XML document import not working for text elements

[perklason]

Hi when I try to import XML with tags like this one (I am using jackalope-doctrine-dbal in Symfony):

<title>HarryPotter</title>

I get the following error:

[PHPCR\InvalidSerializedDataException]             

Unexpected element in stream: #text="HarryPotter"  
  
Exception trace:
 () at /var/www/html/std1/vendor/jackalope/jackalope/src/Jackalope/ImportExport/ImportExport.php:631

If I do a $xml->isValid just before that line, it returns true. So what is wrong? Bug?

This is the XML doc I try to import for testing:

<bookstore>
  <book category="children">
    <title>HarryPotter</title>
    <author>J K. Rowling</author>   
    <year>2005</year>
    <price>29.99</price>
  </book>
  <book category="web">
    <title>Learning XML</title>
    <author>Erik T. Ray</author>    
    <year>2003</year>
    <price>39.95</price>
  </book>
</bookstore> 

[dbu]

are you trying to import this as "system view" or as "document view"?

[perklason]

It gets auto imported as document view (ends up in function importDocumentView).

Seems something weird happens here somewhere:

// get current depth to detect self-closing tag. unfortunately, we do
        // not get an END_ELEMENT for self-closing tags but read() just jumps
        // to the next element, even moving up in the tree,
        $depth = $xml->depth;
        if (! $hasAttributes) {
            // we where on an empty element, thus not inside its attributes. change depth to 1 deeper
            // thanks XMLReader, great work :-(
            $depth++;
        }
        $xml->read(); // move out of current node to next

        // TODO: what about significant whitespace? maybe the read above should not even skip significant empty whitespace...

        // while we are on element and at same depth, these are children of the current node
        while (XMLReader::ELEMENT == $xml->nodeType && $xml->depth == $depth) {
            self::importDocumentView($node, $ns, $xml, $uuidBehavior, $namespaceMap);
        }

        if (XMLReader::END_ELEMENT != $xml->nodeType && $xml->depth != $depth - 1) {
            throw new InvalidSerializedDataException('Unexpected element in stream: '.$xml->name.'="'.$xml->value.'"');
        }

[dbu]

it looks like we missed to implement handling of text nodes in the document view parser. according to https://docs.adobe.com/docs/en/spec/jcr/2.0/11_Import.html we should create a node jcr:xmltext with a property jcr:xmlcharacters.

do you want to implement this? i think we can simply add a check after the while ELEMENT before the if ! END_ELEMENT and create the text child node. would need to keep track of the nodes we create in the self::importDocumentView so we can attach it to the right place.

a first step would be to modify the test in https://github.com/phpcr/phpcr-api-tests/ so that it contains a string content and see it fail, then make it work...

[perklason]

It would be very nice if this can be implemented. Would help out if I can, but this is a bit beyond me, unfortunately.