Merge branch 'webmention-rocks' into 'nightly'

webmention.rocks

I have improved the webmention handling so that all but two of the webmention.rocks compliance tests pass now.  Also improved parsing of time/authors on incoming webmentions.

See merge request !128

classes/Notice.php

@@ -248,10 +248,10 @@ class Notice extends Managed_DataObject
         return common_local_url('shownotice', array('notice' => $this->id), null, null, false);
     }
 
-    public function getTitle()
+    public function getTitle($imply=true)
     {
         $title = null;
-        if (Event::handle('GetNoticeTitle', array($this, &$title))) {
+        if (Event::handle('GetNoticeTitle', array($this, &$title)) && $imply) {
             // TRANS: Title of a notice posted without a title value.
             // TRANS: %1$s is a user name, %2$s is the notice creation date/time.
             $title = sprintf(_('%1$s\'s status on %2$s'),

lib/noticelistitem.php

@@ -179,8 +179,9 @@ class NoticeListItem extends Widget
     function showNoticeTitle()
     {
         if (Event::handle('StartShowNoticeTitle', array($this))) {
+            $nameClass = $this->notice->getTitle(false) ? 'p-name ' : '';
             $this->element('a', array('href' => $this->notice->getUri(),
-                                      'class' => 'p-name u-uid'),
+                                      'class' => $nameClass . 'u-uid'),
                            $this->notice->getTitle());
             Event::handle('EndShowNoticeTitle', array($this));
         }
@@ -348,7 +349,8 @@ class NoticeListItem extends Widget
     function showContent()
     {
         // FIXME: URL, image, video, audio
-        $this->out->elementStart('article', array('class' => 'e-content'));
+        $nameClass = $this->notice->getTitle(false) ? '' : 'p-name ';
+        $this->out->elementStart('article', array('class' => $nameClass . 'e-content'));
         if (Event::handle('StartShowNoticeContent', array($this->notice, $this->out, $this->out->getScoped()))) {
             if ($this->maxchars > 0 && mb_strlen($this->notice->content) > $this->maxchars) {
                 $this->out->text(mb_substr($this->notice->content, 0, $this->maxchars) . '[…]');

plugins/Linkback/LinkbackPlugin.php

@@ -101,14 +101,28 @@ class LinkbackPlugin extends Plugin
         return true;
     }
 
+    function unparse_url($parsed_url)
+    {
+       $scheme   = isset($parsed_url['scheme']) ? $parsed_url['scheme'] . '://' : '';
+       $host     = isset($parsed_url['host']) ? $parsed_url['host'] : '';
+       $port     = isset($parsed_url['port']) ? ':' . $parsed_url['port'] : '';
+       $user     = isset($parsed_url['user']) ? $parsed_url['user'] : '';
+       $pass     = isset($parsed_url['pass']) ? ':' . $parsed_url['pass']  : '';
+       $pass     = ($user || $pass) ? "$pass@" : '';
+       $path     = isset($parsed_url['path']) ? $parsed_url['path'] : '';
+       $query    = isset($parsed_url['query']) ? '?' . $parsed_url['query'] : '';
+       $fragment = isset($parsed_url['fragment']) ? '#' . $parsed_url['fragment'] : '';
+       return "$scheme$user$pass$host$port$path$query$fragment";
+    }
+
     function linkbackUrl($url)
     {
         common_log(LOG_DEBUG,"Attempting linkback for " . $url);
 
         $orig = $url;
         $url = htmlspecialchars_decode($orig);
-        $scheme = parse_url($url, PHP_URL_SCHEME);
+        $base = parse_url($url);
-        if (!in_array($scheme, array('http', 'https'))) {
+        if (!in_array($base['scheme'], array('http', 'https'))) {
             return $orig;
         }
 
@@ -126,13 +140,17 @@ class LinkbackPlugin extends Plugin
             return $orig;
         }
 
