Merge pull request #7 from j0k3r/fix-nbsp

Avoid error with  

README.md

@@ -3,7 +3,7 @@
 [![Build Status](https://travis-ci.org/j0k3r/php-readability.svg?branch=master)](https://travis-ci.org/j0k3r/php-readability)
 [![Code Coverage](https://scrutinizer-ci.com/g/j0k3r/php-readability/badges/coverage.png?b=master)](https://scrutinizer-ci.com/g/j0k3r/php-readability/?branch=master)
 
-This is an extract of the Readability class from the [full-text-rss](https://github.com/Dither/full-text-rss) fork. It kind be defined as a better version of the original [php-readability](http://code.fivefilters.org/php-readability).
+This is an extract of the Readability class from the [full-text-rss](https://github.com/Dither/full-text-rss) fork. It kind be defined as a better version of the original [php-readability](https://bitbucket.org/fivefilters/php-readability/overview).
 
 ## Differences
 

src/JSLikeHTMLElement.php

@@ -47,7 +47,7 @@ class JSLikeHTMLElement extends \DOMElement
     {
         if ($name == 'innerHTML') {
             // first, empty the element
-            for ($x = $this->childNodes->length - 1; $x >= 0; $x--) {
+            for ($x = $this->childNodes->length - 1; $x >= 0; --$x) {
                 $this->removeChild($this->childNodes->item($x));
             }
             // $value holds our new inner HTML

src/Readability.php

@@ -63,6 +63,7 @@ class Readability
     protected $bodyCache = null; // Cache the body HTML in case we need to re-use it later
     protected $flags = 7; // 1 | 2 | 4;   // Start with all processing flags set.
     protected $success = false; // indicates whether we were able to extract or not
+
     /**
      * All of the regular expressions in use within readability.
      * Defined up here so we don't instantiate them repeatedly in loops.
@@ -88,9 +89,9 @@ class Readability
         'output-xhtml' => true,
         'logical-emphasis' => true,
         'show-body-only' => false,
-        'new-blocklevel-tags' => 'article,aside,audio,details,figcaption,figure,footer,header,hgroup,nav,section,source,summary,temp,track,video',
-        'new-empty-tags' => 'command,embed,keygen,source,track,wbr',
-        'new-inline-tags' => 'audio,canvas,command,datalist,embed,keygen,mark,meter,output,progress,time,video,wbr',
+        'new-blocklevel-tags' => 'article aside audio bdi canvas details dialog figcaption figure footer header hgroup main menu menuitem nav section source summary template track video',
+        'new-empty-tags' => 'command embed keygen source track wbr',
+        'new-inline-tags' => 'audio command datalist embed keygen mark menuitem meter output progress source time video wbr',
         'wrap' => 0,
         'drop-empty-paras' => true,
         'drop-proprietary-attributes' => false,
@@ -122,6 +123,7 @@ class Readability
         '!</code>\s*</pre>!is' => '</pre>',
         '!<[hb]r>!is' => '<\\1 />',
     );
+
     // flags
     const FLAG_STRIP_UNLIKELYS = 1;
     const FLAG_WEIGHT_ATTRIBUTES = 2;
@@ -137,6 +139,7 @@ class Readability
     const MIN_ARTICLE_LENGTH = 200;
     const MIN_NODE_LENGTH = 80;
     const MAX_LINK_DENSITY = 0.25;
+
     /**
      * Create instance of Readability.
      *
@@ -192,6 +195,7 @@ class Readability
 
         if (!($parser == 'html5lib' && ($this->dom = \HTML5_Parser::parse($html)))) {
             libxml_use_internal_errors(true);
+
             $this->dom = new \DOMDocument();
             $this->dom->preserveWhiteSpace = false;
 
