From 8594a2ba16a6884bbac6dbe0951c263ce940944f Mon Sep 17 00:00:00 2001
From: Mikael Nordfeldth <mmn@hethane.se>
Date: Wed, 14 Jan 2015 01:16:28 +0100
Subject: [PATCH] FeedPoller plugin, for hubless feeds

---
 plugins/FeedPoller/FeedPollerPlugin.php       | 61 ++++++++++++++++
 plugins/FeedPoller/lib/feedpoll.php           | 69 ++++++++++++++++++
 .../FeedPoller/lib/feedpollqueuehandler.php   | 41 +++++++++++
 plugins/FeedPoller/scripts/pollfeed.php       | 71 +++++++++++++++++++
 plugins/OStatus/classes/FeedSub.php           |  2 +-
 5 files changed, 243 insertions(+), 1 deletion(-)
 create mode 100644 plugins/FeedPoller/FeedPollerPlugin.php
 create mode 100644 plugins/FeedPoller/lib/feedpoll.php
 create mode 100644 plugins/FeedPoller/lib/feedpollqueuehandler.php
 create mode 100644 plugins/FeedPoller/scripts/pollfeed.php

diff --git a/plugins/FeedPoller/FeedPollerPlugin.php b/plugins/FeedPoller/FeedPollerPlugin.php
new file mode 100644
index 0000000000..c844d99c1b
--- /dev/null
+++ b/plugins/FeedPoller/FeedPollerPlugin.php
@@ -0,0 +1,61 @@
+<?php
+/**
+ * GNU social feed polling plugin, to avoid using external PuSH hubs
+ *
+ * @category  Feed
+ * @package   GNUsocial
+ * @author    Mikael Nordfeldth <mmn@hethane.se>
+ * @copyright 2013 Free Software Foundation, Inc.
+ * @license   http://www.fsf.org/licensing/licenses/agpl-3.0.html AGPL 3.0
+ * @link      http://www.gnu.org/software/social/
+ */
+
+if (!defined('GNUSOCIAL')) { exit(1); }
+
+class FeedPollerPlugin extends Plugin {
+    public $interval = 5;   // interval in minutes for feed checks
+
+    public function onEndInitializeQueueManager(QueueManager $qm)
+    {
+        $qm->connect(FeedPoll::QUEUE_CHECK, 'FeedPollQueueHandler');
+        return true;
+    }
+
+    public function onCronMinutely()
+    {
+        $args = array('interval'=>$this->interval);
+        FeedPoll::enqueueNewFeeds($args);
+        return true;
+    }
+
+    public function onFeedSubscribe(FeedSub $feedsub)
+    {
+        if (!$feedsub->isPuSH()) {
+            FeedPoll::setupFeedSub($feedsub, $this->interval*60);
+            return false;   // We're polling this feed, so stop processing FeedSubscribe
+        }
+        return true;
+    }
+
+    public function onFeedUnsubscribe(FeedSub $feedsub)
+    {
+        if (!$feedsub->isPuSH()) {
+            // removes sub_state setting and such
+            $feedsub->confirmUnsubscribe();
+            return false;
+        }
+        return true;
+    }
+
+    public function onPluginVersion(&$versions)
+    {
+        $versions[] = array('name' => 'FeedPoller',
+                            'version' => GNUSOCIAL_VERSION,
+                            'author' => 'Mikael Nordfeldth',
+                            'homepage' => 'http://www.gnu.org/software/social/',
+                            'description' =>
+                            // TRANS: Plugin description.
+                            _m('Feed polling plugin to avoid using external push hubs.'));
+        return true;
+    }
+}
diff --git a/plugins/FeedPoller/lib/feedpoll.php b/plugins/FeedPoller/lib/feedpoll.php
new file mode 100644
index 0000000000..1ac5fad0cf
--- /dev/null
+++ b/plugins/FeedPoller/lib/feedpoll.php
@@ -0,0 +1,69 @@
+<?php
+/**
+ * Store last poll time in db, then check if they should be renewed (if so, enqueue).
+ * Can be called from a queue handler on a per-feed status to poll stuff.
+ *
+ * Used as internal feed polling mechanism (atom/rss)
+ *
+ * @category OStatus
+ * @package  GNUsocial
+ * @author   Mikael Nordfeldth <mmn@hethane.se>
+ * @license  http://www.fsf.org/licensing/licenses/agpl.html AGPLv3
+ * @link     http://www.gnu.org/software/social/
+ */
+
+if (!defined('GNUSOCIAL')) { exit(1); }
+
+class FeedPoll {
+    const DEFAULT_INTERVAL = 5; // in minutes
+
+    const QUEUE_CHECK = 'feedpoll-check';
+
+    // TODO: Find some smart way to add feeds only once, so they don't get more than 1 feedpoll in the queue each
+    //       probably through sub_start sub_end trickery.
+    public static function enqueueNewFeeds(array $args=array()) {
+        if (!isset($args['interval']) || !is_int($args['interval']) || $args['interval']<=0) {
+            $args['interval'] = self::DEFAULT_INTERVAL;
+        }
+
+        $args['interval'] *= 60;    // minutes to seconds
+
+        $feedsub = new FeedSub();
+        $feedsub->sub_state = 'nohub';
+        // Find feeds that haven't been polled within the desired interval,
+        // though perhaps we're abusing the "last_update" field here?
+        $feedsub->whereAdd(sprintf('last_update < "%s"', common_sql_date(time()-$args['interval'])));
+        $feedsub->find();
+
+        $qm = QueueManager::get();
+        while ($feedsub->fetch()) {
+            $orig = clone($feedsub);
+            $item = array('id' => $feedsub->id);
