Add a persistent 'shorturl' key to all links

All existing link will keep their permalinks. New links will have smallhash generated with date+id. The purpose of this is to avoid collision between links due to their creation date.
shaarli · Dec 12, 2016 · cd484e7 · cd484e7
1 parent 8e7c729
commit cd484e7
Show file tree

Hide file tree

Showing 15 changed files with 115 additions and 91 deletions.
diff --git a/application/FeedBuilder.php b/application/FeedBuilder.php
@@ -143,7 +143,7 @@ public function buildData()
      */
     protected function buildItem($link, $pageaddr)
     {
-        $link['guid'] = $pageaddr .'?'. smallHash($link['created']->format('Ymd_His'));
+        $link['guid'] = $pageaddr .'?'. $link['shorturl'];
         // Check for both signs of a note: starting with ? and 7 chars long.
         if ($link['url'][0] === '?' && strlen($link['url']) === 7) {
             $link['url'] = $pageaddr . $link['url'];

diff --git a/application/LinkDB.php b/application/LinkDB.php
@@ -22,6 +22,7 @@
  *              Can be absolute or relative.
  *              Relative URLs are permalinks (e.g.'?m-ukcw')
  *  - real_url  Absolute processed URL.
+ *  - shorturl  Permalink smallhash
  *
  * Implements 3 interfaces:
  *  - ArrayAccess: behaves like an associative array;
@@ -264,6 +265,7 @@ private function check()
             'created'=> new DateTime(),
             'tags'=>'opensource software'
         );
+        $link['shorturl'] = link_small_hash($link['created'], $link['id']);
         $this->links[1] = $link;
 
         $link = array(
@@ -273,8 +275,9 @@ private function check()
             'description'=>'Shhhh! I\'m a private link only YOU can see. You can delete me too.',
             'private'=>1,
             'created'=> new DateTime('1 minute ago'),
-            'tags'=>'secretstuff'
+            'tags'=>'secretstuff',
         );
+        $link['shorturl'] = link_small_hash($link['created'], $link['id']);
         $this->links[0] = $link;
 
         // Write database to disk
@@ -335,10 +338,11 @@ private function read()
             // To be able to load links before running the update, and prepare the update
             if (! isset($link['created'])) {
                 $link['id'] = $link['linkdate'];
-                $link['created'] = DateTime::createFromFormat('Ymd_His', $link['linkdate']);
+                $link['created'] = DateTime::createFromFormat(self::LINK_DATE_FORMAT, $link['linkdate']);
                 if (! empty($link['updated'])) {
-                    $link['updated'] = DateTime::createFromFormat('Ymd_His', $link['updated']);
+                    $link['updated'] = DateTime::createFromFormat(self::LINK_DATE_FORMAT, $link['updated']);
                 }
+                $link['shorturl'] = smallHash($link['linkdate']);
             }
         }
 
@@ -558,7 +562,7 @@ public function getNextId()
      *
      * @param int $id Persistent ID of a link.
      *
-     * @return int Real offset in local array, or null if doesn't exists.
+     * @return int Real offset in local array, or null if doesn't exist.
      */
     protected function getLinkOffset($id)
     {

diff --git a/application/LinkFilter.php b/application/LinkFilter.php
@@ -120,7 +120,7 @@ private function filterSmallHash($smallHash)
     {
         $filtered = array();
         foreach ($this->links as $key => $l) {
-            if ($smallHash == smallHash($l['created']->format('Ymd_His'))) {
+            if ($smallHash == $l['shorturl']) {
                 // Yes, this is ugly and slow
                 $filtered[$key] = $l;
                 return $filtered;