@@ -206,6 +210,7 @@ class Readability
 
         $this->dom->registerNodeClass('DOMElement', 'Readability\JSLikeHTMLElement');
     }
+
     /**
      * Get article title element.
      *
@@ -215,6 +220,7 @@ class Readability
     {
         return $this->articleTitle;
     }
+
     /**
      * Get article content element.
      *
@@ -224,6 +230,7 @@ class Readability
     {
         return $this->articleContent;
     }
+
     /**
      * Add pre filter for raw input HTML processing.
      *
@@ -234,6 +241,7 @@ class Readability
     {
         $this->pre_filters[$filter] = $replacer;
     }
+
     /**
      * Add post filter for raw output HTML processing.
      *
@@ -244,6 +252,7 @@ class Readability
     {
         $this->post_filters[$filter] = $replacer;
     }
+
     /**
      * Runs readability.
      *
@@ -284,11 +293,11 @@ class Readability
         if (!$articleContent) {
             $this->success = false;
             $articleContent = $this->dom->createElement('div');
-            $articleContent->setAttribute('id', 'readability-content');
+            $articleContent->setAttribute('class', 'readability-content');
             $articleContent->innerHTML = '<p>Sorry, Readability was unable to parse this page for content.</p>';
         }
-        $overlay->setAttribute('id', 'readOverlay');
-        $innerDiv->setAttribute('id', 'readInner');
+        $overlay->setAttribute('class', 'readOverlay');
+        $innerDiv->setAttribute('class', 'readInner');
         // Glue the structure of our document together.
         $innerDiv->appendChild($articleTitle);
         $innerDiv->appendChild($articleContent);
@@ -305,6 +314,7 @@ class Readability
 
         return $this->success;
     }
+
     /**
      * Debug.
      */
@@ -325,6 +335,7 @@ class Readability
             syslog(6, $this->debugText); // 1 - error 6 - info
         }
     }
+
     /**
      * Run any post-process modifications to article content as necessary.
      *
@@ -336,6 +347,7 @@ class Readability
             $this->addFootnotes($articleContent);
         }
     }
+
     /**
      * Get the article title as an H1.
      *
@@ -349,6 +361,7 @@ class Readability
             $curTitle = $origTitle = $this->getInnerText($this->dom->getElementsByTagName('title')->item(0));
         } catch (Exception $e) {
         }
+
         if (preg_match('/ [\|\-] /', $curTitle)) {
             $curTitle = preg_replace('/(.*)[\|\-] .*/i', '$1', $origTitle);
             if (count(explode(' ', $curTitle)) < 3) {
@@ -365,15 +378,18 @@ class Readability
                 $curTitle = $this->getInnerText($hOnes->item(0));
             }
         }
+
         $curTitle = trim($curTitle);
         if (count(explode(' ', $curTitle)) <= 4) {
             $curTitle = $origTitle;
         }
+
         $articleTitle = $this->dom->createElement('h1');
         $articleTitle->innerHTML = $curTitle;
 
