MDL-54853 weblib: Properly specify encoding for blanktarget

diff --git a/lib/tests/weblib_format_text_test.php b/lib/tests/weblib_format_text_test.php
index a7aa3bf..d0be09f 100644 (file)
--- a/lib/tests/weblib_format_text_test.php
+++ b/lib/tests/weblib_format_text_test.php
@@ -115,36 +115,47 @@ class core_weblib_format_text_testcase extends advanced_testcase {
      */
     public function format_text_blanktarget_testcases() {
         return [
-            'Simple link' =>
-                [
-                    '<a href="https://www.youtube.com/watch?v=JeimE8Wz6e4">Hey, that\'s pretty good!</a>',
-                    '<div class="text_to_html"><a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" target="_blank"' .
-                        ' rel="noreferrer">Hey, that\'s pretty good!</a></div>'
-                ],
-            'Link with rel' =>
-                [
-                    '<a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" rel="nofollow">Hey, that\'s pretty good!</a>',
-                    '<div class="text_to_html"><a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" rel="nofollow noreferrer"' .
-                        ' target="_blank">Hey, that\'s pretty good!</a></div>'
-                ],
-            'Link with rel noreferrer' =>
-                [
-                    '<a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" rel="noreferrer">Hey, that\'s pretty good!</a>',
-                    '<div class="text_to_html"><a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" rel="noreferrer"' .
-                     ' target="_blank">Hey, that\'s pretty good!</a></div>'
-                ],
-            'Link with target' =>
-                [
-                    '<a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" target="_self">Hey, that\'s pretty good!</a>',
-                    '<div class="text_to_html"><a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" target="_self">' .
-                        'Hey, that\'s pretty good!</a></div>'
-                ],
-            'Link with target blank' =>
-                [
-                    '<a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" target="_blank">Hey, that\'s pretty good!</a>',
-                    '<div class="text_to_html"><a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" target="_blank"' .
-                        ' rel="noreferrer">Hey, that\'s pretty good!</a></div>'
-                ]
+            'Simple link' => [
+                '<a href="https://www.youtube.com/watch?v=JeimE8Wz6e4">Hey, that\'s pretty good!</a>',
+                '<div class="text_to_html"><a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" target="_blank"' .
+                    ' rel="noreferrer">Hey, that\'s pretty good!</a></div>'
+            ],
+            'Link with rel' => [
+                '<a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" rel="nofollow">Hey, that\'s pretty good!</a>',
+                '<div class="text_to_html"><a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" rel="nofollow noreferrer"' .
+                    ' target="_blank">Hey, that\'s pretty good!</a></div>'
+            ],
+            'Link with rel noreferrer' => [
+                '<a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" rel="noreferrer">Hey, that\'s pretty good!</a>',
+                '<div class="text_to_html"><a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" rel="noreferrer"' .
+                 ' target="_blank">Hey, that\'s pretty good!</a></div>'
+            ],
+            'Link with target' => [
+                '<a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" target="_self">Hey, that\'s pretty good!</a>',
+                '<div class="text_to_html"><a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" target="_self">' .
+                    'Hey, that\'s pretty good!</a></div>'
+            ],
+            'Link with target blank' => [
+                '<a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" target="_blank">Hey, that\'s pretty good!</a>',
+                '<div class="text_to_html"><a href="https://www.youtube.com/watch?v=JeimE8Wz6e4" target="_blank"' .
+                    ' rel="noreferrer">Hey, that\'s pretty good!</a></div>'
+            ],
+            'Link with Frank\'s casket inscription' => [
+                '<a href="https://en.wikipedia.org/wiki/Franks_Casket">ᚠᛁᛋᚳ᛫ᚠᛚᚩᛞᚢ᛫ᚪᚻᚩᚠᚩᚾᚠᛖᚱᚷ ᛖᚾᛒᛖᚱᛁᚷ ᚹᚪᚱᚦᚷᚪ᛬ᛋᚱᛁᚳᚷᚱᚩᚱᚾᚦᚫᚱᚻᛖᚩᚾᚷᚱᛖᚢᛏᚷᛁᛋᚹᚩᛗ ᚻ' .
+                    'ᚱᚩᚾᚫᛋᛒᚪᚾ ᛗᚫᚷᛁᚠᛁᛋᚳ᛫ᚠᛚᚩᛞᚢ᛫ᚪᚻᚩᚠᚩᚾᚠᛖᚱᚷ ᛖᚾᛒᛖᚱᛁᚷ ᚹᚪᚱᚦᚷᚪ᛬ᛋᚱᛁᚳᚷᚱᚩᚱᚾᚦᚫᚱᚻᛖᚩᚾᚷᚱᛖᚢᛏᚷᛁᛋᚹᚩᛗ ᚻᚱᚩᚾᚫᛋᛒᚪᚾ ᛗᚫᚷᛁ</a>',
+                '<div class="text_to_html"><a href="https://en.wikipedia.org/wiki/Franks_Casket" target="_blank" ' .
+                    'rel="noreferrer">ᚠᛁᛋᚳ᛫ᚠᛚᚩᛞᚢ᛫ᚪᚻᚩᚠᚩᚾᚠᛖᚱᚷ ᛖᚾᛒᛖᚱᛁᚷ ᚹᚪᚱᚦᚷᚪ᛬ᛋᚱᛁᚳᚷᚱᚩᚱᚾᚦᚫᚱᚻᛖᚩᚾᚷᚱᛖᚢᛏᚷᛁᛋᚹᚩᛗ ᚻᚱᚩᚾᚫᛋᛒᚪᚾ ᛗᚫᚷᛁᚠᛁᛋᚳ᛫ᚠᛚᚩᛞᚢ᛫ᚪᚻᚩᚠᚩᚾᚠᛖᚱᚷ ᛖᚾ' .
+                    'ᛒᛖᚱᛁᚷ ᚹᚪᚱᚦᚷᚪ᛬ᛋᚱᛁᚳᚷᚱᚩᚱᚾᚦᚫᚱᚻᛖᚩᚾᚷᚱᛖᚢᛏᚷᛁᛋᚹᚩᛗ ᚻᚱᚩᚾᚫᛋᛒᚪᚾ ᛗᚫᚷᛁ</a></div>'
+             ],
+            'No link' => [
+                'Some very boring text written with the Latin script',
+                '<div class="text_to_html">Some very boring text written with the Latin script</div>'
+            ],
+            'No link with Thror\'s map runes' => [
+                'ᛋᛏᚫᚾᛞ ᛒᚣ ᚦᛖ ᚷᚱᛖᚣ ᛋᛏᚩᚾᛖ ᚻᚹᛁᛚᛖ ᚦᛖ ᚦᚱᚢᛋᚻ ᚾᚩᚳᛋ ᚫᚾᛞ ᚦᛖ ᛋᛖᛏᛏᛁᚾᚷ ᛋᚢᚾ ᚹᛁᚦ ᚦᛖ ᛚᚫᛋᛏ ᛚᛁᚷᚻᛏ ᚩᚠ ᛞᚢᚱᛁᚾᛋ ᛞᚫᚣ ᚹᛁᛚᛚ ᛋᚻᛁᚾᛖ ᚢᛈᚩᚾ ᚦᛖ ᚳᛖᚣᚻᚩᛚᛖ',
+                '<div class="text_to_html">ᛋᛏᚫᚾᛞ ᛒᚣ ᚦᛖ ᚷᚱᛖᚣ ᛋᛏᚩᚾᛖ ᚻᚹᛁᛚᛖ ᚦᛖ ᚦᚱᚢᛋᚻ ᚾᚩᚳᛋ ᚫᚾᛞ ᚦᛖ ᛋᛖᛏᛏᛁᚾᚷ ᛋᚢᚾ ᚹᛁᚦ ᚦᛖ ᛚᚫᛋᛏ ᛚᛁᚷᚻᛏ ᚩᚠ ᛞᚢᚱᛁᚾᛋ ᛞᚫᚣ ᚹ' .
+                'ᛁᛚᛚ ᛋᚻᛁᚾᛖ ᚢᛈᚩᚾ ᚦᛖ ᚳᛖᚣᚻᚩᛚᛖ</div>'
+            ]
         ];
     }
 }

