X-Git-Url: https://git.ao2.it/tweeper.git/blobdiff_plain/6c535395860b51d2c1debf92437192c6bf862b16..66097964b2c241e20eae5af87fb4a14fd2c1d889:/src/Tweeper.php?ds=inline

diff --git a/src/Tweeper.php b/src/Tweeper.php
index d58b529..7b2654d 100644
--- a/src/Tweeper.php
+++ b/src/Tweeper.php
@@ -6,7 +6,7 @@ namespace Tweeper;
  * @file
  * Tweeper - a Twitter to RSS web scraper.
  *
- * Copyright (C) 2013-2018  Antonio Ospite <ao2@ao2.it>
+ * Copyright (C) 2013-2020  Antonio Ospite <ao2@ao2.it>
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -36,7 +36,7 @@ date_default_timezone_set('UTC');
  */
 class Tweeper {
 
-  private static $userAgent = "Mozilla/5.0 (Windows NT 6.1; WOW64; rv:60.0) Gecko/20100101 Firefox/60.0";
+  private static $userAgent = "Mozilla/5.0 (Linux; U; Android 4.3; en-us; SM-N900T Build/JSS15J)";
   private static $maxConnectionTimeout = 5;
   private static $maxConnectionRetries = 5;
 
@@ -49,10 +49,18 @@ class Tweeper {
    *   Enables showing the username in front of the content for multi-user
    *   sites (enabled by default). Only some stylesheets supports this
    *   functionality (twitter, instagram, pump.io).
+   * @param bool $show_multimedia
+   *   Enables showing multimedia content (images, videos) directly in the
+   *   item description (enabled by default). Only some stylesheets supports
+   *   this functionality (twitter, instagram, dilbert).
+   * @param bool $verbose_output
+   *   Enables showing non-fatal errors like XML parsing errors.
    */
-  public function __construct($generate_enclosure = FALSE, $show_usernames = TRUE) {
+  public function __construct($generate_enclosure = FALSE, $show_usernames = TRUE, $show_multimedia = TRUE, $verbose_output = TRUE) {
     $this->generate_enclosure = $generate_enclosure;
     $this->show_usernames = $show_usernames;
+    $this->show_multimedia = $show_multimedia;
+    $this->verbose_output = $verbose_output;
   }
 
   /**
@@ -79,6 +87,46 @@ class Tweeper {
   }
 
   /**
+   * Convert Twitter mobile date to the date format expected in a RSS document.
+   */
+  public static function twitterToRssDate($date) {
+    // Twitter uses relative timestamps in minutes for recent tweets.
+    if (preg_match('/^(\d+)m$/', $date, $matches)) {
+      $timestamp = strtotime("+" . $matches[1] . " min", time());
+      if (FALSE === $timestamp) {
+        $timestamp = 0;
+      }
+    }
+    else {
+      /*
+       * In case the time is specified put it after the date,
+       * to make it recognized by strptime().
+       */
+      if (preg_match('/(.*) - (.*)/', $date, $matches)) {
+        $date = $matches[2] . " " . $matches[1];
+      }
+
+      $timestamp = strtotime($date);
+      if (FALSE === $timestamp) {
+        $timestamp = 0;
+      }
+
+      /*
+       * The twitter mobile UI usually only specifies the month and the day, so
+       * strtotime($date) may interpret the date as being in the future.
+       *
+       * If the date is in the future it is probably in the same day but in the
+       * previous year.
+       */
+      if ($timestamp > time()) {
+        $timestamp = strtotime('-1 years', $timestamp);
+      }
+    }
+
+    return Tweeper::epochToRssDate($timestamp);
+  }
+
+  /**
    * Convert string to UpperCamelCase.
    */
   public static function toUpperCamelCase($str, $delim = ' ') {
@@ -116,9 +164,8 @@ class Tweeper {
       CURLOPT_CONNECTTIMEOUT => Tweeper::$maxConnectionTimeout,
       // Follow http redirects to get the real URL.
       CURLOPT_FOLLOWLOCATION => TRUE,
+      CURLOPT_COOKIEFILE => "",
       CURLOPT_RETURNTRANSFER => TRUE,
-      CURLOPT_SSL_VERIFYHOST => FALSE,
-      CURLOPT_SSL_VERIFYPEER => FALSE,
       CURLOPT_HTTPHEADER => array('Accept-language: en'),
       CURLOPT_USERAGENT => Tweeper::$userAgent,
     ));
@@ -140,8 +187,6 @@ class Tweeper {
       // Follow http redirects to get the real URL.
       CURLOPT_FOLLOWLOCATION => TRUE,
       CURLOPT_RETURNTRANSFER => TRUE,
-      CURLOPT_SSL_VERIFYHOST => FALSE,
-      CURLOPT_SSL_VERIFYPEER => FALSE,
       CURLOPT_USERAGENT => Tweeper::$userAgent,
     ));
 
@@ -201,10 +246,17 @@ class Tweeper {
     // See http://sourceforge.net/p/feedvalidator/bugs/72/
     $http_url = preg_replace("/^https/", "http", $url_info['url']);
 
+    // When the server does not provide a Content-Length header,
+    // curl_getinfo() would return a negative value for
+    // "download_content_length", however RSS recommends to use 0 when the
+    // enclosure's size cannot be determined.
+    // See: https://www.feedvalidator.org/docs/error/UseZeroForUnknown.html
+    $length = max($url_info['download_content_length'], 0);
+
     $dom = new DOMDocument();
     $enc = $dom->createElement('enclosure');
     $enc->setAttribute('url', $http_url);
-    $enc->setAttribute('length', $url_info['download_content_length']);
+    $enc->setAttribute('length', $length);
     $enc->setAttribute('type', $url_info['content_type']);
 
     return $enc;
@@ -330,8 +382,10 @@ class Tweeper {
       $xmlDoc->loadHTML($html);
     }
 
-    foreach (libxml_get_errors() as $xml_error) {
-      Tweeper::logXmlError($xml_error);
+    if ($this->verbose_output) {
+      foreach (libxml_get_errors() as $xml_error) {
+        Tweeper::logXmlError($xml_error);
+      }
     }
     libxml_clear_errors();
     libxml_use_internal_errors($xml_errors_value);
@@ -362,6 +416,7 @@ class Tweeper {
     $xsltProcessor->registerPHPFunctions();
     $xsltProcessor->setParameter('', 'generate-enclosure', $this->generate_enclosure);
     $xsltProcessor->setParameter('', 'show-usernames', $this->show_usernames);
+    $xsltProcessor->setParameter('', 'show-multimedia', $this->show_multimedia);
     $xsltProcessor->importStylesheet($xslDoc);
 
     return $xsltProcessor;
@@ -370,7 +425,7 @@ class Tweeper {
   /**
    * Convert the site content to RSS.
    */
-  public function tweep($src_url, $host=NULL, $validate_scheme=TRUE) {
+  public function tweep($src_url, $host = NULL, $validate_scheme = TRUE) {
     $url = parse_url($src_url);
     if (FALSE === $url) {
       trigger_error("Invalid URL: $src_url", E_USER_WARNING);
@@ -385,7 +440,7 @@ class Tweeper {
       }
     }
 
-    // if the host is not given derive it from the URL
+    // If the host is not given derive it from the URL.
     if (NULL === $host) {
       if (empty($url["host"])) {
         trigger_error("Invalid host in URL: $src_url", E_USER_WARNING);