         return $articleTitle;
     }
+
     /**
      * Prepare the HTML document for readability to scrape it.
      * This includes things like stripping javascript, CSS, and handling terrible markup.
@@ -388,17 +404,18 @@ class Readability
             $this->body = $this->dom->createElement('body');
             $this->dom->documentElement->appendChild($this->body);
         }
-        $this->body->setAttribute('id', 'readabilityBody');
+        $this->body->setAttribute('class', 'readabilityBody');
         // Remove all style tags in head.
         $styleTags = $this->dom->getElementsByTagName('style');
-        for ($i = $styleTags->length - 1; $i >= 0; $i--) {
+        for ($i = $styleTags->length - 1; $i >= 0; --$i) {
             $styleTags->item($i)->parentNode->removeChild($styleTags->item($i));
         }
         $linkTags = $this->dom->getElementsByTagName('link');
-        for ($i = $linkTags->length - 1; $i >= 0; $i--) {
+        for ($i = $linkTags->length - 1; $i >= 0; --$i) {
             $linkTags->item($i)->parentNode->removeChild($linkTags->item($i));
         }
     }
+
     /**
      * For easier reading, convert this document to have footnotes at the bottom rather than inline links.
      *
@@ -407,14 +424,14 @@ class Readability
     public function addFootnotes($articleContent)
     {
         $footnotesWrapper = $this->dom->createElement('footer');
-        $footnotesWrapper->setAttribute('id', 'readability-footnotes');
+        $footnotesWrapper->setAttribute('class', 'readability-footnotes');
         $footnotesWrapper->innerHTML = '<h3>References</h3>';
         $articleFootnotes = $this->dom->createElement('ol');
-        $articleFootnotes->setAttribute('id', 'readability-footnotes-list');
+        $articleFootnotes->setAttribute('class', 'readability-footnotes-list');
         $footnotesWrapper->appendChild($articleFootnotes);
         $articleLinks = $articleContent->getElementsByTagName('a');
         $linkCount = 0;
-        for ($i = 0; $i < $articleLinks->length; $i++) {
+        for ($i = 0; $i < $articleLinks->length; ++$i) {
             $articleLink = $articleLinks->item($i);
             $footnoteLink = $articleLink->cloneNode(true);
             $refLink = $this->dom->createElement('a');
@@ -427,7 +444,7 @@ class Readability
             if ((strpos($articleLink->getAttribute('class'), 'readability-DoNotFootnote') !== false) || preg_match($this->regexps['skipFootnoteLink'], $linkText)) {
                 continue;
             }
-            $linkCount++;
+            ++$linkCount;
             // Add a superscript reference after the article link.
             $refLink->setAttribute('href', '#readabilityFootnoteLink-'.$linkCount);
             $refLink->innerHTML = '<small><sup>['.$linkCount.']</sup></small>';
@@ -453,6 +470,7 @@ class Readability
             $articleContent->appendChild($footnotesWrapper);
         }
     }
+
     /**
      * Prepare the article node for display. Clean out any inline styles,
      * iframes, forms, strip extraneous <p> tags, etc.
@@ -475,20 +493,20 @@ class Readability
              * which is what they were before.
              */
             $elems = $xpath->query('.//p[@data-readability-styled]', $articleContent);
-            for ($i = $elems->length - 1; $i >= 0; $i--) {
+            for ($i = $elems->length - 1; $i >= 0; --$i) {
                 $e = $elems->item($i);
                 $e->parentNode->replaceChild($articleContent->ownerDocument->createTextNode($e->textContent), $e);
             }
         }
         // Remove service data-candidate attribute.
         $elems = $xpath->query('.//*[@data-candidate]', $articleContent);
-        for ($i = $elems->length - 1; $i >= 0; $i--) {
+        for ($i = $elems->length - 1; $i >= 0; --$i) {
             $elems->item($i)->removeAttribute('data-candidate');
         }
         // Remove unrelated links and other unneded stuff.
         // (not(*) and not(text()[normalize-space()])) or // What's wrong here?
         $elems = $xpath->query('.//a[@rel="nofollow"]', $articleContent);
-        for ($i = $elems->length - 1; $i >= 0; $i--) {
+        for ($i = $elems->length - 1; $i >= 0; --$i) {
             $elems->item($i)->parentNode->removeChild($elems->item($i));
         }
         // Clean out junk from the article content.
@@ -517,7 +535,7 @@ class Readability
         $this->cleanConditionally($articleContent, 'div');
         // Remove extra paragraphs.
         $articleParagraphs = $articleContent->getElementsByTagName('p');
-        for ($i = $articleParagraphs->length - 1; $i >= 0; $i--) {
+        for ($i = $articleParagraphs->length - 1; $i >= 0; --$i) {
             $imgCount = $articleParagraphs->item($i)->getElementsByTagName('img')->length;
             $embedCount = $articleParagraphs->item($i)->getElementsByTagName('embed')->length;
             $objectCount = $articleParagraphs->item($i)->getElementsByTagName('object')->length;
@@ -530,7 +548,7 @@ class Readability
             // add extra text to iframe tag to avoid an auto-closing iframe and then break the html code
             if ($iframeCount) {
                 $iframe = $articleParagraphs->item($i)->getElementsByTagName('iframe');
-                $iframe->item(0)->nodeValue = '&nbsp;';
+                $iframe->item(0)->nodeValue = ' ';
 