-        // XXX: Should handle relative-URI resolution in these detections
-
         $wm = $this->getWebmention($response);
-        if(!empty($wm)) {
+        if(!is_null($wm)) {
+            $wm = parse_url($wm);
+            if(!$wm) $wm = array();
+            if(!$wm['host']) $wm['host'] = $base['host'];
+            if(!$wm['scheme']) $wm['scheme'] = $base['scheme'];
+            if(!$wm['path']) $wm['path'] = $base['path'];
+
             // It is the webmention receiver's job to resolve source
             // Ref: https://github.com/converspace/webmention/issues/43
-            $this->webmention($url, $wm);
+            $this->webmention($url, $this->unparse_url($wm));
         } else {
             $pb = $this->getPingback($response);
             if (!empty($pb)) {
@@ -156,26 +174,26 @@ class LinkbackPlugin extends Plugin
         $link = $response->getHeader('Link');
         if (!is_null($link)) {
             // XXX: the fetcher gives back a comma-separated string of all Link headers, I hope the parsing works reliably
-            if (preg_match('~<((?:https?://)?[^>]+)>; rel="webmention"~', $link, $match)) {
+            if (preg_match('~<([^>]+)>; rel="?(?:[^" ]* )*(?:http://webmention.org/|webmention)(?: [^" ]*)*"?~', $link, $match)) {
-                return $match[1];
-            } elseif (preg_match('~<((?:https?://)?[^>]+)>; rel="http://webmention.org/?"~', $link, $match)) {
                 return $match[1];
             }
         }
 
         // FIXME: Do proper DOM traversal
-        if(preg_match('/<(?:link|a)[ ]+href="([^"]+)"[ ]+rel="[^" ]* ?webmention ?[^" ]*"[ ]*\/?>/i', $response->getBody(), $match)
+        // Currently fails https://webmention.rocks/test/13, https://webmention.rocks/test/17
-                || preg_match('/<(?:link|a)[ ]+rel="[^" ]* ?webmention ?[^" ]*"[ ]+href="([^"]+)"[ ]*\/?>/i', $response->getBody(), $match)) {
+        if(preg_match('~<(?:link|a)[ ]+href="([^"]*)"[ ]+rel="(?:[^" ]* )*(?:http://webmention.org/|webmention)(?: [^" ]*)*"[ ]*/?>~i', $response->getBody(), $match)
-            return $match[1];
+                || preg_match('~<(?:link|a)[ ]+rel="(?:[^" ]* )*(?:http://webmention.org/|webmention)(?: [^" ]*)*"[ ]+href="([^"]*)"[ ]*/?>~i', $response->getBody(), $match)) {
-        } elseif (preg_match('/<(?:link|a)[ ]+href="([^"]+)"[ ]+rel="http:\/\/webmention\.org\/?"[ ]*\/?>/i', $response->getBody(), $match)
-                || preg_match('/<(?:link|a)[ ]+rel="http:\/\/webmention\.org\/?"[ ]+href="([^"]+)"[ ]*\/?>/i', $response->getBody(), $match)) {
             return $match[1];
         }
+
+        return NULL;
     }
 
     function webmention($url, $endpoint) {
         $source = $this->notice->getUrl();
 
+        common_log(LOG_DEBUG,"Attempting webmention to $endpoint for $url from $source");
+
         $payload = array(
             'source' => $source,
             'target' => $url
@@ -191,7 +209,7 @@ class LinkbackPlugin extends Plugin
                 $payload
             );
 
-            if(!in_array($response->getStatus(), array(200,202))) {
+            if(!in_array($response->getStatus(), array(200,201,202))) {
                 common_log(LOG_WARNING,
                            "Webmention request failed for '$url' ($endpoint)");
             }

plugins/Linkback/actions/webmention.php

@@ -40,30 +40,30 @@ class WebmentionAction extends Action
 
         if(!$source) {
             echo _m('"source" is missing')."\n";
-            throw new ServerException(_m('"source" is missing'), 400);
+            throw new ClientException(_m('"source" is missing'), 400);
         }
 
         if(!$target) {
             echo _m('"target" is missing')."\n";
-            throw new ServerException(_m('"target" is missing'), 400);
+            throw new ClientException(_m('"target" is missing'), 400);
         }
 