diff --git a/lib/weblib.php b/lib/weblib.php
index e094828..1e24072 100644 (file)
--- a/lib/weblib.php
+++ b/lib/weblib.php
@@ -1320,8 +1320,9 @@ function format_text($text, $format = FORMAT_MOODLE, $options = null, $courseidd
     }
 
     if ($options['blanktarget']) {
+        $encoding = ini_get('default_charset') ? ini_get('default_charset') : 'utf-8';
         $domdoc = new DOMDocument();
-        $domdoc->loadHTML($text);
+        $domdoc->loadHTML('<?xml encoding="' . $encoding . '" ?>' . $text);
         foreach ($domdoc->getElementsByTagName('a') as $link) {
             if ($link->hasAttribute('target') && strpos($link->getAttribute('target'), '_blank') === false) {
                 continue;
@@ -1336,7 +1337,7 @@ function format_text($text, $format = FORMAT_MOODLE, $options = null, $courseidd
         // $domdoc->loadHTML($text, LIBXML_HTML_NOIMPLIED | LIBXML_HTML_NODEFDTD); however it seems like the libxml
         // version that travis uses doesn't work properly and ends up leaving <html><body>, so I'm forced to use
         // this regex to remove those tags.
-        $text = trim(preg_replace('~<(?:!DOCTYPE|/?(?:html|body))[^>]*>\s*~i', '', $domdoc->saveHTML()));
+        $text = trim(preg_replace('~<(?:!DOCTYPE|/?(?:html|body))[^>]*>\s*~i', '', $domdoc->saveHTML($domdoc->documentElement)));
     }
 
     return $text;