                 $articleParagraphs->item($i)->parentNode->replaceChild($iframe->item(0), $articleParagraphs->item($i));
             }
@@ -547,6 +565,7 @@ class Readability
             }
         }
     }
+
     /**
      * Initialize a node with the readability object. Also checks the
      * className/id for special names to add to its score.
@@ -614,6 +633,7 @@ class Readability
         }
         $readability->value += $this->getWeight($node);
     }
+
     /**
      * grabArticle - Using a variety of metrics (content score, classname, element types), find the content that is
      * most likely to be the stuff a user wants to read. Then return it wrapped up in a div.
@@ -631,7 +651,7 @@ class Readability
             $xpath = new \DOMXPath($page);
         }
         $allElements = $page->getElementsByTagName('*');
-        for ($nodeIndex = 0; ($node = $allElements->item($nodeIndex)); $nodeIndex++) {
+        for ($nodeIndex = 0; ($node = $allElements->item($nodeIndex)); ++$nodeIndex) {
             $tagName = $node->tagName;
             // Some well known site uses sections as paragraphs.
             if (strcasecmp($tagName, 'p') === 0 || strcasecmp($tagName, 'td') === 0 || strcasecmp($tagName, 'section') === 0) {
@@ -649,14 +669,14 @@ class Readability
                         //$newNode->setAttribute('class', $node->getAttribute('class'));
                         //$newNode->setAttribute('id', $node->getAttribute('id'));
                         $node = $node->parentNode->replaceChild($newNode, $node);
-                        $nodeIndex--;
+                        --$nodeIndex;
                         $nodesToScore[] = $newNode;
                     } catch (Exception $e) {
                         $this->dbg('Could not alter div/article to p, reverting back to div: '.$e->getMessage());
                     }
                 } else {
                     // Will change these P elements back to text nodes after processing.
-                    for ($i = 0, $il = $node->childNodes->length; $i < $il; $i++) {
+                    for ($i = 0, $il = $node->childNodes->length; $i < $il; ++$i) {
                         $childNode = $node->childNodes->item($i);
                         if (is_object($childNode) && get_class($childNode) === 'DOMProcessingInstruction') { //executable tags (<?php or <?xml) warning
                             $childNode->parentNode->removeChild($childNode);
@@ -680,7 +700,7 @@ class Readability
          * A score is determined by things like number of commas, class names, etc.
          * Maybe eventually link density.
          */
-        for ($pt = 0, $scored = count($nodesToScore); $pt < $scored; $pt++) {
+        for ($pt = 0, $scored = count($nodesToScore); $pt < $scored; ++$pt) {
             $parentNode = $nodesToScore[$pt]->parentNode;
             // No parent node? Move on...
             if (!$parentNode) {
@@ -735,7 +755,7 @@ class Readability
          */
         if ($this->flagIsActive(self::FLAG_STRIP_UNLIKELYS) && $xpath) {
             $candidates = $xpath->query('.//*[(self::footer and count(//footer)<2) or (self::aside and count(//aside)<2)]', $page->documentElement);
-            for ($node = null, $c = $candidates->length - 1; $c >= 0; $c--) {
+            for ($node = null, $c = $candidates->length - 1; $c >= 0; --$c) {
                 $node = $candidates->item($c);
                 // node should be readable but not inside of an article otherwise it's probably non-readable block
                 if ($node->hasAttribute('readability') && (int) $node->getAttributeNode('readability')->value < 40 && ($node->parentNode ? strcasecmp($node->parentNode->tagName, 'article') !== 0 : true)) {
@@ -744,7 +764,7 @@ class Readability
                 }
             }
             $candidates = $xpath->query('.//*[not(self::body) and (@class or @id or @style) and ((number(@readability) < 40) or not(@readability))]', $page->documentElement);
-            for ($node = null, $c = $candidates->length - 1; $c >= 0; $c--) {
+            for ($node = null, $c = $candidates->length - 1; $c >= 0; --$c) {
                 $node = $candidates->item($c);
                 $tagName = $node->tagName;
                 /* Remove unlikely candidates */
@@ -756,7 +776,7 @@ class Readability
                 ) {
                     $this->dbg('Removing unlikely candidate '.$node->getNodePath().' by "'.$unlikelyMatchString.'" with readability '.($node->hasAttribute('readability') ? (int) $node->getAttributeNode('readability')->value : 0));
                     $node->parentNode->removeChild($node);
-                    $nodeIndex--;
+                    --$nodeIndex;
                 }
             }
             unset($candidates);
@@ -769,7 +789,7 @@ class Readability
         if ($xpath) {
             // Using array of DOMElements after deletion is a path to DOOMElement.
             $candidates = $xpath->query('.//*[@data-candidate]', $page->documentElement);
-            for ($c = $candidates->length - 1; $c >= 0; $c--) {
+            for ($c = $candidates->length - 1; $c >= 0; --$c) {
                 // Scale the final candidates score based on link density. Good content should have a
                 // relatively small link density (5% or less) and be mostly unaffected by this operation.
                 // If not for this we would have used XPath to find maximum @readability.
@@ -796,6 +816,7 @@ class Readability
                     $this->dbg('Setting body to a raw HTML of original page!');
                     $topCandidate->innerHTML = $page->documentElement->innerHTML;
                     $page->documentElement->innerHTML = '';
+                    $this->reinitBody();
                     $page->documentElement->appendChild($topCandidate);
                 }
             } else {
@@ -822,14 +843,14 @@ class Readability
          * Things like preambles, content split by ads that we removed, etc.
          */
         $articleContent = $this->dom->createElement('div');
-        $articleContent->setAttribute('id', 'readability-content');
+        $articleContent->setAttribute('class', 'readability-content');
         $siblingScoreThreshold = max(10, ((int) $topCandidate->getAttribute('readability')) * 0.2);
         $siblingNodes = $topCandidate->parentNode->childNodes;
         if (!isset($siblingNodes)) {
             $siblingNodes = new stdClass();
             $siblingNodes->length = 0;
         }
-        for ($s = 0, $sl = $siblingNodes->length; $s < $sl; $s++) {
+        for ($s = 0, $sl = $siblingNodes->length; $s < $sl; ++$s) {
             $siblingNode = $siblingNodes->item($s);
             $siblingNodeName = $siblingNode->nodeName;
             $append = false;
@@ -864,19 +885,22 @@ class Readability
                     $this->dbg('Altering siblingNode '.$siblingNodeName.' to div.');
                     $nodeToAppend = $this->dom->createElement('div');
                     try {
+                        if ($siblingNode->getAttribute('id')) {
                             $nodeToAppend->setAttribute('id', $siblingNode->getAttribute('id'));
+                        }
+
                         $nodeToAppend->setAttribute('alt', $siblingNodeName);
                         $nodeToAppend->innerHTML = $siblingNode->innerHTML;
                     } catch (Exception $e) {
                         $this->dbg('Could not alter siblingNode '.$siblingNodeName.' to div, reverting to original.');
                         $nodeToAppend = $siblingNode;
-                        $s--;
-                        $sl--;
+                        --$s;
+                        --$sl;
                     }
                 } else {
                     $nodeToAppend = $siblingNode;
-                    $s--;
-                    $sl--;
+                    --$s;
+                    --$sl;
                 }
                 // To ensure a node does not interfere with readability styles, remove its classnames & ids.
                 // Now done via RegExp post_filter.
@@ -896,10 +920,8 @@ class Readability
          * finding the -right- content.
          */
         if (mb_strlen($this->getInnerText($articleContent, false)) < self::MIN_ARTICLE_LENGTH) {
-            if (!$this->body->hasChildNodes()) {
-                $this->body = $this->dom->createElement('body');
-            }
-            $this->body->innerHTML = $this->bodyCache;
+            $this->reinitBody();
+
             if ($this->flagIsActive(self::FLAG_STRIP_UNLIKELYS)) {
                 $this->removeFlag(self::FLAG_STRIP_UNLIKELYS);
                 $this->dbg('...content is shorter than '.self::MIN_ARTICLE_LENGTH." letters, trying not to strip unlikely content.\n");
@@ -922,6 +944,7 @@ class Readability
 
