Merge branch 'testing' into 0.9.x

7 files changed, 363 insertions, 91 deletions

diff --git a/lib/activity.php b/lib/activity.php
index b1744e68f..bd1d5d56c 100644
--- a/lib/activity.php
+++ b/lib/activity.php
@@ -238,17 +238,17 @@ class Activity
             $this->time = strtotime($pubDateEl->textContent);
         }
 
-        $authorEl = $this->_child($item, self::AUTHOR, self::RSS);
-
-        if (!empty($authorEl)) {
+        if ($authorEl = $this->_child($item, self::AUTHOR, self::RSS)) {
             $this->actor = ActivityObject::fromRssAuthor($authorEl);
+        } else if ($dcCreatorEl = $this->_child($item, self::CREATOR, self::DC)) {
+            $this->actor = ActivityObject::fromDcCreator($dcCreatorEl);
+        } else if ($posterousEl = $this->_child($item, ActivityObject::AUTHOR, ActivityObject::POSTEROUS)) {
+            // Special case for Posterous.com
+            $this->actor = ActivityObject::fromPosterousAuthor($posterousEl);
+        } else if (!empty($channel)) {
+            $this->actor = ActivityObject::fromRssChannel($channel);
         } else {
-            $dcCreatorEl = $this->_child($item, self::CREATOR, self::DC);
-            if (!empty($dcCreatorEl)) {
-                $this->actor = ActivityObject::fromDcCreator($dcCreatorEl);
-            } else if (!empty($channel)) {
-                $this->actor = ActivityObject::fromRssChannel($channel);
-            }
+            // No actor!
         }
 
         $this->title = ActivityUtils::childContent($item, ActivityObject::TITLE, self::RSS);
@@ -362,48 +362,3 @@ class Activity
     }
 }
 
-class AtomCategory
-{
-    public $term;
-    public $scheme;
-    public $label;
-
-    function __construct($element=null)
-    {
-        if ($element && $element->attributes) {
-            $this->term = $this->extract($element, 'term');
-            $this->scheme = $this->extract($element, 'scheme');
-            $this->label = $this->extract($element, 'label');
-        }
-    }
-
-    protected function extract($element, $attrib)
-    {
-        $node = $element->attributes->getNamedItemNS(Activity::ATOM, $attrib);
-        if ($node) {
-            return trim($node->textContent);
-        }
-        $node = $element->attributes->getNamedItem($attrib);
-        if ($node) {
-            return trim($node->textContent);
-        }
-        return null;
-    }
-
-    function asString()
-    {
-        $attribs = array();
-        if ($this->term !== null) {
-            $attribs['term'] = $this->term;
-        }
-        if ($this->scheme !== null) {
-            $attribs['scheme'] = $this->scheme;
-        }
-        if ($this->label !== null) {
-            $attribs['label'] = $this->label;
-        }
-        $xs = new XMLStringer();
-        $xs->element('category', $attribs);
-        return $xs->asString();
-    }
-}
diff --git a/lib/activityobject.php b/lib/activityobject.php
index b1e9071ed..0a358ccab 100644
--- a/lib/activityobject.php
+++ b/lib/activityobject.php
@@ -80,6 +80,13 @@ class ActivityObject
     const URI   = 'uri';
     const EMAIL = 'email';
 
+    const POSTEROUS   = 'http://posterous.com/help/rss/1.0';
+    const AUTHOR      = 'author';
+    const USERIMAGE   = 'userImage';
+    const PROFILEURL  = 'profileUrl';
+    const NICKNAME    = 'nickName';
+    const DISPLAYNAME = 'displayName';
+
     public $element;
     public $type;
     public $id;
@@ -149,7 +156,11 @@ class ActivityObject
     {
         $this->type  = self::PERSON; // XXX: is this fair?
         $this->title = $this->_childContent($element, self::NAME);
-        $this->id    = $this->_childContent($element, self::URI);
+
+        $id = $this->_childContent($element, self::URI);
+        if (ActivityUtils::validateUri($id)) {
+            $this->id = $id;
+        }
 
         if (empty($this->id)) {
             $email = $this->_childContent($element, self::EMAIL);
@@ -162,6 +173,15 @@ class ActivityObject
 
     private function _fromAtomEntry($element)
     {
+        if ($element->localName == 'actor') {
+            // Old-fashioned <activity:actor>...
+            // First pull anything from <author>, then we'll add on top.
+            $author = ActivityUtils::child($element->parentNode, 'author');
+            if ($author) {
+                $this->_fromAuthor($author);
+            }
+        }
+
         $this->type = $this->_childContent($element, Activity::OBJECTTYPE,
                                            Activity::SPEC);
 
