Fix EZP-26096: Custom tags' text line parameters not saved correctly (#1249)

Author: dpobel

extension/ezoe/tests/ezoexmltext_regression.php

@@ -9,6 +9,36 @@
 
 class eZOEXMLTextRegression extends ezpDatabaseTestCase
 {
+    /**
+     * Test for EZP-26096
+     * @link https://jira.ez.no/browse/EZP-26096
+     * @dataProvider providerParsingGreaterThanAttribute
+     */
+    public function testParsingGreaterThanAttribute( $html, $expectedXml )
+    {
+        $parser = new eZOEInputParser();
+        $dom = $parser->process( $html );
+
+        self::assertInstanceOf( 'DomDocument', $dom );
+        self::assertEquals( $expectedXml, trim( $dom->saveXML() ) );
+    }
+
+    public function providerParsingGreaterThanAttribute()
+    {
+        return array(
+            array(
+                '<div type="custom" class="ezoeItemCustomTag factbox" customattributes="title|This > is > a > fact > attribute_separationalign|right"><p>This is a fact</p></div>',
+                '<?xml version="1.0" encoding="utf-8"?>
+<section xmlns:image="http://ez.no/namespaces/ezpublish3/image/" xmlns:xhtml="http://ez.no/namespaces/ezpublish3/xhtml/" xmlns:custom="http://ez.no/namespaces/ezpublish3/custom/"><paragraph xmlns:tmp="http://ez.no/namespaces/ezpublish3/temporary/"><custom name="factbox" custom:title="This &gt; is &gt; a &gt; fact &gt; "><paragraph>This is a fact</paragraph></custom></paragraph></section>',
+            ),
+            array(
+                '<div type="custom" class="ezoeItemCustomTag factbox" customattributes="title|<a href=&quot;#test&quot;>Test</a>attribute_separationalign|right"><p>This is a fact</p></div>',
+                '<?xml version="1.0" encoding="utf-8"?>
+<section xmlns:image="http://ez.no/namespaces/ezpublish3/image/" xmlns:xhtml="http://ez.no/namespaces/ezpublish3/xhtml/" xmlns:custom="http://ez.no/namespaces/ezpublish3/custom/"><paragraph xmlns:tmp="http://ez.no/namespaces/ezpublish3/temporary/"><custom name="factbox" custom:title="&lt;a href=&amp;quot;#test&amp;quot;&gt;Test&lt;/a&gt;"><paragraph>This is a fact</paragraph></custom></paragraph></section>',
+            )
+        );
+    }
+
     /**
      * Test for issue #16605: Online Editor adds a lot of Non Breaking spaces (nbsp)
      * 

kernel/classes/datatypes/ezxmltext/ezxmlinputparser.php

@@ -413,7 +413,7 @@ function parseTag( &$data, &$pos, &$parent )
         //  Regular tag: get tag's name and attributes.
         else
         {
-            $tagEndPos = strpos( $data, '>', $tagBeginPos );
+            $tagEndPos = $this->findEndOpeningTagPosition( $data, $tagBeginPos );
             if ( $tagEndPos === false )
             {
                 $pos = $tagBeginPos + 1;
@@ -583,6 +583,59 @@ function parseTag( &$data, &$pos, &$parent )
         Helper functions for pass 1
     */
 
+    /**
+     * Finds the postion of the > character which marks the end of the opening
+     * tag that starts at $tagBeginPos in $data.
+     * It's not as easy as it seems, because some '>' can also appear in attributes.
+     * So we need to iterate over the next '>' characters to find the correct one.
+     * See https://jira.ez.no/browse/EZP-26096
+     *
+     * @param string $data
+     * @param integer $tagBeginPos
+     * @param integer $offset used for recursive call when a > is found in an attribute.
+     * @return integer|false
+     */
+    private function findEndOpeningTagPosition( $data, $tagBeginPos, $offset = 0 )
+    {
+        $endPos = strpos( $data, '>', $tagBeginPos + $offset );
+        if ( $endPos === false )
+        {
+            return false;
+        }
+        $tagCode = substr( $data, $tagBeginPos, $endPos - $tagBeginPos );
+        if ( strpos( $tagCode, '=' ) === false )
+        {
+            // this tag has no attribute, so the next '>' is the right one.
+            return $endPos;
+        }
+        if ( $this->isValidXmlTag( $tagCode ) )
+        {
+            return $endPos;
+        }
+        return $this->findEndOpeningTagPosition( $data, $tagBeginPos, $endPos + 1 );
+    }
+
+    /**
+     * Checks whether $code can be considered as a valid XML excerpt. If not,
+     * it's probably because we found a '>' in the middle of an attribute.
+     *
+     * @param string $code
+     * @return boolean
+     */
+    private function isValidXmlTag( $code )
+    {
+        if ( $code[strlen( $code ) - 1] !== '/' )
+        {
+            $code .= '/';
+        }
+        $code .= '>';
+        $code = '<' . str_replace( '<', '&lt;', substr( $code, 1 ) );
+        $errorHanding = libxml_use_internal_errors( true );
+        $simpleXml = simplexml_load_string( $code );
+        libxml_use_internal_errors( $errorHanding );
+        return ( $simpleXml !== false );
+    }
+
     function parseAttributes( $attributeString )
     {
         $attributes = array();