         return $articleContent;
     }
+
     /**
      * Get the inner text of a node.
      * This also strips out any excess whitespace to be found.
@@ -946,6 +969,7 @@ class Readability
 
         return $textContent;
     }
+
     /**
      * Remove the style attribute on every $e and under.
      *
@@ -961,6 +985,7 @@ class Readability
             $elem->removeAttribute('style');
         }
     }
+
     /**
      * Get comma number for a given text.
      *
@@ -972,6 +997,7 @@ class Readability
     {
         return substr_count($text, ',');
     }
+
     /**
      * Get words number for a given text if words separated by a space.
      * Input string should be normalized.
@@ -984,6 +1010,7 @@ class Readability
     {
         return substr_count($text, ' ');
     }
+
     /**
      * Get the density of links as a percentage of the content
      * This is the amount of text that is inside a link divided by the total text in the node.
@@ -999,7 +1026,7 @@ class Readability
         $links = $e->getElementsByTagName('a');
         $textLength = mb_strlen($this->getInnerText($e, true, true));
         $linkLength = 0;
-        for ($dRe = $this->domainRegExp, $i = 0, $il = $links->length; $i < $il; $i++) {
+        for ($dRe = $this->domainRegExp, $i = 0, $il = $links->length; $i < $il; ++$i) {
             if ($excludeExternal && $dRe && !preg_match($dRe, $links->item($i)->getAttribute('href'))) {
                 continue;
             }
@@ -1011,6 +1038,7 @@ class Readability
             return 0;
         }
     }
+
     /**
      * Get an element weight by attribute.
      * Uses regular expressions to tell if this element looks good or bad.
@@ -1045,6 +1073,7 @@ class Readability
 
         return $weight;
     }
+
     /**
      * Get an element relative weight.
      *
@@ -1065,6 +1094,7 @@ class Readability
 