+            $qm->enqueue($item, self::QUEUE_CHECK);
+            common_debug('Enqueueing FeedPoll feeds, currently: '.$feedsub->uri);
+            $feedsub->last_update = common_sql_now();
+            $feedsub->update($orig);
+        }
+    }
+
+    public function setupFeedSub(FeedSub $feedsub, $interval=300)
+    {
+        $orig = clone($feedsub);
+        $feedsub->sub_state = 'nohub';
+        $feedsub->sub_start = common_sql_date(time());
+        $feedsub->sub_end   = '';
+        $feedsub->last_update = common_sql_date(time()-$interval);  // force polling as soon as we can
+        $feedsub->update($orig);
+    }
+
+    public function checkUpdates(FeedSub $feedsub)
+    {
+        $request = new HTTPClient();
+        common_debug('Enqueueing FeedPoll feeds went well, now checking updates for: '.$feedsub->getUri());
+        $feed = $request->get($feedsub->uri);
+        if (!$feed->isOk()) {
+            throw new ServerException('FeedSub could not fetch id='.$feedsub->id.' (Error '.$feed->getStatus().': '.$feed->getBody());
+        }
+        $feedsub->receive($feed->getBody(), null);
+    }
+}
diff --git a/plugins/FeedPoller/lib/feedpollqueuehandler.php b/plugins/FeedPoller/lib/feedpollqueuehandler.php
new file mode 100644
index 0000000000..c80253dcba
--- /dev/null
+++ b/plugins/FeedPoller/lib/feedpollqueuehandler.php
@@ -0,0 +1,41 @@
+<?php
+
+if (!defined('GNUSOCIAL')) { exit(1); }
+
+/**
+ * Poll a feed based on its urlhash, the full url is in the feedsub table
+ *
+ * @author Mikael Nordfeldth <mmn@hethane.se>
+ */
+class FeedPollQueueHandler extends QueueHandler
+{
+    public function transport()
+    {
+        return FeedPoll::QUEUE_CHECK;
+    }
+
+    public function handle($item)
+    {
+        common_debug('Enqueueing FeedPoll feeds but actually running the queue handler!');
+        $feedsub = FeedSub::getKV('id', $item['id']);
+        if (!$feedsub instanceof FeedSub) {
+            // Removed from the feedsub table I guess
+            return true;
+        }
+        if (!$feedsub->sub_state == 'nohub') {
+            // We're not supposed to poll this (either it's PuSH or it's unsubscribed)
+            return true;
+        }
+
+        common_debug('Enqueueing FeedPoll feeds but actually checking updates');
+        try {
+            FeedPoll::checkUpdates($feedsub);
+        } catch (Exception $e) {
+            common_log(LOG_ERR, "Failed to check feedsub id= ".$feedsub->id.' ("'.$e->getMessage().'")');
+        }
+
+        common_debug('Enqueueing FeedPoll feeds but actually done with '.$feedsub->id);
+
+        return true;
+    }
+}
diff --git a/plugins/FeedPoller/scripts/pollfeed.php b/plugins/FeedPoller/scripts/pollfeed.php
new file mode 100644
index 0000000000..3aa8296416
--- /dev/null
+++ b/plugins/FeedPoller/scripts/pollfeed.php
@@ -0,0 +1,71 @@
+#!/usr/bin/env php
+<?php
+/*
+ * StatusNet - a distributed open-source microblogging tool
+ * Copyright (C) 2010, StatusNet, Inc.
+ *
+ * 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/>.
+ */
+
+define('INSTALLDIR', realpath(dirname(__FILE__) . '/../../..'));
+
+$helptext = <<<END_OF_HELP
+pollfeed.php feeduri
+
+Poll the feed, assuming it has sub_state 'nohub'.
+
+END_OF_HELP;
+
+require_once INSTALLDIR.'/scripts/commandline.inc';
+
+require_once(__DIR__ . '/../lib/feedpoll.php');
+
+if (empty($args[0]) || !Validate::uri($args[0])) {
+    echo "$helptext\n";
+    exit(1);
+}
+
+$uri = $args[0];
+
+
+$feedsub = FeedSub::getKV('uri', $uri);
+
+if (!$feedsub instanceof FeedSub) {
+    echo "No FeedSub feed known for URI $uri\n";
+    exit(1);
+}
+
+if ($feedsub->sub_state != 'nohub') {
+    echo "Feed is a PuSH feed, so we will not poll it.\n";
+    exit(1);
+}
+
+showSub($feedsub);
+
+try {
+    FeedPoll::checkUpdates($feedsub);
+} catch (Exception $e) {
+    echo "Could not check updates for feed: ".$e->getMessage();
+    echo $e->getTraceAsString();
+    exit(1);
+}
+
+function showSub(FeedSub $sub)
+{
+    echo "  Subscription state: $sub->sub_state\n";
+    echo "  Signature secret: $sub->secret\n";
+    echo "  Sub start date: $sub->sub_start\n";
+    echo "  Record created: $sub->created\n";
+    echo "  Record modified: $sub->modified\n";
+}
diff --git a/plugins/OStatus/classes/FeedSub.php b/plugins/OStatus/classes/FeedSub.php
index 338c2f5144..25a826a552 100644
--- a/plugins/OStatus/classes/FeedSub.php
+++ b/plugins/OStatus/classes/FeedSub.php
@@ -435,7 +435,7 @@ class FeedSub extends Managed_DataObject
     {
         common_log(LOG_INFO, __METHOD__ . ": packet for \"" . $this->getUri() . "\"! $hmac $post");
 
-        if ($this->sub_state != 'active') {
+        if (!in_array($this->sub_state, array('active', 'nohub'))) {
             common_log(LOG_ERR, __METHOD__ . ": ignoring PuSH for inactive feed " . $this->getUri() . " (in state '$this->sub_state')");
             return;
         }