@@ -169,7 +189,11 @@ class ActivityObject
             $this->type = ActivityObject::NOTE;
         }
 
-        $this->id      = $this->_childContent($element, self::ID);
+        $id = $this->_childContent($element, self::ID);
+        if (ActivityUtils::validateUri($id)) {
+            $this->id = $id;
+        }
+
         $this->summary = ActivityUtils::childHtmlContent($element, self::SUMMARY);
         $this->content = ActivityUtils::getContent($element);
 
@@ -290,9 +314,39 @@ class ActivityObject
         $imageEl = ActivityUtils::child($el, Activity::IMAGE, Activity::RSS);
 
         if (!empty($imageEl)) {
-            $obj->avatarLinks[] = ActivityUtils::childContent($imageEl, Activity::URL, Activity::RSS);
+            $url = ActivityUtils::childContent($imageEl, Activity::URL, Activity::RSS);
+            $al = new AvatarLink();
+            $al->url = $url;
+            $obj->avatarLinks[] = $al;
+        }
+
+        return $obj;
+    }
+
+    public static function fromPosterousAuthor($el)
+    {
+        $obj = new ActivityObject();
+
+        $obj->type = ActivityObject::PERSON; // @fixme any others...?
+
+        $userImage = ActivityUtils::childContent($el, self::USERIMAGE, self::POSTEROUS);
+
+        if (!empty($userImage)) {
+            $al = new AvatarLink();
+            $al->url = $userImage;
+            $obj->avatarLinks[] = $al;
         }
 
+        $obj->link = ActivityUtils::childContent($el, self::PROFILEURL, self::POSTEROUS);
+        $obj->id   = $obj->link;
+
+        $obj->poco = new PoCo();
+
+        $obj->poco->preferredUsername = ActivityUtils::childContent($el, self::NICKNAME, self::POSTEROUS);
+        $obj->poco->displayName       = ActivityUtils::childContent($el, self::DISPLAYNAME, self::POSTEROUS);
+
+        $obj->title = $obj->poco->displayName;
+
         return $obj;
     }
 