         return $weight;
     }
+
     /**
      * Remove extraneous break tags from a node.
      *
@@ -1076,6 +1106,7 @@ class Readability
         $html = preg_replace($this->regexps['killBreaks'], '<br />', $html);
         $node->innerHTML = $html;
     }
+
     /**
      * Clean a node of all elements of type "tag".
      * (Unless it's a youtube/vimeo video. People love movies.).
@@ -1089,7 +1120,7 @@ class Readability
     {
         $targetList = $e->getElementsByTagName($tag);
         $isEmbed = ($tag === 'audio' || $tag === 'video' || $tag === 'iframe' || $tag === 'object' || $tag === 'embed');
-        for ($cur_item = null, $y = $targetList->length - 1; $y >= 0; $y--) {
+        for ($cur_item = null, $y = $targetList->length - 1; $y >= 0; --$y) {
             /* Allow youtube and vimeo videos through as people usually want to see those. */
             $cur_item = $targetList->item($y);
             if ($isEmbed) {
@@ -1106,6 +1137,7 @@ class Readability
             $cur_item->parentNode->removeChild($cur_item);
         }
     }
+
     /**
      * Clean an element of all tags of type "tag" if they look fishy.
      * "Fishy" is an algorithm based on content length, classnames,
@@ -1127,7 +1159,7 @@ class Readability
          *
          * TODO: Consider taking into account original contentScore here.
          */
-        for ($node = null, $i = $curTagsLength - 1; $i >= 0; $i--) {
+        for ($node = null, $i = $curTagsLength - 1; $i >= 0; --$i) {
             $node = $tagsList->item($i);
             //$class = $node->getAttribute('class').' '.$node->getAttribute('id'); //debug
             $weight = $this->getWeight($node);
@@ -1148,15 +1180,15 @@ class Readability
                 $a = $node->getElementsByTagName('a')->length;
                 $embedCount = 0;
                 $embeds = $node->getElementsByTagName('embed');
-                for ($ei = 0, $il = $embeds->length; $ei < $il; $ei++) {
+                for ($ei = 0, $il = $embeds->length; $ei < $il; ++$ei) {
                     if (preg_match($this->regexps['media'], $embeds->item($ei)->getAttribute('src'))) {
-                        $embedCount++;
+                        ++$embedCount;
                     }
                 }
                 $embeds = $node->getElementsByTagName('iframe');
-                for ($ei = 0, $il = $embeds->length; $ei < $il; $ei++) {
+                for ($ei = 0, $il = $embeds->length; $ei < $il; ++$ei) {
                     if (preg_match($this->regexps['media'], $embeds->item($ei)->getAttribute('src'))) {
-                        $embedCount++;
+                        ++$embedCount;
                     }
                 }
                 $linkDensity = $this->getLinkDensity($node, true);
@@ -1192,8 +1224,8 @@ class Readability
                     } elseif ($input > floor($p / 3)) {
                         $this->dbg('  too many <input> elements');
                         $toRemove = true;
-                    } elseif ($contentLength < 25 && ($img === 0 || $img > 2)) {
-                        $this->dbg('  content length less than 25 chars and 0 images, or more than 2 images');
+                    } elseif ($contentLength < 10 && ($img === 0 || $img > 2)) {
+                        $this->dbg('  content length less than 10 chars and 0 images, or more than 2 images');
                         $toRemove = true;
                     } elseif ($weight < 25 && $linkDensity > 0.2) {
                         $this->dbg('  weight is '.$weight.' lower than 0 and link density is '.sprintf('%.2f', $linkDensity).' > 0.2');
@@ -1214,6 +1246,7 @@ class Readability
             }
         }
     }
+
     /**
      * Clean out spurious headers from an Element. Checks things like classnames and link density.
      *
@@ -1221,25 +1254,39 @@ class Readability
      */
     public function cleanHeaders($e)
     {
-        for ($headerIndex = 1; $headerIndex < 3; $headerIndex++) {
+        for ($headerIndex = 1; $headerIndex < 3; ++$headerIndex) {
             $headers = $e->getElementsByTagName('h'.$headerIndex);
-            for ($i = $headers->length - 1; $i >= 0; $i--) {
+            for ($i = $headers->length - 1; $i >= 0; --$i) {
                 if ($this->getWeight($headers->item($i)) < 0 || $this->getLinkDensity($headers->item($i)) > 0.33) {
                     $headers->item($i)->parentNode->removeChild($headers->item($i));
                 }
             }
         }
     }
+
     public function flagIsActive($flag)
     {
         return ($this->flags & $flag) > 0;
     }
+
     public function addFlag($flag)
     {
         $this->flags = $this->flags | $flag;
     }
+
     public function removeFlag($flag)
     {
         $this->flags = $this->flags & ~$flag;
     }
+
+    /**
+     * Will recreate previously deleted body property.
+     */
+    protected function reinitBody()
+    {
+        if (!isset($this->body->childNodes)) {
+            $this->body = $this->dom->createElement('body');
+            $this->body->innerHTML = $this->bodyCache;
+        }
+    }
 }

tests/ReadabilityTest.php

@@ -48,6 +48,8 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertFalse($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('Sorry, Readability was unable to parse this page for content.', $readability->getContent()->innerHTML);
     }
 