         $response = linkback_get_source($source, $target);
         if(!$response) {
             echo _m('Source does not link to target.')."\n";
-            throw new ServerException(_m('Source does not link to target.'), 400);
+            throw new ClientException(_m('Source does not link to target.'), 400);
         }
 
         $notice = linkback_get_target($target);
         if(!$notice) {
             echo _m('Target not found')."\n";
-            throw new ServerException(_m('Target not found'), 404);
+            throw new ClientException(_m('Target not found'), 404);
         }
 
         $url = linkback_save($source, $target, $response, $notice);
         if(!$url) {
             echo _m('An error occured while saving.')."\n";
-            throw new ServerException(_m('An error occured while saving.'), 500);
+            throw new ClientException(_m('An error occured while saving.'), 500);
         }
 
         echo $url."\n";

plugins/Linkback/lib/util.php

@@ -7,7 +7,7 @@ function linkback_lenient_target_match($body, $target) {
 function linkback_get_source($source, $target) {
     // Check if we are pinging ourselves and ignore
     $localprefix = common_config('site', 'server') . '/' . common_config('site', 'path');
-    if(linkback_lenient_target_match($source, $localprefix)) {
+    if(linkback_lenient_target_match($source, $localprefix) === 0) {
         common_debug('Ignoring self ping from ' . $source . ' to ' . $target);
         return NULL;
     }
@@ -22,7 +22,7 @@ function linkback_get_source($source, $target) {
 
     $body = htmlspecialchars_decode($response->getBody());
     // We're slightly more lenient in our link detection than the spec requires
-    if(!linkback_lenient_target_match($body, $target)) {
+    if(linkback_lenient_target_match($body, $target) === FALSE) {
         return NULL;
     }
 
@@ -56,7 +56,7 @@ function linkback_get_target($target) {
         }
         if(!$user) {
             preg_match('/\/([^\/\?#]+)(?:#.*)?$/', $response->getEffectiveUrl(), $match);
-            if(linkback_lenient_target_match(common_profile_url($match[1]), $response->getEffectiveUrl())) {
+            if(linkback_lenient_target_match(common_profile_url($match[1]), $response->getEffectiveUrl()) !== FALSE) {
                 $user = User::getKV('nickname', $match[1]);
             }
         }
@@ -70,7 +70,7 @@ function linkback_get_target($target) {
 
 function linkback_is_contained_in($entry, $target) {
     foreach ((array)$entry['properties'] as $key => $values) {
-        if(count(array_filter($values, function($x) use ($target) { return linkback_lenient_target_match($x, $target); })) > 0) {
+        if(count(array_filter($values, function($x) use ($target) { return linkback_lenient_target_match($x, $target) !== FALSE; })) > 0) {
             return $entry['properties'];
         }
 
@@ -79,7 +79,7 @@ function linkback_is_contained_in($entry, $target) {
             if(isset($obj['type']) && array_intersect(array('h-cite', 'h-entry'), $obj['type']) &&
                isset($obj['properties']) && isset($obj['properties']['url']) &&
                count(array_filter($obj['properties']['url'],
-                     function($x) use ($target) { return linkback_lenient_target_match($x, $target); })) > 0
+                     function($x) use ($target) { return linkback_lenient_target_match($x, $target) !== FALSE; })) > 0
             ) {
                 return $entry['properties'];
             }
@@ -130,7 +130,7 @@ function linkback_entry_type($entry, $mf2, $target) {
 
     if($mf2['rels'] && $mf2['rels']['in-reply-to']) {
         foreach($mf2['rels']['in-reply-to'] as $url) {
-            if(linkback_lenient_target_match($url, $target)) {
+            if(linkback_lenient_target_match($url, $target) !== FALSE) {
                 return 'reply';
             }
         }
@@ -144,7 +144,7 @@ function linkback_entry_type($entry, $mf2, $target) {
     );
 
     foreach((array)$entry as $key => $values) {
-        if(count(array_filter($values, function($x) use ($target) { return linkback_lenient_target_match($x, $target); })) > 0) {
+        if(count(array_filter($values, function($x) use ($target) { return linkback_lenient_target_match($x, $target) != FALSE; })) > 0) {
             if($classes[$key]) { return $classes[$key]; }
         }
 