diff --git a/lib/activityutils.php b/lib/activityutils.php
index c85a3db55..a7e99fb11 100644
--- a/lib/activityutils.php
+++ b/lib/activityutils.php
@@ -240,4 +240,26 @@ class ActivityUtils
             throw new ClientException(_("Can't handle embedded Base64 content yet."));
         }
     }
+
+    /**
+     * Is this a valid URI for remote profile/notice identification?
+     * Does not have to be a resolvable URL.
+     * @param string $uri
+     * @return boolean
+     */
+    static function validateUri($uri)
+    {
+        if (Validate::uri($uri)) {
+            return true;
+        }
+
+        // Possibly an upstream bug; tag: URIs aren't validated properly
+        // unless you explicitly ask for them. All other schemes are accepted
+        // for basic URI validation without asking.
+        if (Validate::uri($uri, array('allowed_scheme' => array('tag')))) {
+            return true;
+        }
+
+        return false;
+    }
 }
diff --git a/lib/atomcategory.php b/lib/atomcategory.php
new file mode 100644
index 000000000..4cc3b4f4d
--- /dev/null
+++ b/lib/atomcategory.php
@@ -0,0 +1,77 @@
+<?php
+/**
+ * StatusNet, the distributed open-source microblogging tool
+ *
+ * PHP version 5
+ *
+ * LICENCE: This program is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ *
+ * @category  Feed
+ * @package   StatusNet
+ * @author    Evan Prodromou <evan@status.net>
+ * @author    Zach Copley <zach@status.net>
+ * @copyright 2010 StatusNet, Inc.
+ * @license   http://www.fsf.org/licensing/licenses/agpl-3.0.html AGPLv3
+ * @link      http://status.net/
+ */
+
+if (!defined('STATUSNET')) {
+    exit(1);
+}
+
+class AtomCategory
+{
+    public $term;
+    public $scheme;
+    public $label;
+
+    function __construct($element=null)
+    {
+        if ($element && $element->attributes) {
+            $this->term = $this->extract($element, 'term');
+            $this->scheme = $this->extract($element, 'scheme');
+            $this->label = $this->extract($element, 'label');
+        }
+    }
+
+    protected function extract($element, $attrib)
+    {
+        $node = $element->attributes->getNamedItemNS(Activity::ATOM, $attrib);
+        if ($node) {
+            return trim($node->textContent);
+        }
+        $node = $element->attributes->getNamedItem($attrib);
+        if ($node) {
+            return trim($node->textContent);
+        }
+        return null;
+    }
+
+    function asString()
+    {
+        $attribs = array();
+        if ($this->term !== null) {
+            $attribs['term'] = $this->term;
+        }
+        if ($this->scheme !== null) {
+            $attribs['scheme'] = $this->scheme;
+        }
+        if ($this->label !== null) {
+            $attribs['label'] = $this->label;
+        }
+        $xs = new XMLStringer();
+        $xs->element('category', $attribs);
+        return $xs->asString();
+    }
+}
diff --git a/plugins/OStatus/OStatusPlugin.php b/plugins/OStatus/OStatusPlugin.php
index 58f373e45..c985fb4bc 100644
--- a/plugins/OStatus/OStatusPlugin.php
+++ b/plugins/OStatus/OStatusPlugin.php
@@ -947,23 +947,4 @@ class OStatusPlugin extends Plugin
         }
         return false;
     }
-
-    /**
-     * Utility function to check if the given URL is a canonical user profile
-     * page, and if so return the ID number.
-     *
-     * @param string $url
-     * @return mixed int or false
-     */
-    public static function localProfileFromUrl($url)
-    {
-        $template = common_local_url('userbyid', array('id' => '31337'));
-        $template = preg_quote($template, '/');
-        $template = str_replace('31337', '(\d+)', $template);
-        if (preg_match("/$template/", $url, $matches)) {
-            return intval($matches[1]);
-        }
-        return false;
-    }
-
 }
diff --git a/plugins/OStatus/classes/Ostatus_profile.php b/plugins/OStatus/classes/Ostatus_profile.php
index d2e046a60..de5175427 100644
--- a/plugins/OStatus/classes/Ostatus_profile.php
+++ b/plugins/OStatus/classes/Ostatus_profile.php
@@ -708,9 +708,13 @@ class Ostatus_profile extends Memcached_DataObject
     }
 
     /**
+     * Look up and if necessary create an Ostatus_profile for the remote entity
+     * with the given profile page URL. This should never return null -- you
+     * will either get an object or an exception will be thrown.
+     *
      * @param string $profile_url
      * @return Ostatus_profile
-     * @throws FeedSubException
+     * @throws Exception
      */
 
     public static function ensureProfileURL($profile_url, $hints=array())
@@ -731,7 +735,7 @@ class Ostatus_profile extends Memcached_DataObject
         $response = $client->get($profile_url);
 
         if (!$response->isOk()) {
-            return null;
+            throw new Exception("Could not reach profile page: " . $profile_url);
         }
 
         // Check if we have a non-canonical URL
@@ -785,11 +789,20 @@ class Ostatus_profile extends Memcached_DataObject
 
         if (!empty($feedurl)) {
             $hints['feedurl'] = $feedurl;
-
             return self::ensureFeedURL($feedurl, $hints);
         }
+
+        throw new Exception("Could not find a feed URL for profile page " . $finalUrl);
     }
 