@@ -58,7 +60,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('<div readability=', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is the awesome content :)', $readability->getContent()->innerHTML);
     }
 
@@ -69,7 +73,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('<div readability=', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is the awesome content :)', $readability->getContent()->innerHTML);
     }
 
@@ -81,7 +87,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('<div readability=', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is the awesome content :)', $readability->getContent()->innerHTML);
     }
 
@@ -94,7 +102,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('<div readability=', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
         $this->assertContains('readabilityFootnoteLink', $readability->getContent()->innerHTML);
         $this->assertContains('readabilityLink-3', $readability->getContent()->innerHTML);
@@ -109,7 +119,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('<div readability=', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
         $this->assertNotContains('will be removed', $readability->getContent()->innerHTML);
         $this->assertNotContains('<h2>', $readability->getContent()->innerHTML);
@@ -123,7 +135,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('<div readability=', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
         $this->assertContains('nofollow', $readability->getContent()->innerHTML);
     }
@@ -136,7 +150,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('alt="article"', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
         $this->assertContains('nofollow', $readability->getContent()->innerHTML);
     }
@@ -149,7 +165,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('alt="article"', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
         $this->assertNotContains('<aside>', $readability->getContent()->innerHTML);
         $this->assertContains('<footer/>', $readability->getContent()->innerHTML);
@@ -163,7 +181,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('alt="article"', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
         $this->assertNotContains('This text should be removed', $readability->getContent()->innerHTML);
     }
@@ -176,7 +196,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('alt="tr"', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
     }
 
@@ -188,7 +210,9 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('alt="article"', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
         $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
         $this->assertContains('This text is also an awesome text and you should know that', $readability->getContent()->innerHTML);
     }
@@ -201,7 +225,69 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
 
