fix of issue #677: When downloading images, wallabag doesnt respect html "base" tag, tnx to @fivefilters

2 files changed, 15 insertions, 1 deletions

diff --git a/inc/3rdparty/makefulltextfeed.php b/inc/3rdparty/makefulltextfeed.php
index 62c050ec..a081f88b 100755
--- a/inc/3rdparty/makefulltextfeed.php
+++ b/inc/3rdparty/makefulltextfeed.php
@@ -671,7 +671,11 @@ foreach ($items as $key => $item) {
                         $html .= $item->get_description();

                 } else {

                         $readability->clean($content_block, 'select');

-                        if ($options->rewrite_relative_urls) makeAbsolute($effective_url, $content_block);

+                        // get base URL

+                        $base_url = get_base_url($readability->dom);

+                        if (!$base_url) $base_url = $effective_url;

+                        // rewrite URLs

+                        if ($options->rewrite_relative_urls) makeAbsolute($base_url, $content_block);

                         // footnotes

                         if (($links == 'footnotes') && (strpos($effective_url, 'wikipedia.org') === false)) {

                                 $readability->addFootnotes($content_block);

diff --git a/inc/3rdparty/makefulltextfeedHelpers.php b/inc/3rdparty/makefulltextfeedHelpers.php
index 4e985372..ac872ab8 100755
--- a/inc/3rdparty/makefulltextfeedHelpers.php
+++ b/inc/3rdparty/makefulltextfeedHelpers.php
@@ -377,3 +377,13 @@ function debug($msg) {
                 flush();
         }
 }
+
+function get_base_url($dom) {

+        $xpath = new DOMXPath($dom);

+        $base_url = @$xpath->evaluate('string(//head/base/@href)', $dom);

+        if ($base_url !== '') {

+                return $base_url;

+        } else {

+                return false;

+        }

+}