+    /**
+     * Look up the Ostatus_profile, if present, for a remote entity with the
+     * given profile page URL. Will return null for both unknown and invalid
+     * remote profiles.
+     *
+     * @return mixed Ostatus_profile or null
+     * @throws Exception for local profiles
+     */
     static function getFromProfileURL($profile_url)
     {
         $profile = Profile::staticGet('profileurl', $profile_url);
@@ -821,6 +834,14 @@ class Ostatus_profile extends Memcached_DataObject
         return null;
     }
 
+    /**
+     * Look up and if necessary create an Ostatus_profile for remote entity
+     * with the given update feed. This should never return null -- you will
+     * either get an object or an exception will be thrown.
+     *
+     * @return Ostatus_profile
+     * @throws Exception
+     */
     public static function ensureFeedURL($feed_url, $hints=array())
     {
         $discover = new FeedDiscovery();
@@ -849,6 +870,18 @@ class Ostatus_profile extends Memcached_DataObject
         }
     }
 
+    /**
+     * Look up and, if necessary, create an Ostatus_profile for the remote
+     * profile with the given Atom feed - actually loaded from the feed.
+     * This should never return null -- you will either get an object or
+     * an exception will be thrown.
+     *
+     * @param DOMElement $feedEl root element of a loaded Atom feed
+     * @param array $hints additional discovery information passed from higher levels
+     * @fixme should this be marked public?
+     * @return Ostatus_profile
+     * @throws Exception
+     */
     public static function ensureAtomFeed($feedEl, $hints)
     {
         // Try to get a profile from the feed activity:subject
@@ -899,8 +932,40 @@ class Ostatus_profile extends Memcached_DataObject
         throw new FeedSubException("Can't find enough profile information to make a feed.");
     }
 
+    /**
+     * Look up and, if necessary, create an Ostatus_profile for the remote
+     * profile with the given RSS feed - actually loaded from the feed.
+     * This should never return null -- you will either get an object or
+     * an exception will be thrown.
+     *
+     * @param DOMElement $feedEl root element of a loaded RSS feed
+     * @param array $hints additional discovery information passed from higher levels
+     * @fixme should this be marked public?
+     * @return Ostatus_profile
+     * @throws Exception
+     */
     public static function ensureRssChannel($feedEl, $hints)
     {
+        // Special-case for Posterous. They have some nice metadata in their
+        // posterous:author elements. We should use them instead of the channel.
+
+        $items = $feedEl->getElementsByTagName('item');
+
+        if ($items->length > 0) {
+            $item = $items->item(0);
+            $authorEl = ActivityUtils::child($item, ActivityObject::AUTHOR, ActivityObject::POSTEROUS);
+            if (!empty($authorEl)) {
+                $obj = ActivityObject::fromPosterousAuthor($authorEl);
+                // Posterous has multiple authors per feed, and multiple feeds
+                // per author. We check if this is the "main" feed for this author.
+                if (array_key_exists('profileurl', $hints) &&
+                    !empty($obj->poco) &&
+                    common_url_to_nickname($hints['profileurl']) == $obj->poco->preferredUsername) {
+                    return self::ensureActivityObjectProfile($obj, $hints);
+                }
+            }
+        }
+
         // @fixme we should check whether this feed has elements
         // with different <author> or <dc:creator> elements, and... I dunno.
         // Do something about that.
@@ -1042,11 +1107,14 @@ class Ostatus_profile extends Memcached_DataObject
     /**
      * Fetch, or build if necessary, an Ostatus_profile for the actor
      * in a given Activity Streams activity.
+     * This should never return null -- you will either get an object or
+     * an exception will be thrown.
      *
      * @param Activity $activity
      * @param string $feeduri if we already know the canonical feed URI!
      * @param string $salmonuri if we already know the salmon return channel URI
      * @return Ostatus_profile
+     * @throws Exception
      */
 
     public static function ensureActorProfile($activity, $hints=array())
@@ -1054,6 +1122,18 @@ class Ostatus_profile extends Memcached_DataObject
         return self::ensureActivityObjectProfile($activity->actor, $hints);
     }
 
