diff --git a/.gitignore b/.gitignore
index 48b8bf9..d191143 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1 +1,3 @@
vendor/
+coverage/
+composer.lock
diff --git a/phpunit.xml.dist b/phpunit.xml.dist
index f727e46..e6a0903 100644
--- a/phpunit.xml.dist
+++ b/phpunit.xml.dist
@@ -19,11 +19,14 @@
- ./src/TubeLink/
+ ./src/
./tests
+
+
+
diff --git a/src/Readability.php b/src/Readability.php
index 7f1837e..50e5d2f 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -52,6 +52,7 @@ class Readability
public $revertForcedParagraphElements = true;
public $articleTitle;
public $articleContent;
+ public $original_html;
public $dom;
public $url = null; // optional - URL where HTML was retrieved
public $lightClean = true; // preserves more content (experimental)
@@ -179,7 +180,7 @@ class Readability
$this->debugText .= 'Tidying document'."\n";
$tidy = tidy_parse_string($html, $this->tidy_config, 'UTF8');
if (tidy_clean_repair($tidy)) {
- $original_html = $html;
+ $this->original_html = $html;
$this->tidied = true;
$html = $tidy->value;
$html = preg_replace('/]+>/i', '', $html);
@@ -258,7 +259,7 @@ class Readability
if ($this->bodyCache == null) {
$this->bodyCache = '';
foreach ($bodyElems as $bodyNode) {
- $this->bodyCache += $bodyNode->innerHTML;
+ $this->bodyCache .= trim($bodyNode->innerHTML);
}
}
if ($bodyElems->length > 0 && $this->body == null) {