diff --git a/application/LinkUtils.php b/application/LinkUtils.php
@@ -169,3 +169,16 @@ function space2nbsp($text)
 function format_description($description, $redirector = '', $indexUrl = '') {
     return nl2br(space2nbsp(hashtag_autolink(text2clickable($description, $redirector), $indexUrl)));
 }
+
+/**
+ * Generate a small hash for a link.
+ *
+ * @param DateTime $date Link creation date.
+ * @param int      $id   Link ID.
+ *
+ * @return string the small hash generated from link data.
+ */
+function link_small_hash($date, $id)
+{
+    return smallHash($date->format(LinkDB::LINK_DATE_FORMAT) . $id);
+}
diff --git a/application/NetscapeBookmarkUtils.php b/application/NetscapeBookmarkUtils.php
@@ -174,6 +174,7 @@ public static function import($post, $files, $linkDb, $pagecache)
             $newLinkDate->setTimezone(new DateTimeZone(date_default_timezone_get()));
             $newLink['created'] = $newLinkDate;
             $newLink['id'] = $linkDb->getNextId();
+            $newLink['shorturl'] = link_small_hash($newLink['created'], $newLink['id']);
             $linkDb[$newLink['id']] = $newLink;
             $importCount++;
         }

diff --git a/application/Updater.php b/application/Updater.php
@@ -223,6 +223,9 @@ public function updateMethodEscapeUnescapedConfig()
      * Since this update is very sensitve (changing the whole database), the datastore will be
      * automatically backed up into the file datastore.<datetime>.php.
      *
+     * LinkDB also adds the field 'shorturl' with the precedent format (linkdate smallhash),
+     * which will be saved by this method.
+     *
      * @return bool true if the update is successful, false otherwise.
      */
     public function updateMethodDatastoreIds()

diff --git a/application/Utils.php b/application/Utils.php
@@ -31,7 +31,11 @@ function logm($logFile, $clientIp, $message)
  *   - are NOT cryptographically secure (they CAN be forged)
  *
  *  In Shaarli, they are used as a tinyurl-like link to individual entries,
- *  e.g. smallHash('20111006_131924') --> yZH23w
+ *  built once with the combination of the date and item ID.
+ *  e.g. smallHash('20111006_131924' . 142) --> eaWxtQ
+ *
+ * @warning before v0.8.1, smallhashes were built only with the date,
+ *          and their value has been preserved.
  *
  * @param string $text Create a hash from this text.
  *

diff --git a/index.php b/index.php
@@ -566,21 +566,17 @@ function showDailyRSS($conf) {
     /* Some Shaarlies may have very few links, so we need to look
        back in time until we have enough days ($nb_of_days).
     */
-    $ids = array();
-    foreach ($LINKSDB as $id => $value) {
-        $ids[] = $id;
-    }
     $nb_of_days = 7; // We take 7 days.
     $today = date('Ymd');
     $days = array();
 
-    foreach ($ids as $id) {
-        $day = $LINKSDB[$id]['created']->format('Ymd'); // Extract day (without time)
+    foreach ($LINKSDB as $link) {
+        $day = $link['created']->format('Ymd'); // Extract day (without time)
         if (strcmp($day, $today) < 0) {
             if (empty($days[$day])) {
                 $days[$day] = array();
             }
-            $days[$day][] = $id;
+            $days[$day][] = $link;
         }
 
         if (count($days) > $nb_of_days) {
@@ -600,23 +596,18 @@ function showDailyRSS($conf) {
     echo '<copyright>'. $pageaddr .'</copyright>'. PHP_EOL;
 
     // For each day.
-    foreach ($days as $day => $ids) {
+    foreach ($days as $day => $links) {
         $dayDate = DateTime::createFromFormat(LinkDB::LINK_DATE_FORMAT, $day.'_000000');
         $absurl = escape(index_url($_SERVER).'?do=daily&day='.$day);  // Absolute URL of the corresponding "Daily" page.
 
-        // Build the HTML body of this RSS entry.
-        $links = array();
-
         // We pre-format some fields for proper output.
-        foreach ($ids as $id) {
-            $l = $LINKSDB[$id];
-            $l['formatedDescription'] = format_description($l['description'], $conf->get('redirector.url'));
-            $l['thumbnail'] = thumbnail($conf, $l['url']);
-            $l['timestamp'] = $l['created']->getTimestamp();
-            if (startsWith($l['url'], '?')) {
-                $l['url'] = index_url($_SERVER) . $l['url'];  // make permalink URL absolute
+        foreach ($links as &$link) {
+            $link['formatedDescription'] = format_description($link['description'], $conf->get('redirector.url'));
+            $link['thumbnail'] = thumbnail($conf, $link['url']);
+            $link['timestamp'] = $link['created']->getTimestamp();
+            if (startsWith($link['url'], '?')) {
+                $link['url'] = index_url($_SERVER) . $link['url'];  // make permalink URL absolute
             }
-            $links[$id] = $l;
         }
 