+    /**
+     * Fetch, or build if necessary, an Ostatus_profile for the profile
+     * in a given Activity Streams object (can be subject, actor, or object).
+     * This should never return null -- you will either get an object or
+     * an exception will be thrown.
+     *
+     * @param ActivityObject $object
+     * @param array $hints additional discovery information passed from higher levels
+     * @return Ostatus_profile
+     * @throws Exception
+     */
+
     public static function ensureActivityObjectProfile($object, $hints=array())
     {
         $profile = self::getActivityObjectProfile($object);
@@ -1068,35 +1148,45 @@ class Ostatus_profile extends Memcached_DataObject
     /**
      * @param Activity $activity
      * @return mixed matching Ostatus_profile or false if none known
+     * @throws ServerException if feed info invalid
      */
     public static function getActorProfile($activity)
     {
         return self::getActivityObjectProfile($activity->actor);
     }
 
+    /**
+     * @param ActivityObject $activity
+     * @return mixed matching Ostatus_profile or false if none known
+     * @throws ServerException if feed info invalid
+     */
     protected static function getActivityObjectProfile($object)
     {
         $uri = self::getActivityObjectProfileURI($object);
         return Ostatus_profile::staticGet('uri', $uri);
     }
 
-    protected static function getActorProfileURI($activity)
-    {
-        return self::getActivityObjectProfileURI($activity->actor);
-    }
-
     /**
-     * @param Activity $activity
+     * Get the identifier URI for the remote entity described
+     * by this ActivityObject. This URI is *not* guaranteed to be
+     * a resolvable HTTP/HTTPS URL.
+     *
+     * @param ActivityObject $object
      * @return string
-     * @throws ServerException
+     * @throws ServerException if feed info invalid
      */
     protected static function getActivityObjectProfileURI($object)
     {
-        $opts = array('allowed_schemes' => array('http', 'https'));
-        if ($object->id && Validate::uri($object->id, $opts)) {
-            return $object->id;
+        if ($object->id) {
+            if (ActivityUtils::validateUri($object->id)) {
+                return $object->id;
+            }
         }
-        if ($object->link && Validate::uri($object->link, $opts)) {
+
+        // If the id is missing or invalid (we've seen feeds mistakenly listing
+        // things like local usernames in that field) then we'll use the profile
+        // page link, if valid.
+        if ($object->link && common_valid_http_url($object->link)) {
             return $object->link;
         }
         throw new ServerException("No author ID URI found");
@@ -1109,6 +1199,8 @@ class Ostatus_profile extends Memcached_DataObject
     /**
      * Create local ostatus_profile and profile/user_group entries for
      * the provided remote user or group.
+     * This should never return null -- you will either get an object or
+     * an exception will be thrown.
      *
      * @param ActivityObject $object
      * @param array $hints
@@ -1125,7 +1217,8 @@ class Ostatus_profile extends Memcached_DataObject
             throw new Exception("No profile URI");
         }
 
-        if (OStatusPlugin::localProfileFromUrl($homeuri)) {
+        $user = User::staticGet('uri', $homeuri);
+        if ($user) {
             throw new Exception("Local user can't be referenced as remote.");
         }
 