@@ -152,7 +152,7 @@ function linkback_entry_type($entry, $mf2, $target) {
             if(isset($obj['type']) && array_intersect(array('h-cite', 'h-entry'), $obj['type']) &&
                isset($obj['properties']) && isset($obj['properties']['url']) &&
                count(array_filter($obj['properties']['url'],
-                     function($x) use ($target) { return linkback_lenient_target_match($x, $target); })) > 0
+                     function($x) use ($target) { return linkback_lenient_target_match($x, $target) != FALSE; })) > 0
             ) {
                 if($classes[$key]) { return $classes[$key]; }
             }
@@ -243,8 +243,8 @@ function linkback_notice($source, $notice_or_user, $entry, $author, $mf2) {
 
     if (isset($entry['published']) || isset($entry['updated'])) {
         $options['created'] = isset($entry['published'])
-                                ? common_sql_date($entry['published'][0])
+                                ? common_sql_date(strtotime($entry['published'][0]))
-                                : common_sql_date($entry['updated'][0]);
+                                : common_sql_date(strtotime($entry['updated'][0]));
     }
 
     if (isset($entry['photo']) && common_valid_http_url($entry['photo'])) {
@@ -280,9 +280,42 @@ function linkback_notice($source, $notice_or_user, $entry, $author, $mf2) {
     return array($content, $options);
 }
 
+function linkback_avatar($profile, $url) {
+    // Ripped from OStatus plugin for now
+    $temp_filename = tempnam(sys_get_temp_dir(), 'linback_avatar');
+    try {
+        $imgData = HTTPClient::quickGet($url);
+        // Make sure it's at least an image file. ImageFile can do the rest.
+        if (false === getimagesizefromstring($imgData)) {
+            return false;
+        }
+        file_put_contents($temp_filename, $imgData);
+        unset($imgData);    // No need to carry this in memory.
+
+        $imagefile = new ImageFile(null, $temp_filename);
+        $filename = Avatar::filename($profile->id,
+                                     image_type_to_extension($imagefile->type),
+                                     null,
+                                     common_timestamp());
+        rename($temp_filename, Avatar::path($filename));
+    } catch (Exception $e) {
+        unlink($temp_filename);
+        throw $e;
+    }
+    // @todo FIXME: Hardcoded chmod is lame, but seems to be necessary to
+    // keep from accidentally saving images from command-line (queues)
+    // that can't be read from web server, which causes hard-to-notice
+    // problems later on:
+    //
+    // http://status.net/open-source/issues/2663
+    chmod(Avatar::path($filename), 0644);
+
+    $profile->setOriginal($filename);
+}
+
 function linkback_profile($entry, $mf2, $response, $target) {
-    if(isset($entry['properties']['author']) && isset($entry['properties']['author'][0]['properties'])) {
+    if(isset($entry['author']) && isset($entry['author'][0]['properties'])) {
-        $author = $entry['properties']['author'][0]['properties'];
+        $author = $entry['author'][0]['properties'];
     } else {
         $author = linkback_hcard($mf2, $response->getEffectiveUrl());
     }
@@ -315,6 +348,10 @@ function linkback_profile($entry, $mf2, $response, $target) {
         $profile->nickname = isset($author['nickname']) ? $author['nickname'][0] : str_replace(' ', '', $author['name'][0]);
         $profile->created = common_sql_now();
         $profile->insert();
+
+        if($author['photo'] && $author['photo'][0]) {
+            linkback_avatar($profile, $author['photo'][0]);
+        }
     }
 
     return array($profile, $author);