         // Then build the HTML for this day:
@@ -675,7 +666,6 @@ function showDaily($pageBuilder, $LINKSDB, $conf, $pluginManager)
 
         $taglist = explode(' ',$link['tags']);
         uasort($taglist, 'strcasecmp');
-        $linksToDisplay[$key]['shorturl'] = smallHash($link['created']->format('Ymd_His'));
         $linksToDisplay[$key]['taglist']=$taglist;
         $linksToDisplay[$key]['formatedDescription'] = format_description($link['description'], $conf->get('redirector.url'));
         $linksToDisplay[$key]['thumbnail'] = thumbnail($conf, $link['url']);
@@ -829,7 +819,7 @@ function renderPage($conf, $pluginManager)
         // Get only links which have a thumbnail.
         foreach($links as $link)
         {
-            $permalink='?'.escape(smallHash($link['created']->format('Ymd_His')));
+            $permalink='?'.$link['shorturl'];
             $thumb=lazyThumbnail($conf, $link['url'],$permalink);
             if ($thumb!='') // Only output links which have a thumbnail.
             {
@@ -1248,19 +1238,19 @@ function renderPage($conf, $pluginManager)
         }
 
         // lf_id should only be present if the link exists.
-        $id = !empty($_POST['lf_id']) ? (int) escape($_POST['lf_id']) : $LINKSDB->getNextId();
+        $id = !empty($_POST['lf_id']) ? intval(escape($_POST['lf_id'])) : $LINKSDB->getNextId();
         // Linkdate is kept here to:
         //   - use the same permalink for notes as they're displayed when creating them
         //   - let users hack creation date of their posts
         //     See: https://github.com/shaarli/Shaarli/wiki/Datastore-hacks#changing-the-timestamp-for-a-link
         $linkdate = escape($_POST['lf_linkdate']);
         if (isset($LINKSDB[$id])) {
             // Edit
-            $created = DateTime::createFromFormat('Ymd_His', $linkdate);
+            $created = DateTime::createFromFormat(LinkDB::LINK_DATE_FORMAT, $linkdate);
             $updated = new DateTime();
         } else {
             // New link
-            $created = DateTime::createFromFormat('Ymd_His', $linkdate);
+            $created = DateTime::createFromFormat(LinkDB::LINK_DATE_FORMAT, $linkdate);
             $updated = null;
         }
 
@@ -1287,7 +1277,8 @@ function renderPage($conf, $pluginManager)
             'private' => (isset($_POST['lf_private']) ? 1 : 0),
             'created' => $created,
             'updated' => $updated,
-            'tags' => str_replace(',', ' ', $tags)
+            'tags' => str_replace(',', ' ', $tags),
+            'shorturl' => link_small_hash($created, $id),
         );
 
         // If title is empty, use the URL as title.
@@ -1310,7 +1301,7 @@ function renderPage($conf, $pluginManager)
         $returnurl = !empty($_POST['returnurl']) ? $_POST['returnurl'] : '?';
         $location = generateLocation($returnurl, $_SERVER['HTTP_HOST'], array('addlink', 'post', 'edit_link'));
         // Scroll to the link which has been edited.
-        $location .= '#' . smallHash($created->format('Ymd_His'));
+        $location .= '#' . $link['shorturl'];
         // After saving the link, redirect to the page the user was on.
         header('Location: '. $location);
         exit;
@@ -1324,7 +1315,7 @@ function renderPage($conf, $pluginManager)
         $link = $LINKSDB[(int) escape($_POST['lf_id'])];
         $returnurl = ( isset($_POST['returnurl']) ? $_POST['returnurl'] : '?' );
         // Scroll to the link which has been edited.
-        $returnurl .= '#'.smallHash($link['created']->format('Ymd_His'));
+        $returnurl .= '#'. $link['shorturl'];
         $returnurl = generateLocation($returnurl, $_SERVER['HTTP_HOST'], array('addlink', 'post', 'edit_link'));
         header('Location: '.$returnurl); // After canceling, redirect to the page the user was on.
         exit;
@@ -1340,7 +1331,7 @@ function renderPage($conf, $pluginManager)
         // - we are protected from XSRF by the token.
 
