Currently, text with the writing direction “right-to-left” (RTL), e.g. PDF documents with text in Hebrew or Arabic are not processed correctly. (Rotated text with the writing direction left-to-right can be tested without any problem.)

In the current release 2054.10 text overflow in fields can not be detected when the last line of the text starts inside the field but extends outside.

The property hidden of a form field is evaluated incorrectly in some situations. This problem is currently under investigation using appropriate test documents.

Some properties of form fields are not extracted, for example “background color” , “border color” and “border styles” .

Currently not all signature data is exported into XML. It is possible that the XML structure will change in future releases.

In the current release 2054.10 colors cannot be tested directly. If colors have to be tested, the test has to be carried out using rendered pages. The sections 3.17: “Layout - Entire PDF Pages” and 3.18: “Layout - in Clipping Areas” describe such tests.

Tests related to text and images are not restricted to layers.

In one of the next releases PDFUnit will check the PDF/A compliance of a PDF document.

In the current release only the document-level XMP data are extracted and evaluated.

The content of some XML tags depend on each other. Such dependencies can not be specified by XML Schema. So a parser can not validate it. It is intended, to provide a Schematron ruleset for cross constraints. Good information about Schematron is available in Wikipedia (http://en.wikipedia.org/wiki/Schematron).