference (i.e., URL) to a Trackback Ping URL.
*
* Note that YACS also implements the server part of the trackback specification in [script]links/trackback.php[/script],
* which supports POST REST calls.
*
* @see links/trackback.php
*
* We are claiming to fully support the pingback client interface here, as described in the [link=pingback]http://www.hixie.ch/specs/pingback/pingback[/link] specification.
* A foreign page is considered to be pingback-enabled if it has a meta link to a Pingback Ping URL.
*
* Note that YACS also implements the server part of the pingback specification in [script]services/ping.php[/script],
* which supports XML-RPC calls.
*
* @see services/ping.php
*
* This function transforms every YACS codes into HTML before extracting links,
* and before submitting the excerpt to remote site.
*
* @param string the referencing text that has to be scanned
* @param string the local anchor of the referencing text (e.g., 'article:124')
* @return array list($links, $advertised, $skipped)
*
* @link http://www.movabletype.org/docs/mttrackback.html TrackBack Technical Specification
* @link http://www.hixie.ch/specs/pingback/pingback Pingback specification
*/
public static function ping($text, $anchor)
{
global $context;
// render all codes
if (is_callable(array('Codes', 'beautify'))) {
$text = Codes::beautify($text);
}
// suppress all links not coming from anchors (eg, <img src=...)
$text = strip_tags($text, '<a>');
// extract all links from the text, including those that have been encoded by YACS
preg_match_all('/((http:\\/\\/|http%3A%2F%2F)[^ <"]+)/i', $text, $links);
// nothing to do
if (!@count($links[1])) {
return;
}
// process each link only once
$unique_links = array();
foreach ($links[1] as $url) {
// decode raw url encoding, if any
$url = rawurldecode($url);
// strip the clicking indirection, if any
$url = rawurldecode(preg_replace('/^' . preg_quote($context['url_to_home'] . $context['url_to_root'] . 'links/click.php?url=', '/') . '/i', '', $url));
$unique_links[$url] = 1;
}
// analyze found links
$links_processed = array();
$links_advertised = array();
$links_skipped = array();
foreach ($unique_links as $url => $dummy) {
// analyze no more than 7 links
if (@count($links_processed) >= 7) {
break;
}
// skip links that point to ourself, and not to an article
if (preg_match('/^' . preg_quote($context['url_to_home'], '/') . '\\b/i', $url) && !preg_match('/\\/article(-|s\\/view.php)/i', $url)) {
$links_skipped[] = $url;
continue;
}
// skip invalid links
if (($content = http::proceed($url)) === FALSE) {
$links_skipped[] = $url;
continue;
}
// we will use the content to locate pingback and trackback interfaces
$pages[$url] = $content;
// ensure enough execution time
Safe::set_time_limit(30);
// stats
$links_processed[] = $url;
}
// locate the anchor object for this text, we need its url
$anchor = Anchors::get($anchor);
if (!is_object($anchor)) {
return;
}
// build an excerpt from anchor
$excerpt = $anchor->get_teaser('basic');
// find blog name for anchor
if ($parent = $anchor->get_value('anchor')) {
$blog = Anchors::get($parent);
if (is_object($blog)) {
$blog_name = $blog->get_title();
}
}
// build an absolute URL for the source
$source = $context['url_to_home'] . $context['url_to_root'] . $anchor->get_url();
// process each link
if (@count($pages)) {
foreach ($pages as $target => $content) {
// try trackback, if implemented
if (Links::ping_as_trackback($content, $source, $target, $anchor->get_title(), $excerpt, $blog_name)) {
$links_advertised[] = $target;
} elseif (Links::ping_as_pingback($content, $source, $target)) {
$links_advertised[] = $target;
}
}
}
return array($links_processed, $links_advertised, $links_skipped);
}