         // FIXME! We keep `lf_linkdate` for consistency before a proper API. To be removed.
-        $id = isset($_POST['lf_id']) ? (int) escape($_POST['lf_id']) : (int) escape($_POST['lf_linkdate']);
+        $id = isset($_POST['lf_id']) ? intval(escape($_POST['lf_id'])) : intval(escape($_POST['lf_linkdate']));
 
         $pluginManager->executeHooks('delete_link', $LINKSDB[$id]);
 
@@ -1386,7 +1377,7 @@ function renderPage($conf, $pluginManager)
         $id = (int) escape($_GET['edit_link']);
         $link = $LINKSDB[$id];  // Read database
         if (!$link) { header('Location: ?'); exit; } // Link not found in database.
-        $link['linkdate'] = $link['created']->format('Ymd_His');
+        $link['linkdate'] = $link['created']->format(LinkDB::LINK_DATE_FORMAT);
         $data = array(
             'link' => $link,
             'link_is_new' => false,
@@ -1413,7 +1404,7 @@ function renderPage($conf, $pluginManager)
         if (! $link)
         {
             $link_is_new = true;
-            $linkdate = strval(date('Ymd_His'));
+            $linkdate = strval(date(LinkDB::LINK_DATE_FORMAT));
             // Get title if it was provided in URL (by the bookmarklet).
             $title = empty($_GET['title']) ? '' : escape($_GET['title']);
             // Get description if it was provided in URL (by the bookmarklet). [Bronco added that]
@@ -1437,7 +1428,7 @@ function renderPage($conf, $pluginManager)
             }
 
             if ($url == '') {
-                $url = '?' . smallHash($linkdate);
+                $url = '?' . smallHash($linkdate . $LINKSDB->getNextId());
                 $title = 'Note: ';
             }
             $url = escape($url);
@@ -1452,7 +1443,7 @@ function renderPage($conf, $pluginManager)
                 'private' => $private
             );
         } else {
-            $link['linkdate'] = $link['created']->format('Ymd_His');
+            $link['linkdate'] = $link['created']->format(LinkDB::LINK_DATE_FORMAT);
         }
 
         $data = array(
@@ -1667,7 +1658,6 @@ function buildLinkList($PAGE,$LINKSDB, $conf, $pluginManager)
         $taglist = explode(' ', $link['tags']);
         uasort($taglist, 'strcasecmp');
         $link['taglist'] = $taglist;
-        $link['shorturl'] = smallHash($link['created']->format('Ymd_His'));
         // Check for both signs of a note: starting with ? and 7 chars long.
         if ($link['url'][0] === '?' &&
             strlen($link['url']) === 7) {

diff --git a/plugins/isso/isso.php b/plugins/isso/isso.php
@@ -43,9 +43,7 @@ function hook_isso_render_linklist($data, $conf)
         $link = reset($data['links']);
         $issoHtml = file_get_contents(PluginManager::$PLUGINS_PATH . '/isso/isso.html');
 
-        // FIXME! KO thread unique si même date
-        $linkDate = $link['created']->format('Ymd_His');
-        $isso = sprintf($issoHtml, $issoUrl, $issoUrl, $linkDate, $linkDate);
+        $isso = sprintf($issoHtml, $issoUrl, $issoUrl, $link['id'], $link['id']);
         $data['plugin_end_zone'][] = $isso;
 
         // Hackish way to include this CSS file only when necessary.

diff --git a/tests/FeedBuilderTest.php b/tests/FeedBuilderTest.php
@@ -86,7 +86,7 @@ public function testRSSBuildData()
         // Test first link (note link)
         $link = reset($data['links']);
         $this->assertEquals(41, $link['id']);
-        $this->assertEquals(DateTime::createFromFormat('Ymd_His', '20150310_114651'), $link['created']);
+        $this->assertEquals(DateTime::createFromFormat(LinkDB::LINK_DATE_FORMAT, '20150310_114651'), $link['created']);
         $this->assertEquals('http://host.tld/?WDWyig', $link['guid']);
         $this->assertEquals('http://host.tld/?WDWyig', $link['url']);
         $this->assertRegExp('/Tue, 10 Mar 2015 11:46:51 \+\d{4}/', $link['pub_iso_date']);
@@ -140,7 +140,7 @@ public function testBuildDataFiltered()
         $this->assertEquals(1, count($data['links']));
         $link = array_shift($data['links']);
         $this->assertEquals(41, $link['id']);
-        $this->assertEquals(DateTime::createFromFormat('Ymd_His', '20150310_114651'), $link['created']);
+        $this->assertEquals(DateTime::createFromFormat(LinkDB::LINK_DATE_FORMAT, '20150310_114651'), $link['created']);
     }
 
