Merge branch 'master' of github.com:indieweb/link-rel-parser-php

CODE_OF_CONDUCT.md

@@ -0,0 +1,5 @@
+# Contributor Code of Conduct
+
+See the Code of Conduct on the IndieWebCamp wiki.
+
+https://indiewebcamp.com/code-of-conduct

README.md

@@ -1,5 +1,4 @@
-link-rel-parser-php
-===================
+# link-rel-parser-php
 
 [![Build Status](https://travis-ci.org/indieweb/link-rel-parser-php.png?branch=master)](http://travis-ci.org/indieweb/link-rel-parser-php)
 
@@ -12,3 +11,24 @@ To run the tests, run this on your command line:
 ```
 ./phpunit.phar
 ```
+
+
+## Contributing
+
+Bug reports and pull requests are welcome on GitHub at https://github.com/indieweb/link_rel_parser-ruby/issues. This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the [IndieWebCamp Code of Conduct](https://indiewebcamp.com/code-of-conduct).
+
+1. Fork it
+2. Get it running
+3. Create your feature branch (`git checkout -b my-new-feature`)
+4. Write your code and **tests**
+5. Commit your changes (`git commit -am 'Add some feature'`)
+6. Push to the branch (`git push origin my-new-feature`)
+7. Create new Pull Request
+
+If you find bugs, have feature requests or questions, please
+[file an issue](https://github.com/indieweb/link-rel-parser-php/issues).
+
+
+## Code of Conduct
+
+Everyone interacting in the link-rel-parser-php codebase, issue tracker, chat room, and mailing lists is expected to follow the [link-rel-parser-php code of conduct](https://indiewebcamp.com/code-of-conduct).

src/IndieWeb/get_rel_webmention.php

@@ -0,0 +1,133 @@
+<?php
+/* sample functions for using the link rel parser library to get first webmention (and pingback) endpoints */
+
+/* 
+first_linknode_href, get_rel_webmention by Tantek Çelik http://tantek.com/
+license: http://creativecommons.org/publicdomain/zero/1.0/
+depends on: link_rel_parser.php (e.g. head_http_rels)
+depends on: https://github.com/tantek/cassis/cassis.js (e.g. contains, get_absolute_uri, is_html_type, xphasrel, strcat)
+*/
+
+// Could move this function to cassis.js if more broadly useful
+function is_loopback($href) {
+// in: $href URL
+// out: boolean whether host of URL is a loopback address
+  $host = hostname_of_uri($href);
+  if ($host == 'localhost') { return true; }
+  $host = explode('.', $host);
+  return ($host.length == 4 &&
+          $host[0] == 127 &&
+          ctype_digit($host[1]) &&
+          ctype_digit($host[2]) &&
+          ctype_digit($host[3]));
+}
+
+function first_linknode_href($links, $spacedtagnames='a area link') {
+// in: DOMNodeList $links
+//     $spacedtagnames - space separated tag names, null for any
+// out: href attribute as string
+// return href of first DOMNode in $links that is an a, area, link
+//      with href that is not a loopback address
+// else return null
+  if ($spacedtagnames) {
+    $spacedtagnames = strcat(' ', $spacedtagnames, ' ');
+  }
+  foreach ($links as $link) {
+    if (!$spacedtagnames ||
+        contains($spacedtagnames, 
+                 strcat(' ', $link->nodeName, ' '))) {
+      $href = $link->getAttribute('href');
+      if (!is_loopback($href))
+      {
+        return $href;
+      }
+    }
+  }
+  return null;
+}
+
+// in: $url of page that may or may not have a webmention endpoint
+// out: array of 'webmention' URL of webmention endpoint if any, 
+//           and 'pingback' URL of pingback endpoint if any
+function get_rel_webmention($url) {
+  global $debug;
+  $r = array();
+  $r['webmention'] = '';
+  $r['pingback'] = '';
+
+  $httprels = head_http_rels($url);
+  if ($debug) {
+    echo 'head_http_rels STATUS:"'.$httprels['status'].'"<br/>';
+  }
+  if ($httprels['status'] != "200") {
+    return $r;
+  }
+
+  if ($debug) {
+    echo 'HEAD Content-Type: '.$httprels['type'].' '.
+         string(is_html_type($httprels['type'])).'<br/>';
+  }
+  $wm = '';
+  $pb = '';
+  if (array_key_exists('webmention', $httprels['rels'])) {
+    $wm = $httprels['rels']['webmention'][0];
+    // just use the first one.
+  }
+  if (array_key_exists('pingback', $httprels['rels'])) {
+    $pb = $httprels['rels']['pingback'][0];
+    // just use the first one.
+  }
+  if ($debug && $wm) {
+    echo "HEAD LINK webmention: '$wm'<br/>";
+  }
+  if ($debug && $pb) {
+    echo "HEAD LINK pingback: '$pb'<br/>";
+  }
+  if (!$wm && is_html_type($httprels['type'])) {
+    // no webmention endpoint in HTTP headers, check HTML
+    if ($debug) {
+      echo "looking for wm endpoint in HTML $url<br/>";
+    }
+    $ch = curl_init($url);
+//  curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
+// commented out due to:
+// Warning: curl_setopt(): CURLOPT_FOLLOWLOCATION cannot be activated when an open_basedir is set
+	  curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+    $s = curl_exec($ch);
+    curl_close($ch);
+    if ($s != '') {
+      $dom = new DOMDocument();
+      $dom->loadHTML($s); // ** maybe only load part of it?
+      $domx = new DOMXPath($dom);
+      $wms = $domx->query(xphasrel('webmention'));
+      if ($wms) { $wms = first_linknode_href($wms); }
+      if ($debug) {
+        echo "query xphasrel webmention $wms<br/>";
+      }
+
+      if ($wms !== null) {
+        $wm = get_absolute_uri($wms, $url);
+      }
+      if ($debug && $wm) {
+        echo "HTML rel=webmention returned '$wm'<br/>";
+      }
+      $wms = $domx->query(xphasrel('pingback'));
+      if ($wms) { $wms = first_linknode_href($wms, 'link'); }
+      if ($debug) {
+        echo "query xphasrel pingback $wms<br/>";
+      }
+
+      if ($wms !== null) {
+        $pb = get_absolute_uri($wms, $url);
+      }
+      if ($debug && $pb) {
+        echo "HTML rel=pingback returned '$pb'<br/>";
+      }
+    }
+  }
+  $r['webmention'] = $wm;
+  $r['pingback'] = $pb;
+  return $r;
+}
+
+?>

src/IndieWeb/link_rel_parser.php

@@ -4,6 +4,7 @@
 /* 
 http_rels, head_http_rels by Tantek Çelik http://tantek.com/
 license: http://creativecommons.org/publicdomain/zero/1.0/
+depends on: get_absolute_uri in https://github.com/tantek/cassis/cassis.js
 */
 
 /**