@@ -1425,6 +1518,11 @@ class Ostatus_profile extends Memcached_DataObject
     }
 
     /**
+     * Look up, and if necessary create, an Ostatus_profile for the remote
+     * entity with the given webfinger address.
+     * This should never return null -- you will either get an object or
+     * an exception will be thrown.
+     *
      * @param string $addr webfinger address
      * @return Ostatus_profile
      * @throws Exception on error conditions
diff --git a/tests/ActivityParseTests.php b/tests/ActivityParseTests.php
index b6980a6bb..9d8fd47af 100644
--- a/tests/ActivityParseTests.php
+++ b/tests/ActivityParseTests.php
@@ -170,6 +170,51 @@ class ActivityParseTests extends PHPUnit_Framework_TestCase
         $this->assertFalse(empty($actor));
         $this->assertEquals($actor->title, "Joseph Scott");
     }
+
+    public function testExample7()
+    {
+        global $_example7;
+
+        $dom = DOMDocument::loadXML($_example7);
+
+        $rss = $dom->documentElement;
+
+        $channels = $dom->getElementsByTagName('channel');
+
+        $channel = $channels->item(0);
+
+        $items = $channel->getElementsByTagName('item');
+
+        $item = $items->item(0);
+
+        $act = new Activity($item, $channel);
+
+        $this->assertEquals(ActivityVerb::POST, $act->verb);
+        $this->assertEquals('http://evanpro.posterous.com/checking-out-captain-bones', $act->link);
+        $this->assertEquals('http://evanpro.posterous.com/checking-out-captain-bones', $act->id);
+        $this->assertEquals('Checking out captain bones', $act->title);
+        $this->assertEquals(1269095551, $act->time);
+
+        $actor = $act->actor;
+
+        $this->assertEquals(ActivityObject::PERSON, $actor->type);
+        $this->assertEquals('http://posterous.com/people/3sDslhaepotz', $actor->id);
+        $this->assertEquals('Evan Prodromou', $actor->title);
+        $this->assertNull($actor->summary);
+        $this->assertNull($actor->content);
+        $this->assertEquals('http://posterous.com/people/3sDslhaepotz', $actor->link);
+        $this->assertNull($actor->source);
+        $this->assertTrue(is_array($actor->avatarLinks));
+        $this->assertEquals(1, count($actor->avatarLinks));
+        $this->assertEquals('http://files.posterous.com/user_profile_pics/480326/2009-08-05-142447.jpg',
+                            $actor->avatarLinks[0]);
+        $this->assertNotNull($actor->poco);
+        $this->assertEquals('evanpro', $actor->poco->preferredUsername);
+        $this->assertEquals('Evan Prodromou', $actor->poco->displayName);
+        $this->assertNull($actor->poco->note);
+        $this->assertNull($actor->poco->address);
+        $this->assertEquals(0, count($actor->poco->urls));
+    }
 }
 
 $_example1 = <<<EXAMPLE1
@@ -423,3 +468,43 @@ $_example6 = <<<EXAMPLE6
 </rss>
 EXAMPLE6;
 
+$_example7 = <<<EXAMPLE7
+<?xml version="1.0" encoding="UTF-8"?>
+	<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:posterous="http://posterous.com/help/rss/1.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:media="http://search.yahoo.com/mrss/">
+	  <channel>
+	    <title>evanpro's posterous</title>
+	    <link>http://evanpro.posterous.com</link>
+	    <description>Most recent posts at evanpro's posterous</description>
+	    <generator>posterous.com</generator>
+	    <link type="application/json" xmlns="http://www.w3.org/2005/Atom" rel="http://api.friendfeed.com/2008/03#sup" href="http://posterous.com/api/sup_update#56bcc5eb7"/>
+	    <atom:link rel="self" href="http://evanpro.posterous.com/rss.xml"/>
+	    <atom:link rel="hub" href="http://posterous.superfeedr.com"/>
+	    <item>
+	      <pubDate>Sat, 20 Mar 2010 07:32:31 -0700</pubDate>
+	      <title>Checking out captain bones</title>
+	      <link>http://evanpro.posterous.com/checking-out-captain-bones</link>
+	      <guid>http://evanpro.posterous.com/checking-out-captain-bones</guid>
+	      <description>
+	        <![CDATA[<p>
+		<p>Bones!</p>
+
+	</p>
+
+	<p><a href="http://evanpro.posterous.com/checking-out-captain-bones">Permalink</a>
+
+		| <a href="http://evanpro.posterous.com/checking-out-captain-bones#comment">Leave a comment&nbsp;&nbsp;&raquo;</a>
+
+	</p>]]>
+	      </description>
+	      <posterous:author>
+	        <posterous:userImage>http://files.posterous.com/user_profile_pics/480326/2009-08-05-142447.jpg</posterous:userImage>
+	        <posterous:profileUrl>http://posterous.com/people/3sDslhaepotz</posterous:profileUrl>
+	        <posterous:firstName>Evan</posterous:firstName>
+	        <posterous:lastnNme>Prodromou</posterous:lastnNme>
+	        <posterous:nickName>evanpro</posterous:nickName>
+	        <posterous:displayName>Evan Prodromou</posterous:displayName>
+	      </posterous:author>
+	    </item>
+	</channel>
+</rss>
+EXAMPLE7;