     /**
@@ -157,7 +157,7 @@ public function testBuildDataCount()
         $this->assertEquals(1, count($data['links']));
         $link = array_shift($data['links']);
         $this->assertEquals(41, $link['id']);
-        $this->assertEquals(DateTime::createFromFormat('Ymd_His', '20150310_114651'), $link['created']);
+        $this->assertEquals(DateTime::createFromFormat(LinkDB::LINK_DATE_FORMAT, '20150310_114651'), $link['created']);
     }
 
     /**
@@ -174,15 +174,15 @@ public function testBuildDataPermalinks()
         // First link is a permalink
         $link = array_shift($data['links']);
         $this->assertEquals(41, $link['id']);
-        $this->assertEquals(DateTime::createFromFormat('Ymd_His', '20150310_114651'), $link['created']);
+        $this->assertEquals(DateTime::createFromFormat(LinkDB::LINK_DATE_FORMAT, '20150310_114651'), $link['created']);
         $this->assertEquals('http://host.tld/?WDWyig', $link['guid']);
         $this->assertEquals('http://host.tld/?WDWyig', $link['url']);
         $this->assertContains('Direct link', $link['description']);
         $this->assertContains('http://host.tld/?WDWyig', $link['description']);
         // Second link is a direct link
         $link = array_shift($data['links']);
         $this->assertEquals(8, $link['id']);
-        $this->assertEquals(DateTime::createFromFormat('Ymd_His', '20150310_114633'), $link['created']);
+        $this->assertEquals(DateTime::createFromFormat(LinkDB::LINK_DATE_FORMAT, '20150310_114633'), $link['created']);
         $this->assertEquals('http://host.tld/?RttfEw', $link['guid']);
         $this->assertEquals('https://static.fsf.org/nosvn/faif-2.0.pdf', $link['url']);
         $this->assertContains('Direct link', $link['description']);

diff --git a/tests/LinkDBTest.php b/tests/LinkDBTest.php
@@ -191,7 +191,7 @@ public function testSave()
             'url'=>'http://dum.my',
             'description'=>'One more',
             'private'=>0,
-            'created'=> DateTime::createFromFormat('Ymd_His', '20150518_190000'),
+            'created'=> DateTime::createFromFormat(LinkDB::LINK_DATE_FORMAT, '20150518_190000'),
             'tags'=>'unit test'
         );
         $testDB[$link['id']] = $link;
@@ -447,17 +447,17 @@ public function testFilterHashInValid()
      */
     public function testReorderLinksDesc()
     {
-        self::$publicLinkDB->reorder('ASC');
-        $linkIdToTest = 42;
-        foreach (self::$publicLinkDB as $key => $value) {
-            $this->assertEquals($linkIdToTest, $key);
-            break;
+        self::$privateLinkDB->reorder('ASC');
+        $linkIds = array(42, 4, 1, 0, 7, 6, 8, 41);
+        $cpt = 0;
+        foreach (self::$privateLinkDB as $key => $value) {
+            $this->assertEquals($linkIds[$cpt++], $key);
         }
-        self::$publicLinkDB->reorder('DESC');
-        $linkIdToTest = 41;
-        foreach (self::$publicLinkDB as $key => $value) {
-            $this->assertEquals($linkIdToTest, $key);
-            break;
+        self::$privateLinkDB->reorder('DESC');
+        $linkIds = array_reverse($linkIds);
+        $cpt = 0;
+        foreach (self::$privateLinkDB as $key => $value) {
+            $this->assertEquals($linkIds[$cpt++], $key);
         }
     }
 }