         $this->assertTrue($res);
         $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
         $this->assertContains('alt="article"', $readability->getContent()->innerHTML);
+        $this->assertEmpty($readability->getTitle()->innerHTML);
+        $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
+        $this->assertNotContains('This text is also an awesome text and you should know that', $readability->getContent()->innerHTML);
+    }
+
+    public function testTitle()
+    {
+        $readability = new ReadabilityTested('<title>this is my title</title><article class="awesomecontent">'.str_repeat('<p>This is an awesome text with some links, here there are the awesome</p>', 7).'<p></p></article>', 'http://0.0.0.0');
+        $readability->debug = true;
+        $res = $readability->init();
+
+        $this->assertTrue($res);
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
+        $this->assertContains('alt="article"', $readability->getContent()->innerHTML);
+        $this->assertEquals('this is my title', $readability->getTitle()->innerHTML);
+        $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
+        $this->assertNotContains('This text is also an awesome text and you should know that', $readability->getContent()->innerHTML);
+    }
+
+    public function testTitleWithDash()
+    {
+        $readability = new ReadabilityTested('<title> title2 - title3 </title><article class="awesomecontent">'.str_repeat('<p>This is an awesome text with some links, here there are the awesome</p>', 7).'<p></p></article>', 'http://0.0.0.0');
+        $readability->debug = true;
+        $res = $readability->init();
+
+        $this->assertTrue($res);
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
+        $this->assertContains('alt="article"', $readability->getContent()->innerHTML);
+        $this->assertEquals('title2 - title3', $readability->getTitle()->innerHTML);
+        $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
+        $this->assertNotContains('This text is also an awesome text and you should know that', $readability->getContent()->innerHTML);
+    }
+
+    public function testTitleWithDoubleDot()
+    {
+        $readability = new ReadabilityTested('<title> title2 : title3 </title><article class="awesomecontent">'.str_repeat('<p>This is an awesome text with some links, here there are the awesome</p>', 7).'<p></p></article>', 'http://0.0.0.0');
+        $readability->debug = true;
+        $res = $readability->init();
+
+        $this->assertTrue($res);
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
+        $this->assertContains('alt="article"', $readability->getContent()->innerHTML);
+        $this->assertEquals('title2 : title3', $readability->getTitle()->innerHTML);
+        $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
+        $this->assertNotContains('This text is also an awesome text and you should know that', $readability->getContent()->innerHTML);
+    }
+
+    public function testTitleTooShortUseH1()
+    {
+        $readability = new ReadabilityTested('<title>too short</title><h1>this is my h1 title !</h1><article class="awesomecontent">'.str_repeat('<p>This is an awesome text with some links, here there are the awesome</p>', 7).'<p></p></article>', 'http://0.0.0.0');
+        $readability->debug = true;
+        $res = $readability->init();
+
+        $this->assertTrue($res);
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
+        $this->assertContains('alt="article"', $readability->getContent()->innerHTML);
+        $this->assertEquals('this is my h1 title !', $readability->getTitle()->innerHTML);
         $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->innerHTML);
         $this->assertNotContains('This text is also an awesome text and you should know that', $readability->getContent()->innerHTML);
     }
@@ -216,4 +302,53 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
     //     $this->assertEquals('/0\.0\.0\.0/', $readability->getDomainRegexp());
     //     $this->assertInstanceOf('DomDocument', $readability->dom);
     // }
+
+    // dummy function to be used to the next test
+    public function error2Exception($code, $string, $file, $line, $context)
+    {
+        throw new \Exception($string, $code);
+    }
+
+    public function testAutoClosingIframeNotThrowingException()
+    {
+        error_reporting(E_ALL | E_STRICT);
+        ini_set('display_errors', true);
+        set_error_handler(array($this, 'error2Exception'), E_ALL | E_STRICT);
+
+        $data = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+            <html xmlns="http://www.w3.org/1999/xhtml" lang="ru-RU" prefix="og: http://ogp.me/ns#">
+
+            <head profile="http://gmpg.org/xfn/11">
+            <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
+            <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+
+            </head>
+            <body class="single single-post postid-22030 single-format-standard">
+              <div id="wrapper">
+                <div id="content">
+                  <div class="post-22030 post type-post status-publish format-standard has-post-thumbnail hentry category-video category-reviews tag-193" id="post-22030">
+                    <h1>3D Touch &#8212; будущее мобильных игр</h1>
+                    <div class="postdate">Автор:  <strong>Сергей Пак</strong> | Просмотров: 1363 | Опубликовано:  14 сентября 2015 </div>
+                    <div class="entry">
+                      <p>Компания Apple представила новую технологию 3D Touch, которая является прямым потомком более ранней версии Force Touch &#8212; последняя, напомним, используется сейчас в трекпадах Macbook Pro и Macbook 2015. Теперь управлять устройством стало в разы проще, и Force Touch открывает перед пользователями новые возможности, но при этом 3D Touch &#8212; это про другое. Дело в том, что теперь и на мобильных устройствах интерфейс будет постепенно меняться, кардинальные перемены ждут мобильный гейминг, потому что здесь разработчики действительно могут разгуляться.<span id="more-22030"></span></p>
+                      <p><iframe src="https://www.youtube.com/embed/PUep6xNeKjA" width="560" height="315" frameborder="0" allowfullscreen="allowfullscreen"></iframe></p>
+                      <p>Итак, просто представьте себе, что iPhone 6S &#8212; это, по большому счету, отличная игровая приставка, которую вы носите с собой, а еще она может выдавать невероятной красоты картинку. Но проблема заключается, пожалуй, в том, что управлять персонажем в играх довольно трудно &#8212; он неповоротлив, обладает заторможенной реакцией, а игровой клиент зачастую требует перегруза интерфейса для того, чтобы обеспечить максимально большое количество возможностей. Благодаря трехуровневому нажатию можно избавиться от лишних кнопок и обеспечить более качественный обзор местности, и при этом пользователь будет закрывать пальцами минимальное пространство.</p>
+                    </div>
+                  </div>
+                </div>
+              </div>
+            </body>
+            </html>';
+
+        $readability = new ReadabilityTested($data, 'http://iosgames.ru/?p=22030');
+        $readability->debug = true;
+
+        $res = $readability->init();
+
+        $this->assertTrue($res);
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getContent());
+        $this->assertInstanceOf('Readability\JSLikeHTMLElement', $readability->getTitle());
+        $this->assertContains('<iframe src="https://www.youtube.com/embed/PUep6xNeKjA" width="560" height="315" frameborder="0" allowfullscreen="allowfullscreen"> </iframe>', $readability->getContent()->innerHTML);
+        $this->assertContains('3D Touch', $readability->getTitle()->innerHTML);
+    }
 }